AN ANALYSIS OF ECONOMIC EFFICIENCY IN PREDICTING LEGISLATIVE VOTING BEYOND A TRADITIONAL LIBERAL-CONSERVATIVE SPECTRUM Except where reference is made to the work of others, the work described in this dissertation is my own or was done in collaboration with my advisory committee. This dissertation does not include proprietary or classified information. __________________________ Benjamin Bryan Boozer, Jr. Certificate of Approval: __________________________ __________________________ Cynthia J. Bowling Anne Permaloff, Chair Associate Professor Professor Political Science Political Science and Public Administration __________________________ __________________________ Gerard S. Gryski Thomas J. Vocino Professor Professor Political Science Political Science and Public Administration __________________________ Joe F. Pittman Interim Dean Graduate School AN ANALYSIS OF ECONOMIC EFFICIENCY IN PREDICTING LEGISLATIVE VOTING BEYOND A TRADITIONAL LIBERAL-CONSERVATIVE SPECTRUM Benjamin Bryan Boozer, Jr. A Dissertation Submitted to the Graduate Faculty of Auburn University in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy Auburn, Alabama May 10, 2008 iii AN ANALYSIS OF ECONOMIC EFFICIENCY IN PREDICTING LEGISLATIVE VOTING BEYOND A TRADITIONAL LIBERAL-CONSERVATIVE SPECTRUM Benjamin Bryan Boozer, Jr. Permission is granted to Auburn University to make copies of this dissertation at its discretion, upon request of individuals or institutions and at their expense. The author reserves all publication rights. __________________________ Signature of Author __________________________ Date of Graduation iv DISSERTATION ABSTRACT AN ANALYSIS OF ECONOMIC EFFICIENCY IN PREDICTING LEGISLATIVE VOTING BEYOND A TRADITIONAL LIBERAL-CONSERVATIVE SPECTRUM Benjamin Bryan Boozer, Jr. Doctor of Philosophy, May 10, 2008 (M.P.A., Jacksonville State University, 1992) (B.S., Jacksonville State University, 1990) 405 Typed Pages Directed by Anne Permaloff Self-interest and ideology are important explanatory variables for human behavior and are the two primary determinants of legislative decision-making. Self- interest usually pertains to a maximization of financial resources and is closely related to the concept of rationality. Utility maximization is a component of self-interest and individuals cannot be expected to pursue public interests unless the individual?s self- interest is met. Ideology reflects deep beliefs about a person and how individual satisfaction derives from improving the lives of others or promoting ideological positions. Ideology is defined as an action oriented model of people and society. Political ideologies depict the preferred states of the world and are often illustrated through a liberal-conservative v spectrum. Liberal positions generally espouse more government intervention and equitable resource distribution. Conservative positions, on the other hand, are less likely to embrace the need for government policies and are more concerned about costs associated with such intervention. Measuring legislative decision-making through a liberal-conservative spectrum includes characteristics of a legislator?s behavior but does not include results of those actions. Through development of an economic efficiency index (E-score) that assigns numerical values to legislative voting, public benefits of a public policy decision are measured vis-?-vis public costs. Higher E-scores are consistent with legislative behavior promoting greater net public policy benefits, while lower E-scores are associated with relatively lower net public policy benefits. The model utilizes two dependent variables: support for increasing the federal minimum wage (an economically inefficient policy) and support for medical malpractice reform (an economically efficient policy). Roll call votes of members of the U.S. House of Representatives and U.S. Senate are analyzed from the 99 th through 108 th Congresses. A multivariate analysis of the model finds that liberal-conservative ideology is a better predictor of legislative behavior than economic efficiency. This study finds that the potential use of economic efficiency is numerous in public policy dialogue and analysis for supplementing liberal-conservative measures with objective criterion for understanding behavior. Application of an E-score transcends legislative voting at the federal level to include state and local government analysis of public policy solutions to private sector needs. vi Style manual used: Publication Manual of the American Psychological Association, 5 th Edition (2003) Computer software used: Microsoft Word and Excel, Statistical Package for the Social Sciences 11.5 vii TABLE OF CONTENTS PAGE LIST OF TABLES .....................................................................................................xi-xiv LIST OF FIGURES..........................................................................................................xv CHAPTER ONE AN EXAMINATION OF ECONOMIC EFFICIENCY IN PREDICTING LEGISLATIVE VOTING BEYOND A TRADITIONAL LIBERAL-CONSERVATIVE SPECTRUM....................1 Introduction...................................................................................................1 Statement of Problem .....................................................................................4 Research Question.........................................................................................29 E-score Decision Rules............................................................................31 Methodology.............................................................................................. 41 Contribution of Study...................................................................................47 Outline of Dissertation .................................................................................49 TWO LITERATURE REVIEW.............................................................................51 Introduction to Economic Efficiency........................................................52 Self-interest and Ideology........................................................................61 Interest Group Theory...............................................................................74 Institutions of Government.......................................................................87 viii Lawmaking...............................................................................95 Political Party Control.............................................................................111 Conditional Party Government...............................................................114 Summary.......................................................................................121 THREE CONCEPTUAL DESIGN AND RESEARCH METHODOLOGY...........................................................124 Dependent Variables.................................................................................126 Vectors of Analysis.................................................................................133 E-score Development..............................................................................146 Methods of Analysis................................................................................150 Hypothesis Development........................................................................155 Limitations of Research..........................................................................161 FOUR ANALYSIS AND EMPIRICAL RESULTS...........................................163 Descriptive Analysis...............................................................................169 Multivariate Analysis..............................................................................184 Hypothesis Testing..................................................................................189 Hypothesis Testing Summary.................................................................319 Time Series Analysis..............................................................................322 FIVE OVERALL CONCLUSIONS AND POLICY IMPLICATION.............332 Public Policy Development and Economic Efficiency............................334 ix Economic Efficiency and Vote Models...................................................340 Usefulness of E-score..............................................................................343 Conclusion.................................................................................................. 350 REFERENCES..................................................................................................352 APPENDICES.........................................................................................................373 APPENDIX A Medical Malpractice Legislation Selected as Dependent Variable....................................................374 APPENDIX B Minimum Wage Legislation Selected as Dependent Variable........................................................378 APPENDIX C Roll Call Votes Selected by House and Senate in Compiling an E-score for Legislators.............................382 x LIST OF TABLES PAGE 3.1 Control Variables for Party Control of Government????????..............142 3.2 Control Variables for Geographic Conditions?????....????.............144 3.3 Control Variables for State Economic Conditions????????.............. 146 3.4 Summary of E-score Development????????????????.....149 3.5 Hypotheses for Medical Malpractice Reform Policy Area???????......159 3.6 Hypotheses for Minimum Wage Legislation Policy Area???????.......160 4.1 Party Divisions, 99 th ? 108 th Congresses??????????????.....165 4.2 Dependent Variable Legislation for Each Congress: House and Senate???????????????????????....167 4.3 Regression Analysis of Ideology Influences on Legislative Voting in the House?????????????????......178 4.4 Regression Analysis of Ideology Influences on Legislative Voting in the Senate??????????????.??.........180 4.5 Regression Analysis of Scoring Models from 99 th to 108 th Congress of Ideology Influences on Legislative Voting in House??..............181 4.6 Regression Analysis of Scoring Models from 99 th to 108 th Congress of Ideology Influences on Legislative Voting in Senate??.............183 4.7 Hypotheses for Medical Malpractice Reform Policy Area????..................186 4.8 Hypotheses for Minimum Wage Legislation Policy Area???????........187 4.9 Variables Used in Testing Hypotheses for Medical Malpractice and Minimum Wage ???????????????????????190 xi 4.10 Regression Analysis of Base Model for 99 th to 108 th House: Medical Malpractice???????????????????????.191 4.11 Regression Analysis of Base Model and Party Unity Substitution for 99 th to 108 th House: Medical Malpractice????????????......198 4.12 Regression Analysis of Base Model and ADA Substitution for 99 th to 108 th House: Medical Malpractice????????????......202 4.13 Regression Analysis of Base Model and ACU Substitution for 99 th to 108 th House: Medical Malpractice???????????..........207 4.14 Regression Analysis of Base Model and DW Nominate Substitution for 99 th to 108 th House: Medical Malpractice???????......212 4.15 Regression Analysis of Base Model and Legislator Party Substitution for 99 th to 108 th House: Medical Malpractice???????.....216 4.16 Regression Analysis of Base Model for 99 th to 108 th House: Minimum Wage????????????????????????.... 222 4.17 Regression Analysis of Base Model and Party Unity Substitution for 99 th to 108 th House: Minimum Wage??????????????.....228 4.18 Regression Analysis of Base Model and ADA Substitution for 99 th to 108 th House: Minimum Wage?????????????.........233 4.19 Regression Analysis of Base Model and ACU Substitution for 99 th to 108 th House: Minimum Wage??????????????.....238 4.20 Regression Analysis of Base Model and DW Nominate Substitution for 99 th to 108 th House: Minimum Wage.........................................242 4.21 Regression Analysis of Base Model and Legislator Party Substitution for 99 th to 108 th House: Minimum Wage???????..??...248 4.22 Regression Analysis of Base Models for 102 nd to 108 th Senate: Medical Malpractice??????????????????????......256 4.23 Regression Analysis of Base Models and Party Unity Substitution for 102 nd to 108 th Senate: Medical Malpractice????????????......260 4.24 Regression Analysis of Base Models and ADA Substitution for 102 nd to 108 th Senate: Medical Malpractice????????????......265 xii 4.25 Regression Analysis of Base Models and ACU Substitution for 103 rd to 108 th Senate: Medical Malpractice?????????????..269 4.26 Regression Analysis of Base Models and DW Nominate Substitution for 102 nd to 108 th Senate: Medical Malpractice???????......273 4.27 Regression Analysis of Base Models and Legislator Party Substitution for 102 nd to 108 th Senate: Medical Malpractice???????..277 4.28 Regression Analysis of Base Models for 100 th to 108 th Senate: Minimum Wage????????????????????????...282 4.29 Regression Analysis of Base Models and Party Unity Substitution for 100 th to 108 th Senate: Minimum Wage????????........??............288 4.30 Regression Analysis of Base Models and ADA Substitution for 99 th to 108 th Senate: Minimum Wage???????????????..294 4.31 Regression Analysis of Base Models and ACU Substitution for 100 th to 108 th Senate: Minimum Wage????????????............300 4.32 Regression Analysis of Base Models and DW Nominate Substitution for 100 th to 108 th Senate: Minimum Wage??????????..306 4.33 Regression Analysis of Base Models and Legislator Party Substitution for 100 th to 108 th Senate: Minimum Wage???????..........312 4.34 Summary and Confirmation of Hypothesis Tested in House and Senate: Medical Malpractice and Minimum Wage Dependent Variables?????.....319 4.35 Interrupted Time Series Analysis of Time and Changes in Party Control on Mean DW Nominate Scores for Republicans and Democrats in the House?............................................................................325 4.36 Interrupted Time Series Analysis of Time and Changes in Party Control on Mean Medical Malpractice Scores for Republicans and Democrats in the Senate???????????????.................. 326 4.37 Interrupted Time Series Analysis of Time and Changes in Party Control on Mean ADA Score for Republicans and Democrats in the Senate???....................................................................328 xiii 4.38 Interrupted Time Series Analysis of Time and Changes in Party Control on Mean DW Nominate Scores for Republicans and Democrats in the Senate???????????????.....???....330 5.1 Summary of Advantages in E-score Model Relative to Liberal-Conservative Models??????????????????........340 5.2 Application Summary of E-score Model in Public Policy Development and Analysis???????????????????........350 xiv LIST OF FIGURES PAGE 1.1 Production Possibilities Frontier..........................................................................30 1.2 Excise Subsidy and Consumer Surplus (Browning & Zupan, 2002)?...........?.32 1.3 Barriers to Entry of Markets Limit Competition (Kennedy, 2005).......................34 1.4 Wage and Price Controls (Browning & Zupan, 2002)..........................................36 2.1 Net Benefit Maximization (Kennedy, 2005).........................................................54 3.1 Minimum Wage and Labor Analysis (Browning & Zupan, 2002)........................129 4.1 Mean E-score, ADA, and ACU Values -- U.S. House.......................................171 4.2 Mean E-score, ADA, and ACU Values -- U.S. Senate...??.................?.......172 4.3 Mean DW Nominate Values -- U.S. House???????.?..??...........173 4.4 Mean DW Nominate Values -- U.S. Senate???.............???????.174 4.5 Mean E-scores for Republicans and Democrats -- U.S. House ?...........??....175 4.6 Mean E-scores for Republicans and Democrats -- U.S. Senate...........................176 1 CHAPTER ONE AN ANALYSIS OF ECONOMIC EFFICIENCY IN PREDICTING LEGISLATIVE VOTING BEYOND A TRADITIONAL LIBERAL-CONSERVATIVE SPECTRUM INTRODUCTION This dissertation tests whether a measure of economic efficiency known as the E- score is a better predictor of legislative voting than traditional explanatory variables such as ideology (liberal-conservative) and self-interest. This chapter introduces the problem of using measures of liberalism and conservatism in analyzing decision-making ideology and provides an overview of issues encountered in studying legislative voting. Based on this overview the research question is developed. The use of this research question portends the importance of economic efficiency as a variable in this study, and criteria for objective development of an E-score are listed. The theoretical focus of the study and research methodology employed in analyzing hypotheses in the model are introduced. The chapter concludes with an overview of chapters to follow. Overview of Subject Ideology and self-interest are well-established measures of individual decision- making. The variables reflect characteristics of the individual making the decision and not the consequences of such decisions. In legislative voting the consequences of 2 decisions are paramount to that decision serving the interests of the constituency that a legislator represents. Considering the political consequences of each decision is best illustrated through economic efficiency, a benefit-cost comparison of a legislator supporting legislation that expands social benefit. Economic efficiency is measured through an E-score. The extent that social benefit exceeds social cost considers the consequences of the legislator?s voting action that affects his or her constituency. Economic efficiency as the criterion for measuring social benefit-cost of legislative voting differs from administrative processes of efficiency pertaining to the method and means of decision making. Individuals are rational actors and each decision reflects such rationality. With consequences of socially beneficial decision making a focus of this study, how changes in government and the effect of political parties influence decision making are explored. This study considers ten Congresses (99 th -108 th ) in order to evaluate a series of changes in leadership and majority ? minority party relationships that influence the individual legislator as part of a larger institution. Ideology and Self-interest Ideology and self-interest are important explanatory variables for human behavior. Self-interest as a general explanation of human behavior has a long intellectual basis. In economics self-interest usually pertains to a maximization of financial resources and is closely related to the concept of rationality (Arrow, 1963; Buchanan & Tullock, 1962; and Downs, 1957). Buchanan (1972) argues that utility maximization is a component of self-interest. Individuals behave differently as part of group interaction (Olson, 1965; Truman, 1960), but cannot be expected to pursue public interests unless 3 the individual?s self-interest is met (Ostrom, 1989). Direct and political action committee (PAC) contributions are two areas where donations to legislators depict the effect of self-interests on voting decisions. Ideology, on the other hand, reflects deep-seated beliefs about a person. Kalt and Zupan (1984, p. 281) identify two components of ideology. First, the successful promotion of a deeply held belief may give one the satisfaction of improving the lives of others. Second, even if the pursuit of ideology has no effect on another person, satisfaction is possible from promoting these ideological positions. Grafton and Permaloff (2005b, p. 173) define ideology as an ?action oriented model of people and society.? They argue that political ideologies are ?more or less consistent sets of normative statements as to the best or preferred states of the world?about how government can best serve their proponents? conceptions of the public interest? (p. 68) and identify pure ideology, if it actually exists, as ?the manifestation of altruism in the public sector? (p. 69). These best or preferred states of the world often are depicted through a liberal-conservative spectrum. 1 Liberal positions generally espouse more government intervention and equitable distribution of resources. Conservative positions, on the other hand, are less likely to embrace the need for government policies and are more concerned about costs associated with such intervention (Ostrom & Ostrom, 1971). Policy decisions that are made along a liberal-conservative spectrum illustrate attributes of the legislator and how those individual preferences relate to public policy decisions. 1 Liberal-conservative terminology represents numerous areas of ideological reasoning that extend beyond the discussion offered here. For a more complete discussion of this concept, refer to Van Dyke (1995). 4 To better understand political events, adding or substituting ideology and self- interest produces interesting results (Uslaner, 1999; Levitt, 1996, Nelson & Silberberg, 1987; and Kalt & Zupan, 1984). For Alesina and Rosenthal (1995) ideology is a means to self-interested ends. Ideology and self-interest are action oriented, but important differences are present. Self-interest is not concerned with interrelationships between individuals and society, but start and end with policy decisions that benefit or hurt the self-interested person or those that are close to that person (Downs, 1998, p. 19; Buchanan, 1972, p. 19). In terms of ideology, individual behavior and political decisions are often studied along a liberal-conservative continuum. Voting by legislators in support of public policy positions is usually defined along this spectrum. Liberal-conservative ideological positions describe beliefs held by individuals who make legislative decisions. Those beliefs, however, better identify characteristics of a legislator and not the consequence or impact of the voting decision, a potential problem with liberal-conservative measures. Americans for Democratic Action (ADA) is liberal interest group. American Conservative Union (ACU) is a conservative interest group. ADA and ACU each measure liberal-conservative effects through an index of roll call votes made by legislators in the House and Senate. Each score assigned to a legislator is between 0 and 100. For each index ? ADA and ACU ? higher scores represent higher degrees of political liberalism and political conservatism, respectively. Measuring the ideology of a legislator through ADA or ACU type scores of liberalism and conservatism identifies propensities of a legislator to support liberal or conservative policy positions, but does not consider policy effects through policy outcomes. A liberal-conservative spectrum 5 based on the use of interest group scores illustrates the extent to which a legislator embraces policy positions consistent with the interest group?s ideological position. Such a liberal-conservative spectrum does not reflect the aggregate consequences of policy making, the expected result of a policy on social welfare. A consideration of the extent that a legislator espouses policy decisions that maximize social benefits is important in the policy formulation process. Political Parties Political parties also impact the behavior of legislators. The collective impact of party is tied to the level of party unity and the relationships between a legislator?s decision-making and political positions of the political party to which the legislator belongs. Party unity is a reflection of how closely legislators espouse the political positions of the party. The two major political parties in American government are the Democratic Party and Republican Party. In terms of ideology, Democrats are generally more liberal and Republicans are more conservative. Government is unified if House, Senate, and Executive Branch are all controlled by the same political party; otherwise government is nonunified or divided. Unified government should not be confused with party unity, which describes relationships between legislators and their party organization within the legislative body. Differences exist in party influences in the House as opposed to the Senate, with most differences revolving around the institutional effects associated with each chamber. For purposes of this dissertation, united or divided government is considered through the public policy effects of such relationships. Coleman (1999) argues that 6 unified government is more productive as evidenced by the passage of more significant public policies. Unified government offers more incentives to cooperate and is more responsive to the mood of the public (p. 821). To the extent that government is unified or divided might have an impact on legislative shirking by the legislator of those interests of his or her constituency and the self-interests of the legislator. Of particular interest to this study is whether changes in party control affect not only policy formulation but also economic efficiency as a variable that maximizes social benefits. Mayhew (1991) argues that political parties in America are weak and their role is constant regardless of whether government is unified or divided. According to Burns, the structure of Senate rules that require a supermajority to stop filibusters or override vetoes makes party control of that institution especially dubious if party control is less than the proportions needed to direct legislation (as cited in Coleman, 1999, p. 824). Nevertheless, presidents oppose significant legislation more often and much more important legislation fails to pass under divided government as opposed to unified government (Edwards, Barrett, & Peake, 1997). Individual Behavior and Political Party Influence Snyder and Groseclose (2000, p. 193) find that political party affiliation is one of the best predictors of voting behavior across both House and Senate. They state: ?Party influence (reward or punishment tied to a legislator?s vote as a result of direct pressure applied by the political party) appears more frequently on certain type of procedural votes?than on amendments and final passage? (p. 194). Shepsle argues that self-interest benefits are afforded to legislators who vote with party to obtain coveted committee 7 assignments or endorse party leadership (as cited in Strattmann, 2000, p. 666). Rational individuals seek these assignments to support their self-interests. Fiorina (2005, p. 171) explains that until the 1980s parties were motivated to win elections because of values attached to the job and tangible benefits of holding office. Party organizations that once were dominated through patronage are today increasingly identified with training and supporting strong candidates at the local levels. Much of the basis for the nationalization of local issues resulting from the 1994 congressional elections is based on the premise of associating a candidate with a party leader or party issue. Republicans assumed control of the House of Representatives for the first time since the 1950s and nationalized local issues by offering the voters a Contract with America that promised specific legislative proposals would be passed in the first 100 days of the new House. The rise of issue advocates parallels this trend. Coalitions of individuals form groups and espouse positions on issues across those groups and share particular interests. According to Fiorina (2005), ?Incumbents today do not find it as easy to separate themselves from party leaders, party images, and party performance as did incumbents twenty years ago? (pp. 170-171). Party influence links individual behavior of a legislator with national issues that define party activity (Snyder & Groseclose, 2000). Through the effects of party influences on the operation of Congress, efficiency is a function of a legislator to formulate public policy to address state or local concerns when the imagery of party positions and issues is tied to national party leaders. With party influence efficiency is defined through the organization of Congress and the collective responsibility of its members. Values associated with representativeness that 8 exemplified the legislative branch in the early 1970s were later displaced with arguments for efficiency and coherence to combat macroeconomic issues such as the energy crisis and inflation. Organizational reforms associated with Republican control of Congress in the 1990s mirror nationalization trends and increases in efficiency. Lack of boundaries between Congress as an institution and an environment consisting of issue advocates and interest groups exacerbate collective responsibilities of congressional members (Fiorina, 2005, pp. 175-177). Members of Congress may be influenced more by induced preferences from forces in the environment than party effects within the institution (p. 174). According to Aldrich and Rohde, and Sinclair, the argument that political parties through institutional paths explain legislative decision- making appears to hold merit, nevertheless (as cited in Fiorina, 2005, pp. 172-173). Efficiency is a product of numerous constituencies within a legislative district building winning coalitions with the political party. The effect of party is a good measure of preferences. Partisan influences shape a legislator?s decision making and bring efficiency into question. Efficiency in Government The concept of efficiency has a long lineage when applied to government activity. While this dissertation is concerned with public policy formulation maximizing net social benefits available to a constituency, considering other definitions is a necessary basis for this argument. Efficiency until the mid twentieth century focused almost exclusively on process and technique to find a one best way, but has evolved since that time to include rational, self-interests of individuals. According to Rainey and 9 Dimock, definitions of efficiency vary from lowest cost production and allocation to larger ideological roles of what can be accomplished through ideological positions of social initiatives and reforms (cited in White, 1999, p. 8). Ideological roles of efficiency are an illustration of the importance of values in decision-making. ?Max Weber, Frederick Taylor, Henri Fayol, Frank Goodnow, and Woodrow Wilson are among those reinforcing the idea of efficiency as a central instrumental value for the public administrator? (White, 1999, p. 8). Definitions of efficiency can be grouped into four categories. Each definition overlaps to some extent, but a theme appears to be that individual choices play an important role in the use of efficiency to explain a phenomenon. Ideologies and self- interests shape individual behavior, but efficiency can have a role in the process of not only the task but also the outcome. That is, efficiency represents more than a tool of administration and encompasses ideologies and values. Efficient policy outcomes that maximize social benefits are based on the values associated with this end. Definitions of efficiency are as follows. 1. Efficiency is an administrative tool within a bureaucratic structure concerned with best process or procedure 2. Efficiency is an ideology where values are not separate from process and motivation 3. Efficiency is a rational pursuit of goals by self-interested individuals to maximize marginal utility of the choices available to those individuals 4. Efficiency can be expressed as economic efficiency or a maximization of net social benefits from public policy formulation and implementation 10 Efficiency as an Administrative Tool Efficiency in the early 20 th century considered procedure and the best way of accomplishing a task. Taylor and Fayol define efficiency through general administrative techniques in the private sector that can be applied to government. To Taylor, efficiency is a method by which the duty of management structures a production process that maximizes interaction between worker and environment. Fayol argues that techniques to improve administration developed as a result of the control needed to produce efficient results (as cited Parsons, 1995, pp. 313-314). Taylor places the responsibility to devise an efficient work environment on the shoulders of management. He argues that misguided attempts by management to structure work environments force workers to ?soldier? to produce an output of goods (Fry, 1989, p. 52). To soldier refers to the relationship between a worker and the arrangement of tasks necessary to perform a job. Management does not devise a best method of production and workers must perform their jobs without benefit of management direction. Taylor argues that management?s failure to design workplace performance is a primary cause of inefficiency. Organizations should emulate machines, where small tasks are structured and individuals are motivated (as cited in Parsons, 1995, p. 313). The manner by which workers perform jobs and incentives tied to wages to motivate workers are integral factors that define efficiency. From the study of these relationships between worker and the production process, efficiency becomes a tool of administration. Management controls the production process to devise the best method that maximizes output. Mosher finds that this era embraced efficiency as a practical and intellectual concern guided by principles 11 (as cited in White, 1999, p. 10). Decisions are made without concern for organizational or individual values that shape behavior. Associations between workers and management follow strict adherence to procedure for doing a job with management setting the standards that workers follow. Bureaucratic Efficiency The organizational structure of a workplace is an important element in creating the controls necessary for managing workers and enhancing production. Bureaucratic organizations are a frequent unit of analysis to illustrate these effects. Bureaucracy is a hierarchical organizational structure governed by rules and well-defined relationships within the organization. ?Weber defined bureaucracy as the most rational and efficient organization devised by man? (White, 1999, p. 9), where the virtues of efficiency exist within unity and conformity of the bureaucratic structure. As an administrative tool, bureaucratic efficiency considers maximization of effort in producing an output. Weber correlates the growth of an industrialized society with a search for more rational forms of organization within business, industry, and government (as cited in Parsons, 1995, p. 17). Control of individuals through authority is a primary component in Weber?s argument for a bureaucratic structure. As cited in Parsons (1995), Weber identifies the most rational form of bureaucracy as an ?ideal-type? (p. 31) bureaucracy characterized by the ?systematic and deliberate adjustment of economic means? (p. 24) through a price system. The primary effect of capitalism is its ability to disseminate the pursuit of rationality manifested in a bureaucracy. Laws associated with legal precedent and rules are pre-eminent in a capitalist society and provide a stable basis for decision-making 12 (Parsons, 1995, p. 272). The stability and reliability of a bureaucracy offer a foundation upon which rationality produces an efficient output. Bureaucratic efficiency is a function of rules that govern behavior and controls placed on individuals within the bureaucratic structure. Precision and continuity of a bureaucracy correlate with market ethics, and the introduction of rationality is paramount to capitalism. Taylor argues that improvements in efficiency are actually a means to social reforms (Fry, 1989, pp. 47-48), evidenced by changes in the theoretical role of efficiency that paralleled social legislation in the 1930s (White, 1999, p. 9). One ?must consider the extent to which a decision has been the outcome of rational processes? (Parsons, 1995, p. 273) to prevent bureaucracy from exceeding its functions. The virtues of a well-defined organizational structure producing efficiently emanate from a legal basis and rationality defines individual actions shaped by self-interests. ?Bureaucracy is the means for achieving rationally ordered societal action? (Fry, 1989, p. 37), producing outcomes that are not happenstance but predictable and unambiguous. Weber considers bureaucracy and capitalism to be mutually supportive social structures (as cited in Fry, 1989, p. 33). His point is that laws and controls are necessary components to structure relationships among individuals and create predictable market exchanges. Controlling the actions of others is difficult when rationality is introduced. An argument can be made that bureaucratic rules that characterize the modern state usurp individual freedoms rather than promote efficiency. Greater specialization of activities diminishes personal freedom and individual choice within organizations. An efficient organization offers the structure necessary to expedite decision-making and 13 control individuals from the top, but values and motivations that guide individual behavior are discounted. The same rules that introduce impersonality into worker relationships form a barrier around effective communication within organizations. Robert Merton argues that a paradox exists between the impersonal rules of a bureaucratic organization and greater efficiency and rationality (as cited in Fry, 1989). Problems arise as reliability and conformity become exaggerated. Displacement of goals through adherence to formalized procedures shifts objectives to the process of conforming to rules rather than the outcome resulting from an application of the rules. Efficiency as a Value Economic despair associated with the Great Depression brought into question many of the principles that linked public policy making and economics to management techniques and individual ideology. How issues are addressed and if efficiency should be a standard that drives decision-making are brought into question. Strict definitions of efficiency as narrow processes associated with process gradually evolved to include broad associations of needs of an organization and individual. As cited in Parsons (1995, p. 314), Barnard finds that humans have a mix of emotions and differing roles and goals. Conflicts occur between individuals in an organization and the role of leadership is to promote cooperation between individuals. Managing conflict introduces the effect of values in shaping individual behavior and fostering human motivation. An examination of the ideological role of efficiency came later in the twentieth century and became part of cost-benefit analyses at federal, state, and local levels of 14 government (White, 1999). Including an ideological component for efficiency is an argument that efficiency is a value and whether one espouses efficient processes indicates that an organizational environment is not value free. Efficiency as measured through policy costs and benefits should not be confused with the classical expression of one best way and technique. Rather, efficiency in terms of cost-benefit introduces an ideological correlation with efficiency to include maximization of pecuniary, aggregate benefits. Weimer and Vining (2005, p. 338) explain that monetization allows for analysis into common units of currency (dollars) for evaluation of public policy effects through positive or negative efficiency impacts. Efficiency is thus not merely an administrative process but also a component of expanding policy goals and setting policy objectives. In government choosing policy positions based on measures of costs and benefits is an argument that a legislator?s values guide the decision making process. Simon argues that ideology may be separated from efficiency through values to illustrate the impact. According to Simon, two kinds of science should be developed: ?a practical science with the objective of developing more efficient administrative procedures, and a pure science, which is to examine the basic processes of human behavior as they relate to decision making? (as cited in Fry, 1989, p. 15). With values the emphasis on the organizational environment is gradually replaced by individual decisions (Fry, 1989, p. 184). Individual decisions consist of data input along decision premises that are pieces of a decision made at different points in time ?and involve the process of alerting, exploring, and analyzing, which precede the act of choice? (p. 185). 15 Using a fact-value dichotomy Simon argues that values are an essential criterion in decision-making. Efficiency is represented by facts as a primary measurement tool; that is efficiency is presented as a linear process to produce the best output with the least effort. Values introduce qualitative, normative issues into management theory that was heretofore unscientific (White, 1999, p. 15). Values and ideologies strongly correlate. As individuals make decisions, each decision consists of numerous decision premises where values shape each decision and provide direction for the next decision. Efficiency is a process, while values are the basis for policy goals. Efficient decisions might produce results with the least effort expended, but ?(v)alue premises are ethical statements about what should be done? (Fry, 1989, p. 186). Self-Interest Contrasted to Equity Contrasting self-interest and equity is important to make a point that efficiency and equity are both values, but with important differences (LeGrand, 1990). Efficiency as a process correlates with rationality (self-interest) in describing legislative decision- making. Equity is a policy objective often cited in the economics literature and in welfare economics in particular (Hammond, 1976; Boadway, 1976). Equity is typically described through fairness and justness, and exhibits many of the characteristics of democracy. As an equitable policy goal seeks to offer policy solutions with distributional aspects, the values of the legislator form a basis for how the legislator envisions the purpose of government and the role of lawmaking to achieve just ends. Contrasting policy decisions through efficient versus equitable objectives focuses on 16 differing outcomes. A rational legislator chooses policies based on calculating self- interests. Rationality and the correlation of rationality to self-interest are factors in a legislators? support of policies. According to Dimock, an efficient vote in support of a policy must address the ?relation between what is accomplished and what might be accomplished? (as cited in White, 1999, p. 8). Standards of efficiency often include speed, safest, least expensive and numerous others. Rainey finds that efficiency is a function of resources expended and the amount of work produced (as cited in White, 1999, p. 8) and contrasts with equity. Policy decisions that are equitable describe a distribution of goods and services across a spectrum. Seidman makes a point that policy tradeoffs with equity exist in studies of efficiency (as cited in Scholz & Wood, 1999). Although not necessarily mutually exclusive, embracing either efficiency or equity as a policy objective produces differing policy outcomes. Values are a component of choosing either objective. Synergies between efficiency and equity are possible such that economic growth, for example, does not necessarily increase inequalities. ?An improvement in efficiency can accompany more equal distributions of wealth, due to its felicitous effects on effort and educational investment choices? (Putterman, Roemer, & Silvestre, 1998, pp. 866-867). Mintzberg argues that efficiency is heavily dependent on the measurable aspects of benefit, cost, and process (as cited in White, 1999, p. 19). Equity is concerned with distributive effects of policies. 17 Rational Actor Model Individuals are guided by self-interests. Rational decision-making is a product of individual self-interests. Formal structures provide rules necessary for execution of tasks in the workplace. Individual values are a component of interrelationships within organizations that are both formal and informal. Applying efficiency through individual self-interests considers how rules that guide tasks affect how one performs those tasks. Human behavior is a function of motivation and relationships between superiors and subordinates in organizations While the ideological role of efficiency has become an impetus for benefit-cost analysis in government, the rational actor model offers clear justification for self-interest as a motive for behavior. The model is a synonym for public choice economics. Market economies are the logic behind public choice theory. According to public choice theory a rational individual pursues goals in the most efficient means possible. These goals are self-interests (Lindblom, 1959; Downs, 1957, pp. 3-4). Individuals are actors in society and have consistent preferences that guide behavior in the pursuit of self-interested goals. When an individual has a choice among alternatives, he or she will choose that alternative which yields the highest expected marginal utility; that is a rational person will make a conscious choice to choose a goal that maximizes the utility of choices that are available (Monroe, 1991). The public choice school is concerned with ?how could this self interest be constrained and directed to more efficient and effective choices in the interests of taxpayers rather than to ever-bigger budgets and more government? (Parsons, 1995, p. 313). Individual choices that maximize self-interests of the decision maker may not 18 maximize budgetary resources. Considerations of selfishness or selflessness (Arrow, 1963) are part of the answer but do not fully explain rationality. Downs argues that motivations of individual officials in a bureaucracy are diverse and produce a typology of bureaucrat (as cited in Parsons, 1995, p. 309). Downs labels individual officials across a life cycle spectrum from pure self-interest to a sense of public interest. Bureaus experience a similar life cycle of growth but maximize self-interest by growing bigger. The self-interests that motivate an official are a function of the role of the official in the organization. The official may be motivated by power, prestige and money or by a commitment to work performance and loyalty to the larger organization. Maximizing resources flowing to the bureaucracy and increasing the power of the bureaucrat to manage those resources illustrate motivating factors of those officials that are at a higher stage in the bureaucratic life cycle (as cited in Parson, 1995, pp. 309-310). Simon, Thompson, and Smithburg argue that efficiency and rationality are synonymous (as cited in White, 1999, p. 9). Applying the effect of the relationship between an individual and society shows the influence of self-interests and ideology that motivate each individual. Fulfilling individual motives is a prerequisite to achieving organizational efficiency (p. 11). Simon states: ?Although each individual seeks efficiency in terms of his or her values, those values should be the product of an organizational role, and, to the extent that the organization has been successful in establishing objectives, efficiency will be measured to the goals of the management group? (as cited in Fry, 1999, pp. 14-15). A management group is made up of individuals with self-interests who make decisions consistent with the values and interests of the organization. 19 Differences in motivations produce differences in the bureaucrat?s commitment to personal and organizational goals and commitment to serving his or her role in a bureaucracy. Cooperative relationships between organizations and individuals are necessary to maintain equilibrium between the organization and the individual within the organization (White, 1999, p. 11). Barnard says a relationship can be informal or formal (as cited in Fry, 1989, pp. 161-164) or, as Follett argues, a ?reciprocal relationship between the individual and society in which the individual both shapes society and is shaped by it? (p. 101). Conflict between individual self-interest and the organizational interests of the bureaucracy to grow bigger is a concern. Supervision of bureaucrats is necessary to control self-interests if larger public interests are threatened (Parsons, 1995, p. 310). Public interests are those interests served by the bureau and may not be analogous to each bureaucrat?s self-interest. Rationality within an organizational setting is one example of individual motivation and decision-making. Capture is said to occur when bureaucrats are supposed to be acting in the public?s interest fail to comply and rather act in their own self-interest (Stigler, 1971). For the purposes of this dissertation self-interests that influence legislators provide insight into legislative decision-making. Roll Call Voting In the legislative arena roll call voting is an illustration of self-interests of a legislator. The self-interests of an individual are often expressed in economic terms as a derivation of choice. Rationality is a key component in analyzing roll call voting in Congress. Self-interest and rationality are tied to efficiency in explaining legislator 20 behavior when casting roll call votes. A roll call vote guarantees that every member?s vote is recorded. According to Krehbeil, roll call voting provides a forum for considerations of why a legislator is in agreement or disagreement with policy proposals (as cited in Snyder & Groseclose, 2000, p. 193). To the extent that factors affect a legislator?s behavior, efficiency of Congress is impacted. A roll call vote represents a record of a legislator?s behavior at any point in the legislative process. The importance of recording the vote is particularly significant when that vote is a record of that legislators? action on a particular policy proposal. Rational legislators are self-interested legislators. Support or lack of support for a public policy could impact reelection of the legislator, arguably the strongest self-interest indicator (Downs, 1957); congressional district or state self-interest and those of the legislator; and relationships between constituent characteristics and decision making by a legislator. A constituency has a specific set of concerns that a legislator must address as an agent of those constituents (Moe, 1984). Supporting public policies that increase a legislator?s personal wealth is another factor that affects his or her behavior (Caro, 2002). From the self-interest perspective of rationality, support for policies that are in the public?s interest are functions of the economic interests of a constituency and ideological predilections of a legislator (Kalt & Zupan, 1984, p. 280). Economic Efficiency as Social Benefits Maximization Economic efficiency as a maximization of net social benefits differs from efficiency as a process in reaching public policy objectives. While efficiency as a process considers rational self-interests, economic efficiency is a component of ideology 21 that addresses the aggregate effects of decision-making. In short, economic efficiency is maximization of aggregate social benefits of policy decisions to aggregate social costs. Efficiency as a motivator of self-interest is not concerned with the societal effect of policy making, but rather the steps in the process of policy making. This is an important distinction between an application of efficiency as self-interest and economic efficiency tied to maximization of social benefits. Efficiency of self-interest focuses on individual human motivation as a component of a decision rather than outcome of a decision. Human motivation ties to self-interest efficiency and outcomes tie to efficiency as social benefits maximization. Waldo contends that efficient decision-making is a process that maximizes resources and ?leads to a responsive and responsible government better able to serve the needs of the people? (as cited in White, 1999, p. 11). Waldo?s argument is that maximizing resources through emphasis on efficient techniques correlates with maximizing social benefits. Decisions are a function of the organizational environment and individual motives that guide decision-making. White?s (1999, p. 18) argument that ?public management is the study of how to manage government, whereas public policy focuses on why? further differentiates efficiency as a process governed by individual self-interests from economic efficiency that focuses on the how and why of public policy initiatives. Economic efficiency strives for definite policy outcomes or ends. According to Harmon and Mayer, efficiency in government decision processes is ?relevant only when the ends of action (social benefit maximization, for example) are known in advance? (as cited in White, 1999, p. 22). Rational individuals make decisions by ?judging how quickly and how cheaply a 22 particular end is achieved, but it cannot decide what the end should be? (White, 1999, p. 22). Maximization of net social benefits is a policy end that an economically efficient decision maker seeks. Production Aspects of Economic Efficiency Economic efficiency as a concept used in the economics literature appears to transcend traditionally accepted definitions of ideology and self-interest as utilized in the legislative voting literature, but it does not replace those factors as determinants of human behavior. Economic efficiency in the broader, social benefits context of its definition is best addressed through production and allocation. Kennedy (2005, p. 45) argues this point where ?efficiency in production is a measure of the effectiveness of an input with respect to the production of some output.? Considering the production process through a chain of events is necessary for illustration. Bennett 2 offers a model for program evaluation that considers events that occur (e.g. inputs, processes, outputs, outcomes) and evaluative procedures to analyze those events (as cited in Patton, 1997, pp. 233-236). Adapting Bennett?s model, Patton (1997, pp. 233-234) shows inputs as resources expended to start an initiative; activities processed with available inputs; outputs a decision or action realized from those activities; and outcomes a measurable result or consequence of the output. For purposes of this dissertation, the input under consideration is the information available to legislators to discuss a problem area that might require a public policy. Processing the input through political debate or discussion is a step needed to produce an output. The 2 Refer to Bennett (1982, 1979). 23 roll call vote on this debate is the output, the decision or action taken. Outcomes are the consequences of the vote (Kau, Keenan, & Rubin, 1982). Depending on the policy area(s) chosen for analysis, an outcome could be a reduction in malpractice claims and awards or greater access to health care in America, for example. In economic terms, the output produced from policy formulation is a function of the inputs or resources that go into developing that output. According to Browning and Zupan (2002, p. 166), a production function can be expressed mathematically as Q = f (X, Y, Z) Where, Q is the output or policy decision that is made and combinations of X, Y, Z are factors of production or inputs that are employed through the political process to produce a public policy solution or vote. According to Wildavsky, outcomes that produce favorable policy consequences in policy area(s) are the goal of outputs, with efficiency a means of arriving at a destination with the least possible effort (as cited in White, 1999, p. 15). While maximization of voting output from given inputs is possible with any given policy decision, an economically efficient output seeks to increase social benefits relative to social costs in the aggregate. Maximizing the use of resources and disseminating services to a constituency lessens narrow perspectives of benefit-cost measures (Maass, 1966, p. 209). Consideration of aggregate consequences allows one person to consume more without affecting that available for another to consume (Kennedy, 2005, p. 46). Simon ?defines efficiency as the maximization of the ratio of net positive results 24 (positive minus negative results) to opportunity costs? (as cited in White, 1999, p. 14). More efficient positions produce greater net positive results. Efficiency and Public Policy Implications Definitions of efficiency are varied and disparate. For purposes of this research, I consider traditional definitions of efficiency as a process where speed and quickness underlie the process to maximize what is accomplished from the effort put into an endeavor. Applied to public policy formulation, this traditional definition of efficiency considers the process of lawmaking through which legislators? ideologies and self- interests correspond to those ideologies and self-interests of his or her constituency to achieve policy outcomes. Economic interests drive legislative voting behavior to maximize the probability of re-election (Downs, 1957) and can explain individual voting behavior (Silberman & Durden, 1976). Pool and Rosenthal (1997), on the other hand, embrace ideology as a determinant of voting behavior that is more important than constituent economic interests. The less divergent a legislator?s ideology and self- interest is from the ideology and self-interest of his or her constituency, the greater the efficiency of the relationship between the legislator acting as agent and the policy objectives of such constituency. Explaining legislative behavior in efficient terms is not stating that efficiency is a means to an end but that representatives in government embrace efficiency through values. White (1999, p. 16) states: ?Although efficiency itself is not a value, efficiency is only useful within a framework of consciously held values.? That Selznick finds efficiency paradigms overemphasizing a means of operation rather than a value-laden 25 end supports White?s argument (as cited in White, 1999, pp. 16-17). Many of the basic principles of democracy are dependent upon moral behavior that does not quite fit with technical efficiency for the sake of efficiency (see Waldo, 1947, 1965, 1980). For a legislator, policy action incorporates values into the equation that drives support or rejection of policy initiatives. Achieving an efficient end for the sake of efficiency should not be a legislative goal. Costs associated with a legislative decision, both actual costs of implementing a policy and opportunity costs of utilizing resources elsewhere do not address policy goals and outcomes in the aggregate. Individual decisions (e.g., self- interest based rational decisions) might produce aggregate consequences that may or may not be favorable to others affected by the decision. A measure of economic efficiency captures these aggregate consequences as public policy decisions are made. A measure of economic efficiency seeks to identify policy that maximizes aggregate social benefits relative to aggregate social costs. Economic efficiency should not be confused with procedural effects associated with technique. Procedural areas are characteristics of efficiency where process is important to produce the best outcome through the least effort. Economic efficiency encompasses an ideological component where values are not separate from efficiency as a technique but rather are an integral part. For policy making, economic efficiency is a product of the ideological element of individual behavior and decision-making, as a representative acts as an agent to a constituency, but the constituency is society, not the legislator?s geographic constituency. The fact that an economically efficient policy expands economic output demonstrates that values are a part of the concept. Greater economic output is positively correlated with higher standards of living and thus greater 26 aggregate social benefits, but does not specifically address distribution of resources among constituents or representation in government, for example. Support for policies that do not maintain or increase economic output might not be economically efficient. An economically efficient legislator supports higher economic output to maximize aggregate social welfare as opposed to situational aspects of support for individual public policies. Stigler (1971) argues that economic interests are a key component of why legislators sometimes adopt inefficient policies. The economic interests of a constituency must be considered with the self-interests of a legislator and the public interest attributes of policy decisions. Kennedy notes: ?The expectation is that economic interests play a greater explanatory role as the injurious nature of a given policy is more obvious? (Kennedy, 2005, p. 32). Injurious policy is that policy which reduces economic efficiency and beneficial policy enhances welfare (p. 33). Maximization of social benefits might or might not be consistent with those economic interests of a constituency. While traditional definitions of efficiency consider resources and effort expended, the aggregate effects of economic efficiency consider those efforts through benefits as well as costs. Economic Efficiency as a Research Area Economic efficiency is an objective criterion, unlike a liberal-conservative spectrum of ideology that is more subjective (Kennedy, 2005, p. 65). Ideology and self- interest identify characteristics of a legislator (and possibly his or her constituency) but do not address the impact of a voting decision in the formulation of public policies in vote models. Introducing an economically efficient component to vote models alleviates 27 some of the concern with traditional measures of ideology such as Americans for Democratic Action (ADA) and American Conservative Union (ACU) scores. Economic efficiency should not be considered as a replacement for ideology and self-interest within vote models. Rather, the question is: Does economic efficiency offer a preferable measure of behavior in some situations? Does the variable represent an area of commonality that crosses ideology? Perhaps the greatest contribution a measure of economic efficiency might make to the literature on legislative voting is determining whether traditional measures of ideology such as the ADA or ACU scores are a proxy for economic efficiency. That is, is conservative-liberal ideology really masking elements of economic efficiency? Adding economic efficiency to liberal-conservative ideology allows for measurement of voting decisions that not only follow a traditional liberal-conservative spectrum, but also maximize aggregate benefits of the voting decision. The implication for policy making is extensive. Aggregation of decision-making is important to avoid unnecessary costs of policy development, but also considers social benefit of policy formulation. Not only does ideological positioning along a liberal-conservative spectrum offer insight into a legislator?s inclination toward general policy positions, but it may also include the social welfare that such policies produce through economic efficiency. Kennedy E-score Model Kennedy (2005) measures the extent that legislators consider economic efficiency in policy decisions through the compilation of an economic efficiency score (E-score). E-score is a measurement of votes made by a legislator in Congress where it 28 is possible to identify and measure votes that consider economic efficiency. The analysis that Kennedy considers includes roll call votes in the House and Senate in the 106 th and 107 th Congresses. Through the generation of an E-score it is possible to identify those legislators that embrace economic efficiency. Analogous to ADA and ACU scores for measuring liberal-conservative ideology, E-scores measure economic efficiency. The model for computation of the E-score is: N E-score = ? (P i / N) ? 1 i=1 Where, Pi = one if legislator voted in support of enhancing efficiency and zero otherwise N = number of votes considered in the analysis of each legislator As a method of measuring economic efficiency, E-score captures aggregate effects of voting behavior. This fact is an important distinction to separate the effect of economic efficiency as a policy objective from traditional measures of efficiency identified through individual self-interest. E-score ties more closely with ideology as the measure considers those policies where social welfare implications are at stake, a value that is reflective of policy goals. Efficiency as self-interest, conversely, is not a measure of aggregate social benefit of a policy. Rational actors pursue policies that produce the most benefit with the least effort, which reflect individual characteristics and efforts of the individual making the decision and not collective effect of that decision on social welfare. 29 Research Question The major research question that this dissertation addresses is: In some instances does economic efficiency through an E-score function better than a traditional spectrum of liberal-conservative ideology in explaining the ideological position of a representative (House and / or Senate member), congressional activity, and public policy formulation? The analysis modifies and extends the existing Kennedy (2005) model beyond the two Congresses that he studied. The dissertation includes the 99th Congress through the 108th Congress, inclusive. The discussion that follows outlines the components of economic efficiency to explain not only the concept, but also the criteria that must be applied to identify the votes necessary to develop an E-score for this dissertation. Pareto Optimality Public policy making that is economically efficient is consistent with Pareto improving positions. Weimer and Vining (2005, pp. 55-56) define Pareto optimality (see Figure 1.1) as a distribution of goods and/or services where it is not possible to make someone better off without making someone else worse off. 3 A Kaldor-Hicks 4 improvement is closely related to Pareto optimality. A Kaldor-Hicks improvement is indication of a policy that produces winners and losers and is consistent with Pareto principles as long as a net gain in welfare occurs. A net gain in welfare implies that 3 Named for Italian economist Vilfredo Pareto (1848-1923), the developer of criterion for such distributions. 4 Pareto improving positions are frequently referred to as Kaldor-Hicks improvements. This criterion is a description where Pareto optimality is not achieved, but potential Pareto improvements are possible. those who benefit from such policies can potentially compensate those that do not to the extent that a Pareto improvement is possible (Kennedy, 2005, p. 53). Economic efficiency is not a synonym for Pareto optimality, however. Public policy formulation that seeks to expand national output and thus increase the size of economic resources available to all may or may not produce a distribution that makes one individual better off without making someone else worse off. Higher output and standard of living in the aggregate is analogous to a distribution that yields higher social benefits as opposed to social costs. A socially responsible legislator that espouses economically efficient policy making seeks to avoid the costs associated with higher taxes and more regulation in an attempt to expand output and positive net benefit. Figure 1 illustrates production choices possible between two goods, Y and X, such that consumption within the given range is an efficient distribution. Figure 1.1: Production Possibilities Frontier, where movement from point B to point A or point C, or any combination of movements within the parameter labeled E, represent an increase in efficiency. 30 31 Criteria for Economically Efficient Public Policies Public policies that are economically efficient must be considered within the aggregate effects of those policies. For a legislator to select policies that are economically efficient does not imply that his or her preference for equity, for example, is mutually exclusive with efficiency. For a legislator to support economically efficient policies in the aggregate, making benefit-cost comparisons is paramount to such decisions. How a legislator weighs the benefits and costs between individual issues is a first step in analyzing aggregate consequences. The first step in developing criteria to identify economic efficiency is through an analysis of those policies that either enhance efficient policy output or are injurious to social welfare. An efficient policy output is a function of the social benefit produced by that policy. Policies that are injurious to social welfare produce higher social costs relative to the benefit achieved. By maximizing positive policy outputs and minimizing policies that produce higher social costs, maximization of net social benefit is possible. Legislators vote for policies that either increase efficiency or block proposals that reduce net welfare. Either scenario represents a potential Pareto improvement and allows for selection of votes where efficiency is clearly at stake (Kennedy, 2005, p. 56). Stigler (1971) analyzes categories of policies that tend to signal efficiency reduction. Each category involves regulation in private markets or direct intervention where a market failure does not exist (Kennedy, 2005, p. 56). Four categories that Stigler (1971) and Kennedy (2005) analyze follow. Each represents a decision rule that will be followed and criterion for vote selection in this dissertation where a reduction in efficiency would result. 1. Excise or direct monetary subsidies lead to a misallocation of resources. Too many resources are allocated to the production of a good or service, where the marginal cost (MC) of the last unit produced exceeds the marginal benefit (MB) measured by what consumers would pay. As a subsidy increasingly allocates resources to a good or service, consumer surplus increases, but there is a net loss of efficiency from the allocation. Consumer surplus is difference in the price the consumer is willing to pay and the actual price the consumer will pay. Figure 1.2 illustrates deadweight loss represented by points XYZ, or loss in well being associated with an excise subsidy that lowers price but increases cost to government. Price supports to the farming industry are widely argued to be a method of increasing a farmer?s income in lieu of few positive externalities associated with increased farm output (Acemoglu & Robinson, 2001). Figure 1.2 illustrates the effect of an excise subsidy on reducing price from P to P?, but in turn generating a deadweight loss, a loss in well-being as a result of the subsidy is represented by area XYZ (adapted in part from Browning & Zupan, 2002). 32 33 2. Limits to competition push price upward and reduce consumer surplus. When industries seek to limit the addition of new, rival firms, producer surplus to the protected firms increase but there is a net loss of efficiency. Producer surplus is the amount that a producer receives for a good or service that exceeds the price the producer would be willing to accept for that good or service. Examples of limits to competition that Stigler (1971) uses are Federal Deposit Insurance Corporation (FDIC) regulation, which reduces new entry of potential commercial banks into the banking industry, and the inverse relationship between increasing demands for hauling and the number of trucks that can enter the industry, as a result of limits to interstate motor carriers. Voting to regulate hospital payments or competition between providers of health care services is inefficient (Oliver, 1991). Kennedy (2005, p. 58) adds that import or production quotas, as well as protective tariffs all are means of reducing competition. Figure 1.3 depicts how limits on market competition (quotas and tariffs are two examples) affect economic efficiency by disturbing market equilibrium such that new equilibrium levels occur at a higher price level (P?) and a lower quantity of output (Q?). Points ABC represent a net loss in efficiency. Figure 1.3: Barriers to entry of markets limit competition. With fewer firms in the market supply is reduced and market equilibrium occurs at a higher price and lower level of output. The producer surplus afforded to protected firms increases to price level P??, but a net loss in efficiency is represented by area ABC (Kennedy, 2005, p. 58). 3. Policies that affect substitutes and complements are inefficient. Special interests demand policies that support products and services related to the industry of that special interest. Labor unions oppose technology that reduces the need for labor. Rail and air providers of goods support regulations on the trucking and hauling industry. Manufacturers of emission control and related devices support regulations on automobile manufacturers requiring such devices. Public transit subsidies are another example of this effect. Allocating resources to public transit has not countered the effects of increased automobile ownership and commuting time (Wachs, 1989). Each scenario is an example of public policies that affect consumption of goods that are substitutes or complements. Legislative support for policies that affect consumption of goods or services that are substitutes or complements is an inefficient act. 34 35 4. Wage and price controls are inefficient. Control over wage and price levels is best illustrated through a floor under which wage or price cannot fall or a ceiling that wage or price cannot exceed. Producers favor price floors to protect the price of their produce or services but a price ceiling on inputs that the producer must utilize in the production process. Kennedy (2005, p. 59) argues that price controls through price floors allow producers an opportunity to enjoy higher than market prices, but contribute to surpluses from items that are not sold. A price ceiling limits the extent that prices can increase and leads to shortages as consumers demand more product or service at prices that are less than market equilibrium. Deregulation of energy, financial services, and communication are examples of efficiency enhancing policies (Winston, 1993). Figure 1.4 illustrates the effect of wage and price controls that disturb market equilibrium through the effect of a price floor (PF) or price ceiling (PC). At price PF quantity demanded of Q?? exceeds supplied of Q? and shortages result. At a price of PC, quantity supplied of Q?? exceeds quantity demanded of Q? and an overabundance results. A price floor increases consumer surplus and a price ceiling increases producer surplus, but each scenario results in a net loss in efficiency, as the market is not able to clear. Policies that place controls on wages and prices are economically inefficient. Figure 1.4: Wage and price controls create price ceilings (PC) and price floors (PF), where market prices are either above or below, respectively, market equilibrium, resulting in a net loss in efficiency clear (adapted in part from Browning & Zupan, 2002, p. 28). Developing an E-score Using the decision rules above, roll call votes within the House and Senate for all Congresses between the 99 th and 108 th Congress, inclusive, were examined. Only those votes where efficiency is clearly at stake were chosen, and how the legislator voted on each selected roll call vote was recorded. Scoring models were developed for each Congress, House and Senate, which measured the legislator?s support for economic efficiency. For each Congress, House and Senate, legislation was chosen where efficiency was clearly at stake. That is, the legislation would potentially produce either positive or negative results through greater social benefits or higher social costs, respectively. From the total legislation analyzed the legislator?s votes in the House and Senate for efficiency enhancing policies were tallied as a percent of the total votes in 36 37 each chamber analyzed. This process for developing the E-score was repeated for each Congress in the study. Kennedy (2005) argues that votes on amendments offer the best opportunity for analytical precision. In contrast, bills ?often include many provisions making definitive judgments with respect to their impact on efficiency problematic? (p. 60). Votes where efficiency is ambiguous are omitted; otherwise votes where the impact on efficiency is ambiguous, subjective consideration for the major intent is necessary if that vote is recorded within the E-score. A vote is ambiguous if determining the impact on efficiency is impossible or unlikely to clearly delineate between increases or reductions in efficiency. Analyzing a voting decision is required to determine if the vote enhances efficiency. A vote in support of a policy that appears to reflect both efficient and inefficient positions must be considered by impact of the total effect of that position. Total effects of a voting decision are a consideration of the impact of a vote on not only the policy area of the vote, but also unintended consequences, such as higher internal costs or misallocation of resources in other policy areas. Votes that increase economic efficiency in one area but decrease it in another must be judged by the net effect of the vote. The net effect of a vote includes the sum of all positive effects (efficiency enhancing) and the sum of all negative effects (efficiency decreasing) of the voting decision to equal total effects. Votes that have multiple components must be judged by the total effect of the vote for a policy to the outcome of that vote on economic efficiency. Votes to address market failures are an example of an ambiguous policy. While policy solutions to correct an externality often appear to be inefficient, differences 38 in judgment are a problem as legislators make rough cost-benefit assessments of formulating a policy to correct the market failure, but introduce other costs or unintended consequences that diminish efficiency (Kennedy, 2005, p. 61). E-scores are derived for each legislator using Kennedy?s formula discussed earlier. The E-score is applied as an independent variable to measure whether it or ADA and ACU scores are better predictors of the vote in the policy areas identified. By employing E-scores, the study measures the explanatory power of economic efficiency vis-?-vis self-interest, party unity and other variables traditionally used with ideology to predict votes. Recalculating Interest Group Ratings Roll call votes that are tabulated for use in scoring models for legislative support of policy areas (dependent variable) are not included in roll call votes that are tabulated to devise interest group ratings (e.g. ADA, ACU) used for independent variables. The same holds true for E-Scores. Wattier and Tatalovich?s (2005) model is utilized to address this issue. According to their model, for any interest group rating, votes tallied within such ratings that are also tallied within one or more of the policy areas are removed and the rating recalculated. For example, if 20 votes were considered to yield an ADA rating and one of those votes also represented a policy area such as medical malpractice used as a dependent variable, that vote would be removed from the ADA rating and the rating recalculated based on 19 votes rather than 20. 39 Standardizing interest group scores An examination of votes involves a quantifiable and systematic means of data analysis. A problem with comparing votes over time is that each vote is time-bound (Shipan & Lowry, 2001, p. 247). For example, a vote for specific legislation in 1985 may not be readily compatible to a vote for a separate bill in 2004. Groseclose, Levitt, and Snyder (1999) also believe that raw scores are not comparable when considered outside of the immediate time period in which they are tabulated and need to be adjusted. They argue that parties have diverged, as evidenced by increasing polarization, and interest group scores are not accurate when considered over a period of time. Their analysis solution is similar to basing a price index on some arbitrary year (1985=100), but involves ?shift? and ?stretch? parameters that are utilized in adjusting the scores. Poole and Rosenthal (2001) developed a technique for measuring legislative liberalism and conservatism through a process called DW Nominate. Their model adjusts the effect of time on scores through necessary weighting within each score. An improvement of the Poole and Rosenthal model over the Groseclose, Levitt, and Snyder model for adjusting interest group scores lies in the fact that the former continually adjusts the scores for liberalism and conservatism over time as additional votes are cast. Thus, the Poole and Rosenthal approach constantly changes the relative position of the legislator within each Congress vis-?-vis other legislators. Poole and Rosenthal?s approach for capturing the time aspects of liberalism and conservatism was employed in this study. Other interest group ratings (e.g., ADA, ACU scores), E-scores and other time impacted variables are computed nominally as raw scores but standardized to make 40 comparable across the years of the study. To standardize these measures the value of each variable was computed for each legislator and the mean value for the House and Senate chambers in each year of the study. The computed values for each legislator are compared to mean values of that variable for each Congress. Analyzing values for each legislator to mean values for Congress offers a relative comparison of each legislator to respective scores for each chamber. Changes in the differences between the computed value for the legislator and mean value for Congress are indicative of changes in legislative behavior for that legislator across Congresses. Measuring differences between computed scores for each legislator and median scores for the chamber standardizes the analysis and alleviates issues associated with accepting nominal scores that can be impacted by time. Sample Units of analysis for this study are legislators in the U.S. House of Representatives and U.S. Senators. The sample includes all legislators of each house for the 99 th ? 108 th Congress, inclusive. Legislative voting was analyzed for all roll call votes supporting or opposing the bills identified and included in each policy area. The votes that are gathered on the policy area are the dependent variable. Comparisons were made between those legislators voting on all legislation included in each policy area and those legislators of the overall Congress in analyzing the impact of economic efficiency. 41 Regression Model The regression model for this dissertation employs the following methodologies: multiple regression analyses of the effect of various predictor (independent) variables on several dependent variables and time series analyses. Regression analysis measures the direct effect of each independent variable on the dependent variable. The association between two variables in a sample might or might not exist in the entire population. Measured by an F-test for the entire model and a t-test for each sample, tests of significance indicate how likely such association exists. Each variable in the model has a predicted association (+ or -) between independent and dependent, allowing use of a one-tailed test. Statistical significance was determined at the 0.05 level. Multiple regression analysis tests for direct associations between variables. To test for indirect effects, several regressions were run between the variables in the model. Of particular interest is the relationship between E-score and other independent variables and the extent that E-scores appear to transcend liberal-conservative ideology. Time series analysis assumes that successive values in the dataset represent consecutive measurements taken at equally spaced intervals of time. Interrupted time series considers whether an outside event affects successive observations. With political party control of Congress and White House changing on several occasions in the span of this study, interrupted time series analysis is utilized as a methodology. For example, could a change in political party control of the institution affect economically efficient voting behavior? 42 The regression model for this study includes the following: 1. Dependent variable: two distinct policy areas are studied ? medical malpractice tort reform and federal minimum wage. For each policy area a regression equation was developed to measure the effect of independent variables on the dependent variable for that policy area. The dependent variable was developed from a scoring model of roll call votes of all final bills in that policy area within the 99 th -108 th Congresses, inclusive. Votes tabulated within the scoring model for the policy area are not the same data set of votes tallied as interest group ratings. The scoring model represents the percentage of roll call votes cast by that legislator in support of the policy. For example, if five roll call votes were cast for bills in the policy area and the legislator voted in support of four of the bills, the legislator?s score would be 0.80. The legislator?s score of support for the policy position will be coded as a value from 0 to 1, with higher values indicating greater support. 2. Independent variables: the model consists of three vectors representing ideology, self-interest, and legislator or legislator?s chamber. Independent variables are sorted according to one of the respective vectors. Independent variables in the ideology vector include measures of liberalism (ADA score), conservatism (ACU score), a time adjusted range of liberalism and conservatism (DW-NOMINATE), and economic efficiency (E- Scores). Measures of ideology are coded as an actual value that depicts each measure. The value for each measure is a number from 0-100 with 100 indicating total support and 0 indicating total opposition. Ideology variables are converted 43 to natural logarithms when non-converted data do not appear to follow a normal distribution. Independent variables in the self-interest vector include contributions to legislators from interest groups with ties to the policy area of each dependent variable. For the medical malpractice policy area the self-interest variables are contributions from ?Lawyers and Lobbyists? and ?Health? related groups. Self- interest variables in the minimum wage policy area include contributions from ?Business? and ?Labor? groups. Contributions can be direct or through political action committees (PACs). Self-interest variables of contributions to legislators are coded as the actual dollar amount contributed and as a percentage contributed in the policy area to the total contributions received by the legislator. Self- interest variables are converted to natural logarithms when non-converted data do not appear to follow a normal distribution. Independent variables in a vector for chamber environment include party unity and ideological divisions between legislative and executive branch. Party unity is a measure of how closely a legislator votes in accordance with his or her political party and is coded along a continuum from -100 to +100. Negative numbers arbitrarily represent Republicans and positive numbers Democrats. More extreme party unity values (scores closer to ?100 or +100) indicate greater voting unity between the legislator and political party. The variable reflects to what extent Republican or Democratic legislators support the legislation in relation to party support. 44 Ideological divisions between the legislative and executive branch capture how party control of each institution affects legislative voting. This measure considers minority-majority party relationships that affect legislative and executive decision-making. The relative legislative-executive ideological position is compared to mean values for the institution he or she represents in exploring the effects of ideological divisions. 3. Control variables included in the model hold constant the potential effects of party control of government, geographical conditions, and state economic conditions. These variables do not causally impact the dependent variable, but rather are constant variables representing extraneous factors. Each variable is coded dichotomously (0 and 1). Control variables for party control of the institution (House or Senate) are compared to control of the legislator?s party. If the party of the legislator is the same as the party that controls the institution, the variable will be assigned a 1; if the party of the legislator is not the same as the party that controls the institution the variable is assigned a 0. Independent legislators are assigned a value according to the party with which the legislator caucuses. Whether the chamber and the presidency are in the hands of the same party and whether the House and Senate are of the same party or not represent two other variables for examination. Geographical effects of North, East, South, and West are controlled by assigning a 1 to the legislators who represent a state or region in one of the four categories and 0 otherwise. For example, a legislator from Alabama is assigned a 1 for South and 0 for East, North, and West. Measurements of per capita income or percent minority (African American and 45 Hispanic as separate variables) control for economic conditions. Federal spending going to a state and the ratio of federal spending to tax revenue generated by the state are proxies for state economic conditions. The specific policy area used for the dependent variable (malpractice reform or minimum wage) also requires control variables representing conditions in the legislator?s state or district that might impact her/his roll call vote. For example, in the case of malpractice reform, whether a legislator?s state is in a malpractice crisis or not can be used to test for constituency self-interest impact on the legislator?s vote. The same holds for state minimum wage policies that equal or exceed the federal minimum wage for that policy area. The proposed multivariate regression equation used for each policy area in the model is shown as follows: VOTEit = a 0 + b 1 ECONOMIC EFFICIENCY + b 2 IDEOLOGY i + b 3 SELF INTEREST i + b 4 CHAMBERENVIRONMENT i Where, VOTEit is the dependent variable representing a scoring model of final, roll call votes by a legislator on a policy position; a 0 is a constant term; and b 1 ?.b 6 denote the regression coefficients of the independent variables. Each vector in the model is as follows: f(x,y,z); where vector x depicts ideology, vector y self-interest, and vector z party unity. For each policy area (medical malpractice and minimum wage) hypotheses are developed, with the predicted direction of the regression coefficient indicated, and analyzed through multivariate regression. 46 Interrupted time series design is utilized to study the impact of a change in political party control of both institutions (House and Senate) in 1994 and political party control of the Senate in 1986. In anticipation of changes in the relationships between independent variables and dependent variable(s) as a function of political party control of the respective institution, an interrupted time series design allows for measurement of these effects. With interrupted time series, it is possible to measure the impact of changes in political party control of the institution on the support for public policies analyzed in the policy area(s). The model also allows for separate analysis of each independent variable. Kellough (1990) offers a methodology that will be employed in this analysis. Limitations of interrupted time series in this analysis are that roll call votes on policy areas are not necessarily linear from year to year. That is, policy area legislation that is analyzed as the dependent variable is from a spectrum of years, where legislation is considered multiple times in some Congresses and rarely or none in other Congresses of the study. While selecting legislation for analysis before and after the base year of the interrupted time series analysis (e.g., 1994 or 1986) would be ideal, making such selection may not be possible; thus the ability of the design to measure changes in the effect of political party control of houses of Congress would be reduced. Cluster analysis will be employed in analyzing available roll call votes before and after base years and adjusting which base years will be included in the model. 47 Data Sources Sources of data for dependent and independent variables are listed as follows. For the dependent variable(s), the source for the roll call votes that comprise the scoring model is Congressional Quarterly Congress Collection (http://www.cq.com). American Conservative Union (ACU) ratings are available at (http://www.acuratings.org/). Economic efficiency (E-Score) ratings are developed from the E-Score formula above and included in the analysis. Americans for Democratic Action (ADA) ratings are available from http://www.adaction.org/votingrecords.htm. Contributions made to legislators directly or via PACs are available for all Congresses in the study from Federal Elections Commission (http://www.fec.gov/finance/disclosure/ftpdet.shtml), Center for Responsive Politics (http://www.opensecrets.org/politicians/index.asp) for the102 nd ? 108 th Congress, and Political Money Line from Congressional Quarterly at http://www.tray.com/cgi-win/x_pac_init.exe?DoFn= for the 99 th ? 101st Congress. Measures of party unity are available from Congressional Quarterly Congress Collection (http://www.cq.com). Macroeconomic variables including trends in federal spending across states are available from the Northeast Midwest Institute (http://www.nemw.org/) and Tax Foundation (http://www.taxfoundation.org/). Contribution of the Study Considerable research has been conducted on ideology and rational self-interest to explain human behavior (see Downs, 1957; Arrow, 1963; Kalt & Zupan, 1994). Previous research focuses to a large extent on liberal-conservative issues to explain ideology and the rational actor model to explain individual self-interest. While both 48 areas of research offer insight into individual behavior and consequently legislative decision-making, analyzing the impact of public policies as social benefit maximization shifts the focus of the research from individual characteristics of decision-makers to policy outcomes. Kennedy?s (2005) research offers an introduction to economic efficiency as a research topic in studying the 106 th and 107 th Congresses. Kennedy?s research provides a general basis for defining economically efficient policies through development of an E- score, but is limited in explaining to what extent economic efficiency relates to or transcends ideology and self-interest in the policy process. A contribution of this dissertation is the extension of the E-score model beyond the 106 th and 107 th Congresses to encompass a period of 20 years. By analyzing the 99 th through 108 th Congresses, inclusive, the model takes into consideration issues concerning party control and the effect of divided government both within Congress and between the legislative and executive branch during the years of the study. The effect of political party control over the institution in question and economic efficiency associated with the voting of legislators raises important questions concerning liberal-conservative ideology and economic efficiency as a predictors of individual behavior. Introducing an economic efficiency variable separates the effect of the legislator?s vote to enhance aggregate social benefits or diminish welfare from party influence and ideology associated with a liberal-conservative scale. Explaining congressional activity and policy formulation through an E-score might be preferable to other measures of ideology or self-interest. 49 Divided government has been argued as the root of inefficiency within our democracy (Thurber, 1991). Balancing power among separate institutions (Fisher, 1998) is a hallmark of the American political system. Institutions of government fulfill formal and informal roles and have responsibilities to constituents. Considering institutional differences between the House and Senate and applying economic efficiency as a variable that explains individual behavior, vote decisions at different levels of government are explained in part through national or regional responsibilities of each legislator (Stein, 1990). Frymer (1994) argues that divided government is a product of balancing of power and is not a major factor in legislative indecisions. Ideological consistency across districts produces unified representation even if party control between the legislative and executive branches is divided. Traditional measures of liberal-conservative ideology should offer an explanation for behavior, but they do not address aggregate net social benefits of a policy. The role of economic efficiency under unified and divided government is an argument for inclusion of an E-score to explain voting behavior that considers the impact of the public policy decision in the district in a comparison to liberal-conservative similarities across executive and legislative branches of government. Outline of the Dissertation There are five chapters in this dissertation. Chapter One has provided a general overview of the dissertation. 50 Chapter Two provides a literature review of the major literatures covered. These include literatures on American legislative voting, Congress as an institution, rational actor theory, development of the role of ideology and self-interest impacting legislative behavior, and economic efficiency. The literature review offers a body of knowledge on which hypothesis development is based and findings measured through regression coefficients. Chapter Three develops the research hypotheses and methodologies and describes the sample. While Chapter Two describes relationships between various variables that have been studied, those relationships do not include the properties to determine if such relationships exist with the variables in this model. Thus, properties are operationalized in order to measure those variables. This chapter includes discussion of data collection, E-score development, coding of the variables, and types of analysis conducted (regression and interrupted time series analysis). Chapter Four includes presentation of the results of the analysis through textual and tabular methods. This presentation includes both multiple regression analysis results and interrupted time series results. Chapter Five offers overall conclusions and implications. Of particular importance is to what extent economic efficiency predicts human behavior and what that portends for public policy development and formulation. References are placed after Chapter Five along with Appendices of the data searches and data calculations used in developing the major measures employed in this study. 51 CHAPTER TWO LITERATURE REVIEW This chapter examines the major literatures on American legislative voting by considering Congress as an institution and the impact of a multitude of variables that drive decision-making and shape the legislative process. For purposes of this study legislative voting is limited to roll call voting in an attempt to analyze voting decisions by individual members as opposed to the chamber collectively. A variety of variables have been used to explain roll call voting. Some researchers focus on the self-interests of the individual legislator. These include financing of congressional campaigns and interest group lobbying. Other researchers focus on the constituency and emphasize the relationship between the legislator and the individuals within the district that he or she represents. Ideology is a common value that defines these relationships. Examples include degrees of liberalism or conservatism and represent longer term forces that affect a group over time. Still others examine variables related to the institution and its internal and external environment. External factors in part relate to changes in the constituencies. Examples include alignment of a constituency with a political philosophy and the role of government to formulate suitable public policy solutions. Political parties are an important link between legislator and constituent. Internal factors look at changes in the institution itself. Examples include organization of Congress by committee structure and rules, changes in leadership and 52 message over time, and institutionalization of the body, especially the House of Representatives. Ignored in these studies is economic efficiency, the element that is the focus of this dissertation. Relationships between legislator and constituency and between internal and external environments describe the factors that influence the lawmaking process, but they do not address the aggregate, benefit-cost consequences of policymaking. A focus on economic efficiency quantifies the impact of public policies in maximizing and expanding social welfare. The institutions of Congress change over time and linkages between those institutions and constituents, legislators, interest groups, and political parties affect the legislative decision process. This literature review examines economic efficiency in roll call analysis exploring to what extent a legislator?s affinity for economic efficiency is a function of changes between these linkages. It also presents economic efficiency as an alternative variable for predicting legislative voting in addition to accepted definitions of ideology and self-interest. Economic Efficiency Economic efficiency is maximization of net social benefits resulting from public policy decisions. Through the use of an index, Kennedy (2005, p. 2) finds that studying economic efficiency is important in examining why legislators support the economic interests of a constituency as opposed to public interests of a greater society. In short, Kennedy seeks to identify why legislators support policies that reduce efficiency. He references Stigler?s (1971) contention that constituency economic interests explain political behavior, but also finds that the act of voting is ideologically driven. 53 Kennedy?s (2005, pp. 12-13) research does not explicitly explain weaknesses in existing vote models, but in referencing Bender and Lott 1 , it questions ADA scores in explaining no more than the ideological position of a legislator, not why he or she might shirk or vote contrary to a constituent?s economic interests. Using Pareto optimality and Kaldor-Hicks improvements as a basis, support for policies that are in the public interest is logical. Voting to expand the welfare of a public policy and increase utility to more individuals suggests that a legislator?s ideology offers stronger support for his or her actions than constituency concerns alone. Aggregation of Preferences The principle behind an economically efficient policy rests on the premise of maximization of total policy benefits to total policy costs. As an extension of ideology, representatives seek policies that maximize net benefits. Referring to Figure 2.1, legislators seek to formulate policies at point Q*, the maximum point between total benefits and total costs. As Q x increases increasing opportunity costs push up costs in relation to benefit and produce an upward sloping TC curve. Diminishing marginal utility from each additional increase in quantity consumed produces a downward sloping TB curve, as each successive quantity produces fewer and fewer units of satisfaction (Kennedy, 2005, pp. 48-49). A legislator voting in support of economically efficient positions considers the aggregate implications of total benefit and total cost policy comparisons. 1 See B. Bender and J. R. Lott (1996), Legislator voting and shirking: A critical review of the literature. Public Choice, 87:1-2, 67-100. Their research found that a variable, such as ADA, used to capture a legislator?s ideology does not explain legislative voting, but rather confuses the economic interests of a legislator?s constituency. Figure 2.1: Net benefit maximization. Point Q* illustrates a maximization of total benefit (TB) over total cost (TC). Points to the left or right of Q* are associated with relatively lower total benefits or relatively higher total costs, respectively. Net benefits are maximized at Q*. The intent of Kennedy?s study was to derive a voting index that measures the strength of a legislator?s preference for economic efficiency based on total benefits and total costs of a policy. The extent that legislators forsake public interests and embrace economic interests of a constituency is a function of variables such as party affiliation, constituency demographics, and ideological characteristics of the legislator. Kennedy argues that economic theory must be combined with non-economic variables to better understand individual behavior and policy selection. Building an efficiency index requires inclusion of factors that impact behavior. Etzioni?s (1990) argument that individuals and society represent a collective effect driving individual decisions is important to note. Macroeconomic policies combined with structured policies that consider shifts in values or preferences over time are part of the paradigm. 54 55 How changes in behavior or changes within the institutional structure of government or industry affect the use of public policies and the aggregate consequences of developing and implementing such policies on social benefits to a constituency raises an important argument that Kennedy does not fully address. That is, to what extent do changes in the internal and external environment shape decision-making such that economic efficiency is enhanced or reduced? Ideological Basis of Economic Efficiency An appropriate starting point for considering this question is the degree that economic efficiency emanates from policy preferences of a legislator and the constituency he or she represents. Legislators and constituents each possess interests that drive behavior. Tullock (1983) finds that a prevailing function of modern government is redistribution, which is a transfer laced with benefits and costs (pp. 1-3). If a legislator adheres to personal ideology and votes accordingly, the extent to which such a vote is contrary to constituents? interests represents shirking. Kennedy (2005) finds that shirking manifests as ideologically based, where legislators vote against the economic interests of their constituents due to ideological considerations by consuming personal ideology at the expense of those constituents (p. 14). When legislators support policies that are not in the economic interests of a constituency, there is a risk of alienating a constituency and losing re-election to office. Supporting a pork project is an example of legislators embracing narrow economic interests at the expense of the overall constituency he or she represents (Hird, 1991). 56 Drawing a distinction between a legislator?s support for constituency economic interests and public interests, the point in time of a legislator?s political career affects voting decisions. Lott posits that it is ?a legislator?s ideology that tends to keep him or her from engaging in opportunistic behavior [shirking] when he or she is no longer competing for re-election? (as cited in Kennedy, 2005, p. 17). Thus, it is expected that legislators who are in their last term and no longer facing re-election have lower opportunity costs (Kalt & Zupan, 1984, p. 283) and will vote to a greater extent on their personal ideology once the threat of defeat, a cost of consuming ideology, is removed (Rothenberg & Sanders, 2000b). Supporting an expansion of Kennedy?s model for increasing the number of years analyzed, Stratmann?s (2000) research concurs that voting decisions change systematically over the course of a legislative career with party line voting inversely related to congressional seniority and changes in voting behavior in accordance with preference of the median voter in the district. To the extent that a legislator supports economically efficient policy positions the apparent effect of self-interest, as the elected official continues to be reelected, represents a decision-making struggle between policies that consider net benefits and legislative independence. Public Interests Public interests justify Kennedy?s analysis of economic efficiency, but they also open important dialogue for the role of government in providing a better society in which to live, loosely defined around democratic objectives and majority rule (Kernell & Jacobson, 2006, p. 506). Solely applying the principles of efficiency to achieving public 57 policies that enhance public interest is problematic, as market models fail to address other values (Bozeman, 2002). Vogel (1980) states: ?Public interest activism is based on the mistrust of both business and government? (p. 609) and is consistent with unorganized interests having access to privileges and prerogatives as organized interests have in shaping regulatory interaction between business and government. Political forces are strong and shape legislative behavior as rational individuals seek maximization of their private interests at the expense of public interests (Tullock, Seldon, & Brady, 2002, p. 16). But in a context of economic efficiency, Stigler defines public interests as promoting policies that enhance efficiency or opposing policies that diminish efficiency, where public policy regulations are often promoted as the cost of pursuing noble, national goals (as cited in Kennedy, 2005, p. 27). This definition introduces values of the legislator for pursuing this end as the relative impact of a decision, such as intended or unintended consequences associated with higher regulations, for example, translates into changes in social benefits (Putterman, Roemer, & Silvestre, 1998). Struggles between public interests and private interests of rational individuals are of concern to this analysis in that forces within the environment that impact decision- making must be considered in analyzing policy consequences. Redistribution of goods and services is a commonly accepted practice in serving the public interest and expanding social welfare. The impact of policies that distribute goods to areas in need may or may not produce the most economically efficient outcome and could reduce social welfare from such actions. Thus, redistribution seeking to expand social welfare must be based on economically efficient considerations. 58 Redistribution Distribution of goods and services is an important consideration for studying the best arrangement for serving the needs of the public. Kalt and Zupan (1984) argue: ?Since every economic policy decision produces transfers of wealth, it is always possible to infallibly relate political outcomes to distributional impacts? (p. 280). With Tullock (1983, p. 1) finding that redistribution is arguably the most important function of modern government, policies that serve the public?s interest are a solid basis for redistribution of goods from one group or category to another. Appleby, Flathman, and Goodin each find that a distinguishing factor of government in contrast to private organizations is that the former should strive to serve the public?s interest through the policies that it formulates (as cited in Bozeman, 2002, p. 147), although the latter introduces self-interests that affect each decision premise in making such policy decisions. Individuals support income distribution as a minimal level of helping the poor on moral grounds, but the bulk of transfer payments goes to the politically influential and organized (Tullock, 1983, p. 5) and affects the efforts of private interests (i.e., interests groups, organized coalitions, etc.) on the political process. That the primary motives behind income transfer are greed, desire to help others, and to a lesser extent, envy, the moral responsibility of protecting human dignity and providing a level of subsistence to members of American society (Bozeman, 2002, p. 154; Weimer & Vining, 2005, pp. 143-144) is a foundation for building an argument that ideological components of a decision compete against individual self-interests that motivate individuals. 59 Social welfare implications The social welfare implication of economic efficiency is a solid endorsement for analyzing economic efficiency as a component of ideology. With social welfare dependent upon economically efficient policy-making, promotion of greater social welfare and individual freedom through the use of public policies follows a utilitarian argument that seeks the greatest happiness for the greatest number of individuals. The rights of individuals and distributional aspects of policy-making are illustrations of ideology that favors equity or redistribution as opposed to efficiency (Kalt & Zupan, 1984, p. 281). The premise of the theory is that good decisions lead to good consequences as a method of quantifying human welfare (Parsons, 1995, pp. 45-46; Weimer & Vining, 2002, pp. 384-385). Quantification of changes in social welfare through social surpluses of net benefit over net cost is necessary for measurement (Weimer & Vining, 2002, pp. 138- 139) and supports an economic efficiency index as a tool for analysis. Rational decision makers maximize utility by surveying opportunities and costs and taking action to achieve the greatest gain within those constraints (Jones, 2001, p. 26). To determine if a policy initiative is desirable one may compare utility before or after the change or lump sum redistribution from beneficiaries of the change to the losers as a Kaldor-Hicks improvement (Coate, 2000, p. 437). A policy change to achieve distributional outcomes is efficient if no alternative policy change exists that is better for all, but may not produce social welfare without enhancing societal well being (Coate, 2000, pp. 439- 440). 60 Improving Kennedy Economic Efficiency Model Kennedy?s research 2 correctly analyzes policies that measure national benefits and national costs in deriving an economic efficiency index. Separating economic efficiency from other measures of ideology (e.g., ADA and ACU scores) is an effective method for measuring not only legislative decisions, but also how those decisions reflect stewardship of resources entrusted in the hands of a legislator by a constituency. A major weakness with the Kennedy model is not that it is intuitively flawed, but rather inadequately addresses myriad forces that push and pull political decisions that are constantly in flux over a period of time. Examples are population demographics and migratory trends, lobbying and campaign contributions, and political party engagement and party message, to name a few. Expanding Kennedy?s analysis from two congresses to ten, with the 1994 election producing changes in party control, is an important component for this argument. Even though ideology, as a predictor of behavior, is less likely to change in the short term, a longer term analysis introduces links to political parties and the institutional impact of changes in political power bases in Washington, DC. This dissertation considers if internal and external changes affect economically efficient decision-making or if economic efficiency as a variable is merely a component of ideology and is less affected by the aforementioned forces. 2 Refer to http://www.lerner.udel.edu/econ-e/ for further information about this model. 61 Self-interest and Ideology Self-interest and ideology (non self-interests) are included in this study as independent variables in a model testing the effect of each as predictors of legislative voting vis-?-vis an E-score. To the extent that an efficiency index or E-score predicts voting behavior, including self-interest and ideology is a necessary means of analyzing and comparing. Self-interest and ideology explain most individual behavior (Buchanan, 1972, p. 19; Downs, 1998, p. 19, Alesina & Rosenthal, 1995), but for the purposes of this dissertation, concepts of self-interest and ideology are used in explaining political behavior. Their importance in this analysis reflects the labeling of a legislator?s motivation and political behavior around these general principles, although the measures do not embrace the consequences of political decisions and, alone, are insufficient in justifying policy action. As a component of ideology, analyzing economically efficient policy positions requires careful consideration of self-interest and ideology in explaining human decision-making. Legislative decisions are best illustrated through voting decisions. How a legislator considers and defines a potential problem is a function of the ideology that shapes his or her interpretation of the problem. 3 The extent to which an issue is actually a problem requiring a public policy solution is a product of the legislator?s ideology and also self-interests that result from supporting or opposing public policy formulation. 3 Refer to Parsons (1995, pp. 77-78) for a discussion of problem recognition and problem definition in stages of a policy life cycle. 62 Self-interest and ideology affect public policy inputs, outputs, and outcomes by shaping all areas of individual decision-making that in turn influence voting decisions. Many authors have studied voting behavior through myriad self-interested and ideological interactions (e.g., Schneider, 1979; Kalt & Zupan, 1984; Nelson & Silberberg, 1987; McArthur & Marks, 1988; Koford, 1989; Richardson & Munger, 1990; Segal, Cameron & Cover, 1992; Levitt, 1996; Uslaner, 1999). The struggle between self-interest and ideology is as obvious as it is complex. Representatives are elected to serve the interests of their constituents but must anticipate and balance those beliefs with their own self-interest and self-preservation. In part, decisions made by each representative are shaped by his or her responsibility to the people of each district (Kalt & Zupan, 1984). How the legislator envisions the role of government to address public concerns is a function of not only feedback from the ones that he or she represents, but also the ideological tendencies to view the world according to a set of beliefs. Considering self-interest theory and the impact of ideology on behavior, it is possible to learn more about the characteristics of each legislator before addressing consequences of behavior. Self-Interest Theory Self-interest is a primary explanation for individual behavior and is a common area in understanding political events, such as public policy formulation. Self-interest theory rests on the premise that a decision that a person makes is a result of a benefit to that person. In political science, self-interest is the application of rational choice to the political process, with decisions made in support or opposition to public policies (Arrow, 63 1963; Almond, 1991). For a legislator the choice will maximize his or her personal resources or the resources flowing to the district served by increasing probability of re- election (see Downs, 1957). In economics resource maximization is a basis for self- interested behavior, as pecuniary aspects of making a decision play a role in choices selected by the individual. Each decision presents benefits and costs to the decision- maker that are part of the decision process. Considered very broadly self-interest could encompass any activity and explain all behavior. Such definition of self-interest is circular, however; it fails to consider important points of behavior to distinguish between selfish or selfless behavior, or self- interest and ideology (Mansbridge, 1990, pp. 254-263). As an example of this flaw, Buchanan (1972) discusses self-interest in terms of utility maximization, where individuals assign a value to goods (pp. 16-20). His research does not establish the time period in which these goods can be consumed or enjoyed, nor indicate to what extent consumption of such goods will affect that person?s financial profile. Utility that one receives from financial considerations is easily understood; utility generated from non- tangible activities, such as satisfaction from charitable endeavors (Sen, 1990), does not offer a clear distinction between selfish and selfless behavior. Arguing that self-interest can be based on utility and also non-tangible behavior is problematic at best. Christian virtues of faith and love applied to helping one?s fellow man, for example, do not equate with furthering one?s financial position (Buchanan & Tullock, 1962; Sen, 1990, p. 29). Distinguishing between ideologies of community service and charitable works from longer term financial benefits that an individual may receive is very difficult. Downs? (1957, p. 297) contention that rational citizens attempt 64 to maximize utility income emphasizes this distinction. Monetary considerations appear to be a central component when differentiating self-interested from non self-interested behavior. Voting decisions made by constituents or legislators are often compared to non self-interested behavior for illustration. According to Stigler (1971) and Downs (1957, pp. 6-7) nonmaterial aspects of well being, such as spiritual commitment and moral righteousness, social status and adjustment, self esteem, and other ethical beliefs are non self-interested behavior. Choices made by individuals involving these issues do not expand their personal resources or increase their political acumen. Self-interest and ideology appear to have overlapping boundaries that make an accurate definition of self-interest confusing. Sears and Funk (1990) offer a definition of self-interest that makes this distinction and is especially important when measuring variables that affect economic efficiency over time, a key component of this study. Their research considers public opinion but applies equally as well to an understanding of the legislative process. They find that self-interest includes shorter-term forces impacting an issue and is often an immediate reflection of individual choice benefiting the material well being of an individual or his or her family (p. 148). Using Sears and Funk?s definition of self-interest alleviates many of the problematic issues associated with self-interest as all encompassing behavior. To the extent that a force affects an individual immediately or over a longer period of time is an important basis for considering if that force influences the self-interests or non self- interests of that individual?s political decisions. Narrowing the definition to include only the material well being of a person or immediate family, which appeals to pecuniary 65 considerations, is necessary for avoiding the pitfalls faced by Buchanan and Sen in their considerations of self-interest through utility maximization and sympathy, respectively. Ideology and Behavior Ideology is an action-oriented model of people and society (Friedrich, 1965; Grafton & Permaloff, 2005b, p. 173) and offers a solid basis for most legislative decisions. It is important for this study as a tool in explaining the step between individual behavior and consequences of political decisions. An action-oriented model contains directions for resolving political and economic issues, where ideology offers a prescription for solving these issues within a society (Minar, 1961; Drucker, 1974; Reichley, 1981; Van Dyke, 1995). Apter?s (1964) argument that considers a more honorable and dignified social conduct that results from an application of general ideas in specific situations is consistent with this prescription (pp. 16-17). By analyzing and selecting public policies to solve political, economic, and even social issues, the effect of ideology on congressional voting behavior makes explicit a moral basis for action. Ideology is often used in achieving self-interested goals and is the primary means of explaining non self-interested behavior in executive and legislative policy formulation. Non self-interest is synonymous with a standard liberal-conservative spectrum of ideology. In applying these prescriptions to a relationship between constituents and legislators, LaPalombara (1966) cites the historical aspects of ideology developed by L. H. Garstin 4 by chronologically linking sets of values of mankind to 4 See L. H. Garstin (1954). Each age is a dream: A study in ideologies. New York: Toronto Ryerson Press. 66 actions of individuals or government that maintain an existing state of affairs or hasten development for future prescriptions for mankind (p. 7). Garstin?s analysis suggests that ideology must be considered in relative terms, where sets of values that identify a group are not necessarily static but shift over longer periods of time and with institutional changes in government and the economy (Hoover, 2003, pp. 259-260). For the parameters of this dissertation, shifts in values are important in explaining if economically efficient decision-making, when it exists, is supported universally by a legislator or is rather a function of the larger political, social, and economic environment. Borrowing from Groseclose, Levitt and Synder (1999) and Poole and Rosenthal (2001), shifts are controlled in an attempt to standardize measurement. Accepting Sears and Funk?s definition of self-interest, a definition of non self- interested behavior or ideology also includes the time period of the forces acting on an issue and the person or group affected by such forces. Longer-term forces, interests that impact the well being of a group rather than an individual, and nonmaterial components of well-being represent ideologies and offer a clear distinction to self-interested behavior (Stigler, 1971; Quirk, 1990; Downs, 1957, pp. 6-7). Mullins (1972) equates these distinctions with boundaries between ideology and cultural phenomena and finds that ideology molds cognitive ideals among members of groups and enables those members to appraise their political condition and its prospects for the future, facilitating the mobilization of energy and resources for common political undertakings. The significance of ideology in mobilization is not that it ?causes one to do? but that it ?gives one cause for doing? (p. 509). Ideology thus represents the basis for 67 mobilization of actions around common political problems by grouping thoughts and ideals of those sharing a common ideology to reach similar outcomes. Political ideologies allow one to understand reality or generalizations through simplifications that reduce excess information to a manageable size (Stokey & Zeckhauser, 1978, pp. 7-8). Sharing characteristics with fields of science, an analogy to engineering and other scientific disciplines holds merit. The political arena contains an abundance of information flowing through several channels involving constituents, elected officials, and media that require structure before processing. The basis of science is deductive logic, with general principles explaining specific phenomena. Political ideologies involve similar logic, where principles of political thought shape the context in which social, economic, and political matters are considered, requiring political ideologies to structure and simplify those principles into a manageable, recognizable form. Without such structure legislators are not afforded generalizations on which to formulate policies and decision-making is at best problematic. In contrast to Stokey and Zechkauser, LaPalombara (1966) concludes that ideology may not be dogmatic or utopian and is not compatible with science, suggesting that political ideologies contain the necessary ingredients for simplifications of realities within a model but are less concrete and are subject to change over time. The research does not dispute that like science ideologies simplify reality, making considerations for complex, specific events or forces possible from general principles. Rather the historical basis associated with ideologies and potentially numerous outputs possible from myriad political forces, generate outcomes less predictable than scientific fields. 68 To the extent that ideologies shift over time and can be analyzed on the basis of scientific principles, including ideology along with self-interest in this study is a necessary prerequisite for studying economic efficiency. Legislators make decisions with political and economic consequences and such decisions are a product of many factors in the environment, but if economic efficiency truly transcends a liberal- conservative spectrum of ideology, deeply held values not only reflect characteristics of a legislator but also the macroeconomic consequences of his or her voting record. While liberal-conservative ideology is a basis for most of the thought processes that define policy direction for most voting decisions, economic consequences may be what ultimately guide legislative voting, and those economic consequences should be reflected through economic efficiency. Public Choice Theory Including in this study a discussion of public choice theory and the rational behavior of political actors provides a logical basis for political decision-making. The economic principles surrounding such decisions are an illustration of self-interests that affect legislators. To the extent that legislators are rational and their self-interests affect policy decisions, predicting economically efficient outputs must consider not just efficiency as an extension of ideology, but also the rational principles behind each decision. Of critical interest for this analysis is the extent to which legislators follow their self-interests rather than supporting the public interests of a constituency. 69 Public Choice and Economics The public choice or rational actor approach applies economic and political theory to decision making by considering aggregate effects of self-interest on individual behavior. Tullock (2002) defines public choice as a scientific analysis of government behavior and, in particular, the behavior of individuals with respect to government (p. 3). A voter in a voting booth is analogous to a customer in a supermarket making decisions rationally. The unit of analysis for the approach is the individual, and the approach is based heavily on how self-interests motivate individual actions (Barry & Hardin, 1982, pp. 19-20). An economic market and a political arena are devices in which individuals further their self-interests by entering into exchange relationships that are of direct benefit to other individuals on the other side of the transaction. Market exchanges consist of goods and services, while political exchanges involve inputs to secure a common output (Buchanan & Tullock, 1962, p. 19). In explaining policy formulation based on the same assumptions used in explaining the behavior of a firm, Parsons (1995, p. 307) finds that parties make excessive promises to win votes. In democratic societies parties are analogous to profit- seeking entrepreneurs. Just as entrepreneurs produce goods and services and engage in market exchange, parties play a role in policy formulation to produce the most votes to serve private ends (Downs, 1957, p. 295). Rational Decision-Making Maximization of individual self-interests explains rational behavior (Klosko, Muller, & Opp, 1987) and is easily illustrated through trade offs in individual utility and 70 how choices involve opportunity costs that must be considered over time. Rational decision makers face a utility function that relates decision alternatives that are available with benefits of choosing one alternative over another. Calculating costs in terms of forgone opportunities allows one to consider future and present benefits of a decision in accordance with his or her preferences. Each decision is made along indifference curves that constrain individual choice in accordance with his or her preferences (Downs, 1957, pp. 4-6; Jones, 2001, pp. 35-37). For this study, the premise behind economic efficiency as a maximizer of total benefits is rooted in individual rationality. With individual self-interest and ideology influencing the decision process, how one considers the utility of each decision in accordance with his or her preferences affects support for public policies. The logic of economics offers an illustration of the competition between forces that affect each decision (Parsons, 1995, pp. 307-308). That is, legislators are driven by their self- interests while attempting to serve the public?s interests. This presents an opportunity for manipulation of those self-interests (e.g., through campaign contributions) to alter policy outputs. To paraphrase Barnard, the extent that a legislator subjectively evaluates each inducement produces a decision only to the point where the marginal benefits equal or exceed the marginal costs of supporting such decision (as cited in Fry, 1989, p. 8). Capture Theory In politics capture occurs when bureaucrats act according to self-interests instead of in the public?s interest in the framing and passing of legislation. This idea has been extended to legislative voting behavior. The premise behind capture is constituents? 71 ideological interests and self-interests should influence federal spending (Stigler, 1971; Atlas, Gilligan, Henderson, & Zupan, 1995). Kalt and Zupan (1984) pose the issue as competition between a legislator?s ideology and local economic interests. Peltzman?s (1985) view is contrary, as he finds the relationship between legislator ideology and local economic interests to be interrelated. Legislators who engage in ideological voting are shirking their responsibilities to constituents. Intensity and impact of ideological position to other variables such as career advancement, party loyalty, and quest for reelection to name a few, are important factors in the legislative process (Jackson & Kingdon, 1992), findings consistent with constituent and party dimensions that explain congressional voting behavior (Koford, 1989). The legislator?s interest and the constituents? interest are also a function of whether the elected official is leaving office or continues to serve his or her constituents. Legislators often change their voting decisions when exiting political office, especially if the legislator is an ideological centrist as opposed to liberal-conservative extremes, finding ?sufficient evidence to suggest that members do, indeed, pay attention to what their constituencies want when they are subject to reelection but give less attention to such desires when they are not? (Rothenberg & Sanders, 2000a, p. 322). Electoral ties appear to be particularly important when considering departing House members relative to those members who were reelected. By leaving office elected officials are less influenced by constituency concerns. Re-elected incumbents weigh district preferences more heavily than members leaving office (Rothenberg & Sanders, 2000b). Controlling for this aspect of a legislator?s membership to Congress is necessary in analyzing these effects. 72 Collective Action This section identifies the collective aspects of decision-making and external and internal costs associated with the number of individuals making a decision. An example includes an optimal level of individuals necessary to minimize expected costs of making the decision. The costs of making a policy decision are important to this research because they add another dimension to policy-making that transcends an individual legislator making an isolated decision. How the collective aspects of voting behavior affect support for economically efficient public policies suggests that rational individuals are expected to maximize self-interests before maximizing the group?s interests. Decision Science Characteristics of legislators are described on a liberal-conservative basis denoting that person?s ideological predilections. Public policies are debated and formulated around a multitude of complex dimensions. Dividing the stages of the policy process into debate space, decision space, and outcome space, the decision space collapses a multifaceted debate space into a single dimension in which legislators cast their votes, but branches again into many dimensions as policy outcomes affect various groups, individuals, and institutions (Jones, 2001, pp. 155-156). Although legislators are individuals and therefore bounded by limited knowledge with which to understand complex associations of variables, the consequence of policy decisions illustrates strengths and weaknesses in the agenda process. 73 Consistent with Rogers and Dearing?s contention that agenda setting includes media, public, and policy components that are interactive (as cited in Parsons, 1995, p. 114), arguing that lawmaking reflects relative degrees of liberalism or conservatism separates the ideological extremes with very little overlap (Jones, 2001, p. 154). With party unity scores increasing (Leyden & Borrelli, 1990, p. 343) and ideological polarization widening (Collie, 2000, pp. 219-227), a liberal-conservative spectrum appears to lack the necessary cognitive tools to analyze a plethora of multidimensional variables as part of policy debate and outcomes. External and Decision-Making Costs Multiple decision dimensions are consistent with Simon?s underlying decision premises ?and involve the processes of alerting, exploring, and analyzing, which precede the act of choice? (as cited in Fry, 1989, p. 185). Costs are a component of the options that a legislator must consider in making choices. Two types of costs are external costs and decision-making costs. External costs are those costs that an individual bears when a decision deviates from his or her preferences. These costs are highest when any one person can take action for a group collectively. Greater participation in decision-making reduces external costs as the decision will be closer to any individual?s preferences. Decision-making costs are opportunity costs of expending resources on a decision that could have been applied elsewhere (Ostrom, 1989, pp. 58-59). To minimize costs rational legislators seek the decision point where each cost intersects. The political process with greater involvement in decision-making reduces 74 external costs as the decision is a product of greater numbers of individuals and will be less likely to deviate from preferences. The smaller the size of the group, the greater is the paradox between large and small groups, as increased voluntary action exists for common purposes of the individuals in the group (Olson, 1965, pp. 2-3). When the number of individuals making a decision increases collectively, the expected cost decreases to a point. But unanimity raises decision-making costs as resources are expended in reaching a decision that could have been utilized elsewhere (Buchanan & Tullock, 1962, p. 89; Stigler, 1975, pp. 123-126; Ostrom, 1989, pp. 59-60). Interest Group Theory Interest groups are important to this study through the effect of their activity on legislative voting. Specifically, lobbying and contributions to campaigns influence self- interests of legislators and impact policy decisions. The message from an interest group is a unified voice that shapes how a legislator considers an issue or problem and is a factor in how a problem is defined. In this dissertation, campaign and lobbying dollars flowing to legislators are measured as independent variables influencing voting decisions. Parsons (1995, p. 30) defines interest groups as pressure or lobby groups, which seek to influence policy by monitoring existing policy and developing alternative ideas and proposals or shared attitudes (Truman, 1960, p. 33). The effect of these groups on shaping policy can be immense as such groups wield influence through the giving of campaign contributions and distribution of specialist information (Austen-Smith, 1993, p. 799). 75 Interest group theory reflects the self-interests of interest groups in decision- making processes and should be considered along with public choice theory because self-interest is the sole motivator of behavior in each theory. The outcomes produced by the interest group are of benefit to others, but only those participating in the group?s activities face the cost associated with those benefits. The larger the group the greater is the ratio of privately borne costs to privately accrued returns of individual action. Smaller groups provide greater net benefits to individuals and are more likely to persist over time (Schuessler, 2000, pp. 33-34). Although individuals are the unit of analysis for public choice theory and groups are the unit of analysis when considering interest groups (Truman, 1960, p. 502), Bentley 5 finds that within an analysis each term, group or individual, may be used interchangeably without significantly affecting the results (Buchanan & Tullock, 1962, p. 9). Interest groups are dynamic and are not effective without varying with business cycles and issue areas in an attempt to generate political outcomes (McFarland, 1991). Using economic theory in arguing a positive relationship with group activity and modernization, the higher the complexity and differentiation of society over time, the greater will be the proliferation of interest groups. With the ebb and flow of business cycles groups that seek the status quo benefit from political and economic stability, while groups seeking reform will challenge those groups. But after a few years unchecked groups will fail to maintain values responsible for stability, leading to greater political participation by reformers. As the cycle continues, the reform group loses interest in a few years and the cycle repeats. 5 See A. Bentley (1935). The process of government. Bloomington, IN: The Principia Press. 76 Campaign Finance Campaign financing is a source of capital provided to candidates for office to develop a message and communicate with voters. It is an independent variable affecting voting behavior. The cost of congressional campaigns is escalating as the role and scope of the federal government is increasing (see Reichley, 1992, p. 5). In an age of electronic media necessitating the use of specialized campaign strategies and expert consultants, campaign financing occupies a critical if not controversial component of the political process (Kernell & Jacobson, 2006, pp. 443-444). Hamilton (2004) states: ?Many Americans feel that it is money, not ideas and not principles, that reigns supreme in our political system? (p. 115). With all monies spent on congressional elections coming from private sources, a primary concern is evident: privately financed elections create an incentive for elected officials to serve as agents of their contributors rather than of their constituents, with the pursuit of money subverting the purpose of a campaign (Kernell & Jacobson, 2006, p. 444). Democracy demands political equality of one person, one vote (Baker v. Carr 1962; Reynolds v. Sims 1964; Wesberry v. Sanders 1964), but an unequal distribution of contributions to political candidates threatens democratic principles. Ansolabehere and Snyder (2000) state: ?One of the most striking features of congressional elections is the advantage that the typical incumbent enjoys in financing campaigns? (p. 65). Campaign financing increases the political pull of organized interests at the expense of the constituents? interests by purchasing influence, raising issues with legislative accountability. 77 Lobbying Self-interested decisions are the basis for lobbying efforts by interest groups to influence behavior. In the present study, lobbying dollars are an independent variable affecting legislative voting. Hojnacki and Kimball (1998) surmise that organized interests shape the policy decision agenda through careful considerations of which groups or individuals they will target in the legislative process. Lobbying can be friendly to reinforce existing policy preferences, confrontational if political enemies are deciding the fate of a policy issue, or some combination (Kollman, 1997). Lobbying across multiple, institutional venues is common with much variation between lobbying organizations and venues, where expectations of opposition from other interests are a major factor in lobbying decisions within any given venue (Holyoke, 2003). In political institutions, Democratic candidates who receive relatively greater assistance in developing campaign messages vote in higher numbers along party lines (Cantor & Herrnson, 1997). Interrelationships exist among committees, government agencies, and issue networks (Cater, 1964; Hall & Evans, 1990; Heinz, Laumann, Salisbury, & Nelson, 1990). For congressional committees, Hojnacki and Kimball (1998) add that organized interests seek an expansion of their coalitions, affecting the content and fate of bills that are referred to committees. Their research considers three perspectives in terms of units of analysis to integrate their findings: individuals, groups, and the context of the issue. Individuals are sometimes targeted through characteristics of the legislator, policy preferences, or legislative position in Congress, while groups that provide financial or 78 other resources to influencing government policy offer another model of interest group behavior. Lobbying efforts at the very least increase discourse and expand discussion surrounding an issue. Smith conceives that the amount of time Congress deliberates on an issue correlates with the success that lobbyists experience in influencing congressional decisions and Bacheller concludes the public?s perception of an issue and the level of controversy of the issue affect lobbying strategies (as cited in Hojnacki and Kimball, 1998, p. 776). The relative importance of an issue is part of problem definition but also a product of recognizing a problem. Lobbying initiatives play a role in prompting attention to an issue (Parsons, 1995, p. 127). The extent of controversy surrounding an issue is an incentive for lobbying groups to push the issue toward a policy agenda and minimize conflict. Roll Call Voting A roll call vote in Congress specifically identifies the position of the legislator casting the vote in contrast to a voice vote where legislative votes remain anonymous. The roll call votes usually studied are votes on the House and Senate floor because all members may participate in these votes. Interest groups track these votes in order to determine whether their campaign contributions and other support for a legislator have returned benefits in the form of policy decisions (votes) supporting their positions. Some interest groups track a variety of these votes and combine them into indexes that measure overall levels of support received from each member. Knowing whether the 79 legislator supports or opposes a public policy as opposed to a tally for the entire body is paramount for this study in measuring voting behavior and understanding the legislator?s regard for economically efficient outcomes. In this section an examination of roll call voting reveals the dimensions of ideology that are part of each decision premise and compares to legislative self-interests in developing economically efficient voting outputs. Explanation of Behavior Roll call voting provides a permanent record of a legislator?s support or opposition to a public policy, but it does not explain why the elected individual took such action. Strategies for casting roll call votes include a combination of factors between self-interests and ideology and congressional members and their constituency. With legislators seeking reelection to public office (Downs, 1957) or logrolling (Buchanan & Tullock, 1962, p. 132) to trade votes to gain political advantage, inducements shape behavior. Barnard finds that utilities attached to inducements and contributions explain much political behavior (as cited in Fry, 1989, p. 196). Simon?s argument is that these decisions are shaped by information available to the legislator and constrained by bounded rationality that constricts his or her limited cognitive capacity (as cited in Fry, 1989, p. 192). Jones (2001, p. 26) does not dispute human limitations, but rather justifies that mankind is goal oriented and is not always successful in adjusting to a changing world and satisfices with acceptable rather than optimal alternatives (Simon, 1996, p. 30). The effects are most pronounced when voting decisions involve multiple consequences or congressional bills are assembled to include 80 several unrelated items involving different dimensions and levels of support (Jones, 2001, p. 43). Clausen and Van Horn find that shifting of policy domains under new or different rubrics or vectors that demand attention illustrates the power of roll call voting on changing behaviors. Changing policy domains shifts legislative support through new clusters of issues (as cited in Shaffer, 1989, p. 36). Flanigan and Zingale follow that a correlation exists between shifts along the liberal-conservative ideological spectrum and changing perceptions of the ideological distance between the voters and institutions of government (as cited in Grafton & Permaloff, 2005b, p. 174). Adding to this argument, Clausen finds that constituents strongly impact a congressional members? position in civil liberties and foreign affairs but not in domestic areas of social welfare policy and agricultural support, although Peltzman and Kalt and Zupan dispute whether the members? ideology is a significant cause of variation in roll call behavior (as cited in Vandoren, 1990, pp. 311-312). Evidence exists that House members with more moderate ideological positions are more likely to be reelected (Erikson & Wright, 2005, pp. 95-97). Of the myriad forces impacting legislative behavior, Bullock and Brady (1983) acknowledge that party has the largest direct effect, but constituency characteristics have an even a larger effect than party when measured both directly and indirectly in the Senate. Senatorial voting responds to effects of party, as senators move toward the ideological center a few years before the end of their six-year election cycle. Moderation of ideological positioning characterizes members in the House and Senate preceding an election, but more heterogeneous constituencies and a longer term of office magnify 81 changes in behavior in the Senate relative to shorter terms of office in a more homogenous House (Erikson & Wright, 2005, pp. 100-101). Roll Call Votes and Measures of Ideology Ideology is often measured along a liberal-conservative spectrum by considering roll call votes that legislators cast in the House and Senate. Most votes that comprise the measure are compiled by interest groups in an attempt to identify how closely legislators vote in support of legislation consistent with the ideological profile of that group. The exercise presumably allows one to determine the extent to which a representative or senator is conservative, liberal, pro-family, pro-business, etc. and says something about the behavior of the legislator (Shaffer, 1989). In addition to E-score, Kennedy (2005) identifies at least 11 models that represent ideology (p. 66) and Shaffer (1989) finds as many as 15 groups providing ratings for members of Congress, with most measures assembled through roll call voting records. A consensus exists that ADA and ACU roll call votes offer a distinctive measure of a legislator?s ideological location along a liberal-conservative spectrum (Burden, Caldeira, & Groseclose, 2000; Erickson, 1990; Shaffer, 1989; Schwab, 1988). Researchers find that a single liberal-conservative dimension explains as much as 80 percent of voting decisions (Schneider, 1979; Poole, 1981, 1988; Poole & Rosenthal, 1985; Poole & Daniels, 1985), leaving fully 20 percent not explained within a single dimension. E-scores utilize roll call votes but introduce another dimension to voting models by extending traditional liberal-conservative spectrums to incorporate net social benefits. A vote in support or opposition to legislation is a pragmatic measure of the 82 legislator?s ideology, but also presents problems in using voting indexes to understand voting behavior (McRae, 1954). Flower (1982) contends that group emphasis tends to weigh interest group measures toward a few indices and might present a misleading polarized view of the legislator?s ideology. Anderson follows that no external checks are in place to protect the validity of the measure from a researcher?s judgment as warranted, especially if a standard liberal-conservative spectrum oversimplifies dimensions of ideology that Matthews and Stimson reveal exist over several dimensions of conflict (as cited in Shaffer, 1989, pp. 34-36). Policy domains, namely international involvement, civil liberties, government management, social welfare, and agricultural assistance, are examples of multiple dimensions of ideology (Clausen & Cheney, 1970). Shaffer (1989) states: If congressional ideology is indeed multidimensional, then a single index would be a highly inappropriate measure to employ in legislative research. This might be especially true for a rating like the ADA?s, which incorporates a wide range of both domestic and foreign policy roll-call votes. (p. 36) While these concerns are warranted, the fact that roll call analysis appears to depict dimensions of ideology and shifts over time in ideological positioning suggests that voting decisions may change as a result, with the impact on economic efficiency of foremost concern. These shifts in ideology may reflect a changing policy agenda as larger social, political, and economic changes occur. Simple indices of liberal- conservative divisions are inadequate over time when significant transformations occur (Deckard & Stanley, 1974; Bethell, 1979; Shaffer, 1989). Identifying shifts in ideology correlates with shifts in policy domains that Clausen and Cheney (1970) identify. Issues 83 may change as political situations change with the end of the Cold War and beginning of an age of terrorism, or the meaning of liberal and conservative may be altered due to changes in interpretation manifested through the rating agency (e.g., ADA or ACU). Perspective to Self-Interests Legislators rationally make policy decisions according to their self-interests. Roll call votes are important in that voting decisions are made available to constituents, contributors, and other congressional members of the same or opposing political party. Maximizing self-interests is the hallmark of rationality and legislators will carefully cast votes or abstain from voting in a manner consistent with Bachrach and Baratz?s non- decisions (as cited in Parsons, 1995, pp. 135-136). Associating or disassociating with the label of a political party reminds voters of the legislator?s ideological position vis-?-vis the party to which he or she belongs (Snyder & Groseclose, 2000; Bullock & Brady, 1983). Considering the extent that constituents? preferences (Fiorina, 1974, p.30) influence a congressional members? policy position, McRae and Clausen?s research reaffirms the importance of dimensions in congressional decision-making in which alignments can vary from policy area to policy area (as cite in Poole, 1988, pp. 119- 120). This is not inconsistent with a belief system that Converse (1964) defines as ?a configuration of ideas and attitudes in which elements are bound together by some form of constraint or functional interdependence? (p. 207), allowing the legislator to encompass a wider range of information than he or she would find possible without an organization of ideas (p. 214). 84 Policy Implications Numerous public policy implications arise from a single liberal-conservative dimension or multiple dimensions of individual behavior. Wilcox and Clausen (1991) cite research (see MacRae, 1970; Clausen, 1973; Sinclair, 1977) where ?members of Congress distinguish among a variety of substantive dimensions in reaching voting decisions? (p. 393). Vandoren (1990) argues that considering policy dimensions alone is not adequate and in order to fully understand congressional behavior requires a pooling of data through time series analysis. This contention not only offers support for an E- score in capturing a multidimensional policy decision, but also attributes its virtues through a time series perspective. A cause for concern with using roll call floor votes in a study of public policies surrounds the importance of committee votes. Roll-call votes occur only if policy proposals receive committee approval or extraordinary measures (e.g., use of discharge petitions or action by the leadership) are taken to bring legislation to the floor. Legislation may languish on the calendar or an inadequate number of members assemble for a roll call. Hence the effects of factors estimated from roll-call data are actually conditional on the occurrence of committee approval and member support for a roll-call vote (Vandoren, 1990, p. 332). Ignoring committee votes fails to consider aggregate congressional decision-making if time series analysis is not employed for roll-call votes. Determining why a legislator voted in support or opposition to a policy must consider the effect of constituency influence, political party, self-interest, and ideology. The policy implication is that roll call voting analysis too often considers only a small dimension of behavior unless pooled over a period of time, supporting an 85 argument of this dissertation that forces affecting political behavior are shifting and must be considered not as static entries, but rather over time in predicting how closely a legislator espouses economically efficient policy making. American Legislature In this section a discussion of factors internal and external to the American legislative system are considered because of their impacts on policy making. The external forces such as elections impact internal operations (e.g., chamber control by party, nature of the leadership, and policy agendas) in ways that influence individual legislators and their decision-making. Federal System The U.S. Constitution establishes the basic framework for dividing governmental responsibilities among levels of government (Hamilton, 2004, p. 12), where independent interaction across the levels includes the same people and territory and is a necessary ingredient for mutual influence (Kernell & Jacobson, 2006, p. 81). The federal system is one national level of government and 50 state levels that cede powers to local units in a ?blend of elected and appointed officials from all levels of government sharing policy and program duties? (Hamilton, 2004, pp. 11-12). This view is consistent with shared federalism, where levels of government cooperate in jointly providing services that its citizenry expect than neatly divided spheres of sovereignty discussed by Madison in Federalist Paper 45 (Kernell & Jacobson, 2006, pp. 82-83). Modern policy decisions involve complex associations that require involvement of the federal government in 86 policy areas once served exclusively by state or local concerns, a scenario that perpetuates an increasingly active role of the federal government. In agreement Neustadt recognizes that a growing government becomes increasingly complex, with sharing of power proliferating in an attempt to serve the needs of its citizens (as cited in Cater, 1964, p. 11). Sharing powers does not usurp the constitutionally mandated function of an institution, but rather facilitates institutional effectiveness in carrying out those roles identified by the framers by diminishing the risk of concentrated power in the hands of any one person or small group. As an institution Congress represents both the states (Senate) and districts within the states (House). The states control the election process for the legislature and through the Electoral College, for the president as well. Legislators act as agents to the needs of a constituency, but also serve as a statesmen linking local issues with state or regional concerns and managers of political resources and opportunities necessary for cooperation between political institutions within their respective districts (Frantzich, 1979). Democracy Ideals that a society desires correlate with the practices of political representation. Cohen states: ?Democracy is that system of community government in which, by and large, the members of a community participate or may participate, directly or indirectly, in the making of political decisions which affect them all? (as cited in deLeon, 1995, p. 889). Haynie (2005) adds that legitimacy and trust in the political system by the citizenry are necessary to ensure that political institutions meet democratic ideals. ?In the United States, legislatures, more than any other political institution, 87 embody these important principles of democracy? (p. 406). To the extent that democratic values produce public policies that do not infringe upon individual liberties an important link exists between democracy and the policy sciences. The policy sciences and democracy Democratic ideals are the foundation on which policy decisions are made. Policy decisions represent a science that defines political decision-making between an elected official and the constituency he represents. Lasswell 6 clarifies the science of policy making by recognizing how the policy process can expand basic democratic values through the methods and results of an investigation of policy and the findings from a study of political problems (as cited in Parsons, 1995, pp. 18-19). Improving the knowledge of decision makers in the policy process and expanding the contextual framework for policy discussion is consistent with democratic principles and illustrates the importance of considering the individual (whether a constituent or a legislator) and the role of that person in affecting policy decisions. Initiating a problem-oriented focus to synthesize disparate ideas is an important step in eventually realizing an expansion of human dignity. Institutions of Government Institutions are organizations that exist in providing stability through developed procedures and rules. Parsons (1995) finds that ?institutions do not exist in isolation 6 See H. D. Lasswell (1948). The analysis of political behavior: An empirical approach. London: Kegan Paul; H. D. Lasswell (1951). The policy orientation. In D. Lerner and H. D. Lasswell (Eds.), The Policy Sciences: Recent developments in scope and method. Stanford, CA: Stanford University Press; H.D. Laswell (1970). The emerging conception of the policy sciences. Policy Sciences, 1, 3-14. 88 from the wider relationship of state to society? (p. 334). Political institutions standardize relationships between decision-makers and constituents; Congress has formal powers and shared responsibilities between Congress and the executive branch impact legislative decision-making. For this research the House of Representatives and Senate are each distinguished as separate institutions to emphasize probable public policy impact of differences between those two chambers of Congress. The effect of political divisions between each chamber and the executive branch will also be considered. Competitive Market Analogy Organizations combine self-interests of individual members and the common interests of the group that it represents. In a perfectly competitive market firms have a common interest in higher prices for the industry?s product. A firm cannot expect a higher price for itself unless all firms in the industry receive the higher price. Firms in a competitive market, however, produce where marginal cost equals marginal revenue (Browning & Zupan, 2002, pp. 229-230) and if the market is not in equilibrium with price exceeding marginal revenue or price, an incentive exists for firms to produce more (Stigler, 1965, pp. 9-11). But as output increases price falls; the self or individual interests of the firms supersede the common interests of the industry. Borrowing from an analogy to a competitive market, institutions of government serve individual interests before common interests of the state. Patriotism in an age of nationalism is a collective force that pulls together common interests within a nation. But a state cannot serve those interests without compulsory taxation of its citizens to provide services and resources. Taxes, however, provide fundamental services available 89 to everyone. These common benefits are often called public goods as the benefit is not excludable and no rival firm provides the same or similar service (Stigler, 1975, p. 107; Weimer & Veining, 2005, p. 72). It is in the common interest of the state for everyone to benefit from additional services, but not in the individual interest of those burdened with higher taxes (Stigler, 1965, pp. 13-15). Congressional Institutional Distinctions Institutional distinctions between the House of Representatives and Senate affect the legislative process from input and processing of resources to policy outcomes. Differences in the formal structure shape the informal roles of legislators, a concern in this study. Changes in these institutional relationships over time are a function to a large extent of factors within the external environment surrounding political parties. Examples include increasing ideological divisions between parties and party unity within each party. Shifts in party alignment are another example; each will be discussed later under lawmaking and are examined in this study as independent variables producing changes in behavior. Checks and Balances Burden and Kimball (2002) argue that the structural differences associated with a constitutional separation of power provide an inherent tendency for citizens to split votes among political parties (p. 17). The American system of government makes possible for checks and balances within the system to apply to not only interaction among institutions of government and its leadership, but also citizen choices deciding who occupies legislative or executive roles. Literature points to gerrymandering, campaign 90 financing advantages and media access for incumbents, and constituency service in explaining Republican dominance of the presidency and Democratic dominance of Congress during most of the period after World War II (Menefee-Libey, 1991, pp. 519- 520; Burden & Kimball, 2002, p. 18). Petrocik finds that issue ownership is a premise behind the reputation that each party attains from prior consideration of campaign issues, while Jacobson posits that voters gravitate toward the political party that offers greater relative strength and expertise in a specific issue area (as cited in Burden & Kimball, 2002, p. 20). Republican strengths are expected in economics and foreign policy and Democrats usually excel with social issues, such as education, welfare, and environment. Issue ownership and institutional matching of political party are examples of short-term forces that vary from one campaign to the next (Flanigan & Zingale, 2002, p. 60). House of Representatives The House of Representatives is of a larger size, hierarchical, with a locally based, homogenous constituency. The decision making process of the House of Representatives is more formal and rigid and the institution receives relatively less media coverage than the Senate. Members of the House are elected every two years and are ?closely connected to the needs, desires, and wishes of the American people and?the voice of public opinion? (Hamilton, 2004, p. 66), by serving a smaller geographical area than senators and working closely with local officials in the district in fulfilling casework requests (Hamilton, 2004, p. 67). Members of Congress are 91 concerned with maintaining their elected position, as Downs (1957) theorized, and will comply to these requests if possible. Contract with America In the 1994 mid term election the Republican Party was successful in uniting members of the House of Representatives around a ?Contract with America? by nationalizing local issues (Brady, Domofrio, & Fiorina, 2000, p. 130). While House elections gradually became more nationalized and less local in focus beginning in the 1970s, 1994 represents the major change point (p. 148). The election produced the first Republican majority in the House of Representatives since the 1950s. Democratic representatives were portrayed as irresponsible for allowing government to grow ?too big, too expensive, and too inept? (Kernell & Jacobson, 2006, p. 219). Hamilton (2004) states: ?Public approval of how Congress is handling its job has typically been low in recent decades, usually hovering around a 40 percent approval rating ? sometimes going higher, sometimes falling below 30 percent? (p. 75). Hibbing and Tiritilli (2000, p. 114) use National Election Studies data to find that disapproval of Congress spiked from 1988 to 1994, to over 70 percent, reaching a level of disapproval comparable to distrust in government in the mid 1970s. They identify ?the public?s willingness to attribute responsibility for the problems of Congress to the majority party (Democrats) and, then, to vote on the basis of that attribution? (p. 115) as evidence that congressional approval linked to majority party identification is relevant to vote choice. 92 National legislation is not always consistent with local norms and ideologies. House members are tugged by the legislative requirements of their position and their responsibilities as a representative of the needs of a constituency more so than senators. Fenno?s observation that many individuals regard congressional institutions as broken and lacking effectiveness paradoxically finds those same individuals tending to favorably embrace their own legislators (as cited in Hibbing & Tiritilli, 2000, p. 110). That the 1994 election produced changes in local races to equal dissatisfaction with the institution as a whole is evidence of nationalization of local issues, with success of Republicans in presidential races and continuing political strength of the South an impetus (Fiorina, 2005, pp. 163-165). Public disapproval of Congress was not extremely elevated in 1994, unlike the disapproval associated with the Democratic Party in Congress. Under the leadership of Speaker of the House Newt Gingrich the newly elected Republican majority convinced voters to hold them responsible within the first 100 days of the session for the provisions of the contract that promised to change the way that government operates, shrink the size government, and reduce taxes to fight collective irresponsibility of members of the House (Riley, 1995, p. 704; Dodd & Oppenheimer, 2005b, 26; Kernell & Jacobson, 2006, p. 219). Budget struggles were the focus of Gingrich?s efforts to gain political dominance. He rallied party leadership support of most bills to ensure cooperation from standing committees as a united push from Republicans in the House for a united legislative agenda ensued (Dodd & Oppenheimer, 2005b, pp. 26-28). 93 Compromise is a big part of making a decision that fits with a representatives? ideology, serves the needs of a constituency, and facilitate the goals of a free society (Hamilton, 2004, p. 87). The Contract with America brought to the fore weaknesses in a candidate-centered electoral process that fails to consider narrow issues or aggregate consequences of policy making. The 1994 election identified the problems of a legislator being individually responsive to a constituency base and neglecting collective responsibility to produce positive aggregate consequences, such as revenue or spending measures (Kernell & Jacobson, 2006, pp. 218-220). This suggests that maximization of aggregate social benefits was a concern of voters in denouncing a Democratic Party that was depicted as a poor manager of financial resources. Senate The U.S. Senate, in comparison, is a smaller body with more prestige, serving a larger, more heterogeneous constituency. Institutional differences exist not only within the structure of the House or Senate as provided through the Constitution, but also from the organization of the chamber as a result of political party effects from the leadership of each chamber. Rules and procedures are largely a function of a political party platform (Kernell & Jacobson, 2006, pp. 228-237). The Contract with America shines a light on many of the institutional differences between the House of Representatives and Senate. By thriving on conservative activism and confrontational political behavior, House members are often at odds with the institutional deliberativeness of wary budget management that characterizes the Senate (Riley 1995, p. 704). One of the reasons for these differences involves the greater scope 94 of the Senate chamber that restrains senators and especially their leadership as party teams, both in the focus of that chamber toward issues facing the entire state as opposed to a particular district and the need to selectively manage problems across a more heterogeneous constituency (Sinclair, 2005, p. 18). Demands placed on a senator reach into foreign affairs and transcend issues directly pertaining to a local or state constituency (Preston, 1969, p. 51). That the Contract applied to the House and its leadership but was merely a glancing blow in the Senate lends evidence to greater institutional effects as opposed to party effects and argues that party control of an institution does not necessarily permeate other institutions in government (Riley, 1995, p. 704). Generating a similar groundswell of support through party effects for legislative agendas is next to impossible in a Senate chamber with staggered terms of office that dilute any immediate effects to organize a party around an issue (p. 705). Sharing of Power with Executive Branch Congress has formal power granted by the Constitution to make laws, but its power also extends into executive matters through the creation of a collection of agencies and bureaus known as bureaucracy. These associations shape values and preferences of legislators and constituents alike and legislative voting is affected by the partisan policies made by agencies. The Constitution gives presidents modest legislative powers but provides a veto as a tool to block or react to congressional proposals (Kernell & Jacobson, 2006, p. 272). It is through the veto pen or the threat of a veto that the executive branch often exercises legislative discretion and impacts the legislative process. Presidents also participate in 95 the legislative process by initiating legislation and submitting mandated budget proposal and working with their party?s legislative leadership for enactment of these policies. Much legislative activity is oriented toward supporting or opposing presidential initiatives passed on political party or ideological considerations. Relationships between executive and legislative roles are increasingly important due to the growth and complexity of the federal government (Kernell & Jacobson, 2006, pp. 309-312). Lawmaking This section introduces and assimilates changes in the formal rules structures and committees of Congress with the proliferation of unorthodox lawmaking resulting from such changes. Constituency concerns are increasingly a factor in formulating a legislative agenda. Media and other forms of communicating between legislator and constituent redefine principle-agent relationships and are a source of political instability and shifting as policy imagery and issue attention changes. An expansion in the role of the presidency and polarization in government characterize ongoing changes in government at the federal level. How these changes affect voting behavior, to the extent that economic efficiency is impacted, is relevant to this analysis. Agent to Constituents Polsby (1968, p. 165) states: ?A United States Congressman has two principle functions: to make laws and to keep laws from being made.? Members of Congress are agents to their constituents and cast votes according to constituency needs, but also in accordance with ideology and self-interest of the legislator. Legislation rarely includes an isolated issue that is considered only once with little or no debate or opportunity to 96 vote on amendments in addition to final passage. A legislator may vote based on opinions and advice of constituents or on how a constituency might perceive an issue if they were to carefully analyze a problem and make informed decisions. Legislators are often unaware of the preferences of a constituency without direct feedback and attempt to balance the desires of a constituency with their self-interest to maintain office. To balance these effects the legislator might straddle issues and vote for killer amendments or strategically cast a vote on final passage when the fate of a bill has been decided, for example (Kernell & Jacobson, 2006, p. 250). Rules and Committees The structure of Congress is designed to enable majorities to make laws and opponents of a bill the opportunity to delay or kill legislation. For this study the partisan influences from House leadership correlate with control of the institution, as evidenced by Republican control after the 1994 election. Whether party control affects only the leadership or also policies that are in the public?s interest is of concern. After a bill is introduced it is assigned a number and referred to a committee (Kernell & Jacobson, 2006, p. 245). Although the most common result is for a bill to die in committee, Sinclair (2000, p. 227) states: ?Congress has long done its serious substantive work on legislation in committees,? a trend that has accelerated since reforms in the 1970s (Hall & Evans, 1990) and is an example of institutional changes that might impact behavior. If a committee decides to proceed with further action, the bill is referred to an appropriate subcommittee for hearings with various groups in attendance testifying about the issue at stake. If a subcommittee decides to act on a bill, 97 it edits the bill line by line and reports to the full committee, which accepts, rejects, or amends a bill. In the House rules or resolutions frequently specify the procedure for limiting debate of legislation through the House Rules Committee (Kernell & Jacobson, 2006, pp. 246-247). Restrictive or closed rules keep unwanted amendments off the agenda and became a tool of partisan leaders in the House in the 1980s and 1990s to limit minority party debate (Schickler & Pearson, 2005, p. 210). Since a majority vote is necessary on the floor to adopt a rule, the effect is for the House to constrain itself. Debate of a bill is divided equally between proponents and opponents of the legislation, who attempt to make a case that the policy serve the public interest. House leaders induce members through control of agenda by framing issues around party principles (Kernell & Jacobson, 2006, p. 251). The Senate has no equivalent of the House Rules Committee, but arranges orderly consideration of debate, allowable legislation, and procedures through unanimous consent agreements. Without such agreements, individuals can filibuster by making endless speeches and blocking action on a bill. Cooperation among senators is necessary to practice mutual constraint and bipartisan cooperation (Kernell & Jacobson, 2006, p. 248). Floor action is more important in the Senate than the House, as Senate leaders, committees, or subcommittees wield relatively less influence in the Senate chamber. Growth in subcommittees has reduced the influence of seniority rule (Preston, 1969, p. 123). 98 Unorthodox Lawmaking Lawmaking does not always follow a linear path from introduction of a bill, to debate, and eventual signature. Sinclair (2000) explains that increases in workloads and strategic behavior of individuals within the political environment are determinants of procedure, especially in the House. Bills are often considered by multiple committees or bypass committees altogether and call for the measure to be brought to the floor. Increases in multiple referral of legislation mirrors attempts to reform committee jurisdiction and increase broad participation in the legislative process. Party leaders initiate compromises and play a proactive role in negotiations. House reforms in the 1970s shifted the distribution of influence from committees to subcommittees, with more emphasis on party leadership (Sinclair, 2000, p. 84), heightening struggles between party and legislative voting behavior. Decentralized power in the Senate shifted to greater individualism after the 1950s, when activism in the number and size of groups expanded. Senators are hesitant to curtail enormous opportunities for influence. Legislating in the Senate is more problematic as internal reforms fail to provide the tools for negotiations and debate necessary for bringing together disparate individuals or groups to compromise on the issues considered, leading to extended debate and mounting gridlock (Sinclair, 2000, p. 235). Constituency Preferences: The Role of Media Most constituents know very little about the specifics of an issue on which a member of Congress will vote and for those constituents that do express an interest in 99 the political process, their opinions vary widely (Kernell & Jacobson, 2006, p. 251). Perceptions are a major determinant of how one interprets an issue and play a role in problem definition. Policy issues are usually complex and, as Simon argues, require structuring through government to contribute to problem definition (as cited in Parsons, 1995, p. 89). Turner and Schneier find that if a legislator votes consistently against his or her party, district pressure from a constituency is a behavioral cue that may be in part responsible for such division of party loyalty (as cited in Fiorina, 1974, p. 3). Media shapes how a constituency considers an event by defining the event as a problem and magnifying the issue to stir response (Parson, 1995, pp. 106-107). Baumgartner and Jones recognize the activities of media as a major source of instability that affects the imagery associated with an issue and shifts attention to an issue, to different aspects of the same issue, or lurches to another issue. As attention shifts within a constituency and within institutions, decisions are reinforced within the institution and policies formulated as a function of those decisions (as cited in Parsons, 1995, pp. 204- 206). Regional Congressional Patterns since World War Two A significant shift in the balance of congressional power from Congress to the presidency has characterized the latter half of the 20 th century (Cooper, 2005, p. 363). The Great Depression was a turning point in the relationship between the executive and legislative branches of government. The New Deal created a strong central state, ?tying individual sectors of the private economy to government regulatory policy and subsidization? (Benzel, 1984, p. 152). A strong committee system was a major force in 100 both houses of Congress, but leadership was relatively weak, allowing rifts between northern and southern Democrats to exacerbate (Cooper, 2005, p. 381). The role of the president increasingly expanded in the mid twentieth century, as congressional party leaders faced with divided parties and limited organizational power relied on the president to set policy goals and guide policy direction (Kernell & Jacobson, 2006, pp. 276-277). Party voting continued to decline as the southern Democratic-Republican coalition continued, weakening majority leadership in the 1950s and 1960s. In the 1970s party politics strengthened and increasing ideological divisions widened between liberal Democrats and conservative Republicans. Congressional parties framed policies in relation to the position of the president, as migrations to southern states solidified Republican strength in those states with an ideological chasm more polarized and party unity votes more consistent (Cooper, 2005, p. 383). Since the late 1980s, party unity, partisanship, message driven politics, and polarization define the features of Congress. Higher levels of party voting and Democratic leadership power in the late 1980s and early 1990s transitioned to a Republican House in 1995. The role of seniority became less important as strong leadership directed party message and committee outcomes, resulting in partisan conflicts that punished minority Democrats for opposing Republican policy initiatives. Narrow majority margins reinforce the role of strong leadership (Cooper, 2005, p. 384). Although very individualistic, partisan hostility is also a characteristic of a contemporary Senate. Actions on the floor of the Senate frequently supersede committee 101 action, as party leaders attempt to balance gaining public favor with bipartisan support (Kernell & Jacobson, 2006, p. 230; Cooper, 2005, pp. 384-385). Institutionalization Institutionalization is a process of evolutionary changes in an organization that are necessary for that organization to survive and fulfill its mission. This study considers institutionalization in Congress in an attempt to measure how changes in the institution affect voting behavior of its members. Correlating with divided or unified control of government to be discussed later, institutionalization addresses internal processes that affect individual decision-making. To the extent that congressional institutions change internally, how these changes affect policy formulation is important to this analysis. As an organization institutionalizes it not only survives but also persists over time by becoming more durable. Polsby (1968, p. 145) finds that an institutionalized organization selectively recruits members from within, increases in complexity with internal functions defined and a division of labor specified, and is governed by universal rules and precedents. The significance of institutionalization involves goal displacement of how changes in external processes correlate with external demands facing an organization. Behavior changes as inputs into an organization are processed to produce outputs. As an organization becomes more institutionalized changes in the internal environment of an organization affects this process and in turn influences how decision makers analyze organizational goals. The goals of the organization might shift as the process to reach those goals shifts and affect outcomes (Canon, 1989, p. 415). 102 While an institutionalized organization is durable and provides a defined structure, in a study of politics institutionalization represents specialization in government in order to protect freedom and democracy. An institutionalized political organization is stronger and more capable of protecting constituency groups and containing political opposition (Polsby, 1968, p. 144). Individual freedoms in a society require structure to legitimize political institutions that represent large and diverse constituencies. Institutionalization of Congress Institutionalization of Congress is best explained through increasing population densities of a society requiring a greater division of labor that varies in direct ratio. Durkheim finds that this phenomenon is evident as agencies of the federal government institutionalize in response to a larger role of the federal government to the national economy (as cited in Polsby, 1968, p. 164). Greater development of society necessitates an integrated, more developed government to provide services to its constituents. Hall supports a macro interpretation of formal rules and procedures that define political institutions and economic consequences of policy-making, in supporting constraints on decision-making provided by divisions of labor (as cited in Parsons, 1995, pp. 333-336). As a legislative body, members of Congress individually perform many roles as lawmakers. Policies that are formulated and outcomes of policies that are implemented are functions of the institutional structure of the legislature. Polsby (1968) considers lack of turnover in the House (p. 146) and increases in years in office before a member becomes Speaker of the House (p. 148) as indications of increasing institutionalization. 103 Several distinctions exist between institutionalization in the House and Senate affecting how members perform their roles and the political party to which the member belongs. Canon (1989) finds that the Senate is affected to a larger extent than the House by which party controls the presidency and leadership in the Senate is less durable over time as a result (p. 418). Deering finds that Senate leadership is more personal as opposed to institutional as committee structures are not well defined and potentially change greatly as leadership changes (as cited in Canon, 1989, p. 419), with Democratic leadership more stable and structured than Republican leadership (p. 424). As Congress becomes more institutionalized, a decentralization of power is a call for specialists with extensive knowledge within identified areas of government, as opposed to generalists with few specialized skills (Polsby, 1968; Canon, 1989). A narrower, more specific focus of a specialist increases the relative and absolute power of the legislator within the House or Senate chamber and attracts legislators to the political process. Incentives that attract specialists are a clue that ideologies and self-interests that motivate lawmakers will differ depending on levels of expertise of the legislator. With greater specialization it is possible to argue that more specialized roles are less likely to be economically efficient as legislators pursue narrow political agendas. A good example of the virtues of institutionalization in Congress is illustrated through the relationship of the institution to the constituency it serves. The House and Senate each are more inclusive of constituents and service oriented to serving their needs. This may indicate that a greater emphasis on social benefits of policies is increasingly important. 104 The Logic of Political Parties A political party is a coalition of individuals seeking to control the mechanics of government through the political process. This section introduces the logic behind political parties that is developed fully in the next section. With parties playing an integral role in structuring ideological positions and defining a political response to an issue, the message transmitted through such a party vehicle has important consequences to interactions in the political arena. Organizing around party principles gives a unified voice to issue events and creates a benchmark for decision-making, especially along party lines where Republican and Democrat extremes are defined along ideological boundaries. Political Parties Defined Jones (2001, p. 152) explains that party discipline is the most important factor in accounting for floor votes, although it is not absolute. Political parties are not specifically identified in the Constitution but translate public preferences into public policies by coordinating group activities (Kernell & Jacobson, 2006, p. 646). Edmund Burke finds that a political party is a group of individuals that unite based on individual principles to jointly promote a national interest (as cited in Kernell & Jacobson, 2006, p. 463). For Downs a definition of political party introduces aspirations for public office. Downs (1957) defines political parties as ?a team of men seeking to control the governing apparatus by gaining office in a duly constituted election? (p. 25). Schatteschneider considers links between political parties and democracies and finds that modern democracy is unthinkable without parties to recruit and train leaders, foster 105 political involvement and action, and collectively organize citizens and leaders in coalitions to hold elected agents accountable to the needs of their constituency (as cited in Reichley, 1992, p. 3; Kernell & Jacobson, 2006, p. 462). In sum, political parties play an important role in bringing together groups of individuals by structuring support for political initiatives. According to Reichley (1992, p. 414), parties are important to democracy by maintaining a productive balance between accountability and effective government. Parties give ordinary citizens a voice in government and provide a political base for elected officials. Parties offer a means of organizing dissent against incumbent administration policies and are natural guardians of civil liberties protecting rights of free expression. Parties increase honesty in the political process by exposing corruption and deception by its opposition. By encouraging voter registration, participation, and recruitment, parties strengthen the democratic process. The Republican and Democratic parties influence congressional voting by measurable ideological differences (Sinclair, 1977; Reichley, 1992, pp. 353-354; Knuckey, 2005). Congressional party leaders also follow Downsian definitions of self- interest to gain election of party faithful to Congress or the presidency by maintaining leadership roles and pushing party platforms. As national organizations, the Republican and Democratic parties play integral roles in shaping voter demands and policy agendas (Parsons, 1995, pp. 220-222). 106 Political Party Development in America Political party development parallels many of the ongoing social and economic changes in America and is a foundation under issue coalitions that grow from ideological divisions. These divisions are important to this study as the effects of party are measured through unity as an independent variable. Liberal-conservative ideology and its relationship to economically efficient policy positions is firmly rooted in political party development. Separating supporters of these coalitions into major camps is consistent with democratic principles giving a voice to those that might not otherwise be heard and defining the expectations of the role of government. To the extent that economic efficiency transcends ideology, ceteris paribus, party unity among legislators is an illustration of increasing support for the party?s ideology relative to the legislator?s ideology or self-interest. Through shared power among separate institutions, Fisher (1998, pp. 4-6) argues that politics is a venue for competition among economic and social interests, where representation at the state and federal levels of government and within executive and legislative branches of government is sufficient to balance these interests and make organized parties unnecessary (Reichley, 1992, pp. 28-29). Parties are an extension of these competitive forces and are not mutually exclusive from shared powers at any level of government. The framers of the U. S. Constitution may not have intended for political parties to be rooted in the American political landscape, but the alliances and coalitions necessary for sharing power among institutions of government naturally spawned party organizations. Political parties begin appearing when opposing visions of the nation?s 107 future required a solid core of support to attract the majority support required to pass legislation (Kernell & Jacobson, 2006, p. 464). To control policy, legislative leaders find forming alliances around supporters reduces transaction costs of devising a winning coalition of supporters. Participants have to agree to cooperate on an action, either out of shared interests and values or self-interests that serve their purposes. From organized competition for votes comes the need for sustained political efforts to hold the coalitions together around issues that the group supports in congressional elections. Adopting a party label offers an informative means of distinguishing candidates and platforms by associating groups with the political position of the party organization (pp. 465-466). Party Camps While most modern democracies have more than two parties, in America national leaders gradually divided into two major camps during the first few Congresses, a pattern that has continued (Kernell & Jacobson, 2006, p. 467). The camps pitted the political philosophies of Alexander Hamilton and Thomas Jefferson against each other. Each espoused political ideas that formed a basis for liberal-conservative ideologies that gradually evolved into a two-party system. Alexander Hamilton argued for a strong national government supporting commerce and manufacturing industries in the long-term interests of the nation and pursued ambitious policy efforts as Washington?s treasury secretary in forming a national bank and fostering the economic interests of the affluent (Reichley, 1992, pp. 38-39). His proposals found allies in supporters of Constitutional ratification with a 108 penchant for strong national government. This group was known as the Federalists (Kernell & Jacobson, 2006, p. 471). Thomas Jefferson and James Madison represented unified opposition to Hamilton. The issue that concerned Jefferson and Madison was not the viability of the national government or the economic interests of the country, but rather social equality that recognizes the rights of individuals in pursuing their interests (Reichley, 1992, p. 66). Members of this party were called the Antifederalists, which were also known as the Democratic-Republicans, a precursor to the contemporary Democratic Party, the oldest political party in the world (Kernell & Jacobson, 2006, p. 471). These major camps separated into traditions defined as republican and liberal, with Republican traditions representing conservative alliances. Each camp is committed to constitutional protection of individual liberties and representative government, but differs on the priorities they assign to public order, economic growth, and social and economic equality (Reichley, 1992, pp. 4-5). Grafton and Permaloff (2005a) cite a model by Janda, Berry, and Goldman to argue this point. According to the model, conservatives rank social and economic order as most important, followed by freedom, and equality; liberals reverse the ranking and favor equality, freedom, and social and economic order, respectively. This model gives credence to Federalist economic positions that seek order and Democratic-Republican desire for equality. By applying the principles of this model to political party development, one can trace the origins of modern two-party ideological distinctions to its roots. 109 Liberal-Conservative Traditions For this study it is important to note the close association between party and ideology. From the two ideological traditions the following major political parties formed in American national politics: the conservative or republican tradition to the Federalists, National Republicans, Whigs, and modern Republicans; and the liberal tradition to the Antifederalists, Democratic-Republicans, and modern Democrats (Reichley, 1992, p. 6). Both liberals and conservatives accept the responsibility of government to promote the general welfare of its citizens, but they recognize differing paths to reach that objective. Reichley (1992) states: The liberal tradition particularly in the twentieth century has tended to identify such promotion with direct government intervention and support, while the republican (conservative) tradition has emphasized government?s role in securing economic and social conditions favorable to individual, family, and community achievement. (p. 5) Conservative positions are not anathema to curing economic malaise or addressing social problems but generally prefer market forces or incentives for private sector involvement to address a problem as opposed to an expanded state for such purposes. Stigler (1965) believes ?that abuses of private power will usually be checked, and incitements to efficiency and progress usually provided, by forces of competition? (p. 53). Coalitions among political parties are important to this study as values and ideals of the party organization pull legislative behavior toward party principles. Political parties are not dichotomous, ideological systems that do not change over time. Rather 110 the liberal-conservative basis that roots the infancy of party government in America derives characteristics from an era?s social, economic, and technology foundation, and the goals and leadership of political players that unite institutional incentives and responses to problems (Kernell & Jacobson, 2006, p. 471). That is, party coalitions consist of those groups with similar political ideas and values, but each public policy debated and formulated is a function of the distinctive set of values of that era. Platforms adopted by political parties change over time as values and institutions also change. Party Unity In the context of roll call voting analysis party unity represents how closely a legislator votes with his or her political party. Why legislators support party positions is an important area of study to explain the legislative process. Voting blocks within Congress are increasing as measured by party unity scores, 7 enhancing the influence of national parties in the congressional arena. That the effect of party is enduring in American politics is well documented (Reichley, 1992; Leyden & Borrelli, 1990; Cantor & Herrnson, 1997; Snyder & Groseclose, 2000). Two trends that appear to correlate with increasing party unity are the role of majority or minority status within Congress and cohesiveness within that political party and campaign finance activity on influencing party line votes. In analyzing the House of Representatives from 1901-1956, Sinclair (1977) considers party unity or cohesion to study the influence of presidential popular vote, size of House majority, divided control between the House and presidency, and change in 7 Refer to Congressional Quarterly Congress Collection (http://www.cq.com) for party unity scores calculated by year. 111 party control of the House. Sinclair?s research found that the influence of party is greater than whether the party is in the majority or minority status. Democrats are more unified within their party the greater the Democratic presidential popular vote and the shorter the time the Democratic majority has been in power. Republican cohesiveness is highest when pressures to deviate away from the party?s conservative base are least. Party unity also has a self-interested component as well, as noted by Leyden and Borrelli (1990) in analyzing linkages between contributions to political parties and party unity. Their research documents how monetary assistance to parties competes with contributions to PACS, with campaign finance laws placing ceilings on contributions to parties. Political Party Control: Divided or Unified Government Divided or unified government refers to control of the legislative and executive branches of government by the same or different political parties. The impact of political party control on the legislative process is an important component to this study. Liberal- conservative ideology is rooted within each party, but can also be measured within each legislator. Divided or unified control of government represents a struggle between party ideology and the propensity for a legislator to espouse economically efficient public policies. Changes in control of government offer an opportunity to analyze the role of economic efficiency as a predictor of legislative behavior. Government is unified if House, Senate, and Executive Branch are all controlled by the same political party; otherwise government is nonunified or divided. Congress is not considered as one body but as two separate institutions of House and Senate. 112 Menefee-Libey (1991) defines divided government as ?partisan conflict between the executive and legislative branches? (p. 643) or when the president?s party does not control the legislative branch. To say that the American system of government provides a structural basis that acts as a check to balance the effect of any institution or group is correct, but does not explain why citizens choose divided government or whether divided government produces more desirable outcomes. To the extent that political party control of Congress affects economically efficient policy making is of particular interest. It is important to reiterate distinctions between party unity and political party control of government. Party unity is a measure of how closely a legislator?s voting record reflects the political positions of his or her party. Political party control of government refers to control of the institution of government (House, Senate, or Executive Branch) through majority party or minority party status. Two primary areas of discussion surrounding divided political party control are divided government as a natural extension of constitutionally separating powers and policy balancing effects. Party Balancing With political control a concern in explaining economically efficient voting, why control changes and how those changes affect legislative voting are important in considering efficient outcomes. Comparisons of voter ideology reveal preferences for unified or divided government. Carsey and Layman (2004) find that party control of government is one of the defining features of contemporary American politics. By examining citizen preferences for unified or divided government, their research supports party balancing as an 113 explanation for divided government. Party balancing is to select legislators from the opposite political party to balance political control of government across legislative and executive institutions. Citizens have an opportunity to select the agents to represent their interests and create unified or divided party control. An abundance of party balancing literature (e.g., Alesina & Rosenthal, 1989, 1995; Fiorina, 1994, 1996; Ingberman & Villani, 1993; Lacy & Niou, 1998; Lacy & Paolino, 1998; Mebane, 2000; Scheve & Tomz, 1999; Tarrance & DeVries, 1998; Smith, Brown, Bruce, & Overby, 1999) suggests that dividing support for candidates across both political parties is a purposeful vote to balance political party control and achieve moderate public policies (Burden & Kimball, 2002, p. 24). Policy balancing theory rests on the assumption that voters prefer a combination of party control of the presidency and Congress that produces outcomes in a liberal- conservative spectrum most closely resembling the voters? ideological spectrum. It is interesting that party polarization in the 1890s unified government, but in the late 1990s divided government (Smith & Gamm, 2005, p. 195). Voters must base this choice on what they believe is the ideological distance between each party and the ideological distance between the voter?s ideology and the ideology of either political party (Carsey & Layman, 2004, pp. 541-542). The theory considers those voters with extreme ideological views and moderate ideological views and holds that voters with extreme ideological views prefer Congress and the president to be of the same party, but those voters who are ideological moderates prefer split government. Those voters with extreme views favor the same party controlling Congress and the presidency in order to produce policies consistent 114 with their extreme ideology, while moderate voters seek compromise, producing relatively moderate policy outcomes, for ?it is easier for voters to cross party lines when they do not have to travel far along the ideological spectrum? (Burden & Kimball, 2002, p. 26). An important issue when considering party balancing literature involves the extent that voters consider the two parties as polarized. If the voters do not perceive a discernable difference between each party?s ideology, voters might use other criteria, such as government efficiency or accountability, when deciding how to vote (Carsey & Layman, 2004, p. 542). Conditional Party Government Model Heightened party conflict beginning in the 1970s and 1980s is an indication of a resurgent party system (Coleman, 1997; Hager & Talbert, 2000). Increasing party unity scores indicate that legislators are voting along party lines, but do not explain the rise and fall of party influence over time or the homogeneity or heterogeneity of a legislator?s preferences. Rohde argues that for a political issue ?the influence of party would be felt when the preferences of members ? either their personal preferences or those induced by their constituents? desires ? divide along party lines and when preferences within the majority are homogenous? (as cited in Moscardelli, Haspel, & Wike, 1998, p. 693). Legislators are willing to cede power to the legislative party when each shares homogenous party preferences, but to a much smaller extent when those preferences are dissimilar or heterogeneous. 115 The conditional party government model argues that parties matter under certain conditions and that an internally unified majority party has much discretion in changing rules over time to tighten agenda control (Schickler, 2000, p. 270; Aldrich & Rohde, 2005, pp. 265-266). The model defines the degree of authority delegated to and exercised by congressional party leaders as conditioned by the extent that this ideological consensus exists among legislators (Kernell & Jacobson, 2006, p. 230). Aldrich and Rohde label a conditional party government through the cohesiveness and polarization that majority party members provide their leaders to pursue median positions on legislation (as cited in Dodd & Oppenheimer, 2005b, pp. 47-48). According to Cox and McCubbins, the relative homogeneity of party members is positively correlated with greater automatic support of policy decisions made by party leaders, where parties are in a sense a ?legislative cartel? that derives its power by forming rules that govern legislative decision-making (as cited in Hager & Talbert, 2000, p. 78). To the extent that ideology and self-interests tie to reelection and lead to greater leadership control, the model is consistent with increasing institutionalization within Congress, especially the House. Why political parties declined in the early twentieth century, but recently experienced resurgence is a function of the majority party?s ability to shape legislative institutions by enacting rules changes that strengthen its agenda. Schickler (2000, p. 270) notes that changes in rules and thus agenda control occur over time, as shifts in the ideological balance of power on the floor have a greater effect than internal characteristics of the majority party. Owing to the conditional aspect of control, Fiorina finds that the ebb and flow of party influences on legislative preferences in 116 congressional politics is a basis for the conditional party government model that explains leadership control at any point in time (as cited in Moscardelli, Haspel, & Wike, 1998, p. 692). Leadership control is a function of changes in the legislative environment and between constituents and their legislators. Increases in party polarization in the 1980s and 1990s offer clues to the conditional aspects of the model after long periods where parties were less important to legislative strategies (Schickler, 2000, 269; Roberts & Smith, 2003). That the model reflects changes in legislative interrelationships over time justifies its inclusion in this research in arguing that the Kennedy E-score is relatively static and based on a criterion ? efficiency ? that is an integral part of the market models supported by liberals and conservatives, Democrats and Republicans. Conditional Governance and Legislative Preferences In the 1970s the resurgence of the role of parties in the political process mirrored not only institutionalization in Congress, but also introduced changes in the electoral functions of the legislator relative to his or her constituency. Partisan ties within the electorate weakened but parties still offered a brand name by which to distinguish legislators (Hager & Talbert, 2000, p. 77). Jacobson and Mayhew find that brand name is expensive to maintain and only one factor in the reelection of a legislator, with members voting their own or constituents? preferences (as cited in Hager & Talbert, 2000, p. 77). The strength of party leadership supports the party brand as a collective good, although each legislator has the incentive to free ride by voting his or her preferences. 117 Sharing collective interests leads to sharing of collective behavior and that behavior is often exhibited through party leadership (Forgette & Sala, 1999, p. 483). Centralized leadership finds the party leader of the majority party controlling the agenda and influencing legislative voting. Decentralized leadership places power disproportionately in the hands of committees, but power in Congress shifts along a centralized- decentralized continuum in the House and often in the hands of the full chamber in the Senate (Smith & Gamm, 2005, p. 182). Polarizing electoral coalitions are a key element in shaping policy preferences as power shifts along the continuum (Roberts & Smith, 2003, p. 306). Leadership and Policy Outcomes Individual members of Congress have preferences for policies in accordance with their ideology and self-interests. To the extent that changes in collective behavior produce legislative decisions that are inconsistent with economically efficient outcomes, the effect of such collective behavior is important to this study. Cooper and Brady (1981) find that the degree of polarization in the electorate determines the strength of polarization in the congressional parties. Leadership style and individual legislative characteristics are also important to their model in contrast to conditional party government that asserts that strong party leadership makes a difference. When party strength is substantial, power is concentrated and leaders are goal oriented, but when party strength is low, power is dispersed and leaders are oriented to building relationships. (p. 424) 118 For example, in comparison to the relatively high internal conflict of Democratic majorities in the 1950s and 1960s, the House after 1994 was polarized and under the control of a unified Republican party membership willing to cede power to a strong leadership. This enabled the Republicans to slow appropriation increases in the late 1990s and enact tax cuts during the George W. Bush presidency. With conditional government theory producing individual-level behavioral expectations, Moscardelli, Haspel, and Wike (1998) find that a legislator is sensitive to his or her ideological distance to the left or right of party leadership in supporting such policies. Extending E- scores to the time period covered in this research may help us to better capture some of the effects at work during this period. Party members have the incentive to vote their constituents? wishes for reelection purposes, but they struggle with the public?s perception of the party in addition to the legislator?s ideological position in relation to the party?s (Hager & Talbert, 2000, p. 76). More centralized leadership after the 1994 congressional election allowed party leaders an opportunity to promote and pass a national agenda as part of the Contract with America. To the extent that legislators disagree with the party agenda and find themselves ideologically incompatible with party positions, those legislators probably support the policy preferences of their constituents. Moscardelli, Haspel, and Wike?s (1998, p. 699) findings illustrate that only those members who share a close ideology with their leaders are willing to support the party?s agenda for the sake of party building, suggesting the role of self-interest (reelection) and ideology occupy important roles in the legislator?s decision agenda. That change in leadership and the extent of leadership 119 control are considered in this analysis, how legislators vote before and after the changes resulting from the 1994 is important to this study. Political Party Alignment A political party alignment is a partisan coalition that identifies with a political party and is loyal to its principles (Flanigan & Zingale, 2002, pp. 57-63). Alignment impacts this study through voter identification with a political party as a long-term force in political campaigns. To the extent that constituents identify with party, liberal- conservative ideology of the legislator and the self-interests for holding office offer an immediate comparison to policy decisions made by each legislator. The primary focus of realignment literature is directed toward partisan identification of voters. Campbell 8 identifies party identification as a long term, stable force that influences the electorate and issues in the campaign are short-term forces (as cited in Abramowitz & Saunders, 1998, p. 634). V. O. Key makes clear that changes or cleavages in constituency bases of voting behavior among groups finds some segments moving toward one of the two political parties and others moving away or maintaining a previous pattern of support for that party (as cited in Lawrence & Fleisher, 1987, pp. 80- 81). By altering the constituency bases of congressional parties, the impact of partisan realignments affects the composition of government, such as seats held by each party in Congress, and policy formulation resulting from partisan influences (Waterman, 1990), which affects institutional roles of government and policy outputs (Schatteschneider, 1960, pp. 78-96) in building congressional majorities (Brady, 1978, p. 80). 8 See Campbell, Converse, Miller, & Stokes (1960) for a complete discussion of long term and short term forces. 120 According to Burnham, significant public policy transformations are consistent with realigning elections arising from emergent issues that are policy driven, leading to changes in policy outputs to alleviate tensions in the electorate (as cited in Brady, 1978, p. 79). Policy discussion often centers on how the new legislators in the majority party set the political agenda. If alignment is a constituent act, then newly elected legislators will push policy changes to alleviate tensions within the electorate (Champagne, 1983), a phenomenon consistent with Fiorina?s finding that changes in congressmen are the only reliable manner to achieve public policy change in Congress (as cited in Champagne, 1983, pp. 246-247). The 1990s: A Decade of Realignment? The alignment of the Democratic Party around New Deal principles produced a major electoral advantage for that party for nearly 50 years after the Great Depression (Abramowitz & Saunders, 1998, p. 635; Flanigan & Zingale, 2002, p. 68). This alignment of political forces is consistent with Berkowitz and McQuaid?s (1978) contention that an expansion of social welfare in the U.S. until the 1960s is behind the larger role for the federal government today. Strong Republican efforts in the presidential campaigns beginning in the late 1960s and intensifying in the 1980s were harbingers for an electoral shift that eventually consumed Congress. These electoral shifts represented dealignment that would eventually be followed by realignment. The 1964 presidential election was a turning point for a conservative shift to the Republican Party, where the ideological differences between each party gradually widened (Flanigan & Zingale, 2002, pp. 72-73; Kernell & 121 Jacobson, 2006, p. 496). Abramowitz and Saunders (1998) explain that the continuing electoral shift is evidence of realignment. Increasing Republican gains in the South and a gradual increase in the proportion of the electorate identifying with the Republican Party vis-?-vis the Democratic Party have solidified an intergenerational pull (p. 638). Conversely, Flanigan and Zingale (2002) recognize the growth of independent voters who pledge no allegiance to political parties as part of a continuing shift that characterizes the electorate today (p. 69), with neither party successfully luring independent voters nor managing to control all branches of government in the 1980s or 1990s for a significant period of time (Abramowitz & Saunders, 1998). Flanigan and Zingale (2002) identify this period is a continuation of dealignment (p. 73), where voters are most susceptible to short term forces such as personal appeal of a candidate or local issues (p. 65). Summary The focus of this dissertation is whether a measure of economic efficiency through an E-score is a better predictor of legislative behavior than traditional measures of ideology and self-interest. Considering evidence analyzed from a review of the literature, it is possible to identify variables that are relevant to this study. It appears that in measuring voting behavior, the dependent variable should reflect an economic factor where benefit and cost can be identified and measured. For example, Tolluck identifies redistribution of resources as a prevailing function of modern government. Serving the public?s interest introduces welfare implications to policy decisions that market models address, but are inadequate without considering 122 resources that maximize net social benefits. Another example is Clausen?s findings of constituency influence on a legislator in social areas, but not concerning social welfare distribution or agricultural support, as evidence of economic measures when analyzing public interests. The review of existing research also identifies several variables important to this study as independent variables producing changes in voting behavior. A liberal- conservative spectrum functions well in explaining characteristics of a legislator or voter, but not in predicting consequences of a legislator?s voting decisions. Developing an efficiency index based on the Kennedy model is consistent with Pareto and Kaldor- Hicks principles for maximizing public policy resources. Controlling for the effects of institutionalization and party control (unified or divided) of government is important to this study in measuring how changes in the institution of Congress changes behavior. Extending the Kennedy model is an answer to limitations in Kennedy?s research that failed to include these institutional or chamber effects. Existing research shows that over time there has been increasing party unity within Congress, ideological differences increasing between the political parties with fewer moderates elected, a greater level of party leadership control over both institutions but the House in particular, and other changes. How these changes affect legislative voting will expand the Kennedy study as will including consideration of length of service and the year of a senator?s term when a vote is cast. Changes in the institution itself as power shifts to the presidency and executive-legislative relationships grow in importance invite inclusion of independent variables that produce changes in behavior. 123 Examples include ideological and/or party divisions measured between the executive and legislative branches. Constituency alignment with political parties and polarization within the electorate are examples of partisan influences on changing the composition of government. Party alignment is an example of constituents pushing policy development. The effect of the Contract with America in nationalizing local issues raises questions concerning not only party unity around a party platform, but also alignment within the constituency in supporting those legislators. The effect of each interaction is important in analyzing changes in voting behavior and its effects on economic efficiency. The next chapter outlines the methodology used to accomplish the research. 124 CHAPTER THREE CONCEPTUAL DESIGN AND RESEARCH METHODOLOY The conceptual design that guides this study is presented in this chapter. The research question that this dissertation seeks to answer is: Does economic efficiency through an E-score function better than a traditional spectrum of liberal-conservative ideology in explaining the ideological position of a representative (House and / or Senate member), congressional activity, and public policy formulation? To answer this question the study centers on the following three areas: 1) ideology and self-interest as measures of behavior, 2) economic efficiency as a macroeconomic goal of policy makers that transcends traditionally accepted measures of liberal-conservative ideology, and 3) extending the e-score to additional time periods to capture the effect of changes (institutional and political party alignment and control) on legislative voting. This chapter begins by developing an argument that identifies positions taken on medical malpractice reforms and increasing the federal minimum wage as two public policy areas where an examination of economic efficiency is possible. Each policy area is considered separately as a dependent variable. Then, independent and control variables are analyzed within vectors that group such variables around related issues. An overview of each vector is presented, along with a discussion of the relationship between the variables within each vector and in relationships between the vectors. The three vectors are ideology, self-interest, and chamber environment. 125 Independent variables in the ideology vector include measures of liberalism (ADA score), conservatism (ACU score), a spectrum of weighted liberalism and conservatism across time (DW-NOMINATE), and economic efficiency (E-scores). Independent variables in the self-interest vector include direct or PAC contributions to legislators from interest groups with ties to the policy area of the dependent variable. The independent variables in the chamber environment vector include party unity, a measure of how closely a legislator votes in accordance with majority of the members of his or her political party, and various party control patterns of the House, Senate and presidency. Control variables are included in the analysis as constants that do not causally influence the dependent variable. These control variables are party control of the institution (House or Senate) compared to the legislator?s party, geographical region of the legislator?s constituency (Northeast, Midwest, South, and West), and measures of economic conditions in each state and nationally. For specific policy areas identified by each dependent variable one or more control variables represent conditions in the legislator?s state or district that might impact her/his roll call vote. For example, in the case of medical malpractice reform, whether a legislator?s state is in a malpractice crisis or not is used to test for constituency self-interest impact on the legislator?s vote. This chapter also includes discussion of data collection, units of analysis, expanded E-score development, coding of the variables, and types of analysis conducted (regression and interrupted time series analysis). Finally, the chapter concludes with hypothesis development. 126 Dependent Variables Medical malpractice tort reform and minimum wage issues are two policy areas analyzed as separate dependent variables within the model tested. Medical malpractice reforms are considered to be economically efficient public policies and increases in minimum wages are considered to be economically inefficient. Medical Malpractice Tort law provides a mechanism to compensate the injured party in medical malpractice issues for the harm inflicted upon them by the tortfeasor (Rubin, 1995, p. 4). In health care this means that patients wrongfully harmed due to malpractice of a physician or other medical professional should be compensated for the losses associated with that injury (Stailey, 2004, p. 198). A second primary function of tort law is to modify the behavior of the transgressor. The idea is that tort law encourages people to act in a responsible manner (Krauss, 2003, p. 357). In the area of medical malpractice, tort law probably helps deter malpractice (Rubin, 1995; Stailey, 2004, p. 198). Tort law should, at least theoretically, balance these two competing interests (Boozer, Westley, & Landry, working paper under review). In economically efficient terms, an argument is made that tort law is wrecking the American economy, where tort liability is a tax on everyone (Krauss & Levy, 2004, p. 2). In 2003 per capita costs of the tort system are as much as $809 per person, which is ?the equivalent of a 5 percent tax of wages? (p. 12). The tort system is estimated to be nearly 2.25 percent of the United States? gross domestic product (p. 2) and represents a drain on society. 127 Defining high medical malpractice insurance premiums and the current state of tort law as a crisis is a matter of interpretation. There are two schools of thought. Proponents of reform, the normativists, argue that tort reform is necessary to achieve greater efficiency, while positivists view tort law as already largely efficient (Note, 1996, p. 1765). Normativists suggest that medical care costs are high due in large part to excessive tort judgments and that medical malpractice reform will lower tort judgments and consequently the cost of medical care via lower malpractice insurance costs. The logic behind this argument is that high tort judgments lead to a greater number of malpractice suits and increased rates for malpractice insurance. The result is economically inefficient outcomes, such as the misallocation of resources, increased awards to rent-seeking behavior, and higher medical costs to reflect the price premium paid for malpractice insurance due to excessive tort judgments. The positivist view opposes tort reform arguing that evidence shows that most tort judgments are modest and over time the magnitude and incidence of large judgments have not varied greatly (Note, 1996, p. 1773). The perceived inefficiency from large tort judgments, the medical malpractice crisis, may be the result of the underwriting cycle. During times of high returns and general economic growth, premiums are lower and insurers over-extend, but as financial conditions worsen the returns diminish and premiums increase (Stailey, 2004, p. 195). Both interpretations, however, explain a misallocation of resources resulting from the malpractice crisis, suggesting an economically inefficient use of resources and a state in which Pareto improvement exists. Decreases in net benefits to society associated with tort law represent economic inefficiency. Legislation to reform tort laws and lessen the 128 drain on society through medical malpractices is economically efficient (Boozer, Westley, & Landry, working paper under review). Minimum Wage The Fair Labor Standards Act 1 of 1938 established, among other public policies, minimum wage standards. A consensus exists that the goal of the legislation was to target poverty. Ellwood finds that the Act targeted raising wage levels to a minimum threshold, while Johnson and Browning conclude that redistribution of income to low- income households was the intent (as cited in Sobel, 1999, p. 763). Manipulation of wage rates upsets labor markets and reduces economic efficiency as employers struggle with higher labor costs. National standards replaced regional standards with legislators from lower wage states, primarily in the South, more likely opposing the measure than legislators from relatively more affluent geographical areas. However, when controlling for agricultural and demographic factors, southern legislators were no more likely to oppose the law than legislators representing other areas (Seltzer, 1995) That increasing wage inequalities in the 1980s were accompanied by a stagnant federal minimum wage ($3.35 per hour from 1981 to 1989) has increasingly illuminated the political aspect of this issue (Lee, 1999). As a redistributive tool, increases in minimum wage have managed to prop up wages for unskilled and lower income workers, as competitive labor markets tend to push downward wage levels for those workers. Freeman (1996) concedes that manipulating wage levels in an attempt to target specific demographics is challenging. Especially important is the effect that artificially 1 Refer to sections 201-219 of title 29, United States Code. setting a wage rate has on the ability for labor markets to clear, thus affecting efficiency (p. 648). Legislation to increase the minimum wage that employers must pay workers is an economically inefficient act for two reasons. One, consistent with Stigler (1971), a minimum wage represents a wage floor that disturbs labor market equilibrium. More workers are willing to work than employers are willing to pay at that wage. Secondly, with increases in labor costs borne by employers some decrease in employment results as employers attempt to lessen the labor burden of paying a higher wage. This scenario is depicted graphically in Figure 3.1. An increase in minimum wage from $4.25 to $5.15 per hour, for illustration, pushes up the supply of labor from a labor market in equilibrium (L E ) to L S at the higher wage rate. Employers demand less labor at the higher mandated wage and the demand for labor falls from L E to L D . Figure 3.1 Minimum wage and labor analysis depicting labor market disequilibrium as a result of a minimum wage increase (adopted in part from Browning and Zupan, 2002, p. 492). 129 130 Sobel (1999, p. 783) argues that long run and short run forces act differently on labor demand in finding that the elasticity of demand, the steepness of the labor curve associated with the extent that a change in wage laws produce changes in the level of employment, is relatively more inelastic in the short run and more elastic in the long run. The more inelastic the demand curve the steeper the curve, and increases in minimum wage will less adversely affect the level of employment. That is, in the short run employers immediately absorb increases in labor costs in an attempt to prevent disruption of business activities, but in the long run employees are laid off, given reduced hours, or reassigned as the employer struggles with a higher labor cost burden. Overall, the effect of increases in minimum wage is mixed, but reductions in economic efficiency are clear. Gramlich and Kelly contend that a relatively weak correlation exists between low wages and low-income households, suggesting that targeting improvements in household earnings by redistributing earnings does not hold (as cited in Brown, Gilroy, & Kohen, 1982, p. 524). Tullock (1983, pp. 6-8) offers that redistributions involve attempting to identify individuals who are efficient utility generators from those that are not. Creating wage floors does not ensure that lower income workers will efficiently utilize marginal revenue (i.e., increases in income from increases in a minimum wage) to a greater extent than any other worker. Labor unions frequently demand increases throughout the wage structure, since the structure is misaligned due to a new, higher wage floor. With increases in household earnings often masking the political pressures behind calls for minimum wage increases, an 131 economically efficient labor market is not impeded by constraints that prevent market forces from setting wage rates. Scoring Model for Dependent Variables In measuring each dependent variable a scoring model is developed for a series of roll call votes during the time period of this study. The units of analysis are the legislators who voted on each bill representing the dependent variables in the House of Representatives and Senate respectively. To isolate the effects of changes in behavior over time on economic efficiency, only those legislators serving in the 99 th ? 108 th Congress, inclusive, are part of this model. The model indicates support for economic efficiency enhancing positions. Therefore, each legislative vote included in the analysis is coded 1 if the vote represents economic efficiency and 0 if it does not. The votes within a policy area are totaled and divided by the total number of possible votes within that area. The resulting scores represent the percentage of economically efficient votes cast by the legislator. Tables found in Appendices A, B and C list each of the votes included in the measurement of the dependent variables, medical malpractice reform and increasing the federal minimum wage, and E-score, respectively. The tables give the date of the vote, the house voting, bill number, short title, a synopsis of the legislation, the total vote and vote by party, and indication of support or opposition as an economically efficient or inefficient position. Medical malpractice reform votes that seek to reduce liability or extend protection from lawsuits are economically efficient; otherwise the vote is 132 inefficient. Minimum wage issue votes supporting an increase in minimum wage are economically inefficient; votes opposing an increase are economically efficient. Borrowing from Kennedy?s (2005, p. 60) model for deriving an E-score, the criterion for vote selection of the legislation representing the dependent variables include only legislation that is unambiguous relative to the intent of the roll call vote. For example, the final vote on legislation that contains multiple components in addition to each policy area is not included because a vote in support or opposition to the bill may not reflect support or opposition to specific components in the legislation. The criteria consider not only the title of the legislation, but also the intent of legislation. Some congressional action (e.g., invoking cloture in the Senate) requires investigation of the legislation and the debate that precedes such action. Understanding the intent of the legislation is crucial to deciding if a vote in support of the policy position is economically efficient or inefficient. The major sources for this information are the Washington Post (http://projects.washingtonpost.com/congress/) and the Congressional Quarterly Weekly Report (http://public.cq.com/). The roll call votes for medical malpractice and minimum wage legislation were identified from the Washington Post (http://projects.washingtonpost.com/congress/) and Congressional Quarterly Congress Collection (http://www.cq.com) with the latter the data source for the votes. This process identified four House votes and four Senate votes related to medical malpractice and five House votes and four Senate votes related to the minimum wage. 133 Vectors of Analysis Independent variables in the model are categorized according to vectors as a method of grouping those variables. The vectors include ideology, self-interest, and chamber environment. Control variables are utilized to hold constant extraneous effects that might affect the relationship between the independent variables and each dependent variable. Campbell, Converse, Miller, and Stokes argue for using frames of reference for processing information, while Hagner and Pierce find that belief systems are structured around behavioral and attitudinal consequences (as cited in Jacoby, 1986, p. 424). Jacoby finds that a liberal-conservative continuum represents levels of cognitive conceptualization that grouping affords to such research (p. 431). Especially in considering variables depicting self-interest and ideology, analyzing each variable within vectors should enhance the model. The following discussion lists the independent variables within each vector, the potential impact of each variable in relation to the research questions, and data collection procedures and coding of each. Self-Interest Variables representing legislator self-interest are direct contributions to legislators or PAC contributions from interest groups with a connection to the each policy area. Organized interests shape the policy agenda by targeting individuals or groups in the legislative process (Hojnacki & Kimball, 1998). To the extent that higher political contributions produce greater support for the policy area in question, self- 134 interest is a factor in legislative decision-making. Separate measures of self-interest are evaluated for each dependent variable policy area. For the medical malpractice policy area measures of self-interest are medical and law or law related contributions to legislators. These are labeled as the separate variables ?Health? and ?Lawyer.? Values of each variable are coded as the actual dollar amount of the contribution received, averaged over the two year period of each Congress. For the minimum wage policy area, contributions from ?Business? groups and ?Labor? groups are measured as independent variables representing legislator self-interest. In both policy areas, medical malpractice and minimum wage for Congresses 102 through 108, PAC and individual contributions are summed for each of the four categories (Business, Labor, Health, and Lawyer). Total receipts are the sum of all PAC and individual contributions received by each legislator for each Congress. Each of the four categories of this study is analyzed in Congresses 102 through 108 as a percentage of those total receipts representing contributions to the legislator. Data for contributions to legislators in the 102 nd through 108 th Congress were accessed from the Center for Responsive Politics (CRP) at http://www.opensecrets.org/politicians/index.asp. Data for the 99th through 101 st Congress were not available in the above format. To standardize measures of self interest for these three Congresses, the largest PACs were identified in each of the four categories (Business, Labor, Health, and Lawyer) and contributions from each of those PACs to each legislator were tallied by Congress. Each PAC is identified within one of the four categories (Business, Labor, Health, and Lawyer) considered for the 102 nd through 108 th Congresses and coded as the actual dollar amount received by the legislator from that PAC. Total receipts represent the total 135 of only those PAC contributions from the four categories in the analysis (Business, Labor, Health, and Law). Contributions from each PAC to legislators were tallied as a percent of the sum of contributions from all four PACs. Only those PACs analyzed were tallied in determining total receipts. The four PACs considered were National Association of Realtors (Business), National Education Association (Labor), American Medical Association (Health), and Association of Trial Lawyers (Lawyer). Of the four PACs only National Education Association is not the largest in its category. Teamsters Union is the largest labor PAC, but is not organized as a single entity (Makinson, 1990, pp.20-21). The source for political contributions for the 99 th through 101 st Congresses was Political Money Line from Congressional Quarterly at http://www.tray.com/cgi- win/x_pac_init.exe?DoFn=. Ideology Measures of ideology indicate another component of individual behavior is included. Three measures of legislator ideology are utilized. ADA and ACU scores measure liberalism and conservatism, respectively, while E-scores measure economic efficiency. DW-NOMINATE scores indicate relative liberal-conservative positions over time. Each measure of ideology is an independent variable in the model. ADA and ACU raw scores are available for each year and are tallied and averaged across each Congress. ADA ratings are available from http://www.adaction.org/votingrecords.htm and ACU scores are available at http://www.acuratings.org/. DW-NOMINATE scores are measured by Congress and are available from http://voteview.com/dwnomin.htm. 136 Values for each variable ? ADA, ACU, and E-score ? are considered as an average over the two years of each Congress along a continuum from 0 to 100. Higher ACU and ADA scores represent greater conservatism and greater liberalism, respectively, while higher E-scores indicate greater legislative economic efficiency. E- score ratings are developed from the E-score formula discussed earlier. Potential problems with using nominal data are addressed and adjustments are made in an attempt to reduce the time sensitively of the measures. DW-NOMINATE scores alleviate time sensitivity by weighting relative measures of liberalism and conservatism over time and adjusting each score across dimensions by including all votes cast by the legislator. As additional Congresses are added to the dataset expanding the voting history analyzed for each legislator, the DW-NOMINATE model more closely approximates the legislator?s relative liberal-conservative position across each Congress. E-scores also measure ideology but capture a different dimension than liberalism and conservatism. That ideology is not a single dimension along a spectrum, but rather multidimensional is important to this analysis (see Jones, 2001, pp. 155-156; Collie, 2000, pp. 219-227; Shaffer, 1989; Deckard & Stanley, 1974; Bethel, 1979). E-score ratings were developed for each Congress (99 th ? 108 th ) in the study based on the E-score formula: N E-score = ? (P i / N) ? 100 i=1 Where, Pi = one if legislator voted in support of enhancing efficiency and zero otherwise N = number of votes considered in the analysis of each legislator 137 Recalculating interest group ratings Roll call votes that are tabulated for use in scoring models for legislative support of medical malpractice and minimum wage policy areas (dependent variables) cannot be included in roll call votes that are tabulated to devise interest group ratings (e.g., ADA, ACU) used for independent variables. The same holds true for E-scores. Wattier and Tatalovich?s (2005) model is utilized to address this issue. For any ADA, ACU, or E- score measure where votes for medical malpractice or minimum wage are included in that measure, those votes are removed and the measure recalculated. For example, when the dependent variable in the model is medical malpractice if 20 votes are considered to yield an ACU rating and one or more votes related to medical malpractice is included, each vote on medical malpractice will be removed from the ACU rating and the rating recalculated based on a new denominator of 20 votes minus the number of medical malpractice votes removed. The effect of time on interest group scores Raw scores for ADA, ACU, and E-score also are impacted by time. ADA, ACU, and E-scores are computed nominally as raw scores for the year in which the score is calculated. Research has shown that if one compares the median values of interest group scores (e.g., ADA or ACU) for each Congress increased polarization by party is evident (Shipan & Lowry, 2001; Ingberman & Villani, 1993). Adjusting interest group scores produces a comparable index for analyzing trends in data analysis. Two models are available for this index. Groseclose, Snyder, and Levitt (1999) employ the use of shift and stretch parameters to analyze directions of change for each score over time. Their 138 study compared median values for Congress with nominal values for each legislator. Their modified model made possible comparisons across Congresses, where scores for legislators are compared to median scores for the chamber. The current study makes a similar comparison. Poole and Rosenthal (1997) utilized a spatial model of congressional roll call voting called DW-NOMINATE for adjusting a liberal-conservative spectrum of ideology. The current study uses the DW-NOMINATE model as one measure of adjusting liberal-conservative ideology, but also employed a comparison of median values in analyzing nominal ADA, ACU, and E-scores. DW-NOMINATE (dynamic, weighted, nominal three-step estimation) scores place House and Senate members within coordinates in a plane. The coordinate for each legislator is two-dimensional and dynamic and is allowed to move as a linear function of time with each Congress. A legislator?s coordinate is constant within a Congress but varies linearly between Congresses. The error term for DW-NOMINATE coordinates is normally distributed and the coordinates within each dimension are weighted in estimating coordinates for each legislator. The weighted parameter makes the calculation of distance between coordinates possible. Coordinates in one Congress are directly compatible with coordinates in another Congress, but cannot be compared across chambers of government. The two dimensions of the model have traditionally accounted for 85 to 90 percent of all roll call voting decisions. The first dimension captures divisions between the two major parties and the second dimension identifies regional distinctions within each major party (McCarty, Poole, & Rosenthal, 1997, p. 5; Pool & Rosenthal, 2001). 139 After 1975 the second dimension does little to explain voting (Poole & Rosenthal, 1997, pp. 6-8). A continuing trend since the 1970s has been for fewer and fewer roll call divisions between each party internally. The average ideological distance between members of the two parties is increasing with each party becoming increasingly homogenous (McCarty, Pool, & Rosenthal, 1997, p. 14). The format for using the DW-NOMINATE model is available at http://voteview.com/dwnomin.htm. Coordinates of each dimension are estimated along with error terms for each dimension. Legislator estimates for voting dimensions for members of the House of Representatives for all Congresses are available at ftp://pooleandrosenthal.com/junkord/HL01109A21_PRES.DAT and for senators for all Congresses at ftp://pooleandrosenthal.com/junkord/SL01109B21.dat. Chamber Environment Party unity is a measure of the chamber environment within which the legislator serves. Congressional Quarterly (CQ) defines party unity through roll call votes in which a majority of a party votes on one side of an issue and a majority of the other party votes on the other side. Data available through Congressional Quarterly (http://www.cq.com) and Voteview (http://voteview.com/default.htm) show a percentage representing the number of times members vote with a majority of their party. To reduce the effect of absences the percentages are normalized and calculated as follows: party unity = unity / (unity + opposition). CQ and Voteview party unity scores fall on a continuum from 0 to 100. Voteview data are used in this study with Republican scores recorded as negative 140 numbers and Democratic scores as positive numbers resulting in a continuum from -100 to +100. Party unity values are increasingly becoming more extreme (scores closer to ? 100 or +100), which indicates greater voting unity between the legislator and political party and reflects the extent of support of the legislation from a Republican or Democratic legislator in relation to party support. Increasing party unity values in Congress represent less legislative divergence with party platforms and fewer independent decisions made by legislators. The conditional party governance model discussed in Chapter Two suggests that increased party unity scores in part reflect self- interest motivations of legislators, especially reelection. The length of time a legislator has served in Congress affects legislative voting, with more variable behavior exhibited earlier in a legislator?s career (Strattman, 2000, p. 665). Junior legislators are more likely to vote with their party than senior members. Legislators who are initially elected to the House or Senate and are subsequently elected to the other chamber exhibit little systematic change in voting behavior after moving to the other chamber (Grofman, Griffin, & Berry, 1995). For purposes of tenure in Congress the year that a legislator is first elected to either the House or Senate is counted as the beginning of his or her term in office. Time served in Congress affects voting behavior of representatives and senators, but for senators how close each is to the end of his or her term in office is an additional determinant of legislative voting (Tuckel, 1983, Strattman, 2000, p. 675). Data pertaining to legislative tenure are available from Congressional Quarterly (http://www.cq.com) and the Center for Responsive Politics (CRP) at http://www.opensecrets.org/politicians/index.asp. 141 Ideological divisions between the legislator?s party and the party of the executive branch produce changes in legislative voting (Kernell & Jacobson (2006). Such ideological divisions are exacerbated in considering minority-majority party relationships within Congress (Schickler, 2000; Fleisher & Bond, 1996). Data available from Voteview (http://voteview.com/default.htm) measure liberal-conservative divisions between the legislative and executive branches. Differences in median party scores within Congress (House and Senate) and executive or presidential scores are gathered in measuring ideological divisions between each branch of government. Control Variables Unlike independent variables, control variables are not intended to produce changes in the dependent variable, but are rather constant variables representing extraneous factors. With changes in legislative behavior over time an important component to this investigation of economic efficiency, the effect of increasing party unity is but one factor that must be considered within the chamber environment. One way to measure the impact of this variable over time is by controlling for the effects of institutionalization and party control of government (unified or divided). Dichotomous relationships are presented between Republican and Democratic parties for each control variable measured. When the legislator is an Independent, he or she sits with the caucus of one of the parties. Independent legislators are placed with the appropriate caucus as identified by Congressional Quarterly Congress Collection (http://www.cq.com). Table 3.1 summarizes the control variables used to examine the relationships between legislators, political party and institutions. 142 Table 3.1 Control variables for party control of government Variable Coding of variables Legislator Republican = 1 Democrat = 0 Legislator?s party Legislator?s party and party in control of institution are the same = 1 Legislator?s party and party in control of institution differ = 0 House Republican majority = 1 Democratic majority = 0 Senate Republican majority = 1 Democratic majority = 0 Congress Both House and Senate have Republican majority = 1 Both House and Senate have Democratic majority = 0 Congress split House and Senate controlled by different parties = 1 House and Senate controlled by same party = 0 President Republican = 1 Democratic = 0 Congress and president Both houses of Congress and president are Republican = 1 Both houses Congress and president Democrats = 0 House and president split House and president controlled by different parties = 1 House and president controlled by same party = 0 Senate and president split Senate and president controlled by different parties = 1 Senate and president controlled by same party = 0 Congress and president split Both houses of Congress controlled by different party than president = 1 Congress and president of same party = 0 Another way to capture chamber environment is by controlling for regional differences and geographical effects. The United States Census Bureau establishes boundaries along four distinct regions 2 ? West, Midwest, South, and Northeast. Differences exist in ideology and attitudes across regions. Knuckey (2005, pp. 43-45) finds that an ideological realignment in the South is a product of partisan changes 2 The United States Census Bureau (http://factfinder.census.gov/servlet/ReferenceMapFramesetServlet?_bm=y&-rm_config=| b=85|l=en|t=420|zf=0.0|ms=ref_legal_05pep|dw=1.9557697048764706E7|dh=1.4455689123E7|dt=gov.ce nsus.aff.domain.map.LSRMapExtent|if=gif|cx=-1159354.4733499996|cy=7122022.5|zl=10|pz=10| bo=1623:1629:1573:1574:1615:1587:1633|bl=1624:1630:1571:1572:1616:1588:1634|ft=1583:1625:1635: 1601:1611:1631:1595|fl=1626:1636:1602:1612:1632:1596:1584|g=01000US&-redoLog=false&-_lang=en ) defines each region as follows: Northeast -- NY, PA, NJ, CT, RI, MA, VT, NH, ME; South -- AL, MS, GA, FL, KY, LA, TX, AR, OK, TN, SC, NC, VA, WV, MD, DE; Midwest -- OH, IN, MI, IL, MO, KS, NE, IA, ND, SD, MN, WI; and West -- AK, AZ, CA, CO, HI, ID, MT, NM, NV, OR, UT, WA, WY. 143 among white conservatives and increasing strength of the Republican Party in that region. Republican ascendancy in the South, a region that was once solidly Democratic, and shifts by the national Democratic Party to an increasingly liberal ideology indicate partisan swings are occurring across regions. According to Flanigan and Zingale (2002) regional distinction is also apparent by differences in attitudes for the role of government in domestic issues, such as health care (pp. 124-125), and changes in partisanship (pp. 69-73) as shifts in the New Deal alignment occurred in the 1970s. Those attitudes, consisting at a primary level of the extent that fiscal policies are necessary, illustrate not only an ideological basis for the size of government, but an economic basis as well (Jones, 1990). Considering each of the two dependent variables, whether a medical malpractice crisis exists in the state of the legislator, as identified by the American Medical Association 3 (AMA), or if that state has a minimum wage law higher than the federal minimum could impact a legislator?s behavior. 4 Control variables for geographical conditions and the coding for each variable are summarized in Table 3.2. 3 The AMA (http://www.ama-assn.org/ama/noindex/category/11871.html) identifies the following states as in a medical malpractice crisis: WA, OR, NV, WY, MO, IL, FL, TN, KY, OH, NY, PA, NJ, MA, RI, CT, NC 4 The United States Department of Labor (http://www.dol.gov/esa/minwage/america.htm) identifies the following states as having minimum wage laws exceeding the federal minimum wage: WA, OR, CA, NV, AZ, MO, AR, FL, PA, NJ, NY, VT, MA, MT, CO, MN, WI, MI, IL, OH, WV, NC, MD, DE, RI, ME, AK, HI. 144 Table 3.2 Control variables for geographical conditions Variable Coding of variables Region Represented Northeast effect Northeast = 1 Not Northeast = 0 South effect South = 1 Not South = 0 Midwest effect Midwest = 1 Not Midwest = 0 West effect West = 1 Not West = 0 Medical malpractice crisis (Used with malpractice dependent variable) State in medical malpractice crisis = 1 State is not in crisis = 0 Minimum wage (Used with minimum wage dependent variable) State has a minimum wage exceeding the federal rate = 1 If state does not have minimum wage law or has a law below the federal rate = 0 Measures of state per capita income, per capita total federal spending, net federal spending, and percent minority population are four controls that are used as proxies for state economic conditions that may impact the independent-dependent variable relationship. Measures of per capita income control for wage disparities that might exist across states, while federal spending measures capture per capita federal net allocation as a function of per capita congressional representation (Atlas et al, 1995). Canto and Webb ?report negative associations between state per capita total spending and growth in per capita income, and between state transfer payments and per capita income growth? (as cited in Jones, 1990, p. 221) indicating the potential unintended consequences of such policies on the economic vitality of the state. Per capita income is coded for each state as an actual dollar amount. Per capita income is available annually from the United States Census Bureau and Bureau of Economic Analysis at http://www.bea.gov/regional/spi/default.cfm?series=summary. Annual data are averaged for the two years of each Congress. Per capita total spending is 145 measured as a ratio of annual total federal spending in each state to the population of that state. Data are available as per capita federal spending received. Annual data are averaged for the two years of each Congress. Federal spending data (including net federal spending) and state per capita total spending data are available at from Tax Foundation at http://www.taxfoundation.org/research/show/347.html and Northeast- Midwest Institute at http://www.nemw.org/fundsrank.htm. Northeast-Midwest Institute (http://www.nemw.org/data.htm#fedspend) defines net federal spending in a state as the state?s return on federal tax dollars or the net inflow of federal dollars into a state. It is determined by dividing an adjusted level of federal spending by an estimated level of federal tax burden. The result is an estimated amount of federal spending returned to each state or region for $1 in federal taxes. Values greater than $1 indicate higher levels of federal spending in a state relative to its federal tax burden and values less than a $1 indicate a higher flow of federal tax dollars out of the state than federal spending into the state. Minority is coded as two variables representing the percentages of African Americans and Hispanics in each state. Data for percentages of African Americans and Hispanics in a population data are available from the United States Census Bureau at http://www.census.gov. Control variables for economic conditions are summarized in Table 3.3. 146 Table 3.3 Control variables for state economic conditions Variable Coding of variables Ethnic Category African-American Percent African-American in state Hispanic Percent Hispanic in state Per Capita Income Dollar value of per capita income in state Per Capita Total Spending Dollar value of per capita total federal spending in the state Net Federal Spending Proportion of federal spending to federal tax revenue in state expressed numerically. E-score Development In addressing the research question of the role of economic efficiency as another dimension of ideology distinct from liberalism and conservatism, E-scores were developed for each legislator voting in the 99 th ? 108 th Congress, inclusive. Criteria for economically inefficient policy established by Stigler (1971) and Kennedy (2005) guided the scoring process. Those criteria signaling a reduction in economic efficiency are subsidies or price supports that lead to a misallocation of resources; barriers to entry of markets or regulations within an industry or pervasively throughout the economy limit competition; policies that affect substitutes and complements of goods within an industry from special interest demands voiced in opposition; and wage and price controls that artificially set wage floors or ceilings. Minimum wage laws are an example of a wage control, but they are not included in compiling the E-score since that policy area is included in the model as a dependent variable. Using roll call votes pertaining to one or more of the above criteria, a nominal E- score value is developed for each House and Senate member serving in Congress during the period under study. Based on the E-score formula discussed earlier, the value is a 147 number between 0 and 100; the higher the number assigned, the greater the legislator?s support for economically efficient policies. Vote Selection Criterion Criteria for vote selection for both the E-score and the dependent variables of minimum wage and medical malpractice legislation follow the same logic developed by Kennedy (2005, p. 60). That is, legislation selected is unambiguous relative to the intent of the roll call vote. For example, the final vote on legislation that contains multiple components in addition to each policy area is not included because a vote in support or opposition to the bill may not reflect support or opposition to specific components in the legislation. The criteria consider not only the title of the legislation, but also the intent of the legislation. Some congressional action (e.g., invoking cloture in the Senate) requires investigation of the legislation and the debate that preceded such action. Understanding the intent of the legislation is crucial to deciding if a vote in support of the policy position is economically efficient or inefficient. Important differences exist, however, between selection of votes for each dependent policy area and selection of votes in developing an E-score. Minimum wage and medical malpractice are specific policy areas considered within a context of how votes in support or opposition to legislation in each area affect economic efficiency. Only those votes pertaining to each of the policy areas is considered in evaluating the economic efficiency of the legislation. But economic efficiency is not limited to these two policy areas and encompasses more issue areas. Economic efficiency is maximization of aggregate social 148 benefits of policy decisions to aggregate social costs. This expands the number of issues that can be considered in devising an E-score to other policy areas where economic efficiency can be measured. Public policy making that is economically efficient is consistent with Pareto improving positions, where it is not possible to make someone better off without making someone else worse off. Stigler (1971) and Kennedy (2005) find that analyzing legislation that is economically inefficient is not only the flip side of economic efficiency, but also enhances the model for identifying and selecting roll call votes. Legislation signaling economic efficiency reduction involves regulation in private markets or direct intervention where a market failure does not exist (Kennedy, 2005, p. 56). According to Stigler, the nature of any policy having clear welfare implications is paramount to analyzing legislation as economically efficient or inefficient (as cited in Kennedy, 2005, pp. 28-29). For this dissertation all roll call votes for each Congress (99 th ? 108 th , inclusive) are analyzed. Only legislation meeting one or more of the four categories developed by Stigler and Kennedy is selected for inclusion in the analyses. Using these categories a roll call vote in support or opposition to the selected legislation is evaluated. All legislation selected is identifiable as either enhancing economic efficiency or signaling a reduction in economic efficiency. Roll call votes are tabulated for each legislator voting. Roll call votes in support of economically efficient legislation are recorded as an economic efficiency-enhancing vote by that legislator; roll call votes in opposition to economically efficient legislation are recorded as an economically inefficient vote by that legislator. Roll call votes in support of economically inefficient legislation are 149 recorded as an economically inefficient vote by that legislator; roll call votes by each legislator in opposition to economically inefficient legislation are recorded as economic efficiency enhancing. Each vote was identified in the Washington Post database (http://projects.washingtonpost.com/congress/) and Congressional Quarterly Congress Collection (http://www.cq.com) with the latter the data source for each vote. Table 3.4 summarizes the development of the E-score model. Tables found in Appendix C list all House and Senate roll call votes by Congress included in devising each E-score. For each piece of legislation the tables provide a title, bill number, narrative description of the legislation, and indicate if support for the legislation represents an economically efficiency enhancing or reducing event. Votes in support of economically efficient positions are coded 1 and votes for supporting inefficient positions are coded 0. Within each Congress a legislator?s votes are totaled, divided by the total number of votes possible, and converted to a percentage. Table 3.4 Summary of E-score model development Congresses considered 99 th ? 108 th (1985 ? 2005) Chambers of Congress House of Representatives and Senate Votes considered Roll call votes in House and Senate where economic efficiency is affected Criteria Economic efficiency standards developed by Stigler (1971) and Kennedy (2005) Coding Votes for economically efficient positions = 1 Votes for economically inefficient positions = 0 Percentage of votes for economic efficiency based on total votes = 0 ? 100; higher scores depict higher economic efficiency 150 Methods of Analysis In considering the research question the model must measure the multiple effects between independent and dependent variables, while controlling for extraneous factors, and analyze if the combination of relationships produces changes across years of the study. This research employs two different methodologies: multiple regression analysis of the effect of various explanatory (independent) variables on two dependent variables and interrupted time-series analysis in measuring changes in behavior over time. Regression Analysis Ordinary Least Squares (OLS) analysis measures the direct impact of each independent variable on the dependent variables. To test for indirect effects, several regressions are run between the variables in the model. Of particular interest is the relationship between E-Score and other independent variables, such as ADA scores and ACU scores, and the extent that E-scores appear to transcend liberal-conservative ideology. While bivariate associations between an independent variable and dependent variable might exist, controlling for the effects of other independent variables is necessary in measuring if the association is direct, indirect, or spurious. Units of analysis for this study are legislators in the U.S. House and Senate. When data on a population are available and the population is rather small, the entire population should be sampled. For this study the sample will consist of all legislators voting on each dependent variable. Tests of significance demonstrate how likely an association between two variables in a sample might or might not exist. These associations are measured by an F- 151 Test for the entire model and a T-Test for each hypothesis. Variables in the model with a predicted association (+ or -) between independent and dependent allow the use of a one-tailed test. Independent and dependent variables where a predicted relationship does not exist require the use of a two-tailed test to measure the association between such variables. Statistical significant is determined at the 0.05 level. The regression coefficient (beta) measures the strength of association between the independent and dependent variables. While tests of significance address the likelihood of causal associations existing, beta values are important in identifying changes per unit in the dependent variable as a result of applying multiple independent variables. The standardized regression coefficients or Betas compare the relative impacts of the independent variables. Two dependent variables are analyzed within the model: minimum wage legislation and medical malpractice legislation. The model considers a legislator?s support for economic efficiency in analyzing each dependent variable separately. The multivariate regression equation for each dependent variable in the model is as follows. Medical malpractice policy area: VOTEit = a 0 + b 1 ESCORE it + b 2 ADA it + b 3 ACU it + b 4 DW NOMINATE it + b 5 HEALTH it + b 6 LAWYER it + b 7 CHAMBERENVIRONMENT it where, VOTEit = dependent variable representing a scoring model of final, roll call votes by a legislator on medical malpractice policy. a o = constant term representing parameter at Y-intercept. b 1 ?.b 7 = coefficients (beta) for independent variables in the model 152 ESCORE i = E-score for legislator i in time t . ADA it = ADA scores for legislator i in time t . ACU it = ACU scores for legislator i in time t . DW NOMINATE it = DW Nominate scores for legislator i in time t . HEALTH it = Health related political contributions to legislator i in time t . LAWYER it = Law related political contributions to legislator i in time t . CHAMBER ENVIRONMENT it = Chamber environment including party unity and party leadership control for legislator i in time t . Minimum wage policy area: VOTEit = a 0 + b 1 ESCORE it + b 2 ADA it + b 3 ACU it + b 4 DW NOMINATE it + b 5 LABOR it + b 6 BUSINESS it + b 7 CHAMBERENVIRONMENT it where, VOTEit is the dependent variable representing a scoring model of final, roll call votes by a legislator on minimum wage policy. a o = constant term representing parameter at Y-intercept. b 1 ?.b 7 = coefficients (beta) for independent variables in the model ESCORE i = E-score for legislator i in time t . ADA it = ADA scores for legislator i in time t . ACU it = ACU scores for legislator i in time t . DW NOMINATE it = DW Nominate scores for legislator i in time t . LABOR it = Labor related political contributions to legislator i in time t . BUSINESS it = Business related political contributions to legislator i in time t . CHAMBER ENVIRONMENT it = Chamber environment including party unity and party leadership control for legislator i in time t . 153 Interrupted Time-series Design With interrupted time series, it is possible to measure the impact of changes in political party control of the institution (House and Senate) on the support for public policies analyzed in the policy areas across years. The goal is to determine whether changes in political control affect subsequent observations of legislative ideology and self-interest. The intent of using this technique is to evaluate the effect of political party changes on legislative decision-making, and if such decision-making impacts economic efficiency. In separate analyses each independent variable (ADA, ACU, and E-score) is regressed on time and two dummy variables. Changes in economic efficiency are especially relevant in considering the role of party and legislative voting. Party and the influence of party in controlling congressional institutions are major factors linking legislative decision-making (Sinclair, 1977; Menefee-Libey, 1991, Carsey & Layman, 2004) with a growing ideological chasm between each political party (Flanigan & Zingale, 2002, pp. 57-63). The extent that party control affects economic efficiency is compared to the impact of party control on liberal- conservative ideology in analyzing the role of economic efficiency in predicting behavior. Changes in E-scores, as a measure of economic efficiency, and ADA, ACU, and DW-NOMINATE scores, as measures of liberal-conservative ideology, are evaluated by comparing the Congresses before and after the following events: the 1986 congressional election with the Democratic Party regaining control of the Senate, 1994 election with Republicans sweeping both houses of Congress, and the 2000 election with closely divided, Republican controlled Congress and a Republican president. 154 The key additions in an interrupted time series design are two dummy counting variables. One dummy variable is coded zero for observations before changes in institutional control (e.g., a Democratic majority in the Senate is replaced by a Republican majority) and one for observations thereafter. It is used an indicator of whether a change in behavior occurred in and around the event in question. The second dummy variable is coded zero for observations prior to the change in party control and one for the first year after the change in party control, two for the next, three for the next, and so forth. This variable is called a post counter. It is used in determining whether any change in pattern detected is long term or short term in duration. The dependent variable for the analysis is median scores on legislation within each policy area by Congress in measuring the magnitude of changes over time. With median scores sensitive to even minor changes in party control, separate regression analyses are run to measure percent change in median scores from one Congress to the next. A separate regression analysis of median scores for political party minus median scores for another political party for each Congress is another method employed in capturing magnitude of E-score changes. Changes in party control of the Senate occurred with the 100 th Congress and for the presidency with the 103 rd and 107 th Congresses. Party control for House and Senate changed with the 104 th Congress. A limitation with using interrupted time series design in measuring changes in legislative behavior around these events is the lack of roll call votes available on legislation within either policy area for each point in time. Too few votes does not allow for accurate analysis of the data before and after a base year. 155 Hypothesis Development The expected associations between the independent (explanatory) variables and each dependent variable in the model are summarized below. They are based on prior research and will assist in answering these key questions: 1) What factors, ideology or self-interest, contribute significantly to legislative decision-making? 2) Is a legislator?s support for economically efficient policy making situational or rather consistent across time? 3) Does political party control affect legislative voting to the extent that economic efficiency is compromised and public policy formulation altered? Hypothetical Associations and Legislative Voting Hypotheses are developed in explaining legislative voting within each of the two dependent variable policy areas of this study, medical malpractice reform and minimum wage. E-scores measure the economic efficiency of legislative voting and legislators are assigned scores representing the economic efficiency of voting decisions. Because this study considers the extent that E-scores are predictors of legislative behavior over time in each house of Congress, changes in economic efficiency are considered as functions of length of term in office and, for senators, the number of years into his or her term. Research shows that voting behavior of legislators is more variable early in their career, with junior members more likely to vote along party lines than senior members (Stratmann, 2000). Political parties are increasingly polarized along a liberal- conservative spectrum (Shipan & Lowry, 2001, p. 247). Considering if senior members 156 vote in relatively higher numbers according to their ideology and if that ideology follows an economically efficient outcome, then differences in time in office are important to this study. Behavioral labels, such as liberal or conservative, describe legislative ideology and offer clues to support or opposition of a policy issue (Shaffer, 1989). Hinich and Pollard find that labeling patterns of political behavior allow a constituency to derive a label of relative conservatism or relative liberalism that exists across issues (as cited in Poole, 1988, p. 118). As proxies of liberalism and conservatism, ADA scores and ACU scores, respectively, are tools for estimating that legislators who are more liberal are less likely to support malpractice reform than conservative legislators. ADA and ACU scores are usually highly, negatively correlated. Inclusion of both scores in a model reduces some of the bias that places the opposing ideology at a polar extreme (Brunell, Koetzle, Dinardo, Grofman, & Feld, 1999; Austen-Smith, 1993). DW-NOMINATE scores adjust the effect of time on behavioral labels that might otherwise fail to identify gradual changes in relative voting positions. Organized interests affect legislative decision making by appealing to a legislator?s self-interests. Downs (1957) finds the desire to win reelection as a strong incentive for legislative actions, while Sears and Funk (1990) cite the short-term impact of an issue and material well-being afforded to an individual as examples of self- interested behavior. That material well-being naturally flows with political contributions, considering the impact of lobbying efforts by interest groups connected to each policy area is important in understanding self-interested behavior in this study. Frequently, these organized interests are channeled to gain access or influence particular groups or 157 individuals (Ainsworth, 1997). Hypothesizing that higher legal contributions are anathema to legislative support for medical malpractice reform but higher health care contributions are consistent with limiting medical liability and reducing insurance premiums are examples of organized interests lobbying legislative allies as the highest priority before expanding supportive coalitions (Hojnacki & Kimball, 1998). Divided government is found to reduce the passage of significant legislation (Edwards, Barrett, & Peake, 1997). A party label provides brand name identification with party leaders commissioned to maintain or enhance the party?s reputation (Hager & Talbert, 2000). Carsey and Layman (2004) examine citizen preferences for divided government as a function of ideological locations and perceptions of the two parties. Applying this principle, legislators also position themselves within political parties. According to conditional party government theory party members support party positions under certain circumstances. Forgette and Sala (1999) argue that majority party members in particular vote with their parties (p. 467). If a legislator votes with party and not according to maximizing net social benefit of the policy position, then economic efficiency associated with his or her vote suffers. If constituents align with a political party to push public policies, legislators, who are driven by self-interest in seeking reelection are less likely to apply economically efficient principles to their decision making in addressing the concerns of citizens. Medical malpractice reform policy area Medical malpractice reform is an economically efficient public policy objective (Knauss & Levy, pp. 2-12; Note, 1996, pp. 1765-1773). That is, medical malpractice 158 reforms that reduce medical liability and burgeoning insurance premiums produce economically efficient social benefits. Medical malpractice reform improves allocation of resources and reduces rent-seeking behavior from excessive tort awards. Table 3.5 presents the hypotheses related to medical malpractice reform votes and the expected regression coefficient sign for each relationship. The positive signs suggest that hypothetical associations between each independent variable and medical malpractice reform expand social benefits. (In the case of Hypothesis 6 the sign designation reflects the arbitrary coding of Republican Party influence as a negative number and the Democratic numbers as positive.) 159 Table 3.5 Hypotheses for medical malpractice reform policy area Hypotheses Expected Sign of Regression Coefficient H 1: Legislators with higher E-scores vote in support of medical malpractice reform. + H 2: Legislators with higher ADA scores vote in opposition to medical malpractice reform. _ H 3: Legislators with higher ACU scores vote in support of medical malpractice reform. + H 4: Legislators with higher legal political contributions to total contributions vote in opposition to medical malpractice reform. _ H 5: Legislators with higher health care political contributions to total contributions vote in support of malpractice reform. + H 6: Republican legislators are likely to vote for malpractice reform more often than Democratic legislators. _ H 7: The closer senators are to the end of their current term in office, the more likely they are to support malpractice reform. + H 8: The longer a legislator has served, the more likely he or she supports medical malpractice reform. + H 9: The greater the ideological division between the legislator?s party and the party of the executive branch, the less likely the legislator supports medical malpractice reform. _ H 10: Legislators from the minority party (House or Senate) are more likely than majority party legislators to support medical malpractice reform. + H 11: The greater the division between the ideology of the legislator and the median ideology of the party to which the legislator belongs, the more likely the legislator supports medical malpractice reform. + Minimum wage policy area Minimum wage legislation is inversely associated with economic efficiency. Table 3.6 presents the hypotheses associated with this policy area. With the exception of Hypothesis 6, a hypothesis arbitrarily coded in measuring political party influence, support for minimum wage legislation is economically inefficient, suggesting that 160 hypothesized associations supporting minimum wage legislation increase public costs at the expense of social benefits. Thus, the negative signs associated with the variables related to economic efficiency. Table 3.6 Hypotheses for minimum wage legislation policy area Hypotheses Expected Sign of Regression Coefficient H 1: Legislators with a higher E-score will oppose increasing the minimum wage. + H 2: Legislators with higher ADA scores will support increasing the federal minimum wage. _ H 3: Legislators with higher ACU scores will oppose increasing the federal minimum wage. + H 4: Legislators with higher business political contributions to total contributions will oppose increasing the minimum wage. + H 5: Legislators with higher labor political contributions to total contributions will support increasing the minimum wage. _ H 6: Democrat legislators are likely to vote for increasing the minimum wage more often than Republicans. + H 7: The closer senators are to the end of their current term in office, the less likely they are to support increasing the minimum wage. + H 8: The longer a legislator has served, the less likely he or she will support increasing the minimum wage. + H 9: The greater the ideological division between the legislator?s party and the party of the executive branch, the more likely the legislator will support increasing the minimum wage. _ H 10: Legislators from the minority party (House or Senate) are less likely than majority party legislators to support increasing the federal minimum wage. + H 11: The greater the division between the ideology of the legislator and the median ideology of the party to which the legislator belongs, the less likely the legislator supports increasing the federal minimum wage. + 161 Since increases in a minimum wage are redistributive, representatives with higher E-scores are less likely to support such legislation. Those legislators who are more liberal and favor redistribution as a public policy are more inclined to support passage. Labor and business contributions are used in measuring self-interest with this variable. While labor represents the interests of the worker, business captures those interests that typically oppose increases in minimum wage as an addition to labor costs. Hypothetical associations within the chamber environment are the inverse of those formulated for medical malpractice reform. With increases in minimum wage an inefficient act, legislators who are recently into their careers or at the mercy of self ? interests from constituents pushing policy development, supporting inefficient legislation is more likely. Limitations of Research A potential concern with comprising an economic efficiency model is the subjectivity associated with choosing roll-call votes for inclusion into an E-score model. Kennedy (2005) argues that efficiency is an objective criterion while liberal- conservative ideology measured through ADA is subjective. Selecting votes for any vote model involves a degree of subjectivity. Each vote in support of a policy position involves a decision made by a legislator that involves interpretation of that decision. Simon argues that legislators are individuals and individuals make decisions based on decision premises (as cited in Fry, 1989, p. 185). Each vote for a public policy consists of a compilation of individual decisions made at each decision premise. Underlying 162 premises of each decision involve an individual assessment of benefits and costs, but must be considered through aggregate effects of social benefits and social costs. Selecting those votes where individual decisions produce a clearly understood impact on net social benefits is a challenge for researchers. Closely selecting each vote based on the four premises offered by Stigler underscores those issues that reflect less ambiguous efficiency positions. Finding enough votes to analyze and select in comprising the E-core is a concern. Limiting selection of votes comprising the E-Score to those roll-call votes for amendments or final passage potentially limits the inclusion of other votes that could signal efficiency or inefficiency. Including traditional measures of liberal-conservative ideology with an economically efficient variable introduces two measures of ideology that correlate in a regression model. Kennedy (2005) found that the E-score for both 106 th and 107 th Congresses correlate with ADA -0.79 in the House and -0.80 in Senate. Multicollinearity is a problem where correlation between variables produces undesirable effects. Multicollinearity refers to correlations among variables where nominally different measures quantify the same phenomenon to a high degree and are redundant (Gujarati, 1988, p. 283-285). Adding or deleting predictor variables changes the regression result. Adding cases or dropping variables is a technique used in this study to lessen the effects of multicollinearity when correlations between variables adversely affect regression results. The next chapter presents the results of the quantitative analyses used to test the hypotheses presented above. 163 CHAPTER FOUR ANALYSIS AND EMPIRICAL RESULTS Chapter Four presents the results of empirical analyses of the relationship between the independent variables and each dependent variable in the study. Medical malpractice reform and minimum wage were each evaluated as dependent variables in measuring economic efficiency as a predictor of legislative behavior. Multiple regression analysis and interrupted time series analysis are employed in making these analyses. Using the Pearson correlation coefficient, initially bivariate associations were run between each independent variable and each dependent variable to investigate the relationship between the variables. Bivariate associations indicate how closely two variables correlate, but are inadequate for analyzing direct effects between variables when multiple variables are part of a model. They are important in identifying those independent variables that are relatively highly correlated and that may produce multicollinearity, where the close association affects the regression model such that analyzing which variables produce changes in the dependent variable is compromised. Pearson correlation coefficients were examined for correlations over 0.60 in order to determine potential multicollinearity between the independent variables. Highly correlated bivariate associations between the independent variables, ADA, ACU, and 164 DW Nominate scores as measures of a liberal-conservative spectrum, necessitate analyzing the effects of these three scores separately. The regression model is run with each of the ideology variables included and then with each measure of ideology independently. Each variable measures different aspects of ideology ? ADA (liberal), ACU (conservative), and DW Nominate (a moving spectrum of liberalism and conservatism). Running separate multiple regression analyses with ADA, ACU, and DW Nominate scores reduces the likelihood that multicollinearity affects the model. Analyzing changes in the economic efficiency of legislative voting over time is important to this study. Measuring legislative voting over each Congress from the 99 th Congress through the 108 th Congress (1985-2004), inclusive, captures sweeping changes in political party control of each chamber of Congress after the 1994 congressional election and also includes changes in control of either chamber and the impact of executive-legislative relationships over the time period of the study. While the period under study began with mixed control of the House and Senate, this was followed by a period of Democrat control of both houses in the 100 th -103 rd Congresses. From the 104 th through 108 th , the Republicans took control over both houses. Split party control between the Congress and the president was the norm. Table 4.1 illustrates executive-legislative party divisions for each Congress covered in the study. 165 Table 4.1 Party Divisions, 99 th -108 th Congresses Congress House Senate President 99 Democrat Republican Republican 100 Democrat Democrat Republican 101 Democrat Democrat Republican 102 Democrat Democrat Republican 103 Democrat Democrat Democrat 104 Republican Republican Democrat 105 Republican Republican Democrat 106 Republican Republican Democrat 107 Republican Varied control* Republican 108 Republican Republican Republican * Control of Senate shifted between both parties Chapter Four is divided into four sections. The chapter begins with descriptions of trends for each measure of ideology in the study ? E-score, ADA, ACU, and DW Nominate ? over the time period of the study. The results are presented. Analyzing variability of each measure of ideology is essential in exploring to what extent E-score is different from traditional measures of liberal-conservative ideology in explaining legislative behavior. The next section is divided into two parts. Each part is a multivariate analysis of each dependent variable (medical malpractice and minimum wage) and their respective hypotheses. The analysis addresses each dependent variable singularly for each Congress across each chamber. With the purpose of the study surrounding the role of E- score as an ideological tool for predicting legislative behavior each dependent variable, to standardize the analysis between each dependent variable roll call votes were coded to represent an economically efficiency enhancing legislative position. Roll call votes for medical malpractice are economically efficiency enhancing and votes for minimum wage are not economically efficiency enhancing. That is, roll call votes in support of 166 medical malpractice and votes in opposition to minimum wage increases are coded as economically efficiency enhancing. Three distinct scoring models are employed in compiling votes for computing the value of each dependent variable. For each dependent variable policy area ? medical malpractice and minimum wage ? when only one piece of legislation is considered in that Congress, the dependent variable is computed based on roll call votes for that single piece of legislation. For those Congresses where more than one piece of dependent variable legislation is considered roll call votes are tabulated within a scoring model to reflect a percent of the total roll call votes within each policy included in this analysis. For example, if four pieces of legislation were analyzed in the policy area and a legislator voted in support of increasing economic efficiency one time out of four a score of 25 was assigned to the legislator. Legislation for each dependent variable policy area was not considered within each Congress of the study and therefore roll call votes are not available for analysis in some Congresses. To alleviate the problem of no dependent variable available for analysis in these Congresses, separate scoring models for each dependent variable policy area are computed in the House and Senate for those legislators serving in each Congress during the time period of the study, 1985-2004. In the same manner as the scoring model for multiple pieces of legislation in the same Congress, a score is assigned for the legislator as a percent of the total roll call votes available for analysis within each policy area across each Congress in the House and Senate. All legislation in each policy area where roll call votes were cast is not included in the scoring models. Only that legislation that specifically addressed the policy area(s) 167 was included. That is, legislation that was part of other bills or failed to clearly distinguish economic efficiency implications of the legislator?s vote on the policy position is not included. Table 4.2 summarizes legislation included in the dependent variable scoring model for each Congress of the study in House and Senate. For each piece of legislation in the table the name of the legislation, bill number, and date when the roll call vote was cast is included. Table 4.2 Dependent variable legislation for each Congress: House and Senate HOUSE Dependent variables Congress Medical Malpractice Minimum Wage 99 Military Medical Malpractice. HR 3174. (10.7.85) No dependent variable legislation 100 High Risk Occupational Liability. HR 162. (10.15.87) No dependent variable legislation 101 No dependent variable legislation Minimum Wage Increase Passage. HR 2. (3.23.89) Minimum Wage Veto Over Ride. HR 2. (6.14.89) Minimum Wage Increase Passage. HR 2710. (11.1.89) 102 No dependent variable legislation No dependent variable legislation 103 No dependent variable legislation No dependent variable legislation 104 Product Liability Medical Malpractice Cap. HR 956. (3.9.95) Employee Commuting / Minimum Wage Increase. HR 1227. (5.23.96) Product Liability Passage. HR 956. (3.10.95) 105 No dependent variable legislation No dependent variable legislation 106 No dependent variable legislation Minimum Wage 2 year Increase. HR 3846. (3.9.00) Minimum Wage Increase Continuance. HR 3846. (3.9.00) 107 Medical Malpractice Award Passage. HR 4600. (9.26.02) No dependent variable legislation 108 No dependent variable legislation No dependent variable legislation 168 SENATE Dependent Variables Congress Medical Malpractice Minimum Wage 99 No dependent variable legislation No dependent variable legislation 100 No dependent variable legislation Minimum Wage Cloture. S837. (9.23.88) 101 No dependent variable legislation Minimum Wage Increase. S4. (4.11.89) Minimum Wage Increase. HR2. (4.12.89) Minimum Wage ? Training Wage. HR2710. (11.8.89) Minimum Wage passage. HR2710. (11.8.89) 102 No dependent variable legislation No dependent variable legislation 103 No dependent variable legislation No dependent variable legislation 104 Product Liability OB. HR 956. (5.2.95) Product Liability $500,000. HR 956. (5.2.95) Product Liability Cloture. HR 956. (9.4.95) No dependent variable legislation 105 No dependent variable legislation No dependent variable legislation 106 No dependent variable legislation No dependent variable legislation 107 Patients Rights Malpractice Liability. S 1052. (6.29.01) No dependent variable legislation 108 Medical Malpractice Cloture. S 2061. (2.24.04) No dependent variable legislation Results presented include regression output for each dependent variable scoring model ? for legislators voting across each Congress and legislators voting on legislation in the Congress and chamber under consideration ? in each policy area ? medical malpractice reform and minimum wage legislation. The third section of this chapter captures changes in legislative behavior associated with changes in political party control of the institution or the presidency. Kellough?s (1990) interrupted time series model is employed in this analysis. Three 169 change points within the 99 th through 108 th Congress are considered for analysis: 100 th Senate, which included Democrats reverting to majority control of the chamber; sweeping Republican majorities in the House and Senate beginning with the 104 th Congress; and a closely divided House and Senate in the 107 th . Only the second case has enough time periods before and after the change to make the interrupted time series results useful. For the other two time periods, a comparison of trends within the first Congress after the change is the major basis of the analysis. Similarly change in the presidency at the start of the 103 rd Congress is examined. The chapter concludes with a summary of analyses conducted and overall findings from each association. Descriptive Analysis Ideology is a key component of this study. ADA, ACU, DW Nominate, and E- score are variables that measure ideology. ADA and ACU measure a liberal- conservative spectrum through liberal and conservative interest group ratings, respectively. DW Nominate also measures liberal-conservative ideology but makes relative adjustments to ideology over time by assigning weights to scores. Each of these three measures of ideology captures legislative voting characteristics of a legislator. E-score transcends these traditional measures of liberalism and conservatism by considering not the characteristics of a legislator but how the vote by the legislator affects society through social benefit and social costs. Higher E-scores are associated with higher levels of social benefit as opposed to social costs and lower E-scores are associated with lower levels of social benefit as opposed to social costs (Kennedy, 2005, 170 pp. 45-49). Whether a legislator is liberal or conservative supporting policies that maximize benefits has many positive public policy implications. Pearson correlations among these four measures of ideology are relatively high. ADA is negatively correlated with ACU, DW Nominate, and E-score. That is, higher (lower) values for ADA are associated with lower (higher) values for ACU, DW Nominate, and E-score. The correlations among these variables are not perfectly positive or negative; variation exists among movement of the variables. What is important to this study is that E-score has variation and is not the same as traditional ideology. E-score variation indicates that economic efficiency is not the same as liberal-conservative ideology. This is true across both House and Senate. Referring to Figure 4.1, mean values for E-score, ACU, and ADA in the House indicate that not only do E-score values experience a wider variation than values for ADA and ACU the trend for E-scores appears to be higher over time. Mean ADA and ACU values are consistent through the time period of the study. ACU scores gradually climb and ADA scores remain virtually flat. Major shifts upward for ACU and downward for ADA occur with Republican control of both chambers of Congress beginning with the 104 th Congress. Figure 4.1 Mean E-score, ADA, and ACU Values -- U.S. House 0 10 20 30 40 50 60 70 80 99 100 101 102 103 104 105 106 107 108 Congress Mean Escore Mean ACU Mean ADA As indicated in Figure 4.2, in the Senate much more variation occurs in each of the variables, especially the E-scores. Throughout the time period examined, ADA and ACU scores shift several times but remain virtually flat. Increases in ACU scores and decreases in ADA scores reflect a shift in Republican control of both chambers of Congress beginning with the 104 th Congress. The shift is not to the extent of much higher conservatism in the House and does not appear to be a harbinger of changes in a long-term trend in ideology in the Senate. 171 Figure 4.2 Mean E-score, ADA, and ACU Values -- U.S. Senate 0 10 20 30 40 50 60 70 80 99 100 101 102 103 104 105 106 107 108 Congress Mean Escore Mean ACU Mean ADA A relatively stronger increase in conservatism in the House is not only apparent from ADA and ACU values from each Congress of the study, but also from weighted DW Nominate measures of liberalism and conservatism that experience relative changes within the institution and among its members. Referring to mean values for DW Nominate in the House in Figure 4.3, the trend in conservatism over the period of this study is generally upward with a sharp spike beginning with the 104 th Congress and gradually continuing thereafter. Positive numbers indicate a conservative orientation and negative numbers a liberal orientation. In each Congress a positive number is found 172 when the majority in charge is Republican and a negative number is found when the majority is charge is Democratic. Figure 4.3 Mean DW Nominate Values -- U.S. House -0.080 -0.060 -0.040 -0.020 0.000 0.020 0.040 0.060 0.080 99 100 101 102 103 104 105 106 107 108 Congress In the Senate DW Nominate values also experience a sharp increase beginning with the 104 th Congress, but the chamber does not become increasingly conservative. Changes in DW Nominate scores from the 99 th Congress to the 108 th Congress show very little absolute change from beginning to end, but much variability from one Congress to the next. A shift in orientation from liberal to conservative with the 104 th Congress is associated with relatively more extreme DW Nominate scores. Immediately preceding the shift in orientation, the Senate was becoming more liberal with higher negative DW Nominate scores. Immediately after the change in party control with the 173 104 th Congress the Senate was relatively more conservative, as DW Nominate scores were increasingly higher, positive numbers. DW Nominate scores further away from changes in party control with the 104 th Congress, i.e. DW Nominate scores for Congresses closer chronologically to the 99 th or 108 th Congress, experience less movement and are generally closer to zero, a point where ideology is relatively balanced between liberal and conservative. Figure 4.4 illustrates mean DW Nominate values for the U.S. Senate. Figure 4.4 Mean DW Nominate Values -- U.S. Senate -0.080 -0.060 -0.040 -0.020 0.000 0.020 0.040 0.060 0.080 99 100 101 102 103 104 105 106 107 108 Congress Mean E-score values generally increased in the House but remained flat in the Senate. Trends in the relationships among E-score values across political parties offer comparisons to traditional liberal-conservative measures of ideology. When comparing 174 changes in E-scores as a function of political party, variability among Republican and Democrat members of Congress is evident. Figures 4.5 and 4.6 indicate that in each chamber, mean E-scores are higher for Republicans than Democrats. The only exception is the 101 st House, where mean E-score values for Democrats were slightly higher than mean E-score values of Republican members. Figure 4.5 Mean E-scores for Republicans and Democrats -- U.S. House 0 10 20 30 40 50 60 70 80 90 100 99 100 101 102 103 104 105 106 107 108 Congress Mean Escore Rep Mean Escore Dem Wider absolute differences generally exist between E-scores of Republican and Democrat members in the House relative to the Senate. The distribution of E-score values for Democrats from the 99 th to 103 rd Congress is very similar. In the House a spike in E-scores was higher for Republicans than Democrats beginning with the 104 th Congress. Mean E-score values for Democrats in the Senate have increased since the 105 th Congress, while E-scores for Democrats in the House slowly fell from the 103 rd 175 Congress until 107 th and then jumped sharply. In comparing House and Senate E-scores this pattern for Democrats indicates more variation within the E-score in later Congresses as opposed to earlier Congresses in the study. Republican E-score values drop more sharply in the Senate than in the House in the 105 th Congress. Figure 4.6 Mean E-scores for Republicans and Democrats -- U.S. Senate 0 10 20 30 40 50 60 70 80 90 100 99 100 101 102 103 104 105 106 107 108 Congress Mean Escore Rep Mean Escore Dem Ideology Vector 176 In analyzing the variability of E-score relative to traditional measures of a liberal-conservative spectrum, the four ideology variables in the model ? ADA, ACU, DW Nominate, and E-score ? were regressed against each dependent variable. For Congresses where legislation was not considered and a dependent variable was not available, a scoring model of votes for legislation considered across all Congresses was used as the dependent variable for each policy area, minimum wage and medical 177 malpractice. Tables 4.3 and 4.4 present the effect of each ideology variable on legislative voting in the House and Senate, respectively, for Congresses where dependent variable legislation was available. For each Congress statistically significant relationships between each independent variable and the appropriate dependent variable are indicated at p < 0.01 and p < 0.05 levels of significance. One or more ideology variables were statistically significant in each House where dependent variable legislation was available for the Congress analyzed. For medical malpractice ACU was statistically significant in the 99 th Congress and ADA in the 104 th Congress. ACU, ADA, and DW Nominate were each statistically significant in the 100 th Congress. For the 107 th Congress ADA and E-score were statistically significant. For minimum wage dependent variable E-score is statistically significant in 101 st , 104 th , and 106 th Congresses. E-score is positively correlated with minimum wage in the 101 st and 106 th Congress, but inversely correlated with minimum wage in the 104 th Congress. An inverse correlation with minimum wage suggests that legislators with higher economic efficiency ratings do not always support economically efficient policies. Overall, the model produced statistically significant results for each ideology variable across the Congresses where medical malpractice or minimum wage legislation was available as dependent variables. Standardized coefficients indicate that the directional impact of each ideology variable on changes in the dependent variables was consistent with each hypothesis in all Congresses analyzed. Two exceptions are an inverse correlation for ACU and medical malpractice in the 99 th Congress and an inverse 178 correlation between E-score and votes in opposition to increasing the minimum wage in the 104 th Congress. Table 4.3 Regression Analysis of Ideology Influences on Legislative Voting in House HOUSE Medical Malpractice Congress Independent Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores ACU -0.637** 0.236 -0.511 -2.697** ADA 0.042 0.223 0.034 0.187 DW Nominate 6.715 15.562 0.059 0.432 99th E-score -0.079 0.076 -0.066 0.301 N 439 R2 0.289 Adjusted R2 0.282 Constant 107.431 ACU 0.639** 0.150 0.463 4.245** ADA -0.413** 0.154 -0.280 -2.682** DW Nominate 12.650* 5.517 0.092 2.293* 100th E-score -0.015 0.048 -0.011 -0.320 N 441 R2 0.633 Adjusted R2 0.630 Constant 49.529 ACU 0.114 0.169 0.094 0.674 ADA -0.777** 0.150 -0.626 -5.178** DW Nominate 12.381 11.218 0.114 1.104 104th E-score 0.070 0.073 0.046 0.969 N 445 R2 0.754 Adjusted R2 0.752 Constant 80.234 ACU 0.035 0.162 0.028 0.215 ADA -0.948** 0.145 -0.775 -6.531** DW Nominate -0.347 10.667 -0.003 -0.033 107th E-score 0.145* 0.068 0.092 2.115* N 444 R2 0.770 Adjusted R2 0.767 Constant 83.700 179 Minimum Wage Congress Independent Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores ACU 0.664** 0.121 0.612 5.495** ADA -0.247* 0.112 -0.224 -2.202* DW Nominate 1.908 5.424 0.019 0.352 101st E-score 0.192 0.052 0.108 3.688** N 442 R2 0.672 Adjusted R2 0.669 Constant 1.905 ACU 0.377 0.221 0.297 1.706 ADA -0.126 0.190 -0.094 -0.662 DW Nominate 54.007** 15.692 0.476 3.442** 104th E-score -0.232 0.100 -0.144 -2.329* N 445 R2 0.549 Adjusted R2 0.545 Constant 35.686 ACU 1.140** 0.203 0.908 5.612** ADA 0.473** 0.154 0.385 3.081** DW Nominate 24.821* 12.515 0.239 1.983* 106th E-score 0.153 0.055 0.111 2.792** N 440 R2 0.770 Adjusted R2 0.767 Constant -49.658 * p < 0.05 ** p < 0.01 In the Senate the model produced no statistically significant results for ideology variables when minimum wage was considered as the dependent variable in the 100 th Congress or medical malpractice in the 104 th Congress. For medical malpractice dependent variable ADA was statistically significant in the 107 th and 108 th Congresses and inversely correlated to movements in the dependent variable. For minimum wage dependent variable, ACU and DW Nominate are each statistically significant in the 101 st Congress and positively related to the dependent variable as hypothesized. 180 Table 4.4 Regression Analysis of Ideology Influences on Legislative Voting in Senate SENATE Medical Malpractice Congress Independent Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores ACU 0.351 0.357 0.271 0.984 ADA -1.124** 0.336 -0.876 -3.347** DW Nominate -25.546 22.129 -0.230 -1.154 107th E-score -0.055 0.094 -0.029 -0.587 N 102 R2 0.821 Adjusted R2 0.813 Constant 89.525 ACU 0.548 0.327 0.375 1.679 ADA -0.915** 0.248 -0.693 -3.682** DW Nominate -21.927 23.303 -0.194 -0.941 108th E-score 0.046 0.139 0.024 0.332 N 100 R2 0.791 Adjusted R2 0.781 Constant 66.940 Minimum Wage ACU 0.655* 0.257 0.572 2.553* ADA 0.107 0.177 0.094 0.609 DW Nominate 43.746* 17.875 0.425 2.447* 101st E-score 0.063 0.067 0.043 0.940 N 101 R2 0.839 Adjusted R2 0.832 Constant -6.505 * p < 0.05 ** p < 0.01 For Congresses where dependent variable legislation was not available a scoring model of votes including all dependent variable legislation for each dependent variable separately was considered. Scores representing each dependent variable are a compilation of scores for each individual Congress for those legislators serving across all Congresses in the model. Tables 4.5 and 4.6 summarize the results of each ideology variable regressed against a scoring model for each dependent variable in the House and Senate, respectively. 181 Table 4.5 Regression Analysis of Scoring Models from 99 th to 108 th Congress of Ideology Influences on Legislative Voting in House HOUSE Medical Malpractice Congress Independent Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores ACU -0.315 0.329 -0.320 -0.958 ADA -0.511 0.298 -0.558 -1.714 DW Nominate 48.998* 19.747 0.616 2.481* 105th E-score 0.033 0.090 0.029 0.367 N 444 R2 0.754 Adjusted R2 0.738 Constant 99.936 ACU -0.132 0.264 -0.124 -0.499 ADA -0.459* 0.186 -0.527 -2.467* DW Nominate 31.218 16.729 0.406 1.866 108th E-score 0.107 0.093 0.105 1.151 N 440 R2 0.772 Adjusted R2 0.757 Constant 83.315 Minimum Wage Congress Independent Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores ACU 0.981** 0.302 1.008 3.255** ADA 0.237 .0310 0.228 0.766 DW Nominate 6.685 7.758 0.069 0.862 100th E-score 0.080 0.075 0.083 1.061 N 441 R2 0.781 Adjusted R2 0.766 Constant -26.254 ACU 0.847** 0.331 0.854 2.560** ADA 0.274 0.306 0.249 0.894 DW Nominate -8.911 14.482 -0.096 -0.615 102nd E-score 0.426** 0.108 0.408 3.944** N 441 R2 0.787 Adjusted R2 0.773 Constant -36.949 ACU 1.198** 0.452 1.186 2.651** ADA 0.774 0.407 0.724 1.902 DW Nominate 27.632 23.277 0.307 1.187 103rd E-score 0.150 0.105 0.125 1.429 N 442 R2 0.734 Adjusted R2 0.717 Constant -72.171 ACU 1.302** 0.368 1.195 3.539** ADA 0.611 0.333 0.604 1.835 DW Nominate 11.790 22.050 0.134 0.535 105th E-score 0.303** 0.100 0.244 3.016** N 444 R2 0.748 Adjusted R2 0.732 Constant -75.137 182 ACU 0.849* 0.405 0.833 2.099* ADA 0.645 0.369 0.682 1.746 DW Nominate 48.552 25.451 0.565 1.908 107th E-score 0.153 0.144 0.127 1.062 N 444 R2 0.683 Adjusted R2 0.662 Constant -50.760 ACU 0.230 0.382 0.210 0.603 ADA -0.084 0.268 -0.087 -0.313 DW Nominate 48.495* 23.316 0.570 2.080* 108th E-score -0.076 0.126 -0.067 -0.602 N 440 R2 0.655 Adjusted R2 0.632 Constant 27.799 * p < 0.05 ** p < 0.01 In the House the model produced no statistically significant variables in the 101 st , 102 nd , 103 rd , and 106 th Congresses for medical malpractice dependent variable, and 99 th Congress for minimum wage dependent variable. For each of the other Congresses where a scoring model was used in the absence of dependent variable legislation for that Congress ? 105 th and 108 th for medical malpractice and 100 th , 102 nd , 103 rd , 105 th , 107 th , and 108 th for minimum wage dependent variables, respectively ? at least one ideology variable was statistically significant in the model. E-score was statistically significant in the 102 nd and 105 th Congresses with minimum wage as the dependent variable. 183 Table 4.6 Regression Analysis of Scoring Models from 99 th to 108 th Congress of Ideology Influences on Legislative Voting in Senate SENATE Medical Malpractice Congress Independent Variables Coefficient Estimates Standard Errors Standardized Coefficients t ? scores ACU -0.636 0.518 -0.714 -1.228 ADA -0.804 0.406 -0.908 -1.977 DW Nominate 69.303 43.413 0.819 1.596 101st E-score -0.475** 0.148 -0.366 -3.210** N 101 R2 0.823 Adjusted R2 0.783 Constant 143.292 ACU 1.198* 0.505 1.450 2.374* ADA 1.235* 0.528 1.556 2.339* DW Nominate 44.623 38.202 0.537 1.168 102nd E-score 0.544* 0.219 0.522 2.480* N 102 R2 0.799 Adjusted R2 0.754 Constant -79.471 ACU 1.523* 0.587 1.884 2.597* ADA 0.241 0.473 0.267 0.509 DW Nominate -62.401 47.850 -0.750 -1.304 103rd E-score 0.003 0.122 0.003 0.024 N 102 R2 0.813 Adjusted R2 0.769 Constant -34.502 Minimum Wage ACU -0.168 0.378 -0.146 -0.444 ADA -0.611 0.396 -0.550 -1.543 DW Nominate 80.651** 28.643 0.694 2.816** 102nd E-score -0.225 0.164 -0.155 -1.369 N 102 R2 0.942 Adjusted R2 0.929 Constant 89.282 ACU 0.952** 0.262 0.908 3.642** ADA -0.216 0.165 -0.195 -1.307 DW Nominate -13.559 28.396 -0.121 -0.477 104th E-score 0.011 0.162 0.005 0.068 N 103 R2 0.961 Adjusted R2 0.953 Constant 3.354 ACU 0.681* 0.278 0.556 2.450* ADA -0.877* 0.348 -0.773 -2.525* DW Nominate -35.295 27.890 -0.320 -1.266 105th E-score -0.184 0.111 -0.077 -1.667 N 100 R2 0.967 Adjusted R2 0.960 Constant 63.902 * p < 0.05 ** p < 0.01 184 In the Senate the model produced no statistically significant results in the 99 th , 100 th , 105 th , and 106 th Congresses with medical malpractice as the dependent variable and in the 99 th , 103 rd , 106 th , 107 th , and 108 th with minimum wage as the dependent variable. In the remaining Congresses ? 101 st , 102 nd , 103 rd for medical malpractice and 102 nd , 104 th , and 105 th for minimum wage dependent variables, respectively ? where a scoring model was used in compiling a dependent variable at least one ideology variable was statistically significant in each Congress. E-score was statistically significant in the 101 st and 102 nd Congresses with medical malpractice as the dependent variable. Multivariate Analysis Ideology is the key component in analyzing variability of E-score relative to legislative liberalism and conservatism. Including self-interest and party environment vectors in a multivariate analysis is important in understanding influences affecting legislative voting. This section is divided into two parts for analyzing each dependent variable in the model ? medical malpractice and minimum wage. The analysis tests the respective hypotheses of each dependent variable in the House and Senate in measuring the impact of each variable in the model on changes in the dependent variables. Self- interest and chamber environment vectors are added to the multivariate analysis with the ideology vector. Self-interest variables are contributions received by legislators. For medical malpractice dependent variable self-interest variables are health and law contributions, and business and labor contributions for minimum wage dependent variable. 185 Pearson correlations were used to identify strong bivariate correlations among ideology variables and party unity between each other and also control variables within the model. Initial regression runs indicated multicollinearity, precluding using highly correlated variables within single regression runs. For this reason analyzing all variables in the model simultaneously is problematic. Strong bivariate correlations required modification to existing hypotheses for testing. A hypothesis for DW nominate was added as a measure of liberalism and conservatism that captures the time aspect of ideology. Ideological differences between the legislator and party were modified to analyze distinctions between the legislator?s personal ideology and the ideology of the party to which he or she belongs. Tables 4.7 and 4.8 summarize each hypothesis for medical malpractice and minimum wage dependent variables, respectively. Hypotheses are modified in the model for testing relationships between variables in answering the research question. In order to standardize the measurement of each dependent variable, hypotheses for minimum wage express the relationship between variables in the model and the dependent variable, but the expected sign of the regression coefficient is modified to reflect economic efficiency. That is, support for medical malpractice legislation enhances economic efficiency, but support for minimum wage legislation suppresses economic efficiency. Coding minimum wage such that a vote against the legislation is a vote for economic efficiency produces a regression coefficient that is consistent across both dependent variables. The expected sign of the regression coefficient in Table 4.8 reflects coding of the variable in economically efficient terms. 186 Table 4.7 Hypotheses for medical malpractice reform policy area Hypotheses Expected Sign of Regression Coefficient H 1: Legislators with higher E-scores vote in support of medical malpractice reform. + H 2: Legislators with higher ADA scores vote in opposition to medical malpractice reform. _ H 3: Legislators with higher ACU scores vote in support of medical malpractice reform. + H 4: Legislators with higher DW Nominate scores vote in support of medical malpractice reform. + H 5: Legislators with higher health care political contributions to total contributions vote in support of malpractice reform. + H 6: Legislators with higher legal political contributions to total contributions vote in opposition to medical malpractice reform. _ H 7: Republican legislators are likely to vote for malpractice reform more often than Democratic legislators. _ H 8: The closer senators are to the end of their current term in office, the more likely they are to support malpractice reform. + H 9: The longer a legislator has served, the more likely he or she supports medical malpractice reform. + H 10: Legislators from the minority party (House or Senate) are more likely than majority party legislators to support medical malpractice reform + H 11: The greater the ideological division between the legislator and the median ideology of the legislator?s party, the more likely the legislator supports medical malpractice reform. + Modifications to hypotheses for minimum wage dependent variable were necessary to avoid multicollinearity associated with using highly correlated variables in testing each existing hypothesis. Differences in the hypothesis for minimum wage remain self-interest variables measuring political contributions and the inverse relationship between support for minimum wage legislation and economic efficiency. 187 Table 4.8 Hypotheses for minimum wage legislation policy area Hypotheses Expected Sign of Regression Coefficient H 1: Legislators with higher E-scores vote in opposition to increasing the minimum wage. + H 2: Legislators with higher ADA scores vote in support of increasing the federal minimum wage. _ H 3: Legislators with higher ACU scores vote in opposition to increasing the federal minimum wage. + H 4: Legislators with higher DW Nominate scores vote in opposition to increasing the minimum wage. + H 5: Legislators with higher business political contributions to total contributions vote in opposition to increasing the minimum wage. + H 6: Legislators with higher labor political contributions to total contributions vote in support of increasing the minimum wage. _ H 7: Republican legislators are less likely to vote for increasing the minimum wage more often than Democrats. _ H 8: The closer senators are to the end of their current term in office, the less likely they are to support increasing the minimum wage. + H 9: The longer a legislator has served, the less likely he or she will support increasing the minimum wage. + H 10: Legislators from the minority party (House or Senate) are less likely than majority party legislators to support increasing the federal minimum wage. + H 11: The greater the division between the ideology of the legislator and the median ideology of the legislator?s party, the less likely the legislator supports increasing the federal minimum wage. + For each dependent variable a base model of independent and control variables was identified that will be consistent for hypothesis testing for each dependent variable. For medical malpractice variables included in the base model are E-score, Health Contributions, Lawyer Contributions, First Elected, Northeast, South, West, Differences 188 in Legislator DW Nominate and Median Party, Medical Malpractice Crisis, Ratio of Federal Spending, Per Capita Income, Percentage African American, and Percentage Hispanic. For minimum wage dependent variable the base models are identical to those variables included in medical malpractice except Medical Malpractice Crisis is replaced by Minimum Wage Laws and Health and Law are replaced by Business and Labor as measures of self-interest. Five variables ? Party Unity, ACU, ADA, DW Nominate, and Legislative Party ? are highly correlated and are not included concurrently in a regression analysis of changes in the dependent variable. These five variables are substituted for testing the appropriate hypothesis. For example, in testing H 2 for either dependent variable, ADA variable was included in addition to the base model, but Party Unity, ACU, DW Nominate, and Legislative Party were not. The same logic applies to testing of H 3. ACU was included in the model but Party Unity, ADA, DW Nominate, and Legislative Party were not. The key issue in standardizing hypothesis analysis is maintaining consistency across observations. The base model of independent and control variables for each dependent variable was used in testing each hypothesis and measuring changes in each dependent variable. Base model variables are identified under each dependent variable category ? medical malpractice and minimum wage ? with each of the five highly correlated variables specified for the appropriate hypothesis test. 189 Hypothesis Testing The model uses base variables across each hypothesis for testing the impact of independent variables on the dependent variable, medical malpractice reform and then minimum wage. Party Unity, ACU, ADA, DW Nominate, and Legislative Party are highly correlated and are used individually in testing the appropriate hypothesis. The House and Senate base models are the same with the exception of Current Term, which is used only in the Senate model and two self-interest variables (Health and Law; Business and Labor) that vary with the dependent variable used. Current Term measures the years served since the senator was last elected or reelected to office. Table 4.9 summarizes the variables used in the model and the corresponding hypothesis tested. Variables in parentheses represent those substituted when minimum wage is the dependent variable. 190 Table 4.9 Variables Used in Testing Hypotheses for Medical Malpractice and Minimum Wage Base Independent and Control Variables Hypotheses E-score First Elected Health (Business) Lawyer (Labor) Percent African American Percent Hispanic Per capita income Ratio federal spending per tax dollar received Difference legislator DW Nominate and median party Northeast West South Medical Malpractice Crisis (Minimum Wage Laws) Variables used in testing H 1, 5, 6, 9, and11 in House and Senate with H 8 tested in Senate only. Current Term H 8 (Senate only) Substituted Variables Tested with Base Variables Party Unity H 7 Tested in House and Senate ADA H 2 Tested in House and Senate ACU H 3 Tested in House and Senate DW Nominate H 4 Tested in House and Senate Legislative Party H 10 Tested in House and Senate For each Congress (99-108) six separate regression analyses were conducted for the House and Senate in order to avoid including highly correlated variables within the same model. The base model was run first and then Party Unity, ADA, ACU, DW Nominate, and Legislative Party were each separately added to the model and the impact of each addition analyzed. House Table 4.10 presents the base model analysis for the House using medical malpractice as the dependent variable. A strong self-interest component to legislative decision exists through statistically significant political contributions to Health and Lawyer groups. E-score is also a strong predictor of legislative behavior, as a measure of 191 ideology. Referring to Table 4.9 for a summary of base and substituted variables used in testing each hypothesis, the following hypotheses were tested in the Base Model with results for each test presented here. In testing Hypothesis 1a: Legislators with higher E- scores vote in support of medical malpractice reform, E-score was found to be a statistically significant base variable for each House in the analysis except in the 101st Congress. In each Congress in which E-score was statistically significant levels of significance of p < 0.01 indicate the results most likely were not the result of random occurrences. Standardized coefficients for the variable indicate positive movement between higher E-scores and economically efficient voting in each Congress except the 99th Congress; in the 99th Congress the correlation is negative. Positive correlations between E-score and support for medical malpractice reform are consistent with hypothesized associations. E-score is a powerful predictor of behavior in the base model. Table 4.10 Regression Analysis of Base Model for 99th to 108th House: Medical Malpractice Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores Year Elected -0.233 0.248 -0.044 -0.940 Northeast -10.125 6.384 -0.101 -1.586 South -1.025 5.921 -0.012 -0.173 West 6.470 6.725 0.060 0.962 Diff DW Nom -28.910 17.647 -0.076 -1.638 Fed Spending 1.979 9.798 0.011 0.202 Per Cap Inc 0.003 0.001 0.132 2.095* African Amer. -77.031 38.172 -0.121 -2.018* Hispanic -38.095 31.563 -0.070 -1.207 Med Mal Crisis 6.851 4.601 0.083 1.489 Health -9.235 10.706 -0.045 -0.863 Lawyer 27.804 11.027 0.132 2.521* 99th E-score -0.485 0.061 -0.410 -7.983** N 439 R2 0.247 Adjusted R2 0.221 Constant 65.407 Year Elected 0.148 0.265 0.024 0.557100th Northeast 1.037 7.542 0.009 0.137 192 Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores South -0.335 7.503 -0.003 -0.045 West -16.710 8.326 -0.135 -2.007* Diff DW Nom 13.117 9.841 0.058 1.333 Fed Spending 19.885 11.644 0.093 1.708 Per Cap Inc -0.004 0.001 -0.176 -2.840** African Amer. 68.050 47.135 0.087 1.444 Hispanic 32.955 38.103 0.051 0.865 Med Mal Crisis 8.430 5.280 0.084 1.597 Health 59.178 13.629 0.220 4.342** Lawyer -62.904 13.275 -0.233 -4.738** E-score 0.468 0.069 0.329 6.769** N 441 R2 0.353 Adjusted R2 0.332 Constant 57.990 Year Elected -0.621 0.806 -0.098 -0.770 Northeast -13.229 17.082 -0.156 -0.774 South 23.363 16.038 0.317 1.457 West 15.828 19.355 0.193 0.818 Diff DW Nom 11.174 37.300 0.043 0.300 Fed Spending -25.908 32.501 -0.145 -0.797 Per Cap Inc 0.002 0.003 0.159 0.753 African Amer. -47.005 91.384 -0.087 -0.514 Hispanic -55.786 70.500 -0.165 -0.791 Med Mal Crisis 11.459 11.073 0.170 1.035 Health 53.182 22.290 0.313 2.386* Lawyer -83.397 37.792 -0.304 -2.207* 101st E-score -0.357 0.209 -0.216 -1.709 N 442 R2 0.353 Adjusted R2 0.182 Constant 59.825 Year Elected 0.699 0.575 0.111 1.214 Northeast -5.377 12.162 -0.062 -0.442 South 9.897 11.951 0.132 0.828 West -3.391 13.828 -0.040 -0.245 Diff DW Nom -13.578 23.790 -0.052 -0.571 Fed Spending 27.640 23.454 0.159 1.178 Per Cap Inc 0.003 0.002 0.215 1.502 African Amer. 8.411 55.786 0.016 0.151 Hispanic -50.497 51.719 -0.151 -0.976 Med Mal Crisis 2.464 6.994 0.036 0.352 Health -14.094 54.300 -0.023 -0.260 Lawyer -287.526 87.119 -0.311 -3.300** 102nd E-score 0.641 0.093 0.679 6.890** N 441 R2 0.674 Adjusted R2 0.592 Constant -36.450 Year Elected -0.429 0.702 -0.071 -0.611103rd Northeast -19.172 14.925 -0.221 -1.285 193 Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores South 1.031 14.340 0.014 0.072 West -1.851 17.427 -0.022 -0.106 Diff DW Nom 45.533 36.598 0.143 1.244 Fed Spending 25.657 26.837 0.156 0.956 Per Cap Inc 0.003 0.002 0.203 1.176 African Amer. -3.059 67.856 -0.006 -0.045 Hispanic -11.615 58.563 -0.036 -0.198 Med Mal Crisis 9.341 8.726 0.138 1.070 Health -38.133 60.397 -0.068 -0.631 Lawyer -405.586 97.666 -0.435 -4.153** E-score 0.458 0.114 0.420 4.000** N 442 R2 0.500 Adjusted R2 0.376 Constant -25.668 Year Elected 0.164 0.186 0.028 0.880 Northeast -6.748 5.379 -0.059 -1.255 South -4.288 5.320 -0.044 -0.806 West -0.786 5.784 -0.007 -0.136 Diff DW Nom 36.544 9.023 0.138 4.050** Fed Spending 7.292 9.748 0.034 0.748 Per Cap Inc 0.001 0.001 0.045 0.853 African Amer. 6.059 25.102 0.010 0.241 Hispanic -26.254 19.093 -0.060 -1.375 Med Mal Crisis -1.114 3.509 -0.012 -0.317 Health 73.428 33.916 0.068 2.165* Lawyer 170.462 40.289 -0.143 -4.231** 104th E-score 1.030 0.053 0.670 19.544** N 445 R2 0.616 Adjusted R2 0.604 Constant -28.535 Year Elected -0.291 0.830 -0.046 -0.351 Northeast 18.581 16.772 -0.214 -1.108 South 10.376 17.647 0.138 0.588 West 7.944 19.865 0.095 .0400 Diff DW Nom 50.702 43.757 0.139 1.159 Fed Spending 21.728 35.930 0.131 0.605 Per Cap Inc 0.002 0.002 0.157 0.764 African Amer. -36.428 74.510 -0.069 -0.489 Hispanic -52.742 67.716 -0.177 -0.779 Med Mal Crisis 8.751 9.715 0.129 0.901 Health 8.137 63.285 0.016 0.129 Lawyer 229.941 124.917 -0.228 -1.841 105th E-score 0.560 0.134 0.499 4.173** N 444 R2 0.374 Adjusted R2 0.218 Constant -16.421 106th Year Elected 0.226 0.642 0.036 0.352 194 Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores Northeast -6.684 13.499 -0.077 -0.495 South 13.880 14.022 0.185 0.990 West 13.533 15.474 0.161 0.875 Diff DW Nom 24.561 34.050 0.068 0.721 Fed Spending 2.639 29.400 0.017 0.090 Per Cap Inc 0.001 0.002 0.147 0.868 African Amer. -37.598 58.120 -0.072 -0.647 Hispanic -59.041 54.768 -0.202 -1.078 Med Mal Crisis 4.752 7.710 0.070 0.616 Health -37.391 49.985 -0.073 -0.748 Lawyer -291.115 84.958 -0.337 -3.427** E-score 0.620 0.090 0.645 6.924** N 440 R2 0.612 Adjusted R2 0.515 Constant 6.449 Year Elected 0.303 0.209 0.046 1.447 Northeast 8.431 5.939 0.068 1.420 South -0.279 5.526 -0.003 -0.050 West -3.452 6.064 -0.028 -0.569 Diff DW Nom 10.295 15.694 0.020 0.656 Fed Spending 4.343 10.186 0.023 0.426 Per Cap Inc 0.000 0.001 -0.016 -0.304 African Amer. 11.822 26.372 0.019 0.448 Hispanic 0.605 20.171 0.001 0.030 Med Mal Crisis 0.023 3.768 0.000 0.006 Health 72.399 30.271 0.075 2.392* Lawyer -229.406 40.162 -0.192 -5.712** 107th E-score 1.104 0.054 0.703 20.627** N 444 R2 0.626 Adjusted R2 0.614 Constant -6.002 Year Elected -0.436 0.643 -0.075 -0.678 Northeast -11.882 13.781 -0.137 -0.862 South -1.479 15.271 -0.020 -0.097 West -5.588 16.676 -0.066 -0.335 Diff DW Nom 47.840 37.096 0.132 1.290 Fed Spending 8.330 22.131 0.074 0.376 Per Cap Inc 0.001 0.001 0.162 1.043 African Amer. -57.796 63.081 -0.110 -0.916 Hispanic -39.327 55.186 -0.146 -0.713 Med Mal Crisis -1.352 8.382 -0.020 -0.161 Health -0.182 39.812 0.000 -0.005 Lawyer -159.504 106.011 -0.157 -1.505 108th E-score 0.751 0.106 0.736 7.062** N 440 R2 0.565 Adjusted R2 0.457 Constant -11.767 * p < 0.05, ** p < 0.01 195 Because DW Nominate is highly correlated with other ideology variables such as the E-score, differences in a legislator?s DW Nominate scores and median DW Nominate scores for his or her party (Diff DW Nom) were analyzed to test Hypothesis 11a: The greater the ideological division between the legislator and the median ideology of the legislator?s party, the more likely the legislator supports medical malpractice reform. In the Base Model the variable was statistically significant (p < 0.01) only in the 104th House and the coefficient in this year was positive. A positive standardized coefficient reflects a positive relationship between higher DW Nominate difference scores, an indication of more legislator conservatism, relative to conservatism of his or her median party. The more division that exists between a legislator?s ideology and median party ideology the greater is the legislator?s support for economically efficient policies. In the years where the variable did not achieve statistical significance in the model, the signs for the coefficients varied from year to year. Results for this test were NOT as hypothesized. Two self-interest hypotheses tested in the Base Model. These are Hypothesis 5a: Legislators with higher health care political contributions to total contributions vote in support of malpractice reform and Hypothesis 6a: Legislators with higher legal political contributions to total contributions vote in opposition to medical malpractice reform. In reporting these results each hypothesis is considered together in illustrating the effect of self-interest on supporting or opposing economically efficient legislation. Health is statistically significant in the 100 th , 101 st , 104 th , and 107 th Congresses and Law is statistically significant the 99 th through 107 th Congresses. Directional movements between each self interest variable were as hypothesized with Health contributions 196 contributing to greater support for economically efficient policymaking and Law contributions negatively related to a legislator?s support for economically efficient policies. These results held for each Congress with the exception of the 99 th House, where Law contributions were positively related to economic efficiency. With lower standardized coefficients self-interest variables generally did not produce higher per unit effects in measuring support for medical malpractice reform. While statistically significant at p < 0.05, higher p values indicate a greater presence of random chance affecting the results. Hypothesis 9a reads: The longer a legislator has served, the more likely he or she supports medical malpractice reform. Accordingly, legislators with longer tenures were expected to take the economically efficient position on this reform more often than legislators with shorter tenures. The variable was not statistically significant in the Base Model for any Congress, and the sign for the coefficient was not consistent. Coefficients of determination (R square and adjusted R square) indicate how well the regression line approximates actual data points. Adjusted R square adjusts for the number of independent variables in the equation and gives a more accurate picture of well the independent variables explain the dependent variable?s behavior. According to adjusted R square in Base Model analysis, over 60 percent of the variation in the vote is explained by the model applied to the 104th and 107th Houses and the least variation is explained for the 101st House (about 8 percent). As the following analyses will indicate, coefficients of determination were generally lower in the base model than when substituting party unity, ADA, ACU, DW Nominate, and legislative party into the model. Each of the substituted variables is a measure of either ideology (ADA, ACU, 197 and DW Nominate) or party (party unity and legislative party) and as substitutions indicate, ideology and party variables are strong predictors of legislative behavior in this model. Adding those variables to the model enhances the fit of the regression and produces higher coefficients of determination. Party Unity was the first variable substituted into the base. Consistent with previous research results reported in Table 4.10 indicate that political party is an important factor in legislative voting. Party unity measures how closely legislators adhere to party positions in voting decisions in supporting or opposing policies. In each Congress coefficients of determination increased when party unity was substituted into the model. Party unity is a statistically significant variable in each Congress in the model and Republicans overwhelming support medical malpractice reform in each Congress except the 99th. In the 99th Congress Republicans are not unified in support of medical malpractice reform. E-score was statistically significant in the base model in the 100th, 102nd, 103rd, 104th, 105th, 106th, and 108th Houses, but in each of these years E-score failed to reach statistical significance when considered with party unity in the model. This is an indication of the strength of political parties in legislative voting that might displace economic efficiency in predicting behavior and lends support to Hypothesis 7a: Republican legislators are likely to vote for malpractice reform more often than Democratic legislators. Introducing party unity into the model produces mixed effects for the self- interests of legislators. In the 100th and 101st Houses both measures of self-interest were statistically significant in the base model but failed tests of statistical significance 198 when measured with party unity. Lawyer was statistically significant in the base model but failed tests of statistical significance in the 102nd, 103rd, and 106th Houses with party unity substitution. Changes in statistical significance for self interest variables when considered with a party variable are indicative of the relative strength of party vis- ?-vis self-interest in explaining legislative behavior in the model. Like E-score as a measure of ideology, changes in statistical significance for self-interest, a strong predictor of behavior, must be considered with the collective effects of party principles that guide legislators. Also the signs for both self-interest variables follow no consistent pattern. Table 4.11 Regression Analysis of Base Model and Party Unity Substitution for 99th to 108th House: Medical Malpractice Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores Year Elected -0.325 0.246 -0.062 -1.323 Northeast -7.289 6.344 -0.073 -1.149 South -2.717 5.856 -0.031 -0.464 West 4.866 6.644 0.045 0.732 Diff DW Nom -31.703 17.410 -0.083 -1.821 Fed Spending 3.004 9.661 0.017 0.311 Per Cap Inc 0.002 .001 0.116 1.857 African Amer. -63.779 37.814 -0.100 -1.687 Hispanic -34.759 31.121 -0.064 -1.117 Med Mal Crisis 7.423 4.538 0.090 1.636 Health -0.966 10.817 -0.005 -0.089 Lawyer 25.752 10.884 0.122 2.366* E-score -0.275 0.085 -0.232 -3.231** 99 th Party Unity 0.124 0.036 0.248 3.471** N 439 R2 0.271 Adjusted R2 0.243 Constant 58.256 Year Elected 0.236 0.200 0.039 1.180 Northeast -3.994 5.683 -0.033 -0.703 South 2.258 5.648 0.021 0.400 West -11.117 6.273 -0.090 -1.772 Diff DW Nom 15.156 7.405 0.067 2.047* Fed Spending 14.856 8.766 0.070 1.695 Per Cap Inc -0.002 0.001 -0.109 -2.320* 100th African Amer. 29.878 35.535 0.038 0.841 199 Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores Hispanic 36.098 28.671 0.056 1.259 Med Mal Crisis 4.453 3.980 0.045 1.119 Health 5.829 10.711 0.022 0.544 Lawyer -11.284 10.427 -0.042 -1.082 E-score -0.006 0.059 -0.004 -0.106 Party Unity -0.446 0.026 -0.743 -17.250** N 441 R2 0.635 Adjusted R2 0.622 Constant 71.400 Year Elected -0.071 0.448 -0.011 -0.159 Northeast -2.115 9.490 -0.025 -0.223 South -5.378 9.260 -0.073 -0.581 West -3.318 10.838 -0.040 -0.306 Diff DW Nom 17.488 20.604 0.067 0.849 Fed Spending 19.632 18.451 0.110 1.064 Per Cap Inc 0.001 0.001 0.057 0.487 African Amer. 65.285 51.554 0.121 1.266 Hispanic -14.090 39.124 -0.042 -0.360 Med Mal Crisis -1.015 6.226 -0.015 -0.163 Health 3.870 13.154 0.023 0.294 Lawyer -26.304 21.549 -0.096 -1.221 E-score -0.010 0.120 -0.006 -0.084 101st Party Unity -0.349 0.033 -0.864 -10.617** N 442 R2 0.807 Adjusted R2 0.751 Constant 29.073 Year Elected 0.044 0.460 0.007 0.096 Northeast -1.805 9.469 -0.021 -0.191 South -5.657 9.649 -0.075 -0.586 West -6.137 10.755 -0.073 -0.571 Diff DW Nom -2.267 18.583 -0.009 -0.122 Fed Spending 23.209 18.239 0.133 1.272 Per Cap Inc 0.001 0.002 0.063 0.549 African Amer. 55.474 44.067 0.103 1.259 Hispanic -2.298 41.001 -0.007 -0.056 Med Mal Crisis -2.992 5.511 -0.044 -0.543 Health -3.653 42.228 -0.006 -0.087 Lawyer -77.717 76.389 -0.084 -1.017 E-score 0.024 0.127 0.025 0.189 102nd Party Unity -0.332 0.056 -0.841 -5.927** N 441 R2 0.807 Adjusted R2 0.754 Constant 26.190 Year Elected -0.360 0.419 -0.060 -0.859 Northeast -4.812 9.042 -0.056 -0.532 South -6.817 8.608 -0.091 -0.792 West -5.383 10.422 -0.064 -0.517 Diff DW Nom 26.651 21.960 0.083 1.214 Fed Spending 24.288 16.041 0.147 1.514 103rd Per Cap Inc 0.001 0.001 0.065 0.621 200 Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores African Amer. 47.375 40.887 0.089 1.159 Hispanic -14.410 35.004 -0.045 -0.412 Med Mal Crisis -1.169 5.326 -0.017 -0.219 Health 7.726 36.405 0.014 0.212 Lawyer -39.510 69.460 -0.042 -0.569 E-score 0.087 0.078 0.079 1.107 Party Unity -0.323 0.033 -0.842 -9.724** N 442 R2 0.825 Adjusted R2 0.777 Constant 19.638 Year Elected 0.177 0.152 0.031 1.164 Northeast -5.377 4.410 -0.047 -1.219 South -5.089 4.361 -0.052 -1.167 West -4.985 4.750 -0.044 -1.049 Diff DW Nom 63.334 7.631 0.238 8.300** Fed Spending 16.480 8.016 0.076 2.056* Per Cap Inc 0.002 0.001 0.105 2.381* African Amer. 20.983 20.602 0.035 1.018 Hispanic -11.799 15.682 -0.027 -0.752 Med Mal Crisis -4.318 2.885 -0.047 -1.497 Health 56.531 27.825 0.052 2.032* Lawyer -67.156 33.809 -0.057 -1.986* E-score 0.131 0.076 0.085 1.711 104th Party Unity -0.375 0.026 -0.714 -14.255** N 445 R2 0.742 Adjusted R2 0.734 Constant -6.616 Year Elected -0.233 0.409 -0.037 -0.570 Northeast -8.207 8.305 -0.095 -0.988 South -8.364 8.820 -0.111 -0.948 West -8.990 9.879 -0.107 -0.910 Diff DW Nom 24.601 21.661 0.067 1.136 Fed Spending 27.772 17.713 0.167 1.568 Per Cap Inc 0.001 0.001 0.136 1.345 African Amer. 14.659 36.937 0.028 0.397 Hispanic 11.394 33.747 0.038 0.338 Med Mal Crisis -1.192 4.850 -0.018 -.246 Health 4.440 31.189 0.009 0.142 Lawyer 3.370 64.215 0.003 0.052 E-score 0.042 0.078 0.038 0.546 105th Party Unity -0.351 0.027 -0.898 -12.771** N 444 R2 0.851 Adjusted R2 0.810 Constant -3.291 Year Elected -.007 0.394 -0.001 -0.019 Northeast -6.682 8.272 -0.077 -0.808 South -3.903 8.800 -0.052 -0.444 West -3.566 9.656 -0.042 -0.369 106th Diff DW Nom 15.909 20.885 0.044 0.762 201 Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores Fed Spending 19.702 18.107 0.128 1.088 Per Cap Inc 0.001 0.001 0.149 1.436 African Amer. 10.952 35.991 0.021 0.304 Hispanic 2.252 34.194 0.008 0.066 Med Mal Crisis 0.440 4.747 0.007 0.093 Health -18.058 30.699 -0.035 -0.588 Lawyer -80.594 56.717 -0.093 -1.421 E-score 0.090 0.079 0.093 1.136 Party Unity -0.315 0.034 -0.816 -9.354 N 440 R2 0.857 Adjusted R2 0.818 Constant 4.273 Year Elected 0.224 0.161 0.034 1.392 Northeast 2.247 4.570 0.018 0.492 South -0.041 4.238 0.000 -0.010 West -4.902 4.652 -0.040 -1.054 Diff DW Nom -1.389 12.057 -0.003 -0.115 Fed Spending 8.767 7.817 0.046 1.122 Per Cap Inc 0.000 0.001 0.009 0.207 African Amer. -32.711 20.397 -0.051 -1.604 Hispanic 8.994 15.479 0.021 0.581 Med Mal Crisis -.300 2.890 -0.003 -0.104 Health 42.491 23.284 0.044 1.825 Lawyer -124.729 31.419 -0.104 -3.970** E-score 0.151 0.070 0.097 2.174* 107th Party Unity -0.411 0.024 -0.750 -16.908** N 444 R2 0.780 Adjusted R2 0.773 Constant 35.137 Year Elected 0.043 0.374 0.007 0.115 Northeast -7.599 7.959 -0.088 -0.955 South -8.569 8.834 -0.114 -0.970 West -11.260 9.633 -0.134 -1.169 Diff DW Nom 16.955 21.605 0.047 0.785 Fed Spending 18.921 12.805 0.168 1.478 Per Cap Inc 0.001 0.001 0.123 1.370 African Amer. 6.736 36.919 0.013 0.182 Hispanic 18.329 32.319 0.068 0.567 Med Mal Crisis -0.192 4.835 -0.003 -0.040 Health -16.004 23.012 -0.042 -0.695 Lawyer -39.942 62.239 -0.039 -0.642 E-score 0.102 0.088 0.100 1.156 108th Party Unity -0.312 0.030 -0.843 -10.263** N 440 R2 0.858 Adjusted R2 .819 Constant 5.645 * p < 0.05, ** p < 0.01 202 Table 4.12 summarizes regression results with ADA add to the base model. Consistent with Hypothesis 2a: Legislators with higher ADA scores vote in opposition to medical malpractice reform, ADA is statistically significant at the p < 0.01 level in each Congress analyzed and negatively related to economically efficient legislation in each Congress except the 99 th . From the standardized coefficients for each regression analysis with ADA in the model changes in the variable produce relatively greater changes in the dependent variable, medical malpractice, than other relationships in the model. The strength of these relationships, as indicated by the standardized coefficient, is evidence of the impact of ideology on legislative behavior. Table 4.12 Regression Analysis of Base Model and ADA Substitution for 99th to 108th House: Medical Malpractice Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t ? scores Year Elected -0.364 0.242 -0.069 -1.505 Northeast -7.048 6.204 -0.070 -1.136 South 0.036 5.748 0.000 0.006 West 3.602 6.531 0.034 0.552 Diff DW Nom -29.103 17.103 -0.077 -1.702 Fed Spending 5.482 9.504 0.030 0.577 Per Cap Inc 0.001 0.001 0.064 1.031 African Amer. -26.410 38.401 -0.042 -0.688 Hispanic -31.213 30.605 -0.058 -1.020 Med Mal Crisis 7.051 4.465 0.085 1.579 Health 5.482 10.730 0.027 0.511 Lawyer 21.993 10.784 0.104 2.039* E-score -0.170 0.084 -0.143 -2.016* 99th ADA 0.488 0.093 0.393 5.249** N 439 R2 0.298 Adjusted R2 0.272 Constant 42.076 Year Elected 0.081 0.203 0.013 0.399 Northeast -0.340 5.753 -0.003 -0.059 South -8.651 5.781 -0.080 -1.496 West -12.198 6.359 -0.099 -1.918 Diff DW Nom 8.625 7.516 0.038 1.148 Fed Spending 16.552 8.884 0.078 1.863 100th Per Cap Inc -0.001 0.001 -0.028 -0.588 203 Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t ? scores African Amer. 0.314 36.405 0.000 0.009 Hispanic 19.423 29.116 0.030 0.667 Med Mal Crisis 2.806 4.049 0.028 0.693 Health -3.070 11.040 -0.011 -0.278 Lawyer -19.021 10.464 -0.070 -1.818 E-score 0.060 0.058 0.043 1.042 ADA -1.095 0.065 -0.739 -16.729** N 441 R2 0.625 Adjusted R2 0.611 Constant 103.966 Year Elected -0.384 0.463 -0.060 -0.828 Northeast -1.522 9.875 -0.018 -0.154 South -4.034 9.602 -0.055 -0.420 West -2.101 11.252 -0.026 -0.187 Diff DW Nom 7.504 21.414 0.029 0.350 Fed Spending 8.960 18.977 0.050 0.472 Per Cap Inc 0.004 0.002 0.296 2.422 African Amer. 20.142 52.881 0.037 0.381 Hispanic -36.952 40.512 -0.109 -0.912 Med Mal Crisis -6.805 6.611 -0.101 -1.029 Health -11.742 14.337 -0.069 -0.819 Lawyer -16.061 22.707 -0.059 -0.707 E-score 0.043 0.126 0.026 0.344 101st ADA -0.879 0.088 -0.937 -10.036** N 442 R2 0.791 Adjusted R2 0.730 Constant 34.099 Year Elected 0.286 0.522 0.045 0.547 Northeast -5.614 10.806 -0.065 -0.520 South -0.711 10.968 -0.009 -0.065 West -5.247 12.296 -0.062 -0.427 Diff DW Nom -6.247 21.222 -0.024 -0.294 Fed Spending 15.740 21.066 0.090 0.747 Per Cap Inc 0.003 0.002 0.244 1.916 African Amer. 10.904 49.569 0.020 0.220 Hispanic -47.993 45.956 -0.144 -1.044 Med Mal Crisis -4.655 6.482 -0.069 -0.718 Health 25.571 49.328 0.041 0.518 Lawyer -161.619 84.007 -0.175 -1.924 E-score 0.195 0.142 0.207 1.372 102nd ADA -0.619 0.161 -0.623 -3.857** N 441 R2 0.747 Adjusted R2 0.678 Constant 14.178 Year Elected -0.257 0.469 -0.043 -0.547 Northeast -5.251 10.117 -0.061 -0.519 South -7.105 9.631 -0.095 -0.738 West -3.426 11.643 -0.041 -0.294 Diff DW Nom 30.961 24.514 0.097 1.263 103rd Fed Spending 14.188 17.983 0.086 0.789 204 Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t ? scores Per Cap Inc 0.003 0.002 0.205 1.777 African Amer. -14.412 45.350 -0.027 -0.318 Hispanic -39.842 39.275 -0.125 -1.014 Med Mal Crisis -1.215 5.973 -0.018 -0.203 Health 49.156 41.761 0.088 1.177 Lawyer -114.176 74.513 -0.122 -1.532 E-score 0.023 0.093 0.021 0.247 ADA -0.824 0.102 -0.852 -8.095** N 442 R2 0.781 Adjusted R2 0.721 Constant 34.924 Year Elected -0.036 0.145 -0.006 -0.249 Northeast -7.323 4.186 -0.064 -1.750 South -3.159 4.155 -0.032 -0.760 West -3.631 4.525 -0.032 -0.802 Diff DW Nom 18.845 7.163 0.070 2.631** Fed Spending 12.072 7.596 0.056 1.589 Per Cap Inc 0.002 0.001 0.146 3.483** African Amer. 4.297 19.693 0.007 0.218 Hispanic -15.417 14.903 -0.035 -1.034 Med Mal Crisis -3.601 2.745 -0.039 -1.312 Health 61.506 26.376 0.057 2.332* Lawyer -70.729 31.977 -0.060 -2.212* E-score 0.067 0.071 0.044 0.940 104th ADA -0.981 0.059 -0.792 -16.559** N 445 R2 0.770 Adjusted R2 0.762 Constant 33.685 Year Elected -0.132 0.499 -0.021 -0.264 Northeast -0.779 10.245 -0.009 -0.076 South -2.457 10.686 -0.033 -0.230 West -2.125 11.981 -0.025 -0.177 Diff DW Nom 27.300 26.403 0.075 1.034 Fed Spending 28.264 21.599 0.170 1.309 Per Cap Inc 0.003 0.001 0.254 2.052* African Amer. -12.171 44.839 -0.023 -0.271 Hispanic -24.574 40.791 -0.083 -0.602 Med Mal Crisis -3.943 5.983 -0.058 -0.659 Health 21.690 38.050 0.042 0.570 Lawyer -48.065 77.387 -0.048 -0.621 E-score 0.106 0.093 0.094 1.131 105th ADA -0.771 0.080 -0.841 -9.646** N 444 R2 0.778 Adjusted R2 0.718 Constant 0.603 Year Elected -0.002 0.443 0.000 -0.005 Northeast -2.878 9.316 -0.033 -0.309 South -1.454 9.868 -0.019 -0.147 106th West 0.458 10.799 0.005 0.042 205 Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t ? scores Diff DW Nom 13.968 23.505 0.039 0.594 Fed Spending 9.600 20.280 0.063 0.473 Per Cap Inc 0.002 0.001 0.224 1.916 African Amer. -3.034 40.306 -0.006 -.075 Hispanic -26.739 37.977 -0.092 -0.704 Med Mal Crisis -2.624 5.400 -0.039 -0.486 Health -5.632 34.695 -0.011 -0.162 Lawyer -119.506 62.698 -0.138 -1.906 E-score 0.152 0.087 0.158 1.744 ADA -0.714 0.093 -0.755 -7.649** N 440 R2 0.819 Adjusted R2 0.770 Constant 31.711 Year Elected 0.319 0.160 0.048 1.988* Northeast 1.990 4.502 0.016 0.442 South -0.524 4.175 -0.005 -0.125 West -4.410 4.574 -0.036 -0.964 Diff DW Nom -14.769 11.963 -0.029 -1.235 Fed Spending 9.805 7.690 0.051 1.275 Per Cap Inc 0.001 0.001 0.062 1.531 African Amer. -38.502 20.130 -0.061 -1.913 Hispanic 7.761 15.220 0.018 0.510 Med Mal Crisis -1.180 2.843 -0.012 -0.415 Health 43.322 22.937 0.045 1.889 Lawyer -133.152 31.650 -0.111 -4.207** E-score 0.110 0.069 0.070 1.587 107th ADA -0.961 0.055 -0.785 -17.597** N 444 R2 0.788 Adjusted R2 0.780 Constant 63.855 Year Elected -0.064 0.436 -0.011 -0.146 Northeast -2.276 9.360 -0.026 -0.243 South 0.289 10.288 0.004 0.028 West -3.908 11.234 -0.046 -0.348 Diff DW Nom 26.003 25.135 0.072 1.035 Fed Spending 1.878 14.928 0.017 0.126 Per Cap Inc 0.002 0.001 0.199 1.895 African Amer. -16.525 42.801 -0.032 -0.386 Hispanic -31.434 37.183 -0.117 -0.845 Med Mal Crisis -6.787 5.687 -0.100 -1.193 Health -14.824 26.878 -0.039 -0.552 Lawyer -106.227 71.714 -0.105 -1.481 E-score 0.213 0.098 0.208 2.162* 108th ADA -0.655 0.082 -0.754 -7.976** N 440 R2 0.807 Adjusted R2 0.754 Constant 40.307 *p < 0.05, ** p < 0.01 206 How the substitution of ADA into the model affects other variables is important. E-score was statistically significant in all Congresses but the 101 st House in the Base Model but failed each test of statistical significance except in the 108th House when ADA was added to the model. Controls for economic conditions appear to play a larger role in the ADA Model, where per capita income became statistically significant in the 101st, 104th, and 105th Congresses. In each Congress the association was positive, indicating that higher per capita incomes in a state are associated with higher levels of support for legislative economic efficiency. Coefficients of determination show a stronger fit in explaining variability for the regression line in the ADA model. Substituting ADA into the model produced results that frequently explained over 70 percent of model variability. Relatively higher coefficients of determination, negative standardized coefficients, and levels of statistically significance indicate a strong inverse relationship between liberal philosophy and economically efficient legislative positions. ACU is a measure of conservative ideology that can be used in comparing legislative behavior predicted through ADA, a measure of liberal ideology. Testing Hypothesis 3a: Legislators with higher ACU scores vote in support of medical malpractice reform, statistically significant results were produced for ACU in each Congress in the model (see Table 4.13). The variable was positively correlated with support for economically efficient policies in each Congress except the 99th Congress, and the regression output produced relatively strong standardized coefficients for that variable. ACU and ADA each indicate a strong presence of liberal-conservative ideology in legislative voting. Inclusion of each variable separately in the base model 207 produced highly significant t scores and standardized coefficients with the relatively strongest impact in the model. Coefficients of determination that explain variability in the model follow trends found with party unity and ADA substitutions. Substituting ACU into the model better explains variability than base model variables alone. Table 4.13 Regression Analysis of Base Model and ACU Substitution for 99th to 108th House: Medical Malpractice Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t ? scores Year Elected -0.406 0.241 -0.077 -1.686 Northeast -6.575 6.163 -0.065 -1.067 South 0.388 5.690 0.004 0.068 West 5.126 6.461 0.048 0.793 Diff DW Nom -22.227 17.036 -0.058 -1.305 Fed Spending 3.771 9.413 0.021 0.401 Per Cap Inc 0.001 0.001 0.062 1.011 African Amer. -23.958 38.150 -0.037 -0.628 Hispanic -29.437 30.337 -0.054 -0.970 Med Mal Crisis 7.704 4.421 0.093 1.743 Health 9.167 10.774 0.045 0.851 Lawyer 20.032 10.735 0.095 1.866 E-score -0.104 0.088 -0.088 -1.180 99th ACU -0.563 0.098 -0.454 -5.729** N 439 R2 0.309 Adjusted R2 0.238 Constant 87.046 Year Elected 0.232 0.199 0.038 1.166 Northeast 2.423 5.664 0.020 0.428 South -6.789 5.647 -0.063 -1.202 West -12.593 6.257 -0.102 -2.013* Diff DW Nom 10.582 7.391 0.047 1.432 Fed Spending 15.201 8.748 0.071 1.738 Per Cap Inc -0.001 0.001 -0.067 -1.428 African Amer. 9.033 35.559 0.012 0.254 Hispanic 23.242 28.618 0.036 0.812 Med Mal Crisis 2.862 3.978 0.029 0.720 Health -5.624 10.896 -0.021 -0.516 Lawyer -21.091 10.257 -0.078 -2.056* E-score 0.017 0.058 0.012 0.287 100th ACU 1.051 0.061 0.759 17.329** N 441 R2 0.636 Adjusted R2 0.623 Constant 17.398 Year Elected -0.032 0.466 -0.005 -0.069101st Northeast 0.979 9.896 0.012 0.099 208 Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t ? scores South -3.433 9.575 -0.047 -0.359 West -4.011 11.272 -0.049 -0.356 Diff DW Nom 0.206 21.415 0.001 0.010 Fed Spending 13.457 19.042 0.075 0.707 Per Cap Inc 0.003 0.002 0.235 1.934 African Amer. 36.021 53.045 0.067 0.679 Hispanic -36.725 40.468 -0.109 -0.908 Med Mal Crisis -6.103 6.585 -0.090 -0.927 Health -11.090 14.291 -0.065 -0.776 Lawyer -14.846 22.717 -0.054 -0.654 E-score 0.102 0.128 0.062 0.796 ACU 0.889 0.088 0.940 10.052** N 442 R2 0.792 Adjusted R2 0.731 Constant -45.962 Year Elected 0.226 0.503 0.036 0.450 Northeast -2.880 10.410 -0.033 -0.277 South -3.423 10.637 -0.046 -0.322 West -5.746 11.832 -0.068 -0.486 Diff DW Nom -15.143 20.338 -0.058 -0.745 Fed Spending 21.702 20.092 0.125 1.080 Per Cap Inc 0.003 0.002 0.223 1.820 African Amer. 8.148 47.684 0.015 0.171 Hispanic -34.248 44.356 -0.103 -0.772 Med Mal Crisis -2.784 6.091 -0.041 -0.457 Health 18.749 46.986 0.030 0.399 Lawyer -129.318 82.378 -0.140 -1.570 E-score 0.131 0.139 0.138 0.942 102nd ACU 0.636 0.142 0.70 4.491** N 441 R2 0.766 Adjusted R2 0.702 Constant -43.448 Year Elected -0.324 0.467 -0.054 -0.694 Northeast -1.944 10.156 -0.022 -0.191 South -8.821 9.620 -0.118 -0.917 West -4.355 11.602 -0.052 -0.375 Diff DW Nom 15.622 24.632 0.049 0.634 Fed Spending 17.225 17.891 0.104 0.963 Per Cap Inc 0.002 0.002 0.167 1.451 African Amer. 8.138 45.180 0.015 0.180 Hispanic -33.051 39.064 -0.104 -0.846 Med Mal Crisis -1.268 5.952 -0.019 -0.213 Health 42.265 41.388 0.075 1.021 Lawyer -82.162 76.158 -0.088 -1.079 E-score 0.084 0.089 0.077 0.940 103rd ACU 0.760 0.093 0.831 8.149** N 442 R2 0.783 Adjusted R2 0.723 Constant -36.398 209 Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t ? scores Year Elected 0.174 0.152 0.030 1.143 Northeast -5.548 4.393 -0.049 -1.263 South -2.989 4.347 -0.031 -0.688 West -5.417 4.755 -0.048 -1.139 Diff DW Nom 18.315 7.565 0.068 2.421* Fed Spending 19.874 8.018 0.091 2.479* Per Cap Inc 0.003 0.001 0.157 3.553** African Amer. -0.482 20.701 -0.001 -0.023 Hispanic -10.805 15.730 -0.025 -0.687 Med Mal Crisis -5.089 2.879 -0.055 -1.767 Health 60.352 27.721 0.056 2.177* Lawyer -50.572 33.951 -0.043 -1.490 E-score 0.192 0.073 0.125 2.637** 104th ACU 0.880 0.061 0.726 14.340** N 445 R2 0.746 Adjusted R2 .737 Constant -76.724 Year Elected -0.166 0.519 -0.026 -0.320 Northeast -1.775 10.647 -0.020 -0.167 South -4.068 11.146 -0.054 -0.365 West -1.406 12.461 -0.017 -0.113 Diff DW Nom 32.537 27.427 0.089 1.186 Fed Spending 30.765 22.483 0.185 1.368 Per Cap Inc 0.003 0.001 0.271 2.105* African Amer. -0.153 46.750 0.000 -0.003 Hispanic -30.209 42.404 -0.102 -0.712 Med Mal Crisis -5.048 6.261 -0.075 -0.806 Health 19.615 39.582 0.038 0.496 Lawyer -21.832 81.398 -0.022 -0.268 E-score 0.201 0.093 0.179 2.162* 105th ACU 0.794 0.088 0.805 9.059** N 444 R2 0.760 Adjusted R2 0.694 Constant -87.910 Year Elected 0.146 0.464 0.023 0.314 Northeast -1.033 9.794 -0.012 -0.105 South -1.611 10.379 -0.021 -0.155 West 1.273 11.326 0.015 0.112 Diff DW Nom -2.777 24.930 -0.008 -0.111 Fed Spending 19.104 21.388 0.125 0.893 Per Cap Inc 0.002 0.001 0.225 1.834 African Amer. -12.050 42.183 -0.023 -0.286 Hispanic -19.667 40.001 -0.067 -0.492 Med Mal Crisis -0.846 5.632 -0.013 -0.150 Health -11.317 36.335 -0.022 -0.311 Lawyer -100.985 67.225 -0.117 -1.502 E-score 0.230 0.086 0.240 2.693** 106th ACU 0.692 0.099 0.696 6.962** 210 Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t ? scores N 440 R2 0.801 Adjusted R2 0.747 Constant -50.810 Year Elected 0.308 0.167 0.046 1.844 Northeast 2.886 4.672 0.023 0.618 South -3.509 4.348 -0.033 -0.807 West -5.436 4.750 -0.044 -1.144 Diff DW Nom -16.466 12.444 -0.032 -1.323 Fed Spending 11.915 7.993 0.062 1.491 Per Cap Inc 0.001 0.001 0.103 2.411* African Amer. -29.157 20.922 -0.046 -1.394 Hispanic 11.207 15.812 0.027 0.709 Med Mal Crisis -0.998 2.953 -0.010 -0.338 Health 55.246 23.777 0.057 2.323* Lawyer -136.498 32.876 -0.113 -4.152** E-score 0.242 0.068 0.154 3.565** 107th ACU 0.901 0.056 0.717 16.099** N 444 R2 0.771 Adjusted R2 0.763 Constant -52.866 Year Elected -0.159 0.457 -0.027 -0.347 Northeast -3.397 9.832 -0.039 -0.345 South -9.679 10.877 -0.129 -0.890 West -4.479 11.814 -0.053 -0.379 Diff DW Nom 23.679 26.489 0.065 0.894 Fed Spending 10.293 15.680 0.091 0.656 Per Cap Inc 0.002 0.001 0.224 2.027* African Amer. -6.146 45.251 -0.012 -0.136 Hispanic -12.717 39.266 -0.047 -0.324 Med Mal Crisis -5.423 5.964 -0.080 -0.909 Health -10.907 28.242 -0.029 -0.386 Lawyer -87.968 75.744 -0.087 -1.161 E-score 0.246 0.103 0.241 2.398* 108th ACU 0.778 0.107 0.733 7.254** N 440 R2 0.786 Adjusted R2 0.727 Constant -49.010 * p < 0.05, ** p < 0.01 Substituting ACU into the model produced effects unlike those in the ADA model, however. Higher federal spending in a state is statistically significant in the 104 th Congress when considered with ACU. Ratio is positively correlated with support for medical malpractice reform and is an indication that legislators support economically 211 efficient policies when federal largess pours into his or her state. The statistical significance of E-score in the base model diminished when substituting ACU and was only statistically significant in the 104 th through 108 th Congresses. Perhaps equally as important, from the base model E-score remains statistically significant in the 104 th , 105 th , 106 th , and 107 th Congresses, while with the ADA substitution it is not. The relation between ACU and E-score could be an indication that modeling economic efficiency through an E-score works better with measures of conservative ideology. Table 4.14 summarizes regression results from substituting DW Nominate into the model. DW Nominate measures liberal-conservative ideology but weighs ongoing roll call votes in Congress as a method of measuring legislative activity over time. By substituting the variable into the model it is possible to analyze changes in ideology over time and compare results from more static measures of ideology such as ACU and ADA. DW Nominate is statistically significant in each Congress in the model at the p < 0.01 level in each test. Directional movements between DW Nominate and support for medical malpractice are positively correlated in each House except the 99th, where associations were negative. Positive associations indicate that greater legislative conservatism leads to greater support for economically efficient lawmaking and are consistent with Hypothesis 4a: Legislators with higher DW Nominate scores vote in support of medical malpractice reform. Standardized coefficients indicate that the effect of changes in DW Nominate on support are relatively strong compared to other variables in the model (especially measures of self-interest), but do not appreciably differ from ADA and ACU scores that do not capture a time element to ideology. While shifts in ideology for the House as an 212 institution occurred beginning with the 104th Congress (see Figure 4.4) relative comparisons to other traditional measures of ideology (ACU and ADA) do not find that capturing a time element produces a better predictor of legislative voting. In substituting DW Nominate with controls for federal spending and differences in legislator ideology from his or her median party ideology mirroring those effects in the 104th House from substituting ACU, similarities exist in using ACU and DW Nominate to model effects of economic efficiency. Table 4.14 Regression Analysis of Base Model and DW Nominate Substitution for 99th to 108th House: Medical Malpractice Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores Year Elected -0.411 0.244 -0.078 -1.683 Northeast -5.817 6.262 -0.058 -0.929 South -1.311 5.747 -0.015 -0.228 West 3.976 6.546 0.037 0.607 Diff DW Nom -30.504 17.167 -0.080 -1.777 Fed Spending 3.768 9.518 0.021 0.396 Per Cap Inc 0.002 0.001 0.078 1.249 African Amer. -56.827 37.559 -0.089 -1.513 Hispanic -31.633 30.659 -0.058 -1.032 Med Mal Crisis 7.503 4.469 0.091 1.679 Health 5.497 10.818 0.027 0.508 Lawyer 23.199 10.803 0.110 2.148* E-score -0.160 0.088 -0.135 -1.815 99th DW Nominate -42.638 8.687 -0.378 -4.908** N 439 R2 0.294 Adjusted R2 0.267 Constant 63.717 Year Elected 0.182 .243 0.030 0.747 Northeast 0.326 6.918 0.003 0.047 South -2.619 6.887 -0.024 -0.380 West -15.999 7.637 -0.129 -2.095* Diff DW Nom 10.560 9.030 0.047 1.169 Fed Spending 18.178 10.681 0.085 1.702 Per Cap Inc -0.003 .001 -0.122 -2.133* African Amer. 60.875 43.238 0.078 1.408 Hispanic 32.182 34.946 0.050 0.921 Med Mal Crisis 6.913 4.846 0.069 1.427 Health 40.655 12.683 0.151 3.205** Lawyer -39.797 12.468 -0.147 -3.192** 100th E-score 0.343 .065 0.241 5.276** 213 Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores DW Nominate 50.953 5.920 0.373 8.607** N 441 R2 0.457 Adjusted R2 0.438 Constant 50.804 Year Elected -0.020 0.500 -0.003 -0.039 Northeast -0.969 10.607 -0.011 -0.091 South -8.342 10.452 -0.114 -0.798 West -7.930 12.175 -0.097 -0.651 Diff DW Nom -12.528 23.224 -0.047 -0.539 Fed Spending 17.173 20.698 0.096 0.830 Per Cap Inc 0.002 0.002 0.174 1.332 African Amer. 27.137 56.742 0.051 0.478 Hispanic 4.383 44.062 0.013 0.099 Med Mal Crisis 0.490 6.923 0.007 0.071 Health 2.562 14.788 0.015 0.173 Lawyer -33.534 23.905 -0.121 -1.403 E-score 0.033 0.135 0.020 0.242 101st DW Nominate 73.302 8.079 0.847 9.074** N 442 R2 0.763 Adjusted R2 0.692 Constant 5.050 Year Elected 0.716 0.507 0.114 1.414 Northeast 4.652 10.997 0.054 0.423 South -0.301 10.826 -0.004 -0.028 West -4.368 12.178 -0.052 -0.359 Diff DW Nom -29.258 21.309 -0.112 -1.373 Fed Spending 23.234 20.681 0.133 1.123 Per Cap Inc 0.002 0.002 0.181 1.433 African Amer. 29.989 49.414 0.056 0.607 Hispanic -28.486 45.869 -0.085 -0.621 Med Mal Crisis -5.668 6.483 -0.084 -0.874 Health 31.044 49.119 0.050 0.632 Lawyer -138.883 85.199 -0.150 -1.630 E-score 0.223 0.133 0.236 1.682 102nd DW Nominate 52.834 13.179 0.631 4.009** N 441 R2 0.752 Adjusted R2 0.684 Constant -16.031 Year Elected -0.099 0.479 -0.017 -0.208 Northeast -4.125 10.324 -0.048 -0.400 South -6.187 9.789 -0.082 -0.632 West -5.750 11.855 -0.068 -0.485 Diff DW Nom 30.644 24.946 0.096 1.228 Fed Spending 26.469 18.240 0.161 1.451 Per Cap Inc 0.003 0.002 0.201 1.710 African Amer. 1.545 46.122 0.003 0.033 Hispanic -17.963 39.811 -0.056 -0.451 Med Mal Crisis -1.510 6.090 -0.022 -0.248 Health 28.834 41.927 0.051 0.688 103rd Lawyer -67.357 79.146 -0.072 -0.851 214 Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores E-score 0.054 0.093 0.049 0.576 DW Nominate 68.586 8.741 0.841 7.847** N 442 R2 0.774 Adjusted R2 0.712 Constant -21.062 Year Elected 0.240 0.158 0.042 1.519 Northeast -6.128 4.573 -0.053 -1.340 South -5.662 4.524 -0.058 -1.252 West -5.069 4.929 -0.045 -1.028 Diff DW Nom 8.578 7.982 0.032 1.075 Fed Spending 17.536 8.326 0.081 2.106* Per Cap Inc 0.002 0.001 0.097 2.119* African Amer. 28.409 21.412 0.047 1.327 Hispanic -10.889 16.276 -0.025 -0.669 Med Mal Crisis -3.563 2.989 -0.038 -1.192 Health 58.314 28.856 0.054 2.021* Lawyer -77.065 35.036 -0.065 -2.200* E-score 0.252 0.076 0.164 3.318** 104th DW Nominate 72.987 5.772 0.672 12.646** N 445 R2 0.723 Adjusted R2 0.714 Constant -13.690 Year Elected 0.112 0.500 0.018 0.224 Northeast -5.850 10.153 -0.068 -0.576 South -8.844 10.777 -0.118 -0.821 West -7.639 12.032 -0.091 -0.635 Diff DW Nom 22.501 26.426 0.062 0.851 Fed Spending 33.733 21.602 0.203 1.562 Per Cap Inc 0.003 0.001 0.239 1.935 African Amer. 0.397 44.885 0.001 0.009 Hispanic 0.959 41.023 0.003 0.023 Med Mal Crisis -2.159 5.939 -0.032 -0.364 Health 22.972 38.016 0.045 0.604 Lawyer -1.521 78.618 -0.002 -0.019 E-score 0.065 0.095 0.058 0.679 105th DW Nominate 69.533 7.197 0.874 9.661** N 444 R2 0.779 Adjusted R2 0.718 Constant -44.561 Year Elected 0.385 0.472 0.061 0.816 Northeast -3.126 9.922 -0.036 -0.315 South -2.806 10.585 -0.037 -0.265 West -0.649 11.550 -0.008 -0.056 Diff DW Nom 14.931 25.033 0.041 0.596 Fed Spending 24.261 21.816 0.158 1.112 Per Cap Inc 0.002 0.001 0.221 1.769 African Amer. -9.183 42.867 -0.018 -0.214 Hispanic -9.012 40.876 -0.031 -0.220 106th Med Mal Crisis -0.221 5.706 -0.003 -0.039 215 Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores Health -3.974 37.021 -0.008 -0.107 Lawyer -122.503 67.179 -0.142 -1.824 E-score 0.171 0.094 0.178 1.827 DW Nominate 56.533 8.379 0.718 6.747** N 440 R2 0.795 Adjusted R2 0.739 Constant -28.102 Year Elected 0.405 0.171 0.061 2.377 Northeast 3.041 4.844 0.025 0.628 South -4.669 4.504 -0.045 -1.037 West -6.199 4.935 -0.051 -1.256 Diff DW Nom -19.806 12.930 -0.039 -1.532 Fed Spending 16.187 8.323 0.085 1.945 Per Cap Inc 0.001 0.001 0.059 1.347 African Amer. -9.453 21.495 -0.015 -0.440 Hispanic 15.556 16.435 0.037 0.947 Med Mal Crisis 1.912 3.067 0.019 0.623 Health 57.332 24.638 0.059 2.327 Lawyer -133.461 33.325 -0.112 -4.005 E-score 0.376 0.067 0.239 5.648 107th DW Nominate 68.086 4.704 0.625 14.473 N 444 R2 0.753 Adjusted R2 0.745 Constant -9.336 Year Elected 0.291 0.470 0.050 0.619 Northeast -5.771 9.863 -0.067 -0.585 South -9.982 10.953 -0.133 -0.911 West -9.346 11.901 -0.111 -0.785 Diff DW Nom 26.817 26.612 0.074 1.008 Fed Spending 20.376 15.869 0.181 1.284 Per Cap Inc 0.002 0.001 0.198 1.781 African Amer. -9.374 45.482 -0.018 -0.206 Hispanic 14.279 40.053 0.053 0.357 Med Mal Crisis -0.808 5.977 -0.012 -0.135 Health 0.588 28.386 0.002 0.021 Lawyer -44.526 77.272 -0.044 -0.576 E-score 0.164 0.112 0.160 1.465 108th DW Nominate 59.028 8.242 0.767 7.162 N 440 R2 0.783 Adjusted R2 0.724 Constant -28.316 * p < 0.05, ** p < 0.01 The effect on other variables of substituting DW Nominate into the model finds self-interest and E-score statistically significant variables disappear when testing DW Nominate in the 101 st , 102 nd , 103 rd , 105 th , and 106 th Congresses. The strength of 216 ideology in these Congresses relative to E-score illustrates the effect of liberal- conservative legislative principles to economic concerns. Controlling for state economic conditions, per capita income is statistically significant in the 100 th and 104 th Houses and the West statistically significant in the 100 th House. Standardized coefficients for West and per capita income indicate a negative correlation with support for medical malpractice reform in the 100 th House, but per capita income is positively correlated in the 104 th House. Legislator party variable is a measure of the effect of party on legislative behavior from the perspective of institutional control. It measures whether the legislator is of the same party as the party in control of the House. Table 4.15 shows the results of substituting this variable into the base model. Table 4.15 Regression Analysis of Base Model and Legislator Party Substitution for 99th to 108th House: Medical Malpractice Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores 99th Year Elected -0.253 0.249 -0.048 -1.015 Northeast -10.226 6.386 -0.102 -1.601 South -1.600 5.957 -0.018 -0.269 West 6.432 6.727 0.060 0.956 Diff DW Nom -30.060 17.696 -0.079 -1.699 Fed Spending 2.183 9.803 0.012 0.223* Per Cap Inc 0.003 0.001 0.130 2.063* African Amer. -76.859 38.182 -0.121 -2.013* Hispanic -37.012 31.593 -0.068 -1.172 Med Mal Crisis 6.930 4.603 0.084 1.505 Health -8.927 10.714 -0.044 -0.833 Lawyer 28.571 11.062 0.136 2.583** E-score -0.485 0.061 -0.410 -7.981** Legislator Party 3.447 3.803 0.041 0.906 N 439 R2 0.249 Adjusted R2 0.220 Constant 64.004 Year Elected 0.135 0.208 0.022 0.648 Northeast -5.955 5.923 -0.050 -1.005 South 4.778 5.884 0.044 0.812 100th West -10.345 6.533 -0.084 -1.584 217 Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores Diff DW Nom 15.781 7.708 0.070 2.047 Fed Spending 13.700 9.127 0.064 1.501 Per Cap Inc -0.003 .001 -0.133 -2.736 African Amer. 34.565 36.972 0.044 0.935 Hispanic 35.085 29.837 0.054 1.176 Med Mal Crisis 5.081 4.140 0.051 1.227 Health 14.072 11.054 0.052 1.273 Lawyer -14.643 10.843 -0.054 -1.350 E-score 0.052 0.060 0.037 0.870 Legislator Party -68.832 4.396 -0.678 -15.657 N 441 R2 0.604 Adjusted R2 0.590 Constant 109.291 Year Elected -0.055 0.447 -0.009 -0.123 Northeast -3.733 9.445 -0.044 -0.395 South -4.601 9.210 -0.063 -0.500 West -1.898 10.783 -0.023 -0.176 Diff DW Nom 24.948 20.573 0.095 1.213 Fed Spending 22.260 18.452 0.124 1.206 Per Cap Inc 0.000 0.001 -0.022 -0.185 African Amer. 63.393 51.358 0.117 1.234 Hispanic -1.351 39.142 -0.004 -0.035 Med Mal Crisis 1.563 6.165 0.023 0.253 Health 9.277 12.942 0.055 0.717 Lawyer -24.833 21.516 -0.090 -1.154 E-score -0.046 0.119 -0.028 -0.386 101st Legislator Party -58.331 5.470 -0.852 -10.664 N 442 R2 0.808 Adjusted R2 0.752 Constant 69.589 Year Elected 0.106 0.463 0.017 0.229 Northeast -2.563 9.567 -0.030 -0.268 South -4.084 9.697 -0.054 -0.421 West -5.323 10.869 -0.063 -0.490 Diff DW Nom 1.663 18.876 0.006 0.088 Fed Spending 25.893 18.429 0.149 1.405 Per Cap Inc 0.000 0.002 0.027 0.229 African Amer. 62.779 44.831 0.116 1.400 Hispanic -0.515 41.547 -0.002 -0.012 Med Mal Crisis -1.849 5.545 -0.027 -0.333 Health -10.397 42.665 -0.017 -0.244 Lawyer -92.990 76.308 -0.101 -1.219 E-score 0.107 0.118 0.114 0.910 102nd Legislator Party -51.507 8.933 -0.758 -5.766 N 441 R2 0.802 Adjusted R2 0.748 Constant 52.679 Year Elected -0.396 0.432 -0.066 -0.916 Northeast -6.987 9.293 -0.081 -0.752 103rd South -6.164 8.872 -0.082 -0.695 218 Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores West -5.708 10.749 -0.068 -0.531 Diff DW Nom 34.423 22.589 0.108 1.524 Fed Spending 25.142 16.541 0.152 1.520 Per Cap Inc 0.000 0.001 0.027 0.245 African Amer. 58.237 42.342 0.109 1.375 Hispanic -6.854 36.099 -0.021 -0.190 Med Mal Crisis -0.235 5.477 -0.003 -0.043 Health -3.587 37.412 -0.006 -0.096 Lawyer -53.643 71.174 -0.058 -0.754 E-score 0.117 0.080 0.107 1.468 Legislator Party -54.861 5.920 -0.807 -9.267 N 442 R2 0.814 Adjusted R2 0.763 Constant 53.966 Year Elected 0.240 0.158 0.042 1.515 Northeast -6.143 4.573 -0.054 -1.343 South -5.663 4.524 -0.058 -1.252 West -5.085 4.929 -0.045 -1.032 Diff DW Nom 81.578 8.456 0.307 9.647 Fed Spending 17.528 8.326 0.081 2.105 Per Cap Inc 0.002 0.001 0.097 2.123 African Amer. 28.247 21.411 0.047 1.319 Hispanic -10.815 16.276 -0.025 -0.664 Med Mal Crisis -3.548 2.989 -0.038 -1.187 Health 58.203 28.856 0.054 2.017 Lawyer -77.282 35.032 -0.065 -2.206 E-score 0.252 0.076 0.164 3.316 104th Legislator Party 58.522 4.628 0.633 12.646 N 445 R2 0.723 Adjusted R2 0.714 Constant -41.948 Year Elected -0.286 0.416 -0.045 -0.686 Northeast -11.136 8.438 -0.128 -1.320 South -7.051 8.965 -0.094 -0.787 West -9.942 10.072 -0.118 -0.987 Diff DW Nom 30.580 22.018 0.084 1.389 Fed Spending 27.607 18.037 0.166 1.531 Per Cap Inc 0.001 0.001 0.098 0.951 African Amer. 15.066 37.618 0.029 0.401 Hispanic 17.555 34.446 0.059 0.510 Med Mal Crisis 0.860 4.916 0.013 0.175 Health -4.155 31.773 -0.008 -0.131 Lawyer 7.213 65.508 0.007 0.110 E-score 0.041 0.079 0.036 0.515 105th Legislator Party 59.937 4.807 0.886 12.470 N 444 R2 0.845 Adjusted R2 0.803 Constant -22.387 106th Year Elected -0.018 0.400 -0.003 -0.045 219 Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores Northeast -8.140 8.402 -0.094 -0.969 South -1.916 8.895 -0.026 -0.215 West -3.772 9.814 -0.045 -0.384 Diff DW Nom 22.302 21.189 0.062 1.053 Fed Spending 19.578 18.388 0.128 1.065 Per Cap Inc 0.001 0.001 0.110 1.046 African Amer. 8.265 36.514 0.016 0.226 Hispanic 5.792 34.812 0.020 0.166 Med Mal Crisis 1.809 4.808 0.027 0.376 Health -23.720 31.140 -0.046 -0.762 Lawyer -87.682 57.373 -0.101 -1.528 E-score 0.102 0.080 0.106 1.286 Legislator Party 53.532 5.865 0.791 9.127 N 440 R2 0.853 Adjusted R2 0.812 Constant -13.216 Year Elected 0.230 0.168 0.035 1.371 Northeast 2.839 4.767 0.023 0.595 South 1.184 4.424 0.011 0.268 West -4.590 4.854 -0.037 -0.946 Diff DW Nom 4.764 12.565 0.009 0.379 Fed Spending 8.262 8.156 0.043 1.013 Per Cap Inc 0.000 0.001 -0.016 -0.368 African Amer. -28.518 21.273 -0.045 -1.341 Hispanic 8.601 16.152 0.020 0.532 Med Mal Crisis 0.198 3.015 0.002 0.066 Health 43.224 24.302 0.045 1.779 Lawyer -134.061 32.753 -0.112 -4.093 E-score 0.259 0.070 0.165 3.677 107th Legislator Party 67.142 4.433 0.671 15.147 N 444 R2 0.761 Adjusted R2 0.753 Constant 5.005 Year Elected 0.033 0.381 0.006 0.087 Northeast -8.336 8.120 -0.096 -1.027 South -7.078 9.007 -0.094 -0.786 West -11.613 9.835 -0.138 -1.181 Diff DW Nom 18.625 22.033 0.051 0.845 Fed Spending 19.621 13.076 0.174 1.501 Per Cap Inc 0.001 0.001 0.104 1.132 African Amer. 5.804 37.678 0.011 0.154 Hispanic 21.759 33.059 0.081 0.658 Med Mal Crisis 0.843 4.939 0.012 0.171 Health -17.848 23.502 -0.047 -0.759 Lawyer -37.800 63.589 -0.037 -0.594 E-score 0.122 0.089 0.120 1.377 108th Legislator Party 55.461 5.572 0.820 9.954 N 440 R2 0.852 Adjusted R2 0.812 Constant -19.724 * p < 0.05, ** p < 0.01 220 Directional impact of Legislator Party must be considered with changes in party control of the House over the time period of this study. Referring to party divisions summarized in Table 4.1, Democrats controlled the House during the 99 th through 103 rd Congresses and Republicans controlled the House during the 104 th through 108 th Congresses. Directional impact of a plus (+) indicates a relationship where the legislator of the majority party is more supportive of medical malpractice reform and a minus (-) indicates support for medical malpractice decrease if the legislator is not a member of the majority party. When Democrats are in the majority standardized coefficients are negative, indicating greater support for medical malpractice reform from Republicans, the party that does not control the House. Republicans are in the majority in the House in the 104 th through 108 th Congresses and vote increasingly for medical malpractice reform regardless of party control, as positive, statistically significant directional impact of Legislator Party indicates. In testing Hypothesis 10a: Legislators from the minority party (House) are more likely than majority party legislators to support medical malpractice reform, legislator party was a statistically significant variable in each Congress with the exception of the 99th Congress. In each Congress, however, results indicate that Republican support for medical malpractice reform as the minority party continued after the party became the majority party beginning with the 104th Congress. After Republican control beginning in the 104th Congress, E-score failed tests of statistical significance in the 105th, 106th, and 108th House, while the variable was statistically significant in the Base Model. That E-score is not statistically significant in those Congresses is a further indication of the importance of party in this model of 221 legislative voting. Legislators vote with party and Republican legislators support economically efficient legislation regardless of influences from party control of the institution. E-score remains statistically significant in the 108th House, but a weaker standardized coefficient relative to its standardized coefficient in the Base Model illustrates a stronger relative importance for party in this model than how closely a legislator supports greater economically efficient lawmaking. When substituting legislator party, coefficients of determination are generally slightly higher than coefficients of determination in analyzing Base Model and substituting ADA and ACU. Higher coefficients of determination for legislator party are a reflection of party line voting; higher coefficients of correlation in later Congresses, e.g. 105th, 106, and 108th Houses, is consistent with an overall trend toward support for party line positions. Minimum wage Unlike medical malpractice legislation, considering minimum wage as a dependent variable illustrates the effect of legislative decision making on an economically inefficient variable. Voting for a minimum wage or minimum wage increase represents economic inefficiency; voting against such legislation is economically efficient. Table 4.16 summarizes the base model analysis when minimum wage votes represent the dependent variable in the House. Three associations are consistent within the Base Model and assist in the testing of hypotheses 1, 6, and 9. E-score is a statistically significant variable in the base model in the 99th, 100th, 102nd, 104th, and 107th House. Associations between a higher E- 222 score and opposition to minimum wage increases are positive as expected in each of these Houses and all others except the 101st, lending support to Hypothesis 1b: Legislators with higher E-scores vote in opposition to increasing the minimum wage. The effect of E-score on changes in opposition to minimum wage is generally weaker than ACU and ADA ideology (to be discussed below) Table 4.16 Regression Analysis of Base Model from 99 th to 108 th House: Minimum Wage Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores Year Elected -0.912 0.584 -0.131 -1.562 Northeast -6.524 9.846 -0.071 -0.663 South -4.065 11.472 -0.049 -0.354 West -1.280 14.255 -0.014 -0.090 Diff DW Nom 52.847 29.364 0.169 1.800 Fed Spending -14.234 19.533 -0.075 -0.729 Per Cap Inc -0.004 0.002 -0.221 -2.308* African Amer. 100.273 68.355 0.144 1.467 Hispanic 30.746 71.728 0.059 0.429 Min. Wage Law 10.636 9.442 0.121 1.127 Business 17.217 13.887 0.119 1.240 Labor -29.369 19.819 -0.177 -1.482 99th E-score 0.592 0.114 0.590 5.184** N 439 R2 0.733 Adjusted R2 0.663 Constant 64.770 Year Elected -0.809 0.606 -0.116 -1.334 Northeast -4.548 11.687 -0.047 -0.389 South -10.349 12.968 -0.128 -0.798 West -5.686 14.094 -0.063 -0.403 Diff DW Nom 0.787 14.967 0.005 0.053 Fed Spending -19.609 20.616 -0.106 -0.951 Per Cap Inc -0.004 0.002 -0.236 -1.763 African Amer. 152.423 78.792 0.221 1.935 Hispanic 85.746 67.101 0.180 1.278 Min. Wage Law 4.750 8.646 0.056 0.549 Business -1.445 20.841 -0.007 -0.069 Labor -141.362 24.644 -0.627 -5.736** 100th E-score 0.271 0.099 0.278 2.730** N 441 R2 0.711 Adjusted R2 0.633 Constant 108.814 Year Elected -0.193 0.166 -0.043 -1.160101st Northeast -4.867 4.603 -0.056 -1.057 223 Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores South -3.664 4.678 -0.048 -0.783 West -2.918 5.291 -0.032 -0.551 Diff DW Nom 11.912 8.192 0.054 1.454 Fed Spending -4.827 8.249 -0.029 -0.585 Per Cap Inc -0.001 0.001 -0.097 -1.562 African Amer. 6.425 23.031 0.014 0.279 Hispanic 62.245 19.091 0.160 3.260** Min. Wage Law 1.744 3.691 0.023 0.472 Business 8.721 4.153 0.081 2.100* Labor -112.838 7.114 -0.643 -15.861** E-score -0.037 0.071 -0.021 -0.522 N 442 R2 0.504 Adjusted R2 0.487 Constant 73.807 Year Elected -0.684 0.549 -0.098 -1.245 Northeast -2.051 10.595 -0.021 -0.194 South 2.772 11.577 0.033 0.239 West -1.510 12.537 -0.016 -0.120 Diff DW Nom 32.456 23.038 0.112 1.409 Fed Spending -10.851 21.779 -0.056 -0.498 Per Cap Inc -0.004 0.002 -0.261 -2.248* African Amer. 55.356 53.704 0.093 1.031 Hispanic -16.661 44.804 -0.045 -0.372 Min. Wage Law 8.460 7.291 0.098 1.160 Business -24.058 38.679 -0.051 -0.622 Labor -60.264 33.521 -0.186 -1.798 102nd E-score 0.776 0.108 0.744 7.172** N 441 R2 0.786 Adjusted R2 0.733 Constant 95.864 Year Elected -2.172 0.676 -0.326 -3.211** Northeast -38.525 14.013 -0.402 -2.749** South -9.725 14.080 -0.117 -0.691 West -14.376 17.576 -0.155 -0.818 Diff DW Nom 65.246 37.797 0.185 1.726 Fed Spending -34.133 26.824 -0.187 -1.272 Per Cap Inc -0.003 0.002 -0.207 -1.410 African Amer. 37.095 66.427 0.063 0.558 Hispanic 35.815 54.611 0.102 0.656 Min. Wage Law 23.208 9.892 0.280 2.346* Business 7.547 51.975 0.016 0.145 Labor -207.214 40.219 -0.610 -5.152** 103rd E-score 0.156 0.147 0.130 1.061 N 442 R2 0.631 Adjusted R2 0.539 Constant 152.292 224 Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores Year Elected 0.085 0.230 0.014 0.371 Northeast -4.480 6.489 -0.037 -0.690 South 0.638 6.879 0.006 0.093 West 17.552 7.316 0.148 2.399* Diff DW Nom 39.356 11.691 0.137 3.366** Fed Spending -18.833 12.284 -0.082 -1.533 Per Cap Inc -0.002 0.001 -0.111 -1.713 African Amer. 49.944 32.807 0.077 1.522 Hispanic -33.529 23.268 -0.072 -1.441 Min. Wage Law -3.597 4.780 -0.034 -0.752 Business -25.111 21.849 -0.048 -1.149 Labor -124.294 21.193 -0.304 -5.865** 104th E-score 0.549 0.085 0.341 6.492** N 445 R2 0.459 Adjusted R2 0.441 Constant 75.644 Year Elected -1.598 0.751 -0.230 -2.129* Northeast -23.290 14.896 -0.243 -1.563 South -0.639 17.111 -0.008 -0.037 West 5.745 18.776 0.062 0.306 Diff DW Nom 56.063 42.904 0.139 1.307 Fed Spending -55.975 34.229 -0.305 -1.635 Per Cap Inc -0.004 0.002 -0.316 -1.788 African Amer. 44.970 69.174 0.077 0.650 Hispanic -9.177 61.995 -0.028 -0.148 Min. Wage Law 7.084 10.535 0.082 0.672 Business 50.508 46.192 0.130 1.093 Labor -167.298 48.013 -0.449 -3.484** 105th E-score 0.180 0.174 0.145 1.033 N 444 R2 0.546 Adjusted R2 0.433 Constant 199.709 Year Elected -1.570 0.739 -0.225 -2.124* Northeast 24.922 14.997 -0.260 -1.662 South -2.508 16.440 -0.030 -0.153 West -4.039 17.574 -0.044 -0.230 Diff DW Nom 47.226 41.333 0.118 1.143 Fed Spending -40.820 33.372 -0.241 -1.223 Per Cap Inc -0.002 0.002 -0.226 -1.254 African Amer. 30.805 68.640 0.053 0.449 Hispanic 18.626 61.869 0.058 0.301 Min. Wage Law 11.803 10.338 0.137 1.142 Business 3.441 47.849 0.008 0.072 Labor -176.556 53.576 -0.489 -3.295** 106th E-score 0.200 0.157 0.189 1.275 N 440 R2 0.566 Adjusted R2 0.457 Constant 163.347 225 Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores 107th Year Elected -1.280 0.689 -0.187 -1.858 Northeast -21.746 13.750 -0.227 -1.581 South 4.009 14.914 0.048 0.269 West -0.593 16.792 -0.006 -0.035 Diff DW Nom 23.304 39.153 0.058 0.595 Fed Spending -37.529 27.594 -0.251 -1.360 Per Cap Inc -0.002 0.001 -0.160 -1.045 African Amer. 34.320 63.691 0.059 0.539 Hispanic 23.467 57.858 0.077 0.406 Min. Wage Law 11.254 9.720 0.131 1.158 Business 70.559 38.432 0.180 1.836 Labor -105.742 39.897 -0.325 -2.650** E-score 0.420 0.140 0.349 2.990** N 444 R2 0.624 Adjusted R2 0.531 Constant 104.925 Year Elected -0.883 0.711 -0.137 1.242 Northeast -27.523 14.616 -0.287 -1.883 South -2.991 16.992 -0.036 -0.176 West -7.247 19.170 -0.078 -0.378 Diff DW Nom 39.402 42.893 0.099 0.919 Fed Spending -23.525 24.423 -0.189 -0.963 Per Cap Inc -0.001 0.002 -0.151 -0.939 African Amer. 26.319 73.846 0.046 0.356 Hispanic 39.616 61.570 0.133 0.643 Min. Wage Law 8.098 10.692 0.094 0.757 Business 63.872 38.803 0.180 1.646 Labor -129.314 54.528 -0.351 -2.372* 108th E-score 0.297 0.157 0.263 1.889 N 440 R2 0.535 Adjusted R2 0.418 Constant 87.218 * p < 0.05, ** p < 0.01 Business and Labor are self-interest variables included in the analysis. Labor is statistically significant in each House except the 99 th and 102 nd . Negative standardized coefficients indicate an inverse relationship with higher Labor political contributions and opposition to increasing the minimum wage. This association is consistent with the expected result for testing Hypothesis 6b: Legislators with higher labor political contributions to total contributions vote in support of increasing the minimum wage. 226 Business contributions were measured in testing Hypothesis 5b: Legislators with higher business political contributions to total contributions vote in opposition to increasing the minimum wage. Business contributions were statistically significant to the model in only the 101st Congress, which is the only Congress where Hispanic is a statistically significant variable. Positive associations between the percentage of Hispanics in a state and higher Business contributions received by legislators in those states provide a link between a higher minority component of a population and economically efficient policymaking. The time in years that a legislator has served in office is a control that appears to have more impact (statistically significant in more Congresses and higher standardized coefficients) when conducting hypothesis testing of minimum wage dependent variable than medical malpractice dependent variable. In testing Hypothesis 9b: The longer a legislator has served, the less likely he or she will support increasing the minimum wage, the variable for first elected is statistically significant in the base model in the 103rd, 105th, and 106th House. However, in each House coefficients indicate that the longer a legislator has served the more likely he or she will support increasing the minimum wage. The base model considers divisions in legislator ideology and the median ideology of his or her party. These divisions are expressed through DW Nominate scores and offer a method of testing ideology in the Base Model without including highly correlated variables that must be substituted into the model for testing. In testing Hypothesis 11b: The greater the division between the ideology of the legislator and the median ideology of the legislator?s party, the less likely the legislator supports 227 increasing the federal minimum wage, the variable is statistically significant in only the 104th House. Positive associations between legislator and party ideology reflect a statistically significant difference in a legislator?s voting preferences when he or she diverges from party positions. A positive coefficient indicates that greater divisions produce more economically efficient policies. These findings are consistent with the stated hypothesis. Coefficients of determination in the model are generally higher in the minimum wage model than the medical malpractice model and less susceptible to swings when other variables are substituted into the model. For most Congresses the Base Model for minimum wage explains 60 to 70 percent of variation in the dependent variable. Higher coefficients of determination for the base minimum wage model indicate a better fit of variables in measuring variance along the regression line. Less upward movement in coefficient of determination values with each substitution is evidence that either the Base Model is a better fit for analysis of the dependent variable or variation in each of the substituted variables adds less to explaining changes in the dependent variable. The various substitutions will be discussed below starting with Party Unity found in Table 4.17. 228 Table 4.17 Regression Analysis of Base Model and Party Unity Substitution for 99th to 108th House: Minimum Wage Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores Year Elected -0.790 0.518 -0.113 -1.524 Northeast -10.414 8.779 -0.113 -1.186 South -6.969 10.188 -0.084 -0.684 West -0.374 12.628 -0.004 -0.030 Diff DW Nom 65.263 26.208 0.209 2.490* Min. Wage Law 10.310 8.363 0.117 1.233 Fed Spending -5.893 17.437 -0.031 -0.338 Per Cap Inc -0.005 0.002 -0.239 -2.817** African Amer. 102.657 60.545 0.147 1.696 Hispanic 62.477 64.065 0.120 0.975 Business 10.817 12.413 0.075 0.871 Labor -6.823 18.510 -0.041 -0.369 E-score 0.266 0.132 0.265 2.011* 99th Party Unity -0.225 0.058 -0.509 -3.839** N 439 R2 0.795 Adjusted R2 0.736 Constant 70.436 Year Elected -0.660 0.481 -0.095 -1.372 Northeast -8.052 9.277 -0.084 -0.868 South -21.253 10.462 -0.264 -2.031* West -6.911 11.162 -0.077 -0.619 Diff DW Nom -7.090 11.939 -0.045 -0.594 Min. Wage Law 2.624 6.857 0.031 0.383 Fed Spending -10.653 16.407 -0.058 -0.649 Per Cap Inc -0.004 0.002 -0.284 -2.678** African Amer. 161.684 62.414 0.234 2.591* Hispanic 127.710 53.691 0.268 2.379* Business -23.575 16.997 -0.114 -1.387 Labor -79.312 22.607 -0.352 -3.508** E-score 0.088 0.085 0.091 1.035 100th Party Unity -0.252 0.046 -0.583 -5.436** N 441 R2 0.823 Adjusted R2 0.770 Constant 121.159 Year Elected -0.137 0.133 -0.030 -1.032 Northeast -2.076 3.662 -0.024 -0.567 South 1.848 3.734 0.024 0.495 West 2.018 4.242 0.022 0.476 Diff DW Nom 9.336 6.585 0.042 1.418 Min. Wage Law 0.790 2.932 0.010 0.269 Fed Spending -6.362 6.551 -0.039 -0.971 Per Cap Inc -0.002 0.001 -0.141 -2.843** African Amer. 41.112 18.485 0.087 2.224* Hispanic 39.722 15.282 0.102 2.599** Business 4.660 3.320 0.044 1.404 Labor -37.603 7.565 -0.214 -4.971** 101st E-score 0.103 0.058 0.057 1.785 229 Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores Party Unity -0.279 0.018 -0.625 -15.161** N 442 R2 0.689 Adjusted R2 0.678 Constant 65.910 Year Elected -0.616 0.529 -0.088 -1.164 Northeast -1.216 10.194 -0.013 -0.119 South -0.613 11.229 -0.007 -0.055 West -0.731 12.060 -0.008 -0.061 Diff DW Nom 38.572 22.313 0.133 1.729 Min. Wage Law 8.623 7.011 0.100 1.230 Fed Spending -8.977 20.958 -0.047 -0.428 Per Cap Inc -0.004 0.002 -0.308 -2.713** African Amer. 63.401 51.760 0.106 1.225 Hispanic 9.930 44.621 0.027 0.223 Business -4.558 38.156 -0.010 -0.119 Labor -33.063 34.354 -0.102 -0.962 E-score 0.539 0.147 0.516 3.671** 102nd Party Unity -0.133 0.058 -0.305 -2.289* N 441 R2 0.806 Adjusted R2 0.753 Constant 106.777 Year Elected -1.121 0.589 -0.169 -1.904 Northeast -19.398 12.034 -0.203 -1.612 South -8.755 11.510 -0.106 -0.761 West -6.826 14.440 -0.074 -0.473 Diff DW Nom 57.749 30.928 0.164 1.867 Min. Wage Law 14.867 8.244 0.179 1.803 Fed Spending -22.911 22.032 -0.126 -1.040 Per Cap Inc -0.005 0.002 -0.311 -2.555* African Amer. 70.291 54.673 0.119 1.286 Hispanic 47.637 44.696 0.135 1.066 Business 36.281 42.844 0.075 0.847 Labor -70.471 42.161 -0.207 -1.671 E-score 0.070 0.121 0.058 0.573 103rd Party Unity -0.237 0.046 -0.560 -5.180** N 442 R2 0.758 Adjusted R2 0.692 Constant 144.991 Year Elected 0.245 .206 .041 1.189 Northeast -4.969 5.783 -.041 -.859 South 1.626 6.131 0.016 0.265 West 12.932 6.535 0.109 1.979* Diff DW Nom 66.464 10.744 0.231 6.186** Min. Wage Law -4.020 4.260 -0.038 -0.943 Fed Spending -11.007 10.973 -0.048 -1.003 Per Cap Inc -0.001 0.001 -0.044 -0.758 African Amer. 55.512 29.242 0.085 1.898 Hispanic -4.509 20.926 -0.010 -0.215 104th Business -26.873 19.472 -0.051 -1.380 230 Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores Labor -23.504 21.263 -0.058 -1.105 E-score -0.179 0.103 -0.111 -1.734 Party Unity -0.401 0.039 -0.730 -10.319** N 445 R2 0.571 Adjusted R2 0.556 Constant 73.823 Year Elected -0.931 0.634 -0.134 -1.469 Northeast -13.963 12.440 -0.146 -1.122 South -4.169 14.148 -0.050 -0.295 West -3.732 15.619 -0.040 -0.239 Diff DW Nom 32.528 35.737 0.081 0.910 Min. Wage Law 7.264 8.700 0.084 0.835 Fed Spending -37.549 28.503 -0.205 -1.317 Per Cap Inc -0.004 0.002 -0.295 -2.022* African Amer. 43.822 57.123 0.075 0.767 Hispanic 42.400 52.213 0.129 0.812 Business 19.165 38.651 0.049 0.496 Labor -48.846 46.125 -0.131 -1.059 E-score 0.088 0.145 0.071 0.608 105th Party Unity -0.246 0.049 -0.571 -5.026** N 444 R2 0.697 Adjusted R2 0.613 Constant 161.084 Year Elected -1.060 0.609 -0.152 -1.741 Northeast -20.847 12.225 -0.218 -1.705 South -4.255 13.379 -0.051 -0.318 West -7.733 14.315 -0.083 -0.540 Diff DW Nom 30.595 33.776 0.076 0.906 Min. Wage Law 9.399 8.423 0.109 1.116 Fed Spending -36.184 27.164 -0.214 -1.332 Per Cap Inc -0.002 0.001 -0.221 -1.502 African Amer. 36.707 55.855 0.063 0.657 Hispanic 52.266 50.740 0.162 1.030 Business 60.139 40.398 0.141 1.489 Labor -73.178 47.829 -0.203 -1.530 E-score -0.162 0.145 -0.153 -1.116 106th Party Unity -0.280 0.053 -0.657 -5.250** N 440 R2 0.718 Adjusted R2 0.641 Constant 144.267 Year Elected -0.810 0.610 -0.118 -1.329 Northeast -20.263 11.964 -0.212 -1.694 South -0.909 13.024 -0.011 -0.070 West -5.712 14.654 -0.062 -0.390 Diff DW Nom 14.895 34.111 0.037 0.437 Min. Wage Law 6.080 8.542 0.071 0.712 Fed Spending -29.318 24.078 -0.196 -1.218 Per Cap Inc -0.002 0.001 -0.177 -1.328 107th African Amer. 19.195 55.510 0.033 0.346 231 Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores Hispanic 44.648 50.570 0.147 0.883 Business 71.410 33.426 0.182 2.136* Labor -57.375 36.549 -0.176 -1.570 E-score -0.003 0.158 -0.002 -0.017 Party Unity -0.226 0.054 -0.546 -4.213** N 444 R2 0.721 Adjusted R2 0.645 Constant 114.741 Year Elected -0.241 0.584 -0.037 -0.412 Northeast -19.772 11.843 -0.207 -1.670 South -5.460 13.676 -0.066 -0.399 West -14.037 15.470 -0.151 -0.907 Diff DW Nom 2.136 35.180 0.005 0.061 Min. Wage Law 3.120 8.649 0.036 0.361 Fed Spending -11.689 19.766 -0.094 -0.591 Per Cap Inc -0.002 0.001 -0.169 -1.305 African Amer. 32.073 59.408 0.055 0.540 Hispanic 80.233 50.089 0.269 1.602 Business 69.354 31.228 0.195 2.221* Labor -36.179 47.107 -0.098 -0.768 E-score -0.113 0.147 -0.100 -0.767 108th Party Unity -0.277 0.051 -0.677 -5.420** N 440 R2 0.705 Adjusted R2 0.624 Constant 87.628 * p < 0.05, ** p < 0.01 Political parties shape legislative voting in each Congress in the model (99 th through 108 th Congresses). Testing Hypothesis 7b: Republican legislators are less likely to vote for increasing the minimum wage more often than Democrats, party unity is a statistically significant variable that indicates strong Republican support for economically efficient policies. Republican legislators were found to be less likely to support increasing the minimum wage, an economically inefficient policy. Substituting party unity into the model increased the statistical significance of Business as a measure of self-interest across the model from Base Model analysis and further exposed links between state economic conditions and legislative voting. Per capita income in a state was negatively related to opposing minimum wage policies. 232 This association is not surprising in that lower levels of per capita income could be a push for supporting wage floors. What is more surprising is that higher percentages of African Americans and Hispanics were found to be positively associated with minimum wage opposition in the 100 th and 101 st Congresses. Higher levels of minorities in the population are usually indicative of lower levels of per capita income in a state. Differences in directional movements between these two control variables and opposition to increasing the minimum wage show that greater Republican unity can be affected by differences in state economic conditions. The effect of years of service to the model disappears when substituting party unity, as ideologies associated with length of tenure are replaced by greater adherence to party positions. Substituting party unity into the model raised coefficients of determination measures to a range of the high sixties to low eighties, offering support to party as a predictor of behavior. Adding ADA to the Base Model (see Table 4.18) finds ADA to be statistically significant in each Congress and negatively correlated with opposition to increasing the minimum wage, results expected from testing Hypothesis 2b: Legislators with higher ADA scores vote in support of increasing the federal minimum wage. Compared to other independent variables in the model, for both minimum wage and medical malpractice dependent variables standardized coefficients for ADA indicate a relatively higher impact of that variable on per unit changes in support for either policy legislation. Controlling for effects of state economic conditions produced mixed results in the ADA model. Hispanic was statistically significant in the 101 st House and per capita income was statistically significant in the 102 nd House. 233 Table 4.18 Regression Analysis of Base Model and ADA Substitution from 99th to 108th House: Minimum Wage Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores Year Elected -0.974 0.497 -0.140 -1.961 Northeast -7.871 8.371 -0.085 -0.940 South -3.116 9.749 -0.038 -0.320 West -0.481 12.113 -0.005 -0.040 Diff DW Nom 74.859 25.425 0.240 2.944** Min. Wage Law 13.349 8.045 0.152 1.659 Fed Spending -9.620 16.629 -0.051 -0.579 Per Cap Inc -0.003 0.002 -0.133 -1.590 African Amer. 29.390 60.178 0.042 0.488 Hispanic 32.751 60.948 0.063 0.537 Business 6.772 12.026 0.047 0.563 Labor -1.879 17.914 -0.011 -0.105 E-score 0.325 0.114 0.324 2.855** 99th ADA -0.574 0.128 -0.537 -4.501** N 439 R2 0.811 Adjusted R2 0.757 Constant 72.679 Year Elected -0.826 0.474 -0.119 -1.744 Northeast -10.910 9.200 -0.114 -1.186 South -19.316 10.255 -0.240 -1.884 West -5.265 11.010 -0.059 -0.478 Diff DW Nom -16.552 12.091 -0.105 -1.369 Min. Wage Law 3.309 6.759 0.039 0.490 Fed Spending -25.275 16.136 -0.137 -1.566 Per Cap Inc -0.002 0.002 -0.112 -1.051 African Amer. 115.774 61.896 0.168 1.870 Hispanic 82.686 52.422 0.174 1.577 Business -33.819 17.268 -0.163 -1.959 Labor -67.323 23.320 -0.299 -2.887** E-score 0.142 0.081 0.146 1.758 100th ADA -0.681 0.121 -0.661 -5.626** N 441 R2 0.827 Adjusted R2 0.776 Constant 133.483 Year Elected -0.196 0.137 -0.043 -1.427 Northeast -1.478 3.772 -0.017 -0.392 South -2.749 3.831 -0.036 -0.718 West -0.176 4.359 -0.002 -0.040 Diff DW Nom 16.466 6.725 0.075 2.449* Min. Wage Law 4.902 3.029 0.063 1.618 Fed Spending -3.003 6.755 -0.018 -0.445 Per Cap Inc 0.000 0.001 -0.031 -0.617 African Amer. 1.684 18.837 0.004 0.089 Hispanic 31.942 15.814 0.082 2.020* Business 5.531 3.411 0.052 1.621 Labor -30.850 8.374 -0.176 -3.684** 101st E-score 0.142 0.061 0.078 2.339* 234 Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores ADA -0.738 0.053 -0.675 -13.863** N 442 R2 0.670 Adjusted R2 0.658 Constant 66.403 Year Elected -0.549 0.511 -0.079 -1.075 Northeast -3.360 9.832 -0.035 -0.342 South -3.109 10.901 -0.037 -0.285 West -0.989 11.625 -0.011 -0.085 Diff DW Nom 34.388 21.369 0.119 1.609 Min. Wage Law 7.175 6.772 0.083 1.059 Fed Spending -11.765 20.194 -0.061 -0.583 Per Cap Inc -0.003 0.002 -0.236 -2.193* African Amer. 55.186 49.791 0.092 1.108 Hispanic 3.903 42.072 0.011 0.093 Business -21.823 35.868 -0.046 -0.608 Labor -27.501 32.847 -0.085 -0.837 E-score 0.496 0.135 0.475 3.657** 102nd ADA -0.424 0.138 -0.386 -3.081** N 441 R2 0.820 Adjusted R2 0.770 Constant 114.778 Year Elected -1.172 0.561 -0.176 -2.091* Northeast -21.649 11.422 -0.226 -1.895 South -10.700 11.090 -0.129 -0.965 West -8.405 13.880 -0.091 -0.606 Diff DW Nom 62.255 29.771 0.176 2.091* Min. Wage Law 18.340 7.836 0.221 2.340* Fed Spending -28.826 21.145 -0.158 -1.363 Per Cap Inc -0.003 0.002 -0.207 -1.794 African Amer. 28.665 52.334 0.049 0.548 Hispanic 38.383 43.011 0.109 0.892 Business 23.664 41.029 0.049 0.577 Labor -88.390 37.856 -0.260 -2.335* E-score -0.019 0.120 -0.016 -0.161 103rd ADA -0.653 0.114 -0.611 -5.731** N 442 R2 0.776 Adjusted R2 0.714 Constant 158.396 Year Elected -0.007 0.211 -0.001 -0.033 Northeast -5.401 5.946 -0.045 -0.908 South 2.438 6.305 0.024 0.387 West 13.859 6.715 0.117 2.064* Diff DW Nom 27.202 10.797 0.094 2.520* Min. Wage Law -3.057 4.380 -0.029 -0.698 Fed Spending -14.957 11.261 -0.065 -1.328 Per Cap Inc 0.000 0.001 -0.021 -0.355 African Amer. 40.722 30.072 0.063 1.354 Hispanic -9.027 21.493 -0.020 -0.420 104th Business -31.184 20.027 -0.059 -1.557 235 Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores Labor -33.084 21.961 -0.081 -1.507 E-score -0.070 0.104 -0.043 -0.670 ADA -0.877 0.099 -0.658 -8.887** N 445 R2 0.547 Adjusted R2 0.531 Constant 105.712 Year Elected -0.758 0.610 -0.109 -1.243 Northeast -8.074 12.050 -0.084 -0.670 South -2.096 13.501 -0.025 -0.155 West 0.892 14.837 0.010 0.060 Diff DW Nom 28.054 34.200 0.070 0.820 Min. Wage Law 3.985 8.328 0.046 0.479 Fed Spending -34.174 27.272 -0.186 -1.253 Per Cap Inc -0.002 0.002 -0.201 -1.430 African Amer. 26.548 54.666 0.046 0.486 Hispanic 26.729 49.310 0.081 0.542 Business 15.799 36.945 0.041 0.428 Labor -39.247 44.026 -0.105 -0.891 E-score 0.098 0.138 0.079 0.709 105th ADA -0.637 0.112 -0.629 -5.706** N 444 R2 0.723 Adjusted R2 0.647 Constant 159.298 Year Elected -0.996 0.586 -0.143 -1.699 Northeast -18.267 11.780 -0.191 -1.551 South -3.836 12.855 -0.046 -0.298 West -4.437 13.740 -0.048 -0.323 Diff DW Nom 21.423 32.616 0.053 0.657 Min. Wage Law 6.884 8.126 0.080 0.847 Fed Spending -44.802 26.099 -0.264 -1.717 Per Cap Inc -0.002 0.001 -0.152 -1.075 African Amer. 26.163 53.670 0.045 0.487 Hispanic 36.261 48.464 0.112 0.748 Business 57.434 38.535 0.134 1.490 Labor -59.775 46.419 -0.166 -1.288 E-score -0.144 0.136 -0.136 -1.057 106th ADA -0.740 0.127 -0.708 -5.837** N 440 R2 0.740 Adjusted R2 0.668 Constant 170.109 Year Elected -0.673 0.592 -0.098 -1.138 Northeast -20.739 11.534 -0.217 -1.798 South -0.056 12.537 -0.001 -0.004 West -2.256 14.087 -0.024 -0.160 Diff DW Nom 4.112 33.081 0.010 0.124 Min. Wage Law 3.156 8.325 0.037 0.379 Fed Spending -32.288 23.168 -0.216 -1.394 Per Cap Inc -0.001 0.001 -0.125 -0.969 107th African Amer. 4.699 53.774 0.008 0.087 236 Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores Hispanic 32.005 48.557 0.106 0.659 Business 70.875 32.233 0.181 2.199* Labor -49.518 35.461 -0.152 -1.396 E-score -0.042 0.152 -0.035 -0.274 ADA -0.578 0.121 -0.611 -4.788** N 444 R2 0.741 Adjusted R2 0.670 Constant 134.072 Year Elected -0.276 0.584 -0.043 -0.473 Northeast -15.401 11.986 -0.161 -1.285 South 1.663 13.715 0.020 0.121 West -6.409 15.443 -0.069 -0.415 Diff DW Nom 1.852 35.246 0.005 0.053 Min. Wage Law -2.255 8.824 -0.026 -0.256 Fed Spending -23.838 19.674 -0.191 -1.212 Per Cap Inc -0.001 0.001 -0.096 -0.734 African Amer. 0.671 59.677 0.001 0.011 Hispanic 42.839 49.602 0.144 0.864 Business 70.579 31.282 0.199 2.256* Labor 7.815 50.744 0.021 0.154 E-score -0.052 0.142 -0.046 -0.368 108th ADA -0.703 0.130 -0.732 -5.398** N 440 R2 0.704 Adjusted R2 0.623 Constant 112.514 * p < 0.05, ** p < 0.01 Directional impact of higher per capita income within the legislator?s state negatively impacted support for economically efficient policy making, while higher levels of Hispanics within the states? population predicted greater support for economically efficient legislation. Legislators from states with higher percentages of Hispanics in the population were more likely to oppose increasing the minimum wage, while legislators from states with higher per capita incomes were more likely to support increasing the minimum wage. The effect of percentage Hispanic in the model produced very small (+0.082) changes in the model, however. Length of service for House members in Congress is statistically significant in the model for ADA. Controls for length of service negatively impact the model, although the effect is a relatively weak 237 0.176 standardized coefficient. E-score remains statistically significant in the 99 th and 102 nd Congresses with no changes in its positive directional impact with opposition to minimum wage increases. As Table 4.19 indicates ACU is a statistically significant independent variable in each House analyzed (99th through 108th) in testing Hypothesis 3b: Legislators with higher ACU scores vote in opposition to increasing the federal minimum wage. A positive association exists between ACU and economic efficiency in each Congress, and based on p values, the variable offers perhaps the strongest ideological basis for predicting legislative behavior in this model. Legislators with higher ACU scores are found to oppose increasing the federal minimum wage, an economically inefficient public policy. Controlling for state economic conditions finds legislators in states with higher percentages of minority population opposing minimum wage increases, while those in states with lower per capita incomes support increasing the minimum wage. Results for state economic conditions are consistent throughout the base model and each substitution. Although these associations were statistically significant in only the 100th and 101st House, the fact that greater differences in legislator ideology from median party ideology in the 100th House supports increasing the minimum wage, an economically inefficient act, suggests that factors associated with a legislators? constituency affects his or her support of economically efficient public policies. E-score was statistically significant in the 101st and 102nd House and positively affected opposition to increasing the minimum wage. In comparison to ADA, ceteris 238 paribus, slightly higher coefficients of determination in each ACU regression suggest advantages of ACU over ADA in this model in explaining model variability. Table 4.19 Regression Analysis of Base Model and ACU Substitution for 99th to 108th House: Minimum Wage Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores Year Elected -0.838 0.491 -0.120 -1.706 Northeast -10.392 8.318 -0.112 -1.249 South -7.993 9.679 -0.097 -0.826 West -3.626 11.993 -0.039 -0.302 Diff DW Nom 55.055 24.687 0.176 2.230* Min. Wage Law 10.473 7.937 0.119 1.319 Fed Spending -5.497 16.526 -0.029 -0.333 Per Cap Inc -0.003 0.002 -0.138 -1.676 African Amer. 43.690 58.724 0.063 0.744 Hispanic 41.594 60.338 0.080 0.689 Business 3.543 12.036 0.024 0.294 Labor -5.893 17.403 -0.036 -0.339 E-score 0.239 0.122 0.238 1.959 99th ACU 0.651 0.139 0.596 4.665** N 439 R2 0.815 Adjusted R2 0.762 Constant 22.941 Year Elected -0.841 0.433 -0.121 -1.945 Northeast -11.630 8.402 -0.121 -1.384 South -22.873 9.430 -0.284 -2.426* West -7.345 10.059 -0.082 -0.730 Diff DW Nom -23.107 11.230 -0.147 -2.058* Min. Wage Law -0.034 6.208 0.000 -0.006 Fed Spending -18.531 14.710 -0.101 -1.260 Per Cap Inc -0.002 0.002 -0.109 -1.117 African Amer. 127.129 56.337 0.184 2.257* Hispanic 111.836 48.026 0.235 2.329* Business -26.112 15.296 -0.126 -1.707 Labor -44.920 22.491 -0.199 -1.997 E-score 0.053 0.078 0.054 0.680 100th ACU 0.755 0.110 0.782 6.877** N 441 R2 0.856 Adjusted R2 0.813 Constant 56.610 Year Elected -0.118 0.131 -0.026 -0.900 Northeast 1.111 3.651 0.013 0.304 South -1.734 3.692 -0.023 -0.470 West -0.327 4.177 -0.004 -0.078 Diff DW Nom 12.329 6.463 0.056 1.908 Min. Wage Law 3.964 2.915 0.051 1.360 Fed Spending -4.811 6.507 -0.029 -0.739 Per Cap Inc -0.001 0.001 -0.079 -1.606 101st African Amer. 13.890 18.174 0.029 0.764 239 Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores Hispanic 31.374 15.191 0.080 2.065* Business 5.368 3.284 0.050 1.635 Labor -25.293 7.963 -0.144 -3.176** E-score 0.193 0.058 0.107 3.311** ACU 0.766 0.049 0.715 15.497** N 442 R2 0.692 Adjusted R2 0.681 Constant 4.745 Year Elected -0.523 0.499 -0.075 -1.049 Northeast -0.924 9.583 -0.010 -0.096 South -4.028 10.639 -0.049 -0.379 West -0.837 11.335 -0.009 -0.074 Diff DW Nom 29.657 20.842 0.102 1.423 Min. Wage Law 7.215 6.600 0.084 1.093 Fed Spending -7.978 19.705 -0.041 -0.405 Per Cap Inc -0.004 0.002 -0.243 -2.318* African Amer. 49.079 48.581 0.082 1.010 Hispanic 10.992 41.243 0.030 0.267 Business -11.743 35.137 -0.025 -0.334 Labor -21.897 32.169 -0.068 -0.681 E-score 0.444 0.135 0.425 3.277** 102nd ACU 0.443 0.125 0.446 3.554** N 441 R2 0.829 Adjusted R2 0.782 Constant 71.846 Year Elected -0.981 0.534 -0.147 -1.836 Northeast -14.733 11.027 -0.154 -1.336 South -11.595 10.469 -0.140 -1.108 West -6.010 13.125 -0.065 -0.458 Diff DW Nom 44.842 28.263 0.127 1.587 Min. Wage Law 14.258 7.477 0.172 1.907 Fed Spending -26.862 19.968 -0.147 -1.345 Per Cap Inc -0.004 0.002 -0.239 -2.186* African Amer. 47.956 49.399 0.081 0.971 Hispanic 40.858 40.597 0.116 1.006 Business 27.195 38.746 0.056 0.702 Labor -62.177 37.165 -0.183 -1.673 E-score 0.029 0.111 0.024 0.259 103rd ACU 0.666 0.101 0.659 6.568** N 442 R2 0.800 Adjusted R2 0.745 Constant 96.880 Year Elected 0.250 0.206 0.042 1.213 Northeast -5.300 5.782 -0.044 -0.917 South 3.732 6.135 0.036 0.608 West 12.748 6.534 0.107 1.951 Diff DW Nom 21.672 10.554 0.075 2.053* Min. Wage Law -4.627 4.260 -0.044 -1.086 104th Fed Spending -7.822 10.995 -0.034 -0.711 240 Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores Per Cap Inc -2.783E-05 0.001 -0.002 -0.028 African Amer. 36.747 29.255 0.057 1.256 Hispanic -0.107 20.979 0.000 -0.005 Business -23.059 19.465 -0.044 -1.185 Labor -31.122 20.922 -0.076 -1.488 E-score -0.136 0.100 -0.085 -1.358 ACU 0.920 0.089 0.725 10.336** N 445 R2 0.571 Adjusted R2 0.557 Constant 6.177 Year Elected -0.578 0.557 -0.083 -1.037 Northeast -6.451 10.935 -0.067 -0.590 South -4.844 12.275 -0.058 -0.395 West 1.672 13.466 0.018 0.124 Diff DW Nom 25.500 31.042 0.063 0.821 Min. Wage Law 0.992 7.597 0.012 0.131 Fed Spending -29.604 24.806 -0.162 -1.193 Per Cap Inc -0.002 0.002 -0.177 -1.380 African Amer. 33.632 49.590 0.058 0.678 Hispanic 25.451 44.688 0.078 0.570 Business 8.917 33.613 0.023 0.265 Labor -28.998 39.546 -0.078 -0.733 E-score 0.159 0.125 0.128 1.275 105th ACU 0.741 0.104 0.680 7.091** N 444 R2 0.772 Adjusted R2 0.709 Constant 77.294 Year Elected -0.733 0.569 -0.105 -1.288 Northeast -12.557 11.406 -0.131 -1.101 South -4.396 12.329 -0.053 -0.357 West -3.316 13.176 -0.036 -0.252 Diff DW Nom 0.973 31.810 0.002 0.031 Min. Wage Law 5.630 7.810 0.065 0.721 Fed Spending -36.233 25.030 -0.214 -1.448 Per Cap Inc -0.002 0.001 -0.160 -1.174 African Amer. 19.600 51.491 0.034 0.381 Hispanic 43.224 46.542 0.134 0.929 Business 50.588 36.612 0.118 1.382 Labor -46.422 44.960 -0.129 -1.033 E-score -0.088 0.126 -0.083 -0.701 106th ACU 0.791 0.123 0.725 6.443** N 440 R2 0.760 Adjusted R2 0.695 Constant 85.460 Year Elected -0.615 0.559 -0.090 -1.099 Northeast -17.488 10.935 -0.183 -1.599 South -3.424 11.906 -0.041 -0.288 West -1.146 13.322 -0.012 -0.086 107th Diff DW Nom 3.359 31.264 0.008 0.107 241 Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores Min. Wage Law 1.304 7.912 0.015 0.165 Fed Spending -31.388 21.919 -0.210 -1.432 Per Cap Inc -0.001 0.001 -0.082 -0.666 African Amer. 3.197 50.832 0.006 0.063 Hispanic 31.079 45.922 0.102 0.677 Business 68.217 30.494 0.174 2.237* Labor -41.315 33.663 -0.127 -1.227 E-score -0.023 0.136 -0.019 -0.170 ACU 0.671 0.119 0.659 5.623** N 444 R2 0.768 Adjusted R2 0.704 Constant 58.729 Year Elected -0.265 0.552 -0.041 -0.480 Northeast -14.884 11.338 -0.155 -1.313 South -9.194 13.005 -0.111 -0.707 West -6.879 14.628 -0.074 -0.470 Diff DW Nom -4.661 33.496 -0.012 -0.139 Min. Wage Law -1.478 8.304 -0.017 -0.178 Fed Spending -16.651 18.670 -0.134 -0.892 Per Cap Inc -0.001 0.001 -0.088 -0.718 African Amer. 27.537 56.352 0.048 0.489 Hispanic 60.670 47.107 0.204 1.288 Business 64.304 29.610 0.181 2.172* Labor -14.740 45.543 -0.040 -0.324 E-score -0.125 0.138 -0.111 -0.908 108th ACU 0.853 0.138 0.779 6.189** N 440 R2 0.734 Adjusted R2 0.661 Constant 31.495 * p < 0.05, ** p < 0.01 Table 4.20 presents the results of substituting DW Nominate into the model. With the exception of the 102nd House, DW Nominate is a statistically significant independent variable in each House lending support to Hypothesis 4b: Legislators with higher DW Nominate scores vote in opposition to increasing the minimum wage. The variable is positively correlated with economically efficient policymaking. That is, legislators with higher DW Nominate scores oppose increasing the minimum wage. This relationship is consistent throughout the testing of the model. 242 Table 4.20 Regression Analysis of Base Model and DW Nominate Substitution for 99th to 108th House: Minimum Wage Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores Year Elected -0.532 0.503 -0.076 -1.058 Northeast -3.428 8.387 -0.037 -0.409 South -4.455 9.740 -0.054 -0.457 West 1.986 12.124 0.021 0.164 Diff DW Nom 61.581 25.004 0.197 2.463* Min. Wage Law 11.013 8.017 0.125 1.374 Fed Spending -5.503 16.696 -0.029 -0.330 Per Cap Inc -0.003 0.002 -0.176 -2.154* African Amer. 65.474 58.542 0.094 1.118 Hispanic 45.405 60.982 0.087 0.745 Business 3.800 12.159 0.026 0.313 Labor -2.315 17.862 -0.014 -0.130 E-score 0.259 0.122 0.258 2.126* 99th DW Nominate 56.317 12.478 0.606 4.513** N 439 R2 0.811 Adjusted R2 0.747 Constant 54.751 Year Elected -0.292 0.573 -0.042 -0.510 Northeast -0.468 10.696 -0.005 -0.044 South -12.244 11.804 -0.152 -1.037 West -3.724 12.827 -0.041 -0.290 Diff DW Nom 9.220 13.841 0.059 0.666 Min. Wage Law 6.036 7.870 0.071 0.767 Fed Spending -12.038 18.881 -0.065 -0.638 Per Cap Inc -0.004 0.002 -0.229 -1.882 African Amer. 139.821 71.735 0.202 1.949 Hispanic 89.580 61.016 0.188 1.468 Business -10.730 19.152 -0.052 -0.560 Labor -123.624 23.031 -0.549 -5.368** E-score 0.242 0.091 0.249 2.679* 100th DW Nominate 28.590 8.593 0.297 3.327** N 441 R2 0.766 Adjusted R2 0.697 Constant 97.077 Year Elected -0.047 0.151 -0.010 -0.312 Northeast -2.744 4.160 -0.032 -0.660 South -1.994 4.222 -0.026 -0.472 West -1.752 4.771 -0.019 -0.367 Diff DW Nom 13.089 7.406 0.059 1.767 Min. Wage Law 3.064 3.337 0.040 0.918 Fed Spending -3.913 7.443 -0.024 -0.526 Per Cap Inc -0.001 0.001 -0.089 -1.579 African Amer. 23.324 20.825 0.050 1.120 Hispanic 48.827 17.274 0.125 2.827* Business 6.684 3.749 0.062 1.783 Labor -65.998 8.032 -0.376 -8.217** 101st E-score 0.069 0.065 0.038 1.053 243 Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores DW Nominate 43.322 4.457 0.433 9.719** N 442 R2 0.599 Adjusted R2 0.584 Constant 54.072 Year Elected -0.441 0.561 -0.063 -0.786 Northeast 0.149 10.514 0.002 0.014 South 0.775 11.459 0.009 0.068 West -0.386 12.359 -0.004 -0.031 Diff DW Nom 27.693 22.860 0.096 1.211 Min. Wage Law 8.105 7.179 0.094 1.129 Fed Spending -9.866 21.444 -0.051 -0.460 Per Cap Inc -0.004 0.002 -0.264 -2.316* African Amer. 55.459 52.857 0.093 1.049 Hispanic 0.317 45.299 0.001 0.007 Business -10.182 39.000 -0.022 -0.261 Labor -42.865 34.661 -0.132 -1.237 E-score 0.624 0.141 0.598 4.410** 102nd DW Nominate 19.229 11.742 0.208 1.638 N 441 R2 0.797 Adjusted R2 0.741 Constant 93.305 Year Elected -0.731 0.565 -0.110 -1.294 Northeast -16.273 11.262 -0.170 -1.445 South -8.235 10.731 -0.099 -0.767 West -6.561 13.450 -0.071 -0.488 Diff DW Nom 60.852 28.807 0.172 2.112* Min. Wage Law 14.284 7.673 0.172 1.862 Fed Spending -18.556 20.592 -0.102 -0.901 Per Cap Inc -0.003 0.002 -0.216 -1.932 African Amer. 36.833 50.613 0.062 0.728 Hispanic 51.444 41.686 0.146 1.234 Business 41.034 39.967 0.085 1.027 Labor -49.681 39.780 -0.146 -1.249 E-score 0.008 0.115 0.006 0.067 103rd DW Nominate 61.089 9.836 0.678 6.211** N 442 R2 0.790 Adjusted R2 0.733 Constant 106.470 Year Elected 0.315 0.206 0.053 1.532 Northeast -5.844 5.768 -0.049 -1.013 South 0.968 6.112 0.009 0.158 West 12.401 6.519 0.104 1.902 Diff DW Nom 2.786 10.960 0.010 0.254 Min. Wage Law -4.694 4.249 -0.045 -1.105 Fed Spending -8.868 10.956 -0.038 -0.809 Per Cap Inc -0.001 0.001 -0.041 -0.703 African Amer. 60.995 29.171 0.094 2.091* Hispanic -3.715 20.871 -0.008 -0.178 104th Business -25.600 19.414 -0.049 -1.319 244 Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores Labor -22.995 21.174 -0.056 -1.086 E-score -0.131 0.099 -0.081 -1.316 DW Nominate 86.018 8.221 0.758 10.463** N 445 R2 0.574 Adjusted R2 0.559 Constant 66.328 Year Elected -0.363 0.601 -0.052 -0.604 Northeast -9.106 11.493 -0.095 -0.792 South -6.481 12.980 -0.078 -0.499 West -4.504 14.300 -0.049 -0.315 Diff DW Nom 20.982 32.936 0.052 0.637 Min. Wage Law 4.933 7.979 0.057 0.618 Fed Spending -26.908 26.306 -0.147 -1.023 Per Cap Inc -0.002 0.002 -0.206 -1.528 African Amer. 33.486 52.373 0.058 0.639 Hispanic 50.546 47.854 0.154 1.056 Business 2.592 35.767 0.007 0.072 Labor -21.708 43.036 -0.058 -0.504 E-score 0.087 0.133 0.070 0.656 105th DW Nominate 62.126 9.845 0.707 6.311** N 444 R2 0.745 Adjusted R2 0.675 Constant 115.542 Year Elected -0.455 0.599 -0.065 -0.760 Northeast -14.587 11.679 -0.152 -1.249 South -5.737 12.677 -0.069 -0.453 West -6.353 13.545 -0.068 -0.469 Diff DW Nom 19.920 32.162 0.050 0.619 Min. Wage Law 5.768 8.027 0.067 0.719 Fed Spending -28.923 25.785 -0.171 -1.122 Per Cap Inc -0.002 0.001 -0.152 -1.086 African Amer. 23.395 52.896 0.040 0.442 Hispanic 55.133 48.045 0.171 1.148 Business 45.026 37.499 0.105 1.201 Labor -40.839 46.977 -0.113 -0.869 E-score -0.127 0.132 -0.120 -0.962 106th DW Nominate 65.628 10.847 0.755 6.051** N 440 R2 0.747 Adjusted R2 0.678 Constant 105.318 Year Elected -.289 .597 -.042 -.484 Northeast -14.457 11.355 -.151 -1.273 South -3.024 12.296 -.036 -.246 West -4.483 13.779 -.048 -.325 Diff DW Nom 11.448 32.163 .028 .356 Min. Wage Law 3.089 8.120 .036 .380 Fed Spending -23.772 22.766 -.159 -1.044 Per Cap Inc -.001 .001 -.124 -.983 107th African Amer. 8.513 52.426 .015 .162 245 Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores Hispanic 48.572 47.656 .160 1.019 Business 63.330 31.521 .161 2.009* Labor -29.554 35.888 -.091 -.824 E-score -.015 .143 -.013 -.108 DW Nominate 57.362 11.152 .667 5.144** N 444 R2 0.753 Adjusted R2 0.685 Constant 82.673 Year Elected .316 .561 .049 .562 Northeast -13.611 11.111 -.142 -1.225 South -7.237 12.694 -.087 -.570 West -11.153 14.314 -.120 -.779 Diff DW Nom -1.412 32.607 -.004 -.043 Min. Wage Law -.971 8.097 -.011 -.120 Fed Spending -8.549 18.365 -.069 -.465 Per Cap Inc -.001 .001 -.100 -.833 African Amer. 21.354 55.097 .037 .388 Hispanic 85.873 46.479 .288 1.848 Business 61.603 28.950 .174 2.128* Labor -4.022 44.998 -.011 -.089 E-score -.175 .138 -.155 -1.272 108th DW Nominate 70.709 10.855 .832 6.514** N 440 R2 0.746 Adjusted R2 0.676 Constant 54.036 * p < 0.05, ** p < 0.01 Per capita income as a control for state economic conditions is statistically significant in the 99 th and 102 nd Congresses, negatively impacting the model. Positive associations continue to exist between percent minority in a population and a legislator?s support of economically efficient policies. Business is a statistically significant variable in the 107 th and 108 th House for all three measures of ideology substitution ? ADA, ACU, and DW Nominate ? and functions along with the substituted variable in explaining legislative voting. That Business is statistically significant in a model with a highly statistically significant substituted variable supports self-interest claims. Standardized coefficients indicate a relatively weaker per unit effect of self-interest on the dependent variable in comparison to ideology. 246 Standardized coefficients indicate moderate movement for DW Nominate values in affecting change per unit support for minimum wage legislation. ACU values show higher standardized coefficients on average than DW Nominate scores, and DW Nominate scores show slightly higher values than ADA scores. Relative high p values support the probability that the association was not the result of chance. Measures of conservative ideology (ACU values and higher DW Nominate scores) are clearly important predictors of legislative behavior in this model. Substituting Legislator Party into the model (Table 4.21) finds that Democrats support increasing the minimum wage while Republicans are opposed. Party control of the House must be defined when analyzing Legislator Party. Referring to party divisions summarized in Table 4.10, Democrats controlled the House during the 99 th through 103 rd Congresses and Republicans controlled the House during the 104 th through 108 th Congresses. Like medical malpractice dependent variable, minimum wage is coded to indicate a positive relationship between the variable and economic efficiency. A vote in opposition to minimum wage increases is an economically efficient act. According to coefficients in the regression output, directional impact of a plus (+) indicates a relationship where the legislator of the majority party is more opposed to increasing the minimum wage and a minus (-) indicates support for increasing the minimum wage increases if the legislator is not a member of the majority party. When Democrats are in the majority coefficients are negative, indicating greater opposition to increasing the minimum wage from Republicans, the party that does not control the House. Republicans are in the majority in the House in the 104 th through 108 th Congresses and 247 vote increasingly against increasing the minimum wage regardless of party control, as positive, statistically significant directional impact of Legislator Party indicates. These results are contrary to the expected distinctions in minority-majority status expressed in Hypothesis 10b: Legislators from the minority party (House) are less likely than majority party legislators to support increasing the federal minimum wage. Legislator party is statistically significant in every Congress in the study except the 99th and 102nd. Per capita income and labor each remain statistically significant in the model after substituting in legislator party. Negative correlation indicates representatives who receive higher labor contributions and represent states that have lower per capita incomes are more likely to support increasing the minimum wage. A positive, statistically significant association between a legislator?s ideology and median party ideology in the base model remains in the 104 th Congress with legislator party included in the model. Divisions in ideology are important in the model and the greater the separation of a legislator?s ideology from the ideology of his or her median party the greater is the effect on supporting economically efficient policies. 248 Table 4.21 Regression Analysis of Base Model and Legislator Party Substitution for 99th to 108th House: Minimum Wage Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores Year Elected -0.853 0.607 -0.122 -1.406 Northeast -6.059 9.992 -0.066 -0.606 South -3.338 11.702 -0.040 -0.285 West -0.936 14.399 -0.010 -0.065 Diff DW Nom 54.450 29.865 0.174 1.823 Min. Wage Law 11.261 9.641 0.128 1.168 Fed Spending -13.363 19.810 -0.070 -0.675 Per Cap Inc -0.004 0.002 -0.220 -2.281* African Amer. 100.605 68.934 0.144 1.459 Hispanic 31.047 72.334 0.060 0.429 Business 17.511 14.022 0.121 1.249 Labor -28.673 20.057 -0.173 -1.430 E-score 0.599 0.117 0.597 5.144** 99th Legislator Party -2.527 6.131 -0.033 -0.412 N 439 R2 0.734 Adjusted R2 0.658 Constant 63.068 Year Elected -0.713 0.513 -0.102 -1.389 Northeast -7.438 9.902 -0.078 -0.751 South -20.165 11.180 -0.250 -1.804 West -6.547 11.918 -0.073 -0.549 Diff DW Nom -3.689 12.693 -0.024 -0.291 Min. Wage Law 3.487 7.315 0.041 0.477 Fed Spending -11.213 17.530 -0.061 -0.640 Per Cap Inc -0.005 0.002 -0.321 -2.799** African Amer. 172.565 66.768 0.250 2.585* Hispanic 131.151 57.627 0.275 2.276* Business -21.443 18.175 -0.103 -1.180 Labor -96.515 23.108 -0.428 -4.177** E-score 0.111 0.091 0.114 1.218 100th Legislator Party -36.368 8.102 -0.484 -4.489** N 441 R2 0.798 Adjusted R2 0.738 Constant 146.377 Year Elected -0.180 0.138 -0.040 -1.307 Northeast -4.014 3.806 -0.046 -1.055 South 1.639 3.887 0.022 0.422 West 1.384 4.386 0.015 0.316 Diff DW Nom 8.993 6.776 0.041 1.327 Min. Wage Law 0.164 3.054 0.002 0.054 Fed Spending -7.341 6.822 -0.045 -1.076 Per Cap Inc -0.002 0.001 -0.158 -3.058** African Amer. 37.661 19.179 0.080 1.964 Hispanic 44.029 15.840 0.113 2.780** Business 4.714 3.446 0.044 1.368 Labor -51.052 7.443 -0.291 -6.859** 101st E-score 0.070 0.060 0.039 1.179 249 Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores Legislator Party -40.260 2.972 -0.546 -13.545** N 442 R2 0.662 Adjusted R2 0.650 Constant 95.570 Year Elected -0.643 0.542 -0.092 -1.185 Northeast -1.580 10.455 -0.017 -0.151 South 0.764 11.491 0.009 0.066 West -0.858 12.374 -0.009 -0.069 Diff DW Nom 38.271 23.027 0.132 1.662 Min. Wage Law 8.815 7.195 0.102 1.225 Fed Spending -9.126 21.511 -0.047 -0.424 Per Cap Inc -0.004 0.002 -0.304 -2.586* African Amer. 63.278 53.216 0.106 1.189 Hispanic 1.906 45.762 0.005 0.042 Business -10.020 39.195 -0.021 -0.256 Labor -42.539 34.955 -0.131 -1.217 E-score 0.628 0.143 0.602 4.399** 102nd Legislator Party -14.873 9.513 -0.198 -1.563 N 441 R2 0.796 Adjusted R2 0.740 Constant 112.242 Year Elected -1.317 0.616 -0.198 -2.140* Northeast -23.288 12.586 -0.243 -1.850 South -8.735 12.148 -0.105 -0.719 West -8.730 15.216 -0.094 -0.574 Diff DW Nom 64.453 32.604 0.183 1.977 Min. Wage Law 16.921 8.655 0.204 1.955 Fed Spending -23.638 23.264 -0.130 -1.016 Per Cap Inc -0.005 0.002 -0.324 -2.506* African Amer. 73.372 57.905 0.124 1.267 Hispanic 49.973 47.220 0.142 1.058 Business 33.721 45.237 0.070 0.745 Labor -91.034 43.798 -0.268 -2.078* E-score 0.092 0.128 0.076 0.720 103rd Legislator Party -36.021 8.289 -0.480 -4.346** N 442 R2 0.731 Adjusted R2 0.657 Constant 169.732 Year Elected 0.315 0.206 0.053 1.530 Northeast -5.859 5.767 -0.049 -1.016 South 0.973 6.111 0.009 0.159 West 12.373 6.518 0.104 1.898 Diff DW Nom 88.859 11.411 0.309 7.787** Min. Wage Law -4.678 4.248 -0.045 -1.101 Fed Spending -8.875 10.954 -0.038 -0.810 Per Cap Inc -0.001 0.001 -0.040 -0.699 African Amer. 60.776 29.165 0.094 2.084* Hispanic -3.633 20.868 -0.008 -0.174 104th Business -25.616 19.411 -0.049 -1.320 250 Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores Labor -22.925 21.171 -0.056 -1.083 E-score -0.131 0.099 -0.082 -1.323 Legislator Party 69.027 6.592 0.708 10.472** N 445 R2 0.574 Adjusted R2 0.559 Constant 32.977 Year Elected -1.048 0.652 -0.151 -1.606 Northeast -16.051 12.817 -0.168 -1.252 South -2.792 14.614 -0.034 -0.191 West -3.952 16.171 -0.043 -0.244 Diff DW Nom 39.369 36.810 0.098 1.070 Min. Wage Law 8.256 8.997 0.096 0.918 Fed Spending -39.672 29.442 -0.216 -1.347 Per Cap Inc -0.004 0.002 -0.320 -2.119* African Amer. 44.445 59.049 0.077 0.753 Hispanic 40.280 54.044 0.123 0.745 Business 24.880 39.838 0.064 0.625 Labor -58.266 47.578 -0.156 -1.225 E-score 0.086 0.150 0.070 0.575 105th Legislator Party 39.168 8.680 0.524 4.512** N 444 R2 0.676 Adjusted R2 0.587 Constant 152.311 Year Elected -1.176 0.634 -0.169 -1.856 Northeast -22.460 12.751 -0.235 -1.761 South -2.992 13.966 -0.036 -0.214 West -8.413 14.960 -0.091 -0.562 Diff DW Nom 39.114 35.157 0.098 1.113 Min. Wage Law 10.732 8.785 0.125 1.222 Fed Spending -36.329 28.366 -0.214 -1.281 Per Cap Inc -0.003 0.002 -0.249 -1.625 African Amer. 35.955 58.320 0.062 0.617 Hispanic 51.983 53.057 0.161 0.980 Business 56.221 42.243 0.132 1.331 Labor -90.357 49.236 -0.250 -1.835 E-score -0.137 0.152 -0.129 -0.897 106th Legislator Party 43.550 9.490 0.583 4.589** N 440 R2 0.692 Adjusted R2 0.608 Constant 132.886 Year Elected -0.910 0.621 -0.133 -1.465 Northeast -20.714 12.243 -0.216 -1.692 South 0.262 13.312 0.003 0.020 West -5.952 15.012 -0.064 -0.396 Diff DW Nom 19.759 34.864 0.049 0.567 Min. Wage Law 7.405 8.710 0.086 0.850 Fed Spending -29.152 24.660 -0.195 -1.182 Per Cap Inc -0.002 0.001 -0.192 -1.405 107th African Amer. 22.229 56.782 0.038 0.391 251 Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores Hispanic 45.703 51.828 0.151 0.882 Business 72.453 34.214 0.185 2.118* Labor -64.808 37.091 -0.199 -1.747 E-score 0.040 0.160 0.033 0.251 Legislator Party 36.450 9.530 0.488 3.825** N 444 R2 0.708 Adjusted R2 0.628 Constant 99.607 Year Elected -0.303 0.603 -0.047 -0.502 Northeast -20.792 12.235 -0.217 -1.699 South -4.190 14.138 -0.051 -0.296 West -14.116 16.008 -0.152 -0.882 Diff DW Nom 6.873 36.290 0.017 0.189 Min. Wage Law 4.239 8.929 0.049 0.475 Fed Spending -12.222 20.447 -0.098 -0.598 Per Cap Inc -0.002 0.001 -0.183 -1.366 African Amer. 33.829 61.450 0.059 0.551 Hispanic 80.113 51.878 0.269 1.544 Business 69.100 32.296 0.195 2.140* Labor -49.571 48.176 -0.134 -1.029 E-score -0.079 0.151 -0.070 -0.523 108th Legislator Party 45.832 9.328 0.613 4.914** N 440 R2 0.684 Adjusted R2 0.598 Constant 68.383 * p < 0.05, ** p < 0.01 Business contributions are statistically significant in the 107 th and 108 th Congresses and positively correlated with opposition to minimum wage increases. Business is statistically significant in each substitution in the model in those two Congresses, indicating a strong self-interest component based on lobbying efforts. That E-score was statistically significant in the 107 th base model but fails the test of statistical significance in legislator party and each of the other substitutions offers support where self interest can trump economic efficiency. Model Summary of House Analysis In analyzing a base model and substituting five variables (party unity, ADA, ACU, DW Nominate, and legislator party) into the base model to measure changes on 252 each of two dependent variables, medical malpractice and minimum wage, the analysis indicates that economic efficiency plays a role in legislative decision making in the House but becomes a less relevant predictor of behavior when considered relative to measures of liberalism and conservatism. Political party plays a large role in legislative behavior, as Republican legislators unite behind policies that expand economic efficiency and Democrat legislators support economically inefficient policies. These results generally hold across each Congress in the analysis for both dependent variables. The 99th Congress with medical malpractice the dependent variable is the only exception. Unlike results for the other Congresses in the model, in the 99th House E- score and conservative ideology were inversely related to support for medical malpractice. With directional impact for self-interest in that Congress also inverse to the relationship between each self-interest variable and medical malpractice in other Congresses in the model, too few observations in compiling the dependent variables and subjectivity associated with coding of that variable magnify the directional effects when making an analysis. Legislator economic efficiency, as measured by an E-score and measures of liberal-conservative ideology are not mutually exclusive. The model found E-score statistically significant in the base model when it was applied to predict each dependent variable. In each Congress where E-score was significant its directional impact positively correlated with economic efficiency of the legislation regardless of the statistical significance of other variables in that Congress. Legislators in the House who embrace economic efficiency as an ideology embrace economically efficient policies. The per unit effect of E-score was stronger than either self-interest variable in explaining 253 support for medical malpractice reform, an economically efficient policy, and was less likely to happen by chance. In this model self-interest was a predictor of behavior consistently in the base model and each substitution in the 101st, 107th, and 108th Congresses. Neither variable of legislative self-interest ? Business and Labor ? affected per unit changes in support for the dependent variable to the extent of measures for ideology and party. With higher coefficients of determination and greater per unit impact on each dependent variable in the House, ACU functioned as the best predictor of legislative behavior in the model. Results for ADA and DW Nominate confirm liberal-conservative extremes between economically inefficient and economically efficient policies. Party line voting increasingly mirrors these extremes with no distinction noted from changes in party control of the institution. Senate The U.S. Senate is a legislative body characterized by a greater level of heterogeneity than found in the U.S. House. Senate members serve a greater number of people, over a longer period of time, and across a more expansive geographical area. Legislative decision-making is affected by these differences. By including the Senate in this analysis of legislative voting the goal is to investigate how closely senators support economically efficient public policymaking. The Senate analysis includes identical independent and control variables as the House analysis. An additional control variable was added to the model to measure the length of time each senator has served in his or her current term in office. This variable 254 distinguishes length of service within a six-year term from length of overall service in the legislative body. In the 99th, 100th, and 101st Congresses for medical malpractice dependent variable Base Model and each substitution were not measurable because of autocollinearity. That is, when running regression analysis the models too closely approximate a perfect linear correlation with the dependent variable, medical malpractice. Lack of variability between independent and control variables in the model and changes in the dependent variable is evidence of a linear relationship as coefficients of determination equal one. Beginning with the 102nd Congress, Table 4.22 presents the effects of the Senate base model using medical malpractice as the dependent variable. For the 105th Congress the Base Model did not produce statistically significant output and was omitted from tabular summaries. Tables 4.23, 4.24, 4.25, 4.26, and 4.27 provide results of regression analysis when party unity, ADA, ACU, DW Nominate, and legislator party are each substituted into the analysis, respectively. ADA and ACU substitutions did not produce statistically significant results in the 102nd Senate and are omitted from tabular summaries of regression output in Tables 4.24 and 4.25. A discussion of each hypothesis corresponding to the regression run follows each table with results for each table summarized. For each substitution a discussion of changes in results invites comparison between the base model and each substitution. How any changes affect the impact of those independent variables are noted with particular emphasis placed on the effect of economic efficiency. 255 Tables 4.28, 4.29, 4.30, 4.31, 4.32, and 4.33 summarize regression analysis of independent and control variables in the Senate for the Base Model and party unity, ADA, ACU, DW Nominate, and legislative party substitutions, respectively for minimum wage dependent variable. Autocollinearity was present from regression results in the 99th Senate for the Base Model and each substitution. As a result the 99th Senate is excluded from all tabular presentations. Each model in the 100 th through 108 th Congresses produced statistically significant results with the exception of the Base Model in the 104 th Senate. Results for models that do not produce statistically significant results are not included in tabular summaries. Absence of statistical significance in the Base Model for the 104th Senate is addressed by analyzing changes in independent and control variables for each substitution according to statistical significance of those independent and control variables attributed to the addition of the substituted variable. A model summary of the Senate analysis explores differences in legislative decision making across House and Senate and how such differences impact economic efficiency. Medical malpractice Table 4.22 presents a summary of the base model analysis of medical malpractice dependent in the Senate. The model finds E-score statistically significant in four of the six Congresses where the overall model produced statistically significant output. E-score was positively associated with medical malpractice reform, consistent with Hypothesis 1b: Legislators with higher E-scores vote in support of medical malpractice reform. 256 Table 4.22 Regression Analysis of Base Model for 102nd to 108th Senate: Medical Malpractice Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores Year Elected 1.326 0.823 0.320 1.611 Current Term -0.661 4.221 -0.038 -0.157 Northeast -5.223 25.789 -0.072 -0.203 South -11.733 18.798 -0.193 -0.624 West -25.948 21.477 -0.383 -1.208 Diff DW Nom 9.929 67.153 0.026 0.148 Fed Spending 5.247 29.560 0.062 0.178 Per Cap Inc -0.003 0.002 -0.404 -1.493 African Amer. -3.156 67.817 -0.010 -0.047 Hispanic 39.121 80.070 0.136 0.489 Med Mal Crisis 37.918 18.747 0.526 2.023 Health 142.773 103.260 0.271 1.383 Lawyer -296.659 159.845 -0.494 -1.856 102nd E-score 0.759 0.195 0.729 3.902* N 102 R2 0.833 Adjusted R2 0.541 Constant 79.423 Year Elected -.210 1.227 -.052 -.171 Current Term 1.106 5.736 .056 .193 Northeast -13.839 37.599 -.192 -.368 South 12.264 30.003 .201 .409 West 15.446 33.401 .228 .462 Diff DW Nom 38.165 91.324 .099 .418 Fed Spending -31.975 40.930 -.362 -.781 Per Cap Inc -.003 .003 -.347 -1.070 African Amer. -26.433 80.022 -.087 -.330 Hispanic -20.927 106.597 -.073 -.196 Med Mal Crisis 16.677 25.087 .231 .665 Health -167.690 84.681 -.475 -1.980 Lawyer -390.053 163.226 -.746 -2.390* 103rd E-score .099 .355 .095 .279 N 102 R2 0.718 Adjusted R2 0.224 Constant 176.078 Year Elected .129 .227 .058 .567 Current Term -.180 1.323 -.014 -.136 Northeast -4.888 6.803 -.093 -.718 South -8.579 6.859 -.197 -1.251 West -11.129 6.155 -.238 -1.808 Diff DW Nom 21.290 17.299 .121 1.231 Fed Spending 7.719 10.841 .097 .712 Per Cap Inc .001 .001 .196 1.285 African Amer. -12.178 31.063 -.056 -.392 Hispanic -2.780 26.378 -.011 -.105 Med Mal Crisis 8.263 4.406 .193 1.876 Health -123.111 40.627 -.310 -3.030** 104th Lawyer 41.227 32.069 .127 1.286 257 Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores E-score -.147 .089 -.168 -1.654 N 103 R2 0.325 Adjusted R2 0.212 Constant 23.489 Year Elected 1.007 .983 .250 1.024 Current Term 7.333 5.179 .370 1.416 Northeast 19.430 34.594 .270 .562 South 17.565 26.957 .288 .652 West -17.823 29.900 -.263 -.596 Diff DW Nom 79.057 136.397 .211 .580 Fed Spending -21.194 49.903 -.257 -.425 Per Cap Inc -.002 .003 -.357 -.731 African Amer. -84.571 105.432 -.285 -.802 Hispanic -4.760 73.805 -.018 -.065 Med Mal Crisis 3.601 27.610 .050 .130 Health 3.224 454.544 .003 .007 Lawyer -86.228 128.600 -.196 -.671 106th E-score .978 .407 .768 2.403* N 102 R2 0.766 Adjusted R2 0.357 Constant 17.543 Year Elected .347 .480 .070 .724 Current Term 1.629 2.953 .052 .552 Northeast 14.531 16.143 .112 .900 South 24.404 14.967 .229 1.631 West 7.695 14.330 .066 .537 Diff DW Nom 3.782 39.321 .009 .096 Fed Spending -5.827 21.870 -.038 -.266 Per Cap Inc -.003 .002 -.294 -1.914 African Amer. -107.413 68.268 -.204 -1.573 Hispanic -80.237 60.032 -.139 -1.337 Med Mal Crisis 13.918 10.452 .134 1.332 Health -18.117 116.833 -.014 -.155 Lawyer -234.078 67.936 -.320 -3.446** 107th E-score .735 .180 .393 4.078** N 102 R2 0.427 Adjusted R2 0.327 Constant 113.561 Year Elected .164 .437 .035 .375 Current Term -2.346 2.651 -.078 -.885 Northeast 2.804 14.917 .021 .188 South 10.223 14.698 .096 .696 West -4.949 13.702 -.043 -.361 Diff DW Nom 4.783 34.867 .012 .137 Fed Spending 10.812 16.568 .079 .653 Per Cap Inc -.001 .001 -.123 -1.061 African Amer. -159.041 63.838 -.303 -2.491* Hispanic -27.591 50.295 -.051 -.549 Med Mal Crisis 5.790 9.897 .055 .585 108th Health -149.285 83.626 -.156 -1.785 258 Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores Lawyer -96.857 65.674 -.130 -1.475 E-score 1.244 .178 .640 6.972** N 100 R2 0.514 Adjusted R2 0.427 Constant 26.456 * p < 0.05, ** p < 0.01 Self-interest variables of Health and Lawyer were statistically significant in the model. Health contributions were statistically significant in the 104 th House but were not positively associated with medical malpractice reform as expected in Hypothesis 5b: Legislators with higher health care political contributions to total contributions vote in support of malpractice reform. Health was negatively related to medical malpractice reform, predicting that senators that received higher health contributions were less likely to support medical malpractice reform policies. Lawyer was statistically significant in the 103 rd and 107 th Congresses and was negatively correlated with support for medical malpractice reform, as expected from Hypothesis 6b: Legislators with higher legal political contributions to total contributions vote in opposition to medical malpractice reform. Senators receiving higher levels of legal political contributions are less likely to support medical malpractice reform. In the 104 th senate percentage of African Americans within a state is the only control variable statistically significant in the Base Model and its negative correlation with support of medical malpractice reform suggests that senators from states with higher percentages of minorities in the state are less likely to support medical malpractice reform. The base model did not produce statistically significant relationships for length of term in office, length of time in current term, and differences in ideology between a 259 legislator and median ideology of his or her party. Thus, offering no support for Hypothesis 8b: The closer senators are to the end of their current term in office, the more likely they are to support malpractice reform, Hypothesis 9b: Legislators from the minority party (Senate) are more likely than majority party legislators to support medical malpractice reform, and Hypothesis 11b: The greater the ideological division between the legislator and the median ideology of the legislator?s party, the more likely the legislator supports medical malpractice reform. Coefficients of determination were relatively low, often in the thirties and forties. Variability in the model explains relatively little of changes in the dependent variable. Another issue is differences in R square values and adjusted R square values that exist within some Congresses in the model (102 nd , 103 rd , and 106 th Congresses). In those Congresses adjusted R square values are much less than R square values and suggest the addition of explanatory variables do nothing to improve the fit of the model beyond random chance in explaining changes in the dependent variable. Substituting party unity into the analysis (Table 4.23) finds that party unity is statistically significant in each Congress where results are measurable (102 nd through 108 th House) with an exception in the 104 th House. In the 104 th House Health continues to be the only variable statistically significant, and its directional impact with support for medical malpractice reform remains negative. The directional impact for party unity is negative (indicating greater Republican support as a result of coding the variable) and consistent with expected association in Hypothesis 7b: Republican legislators are likely to vote for malpractice reform more often than Democratic legislators. 260 Table 4.23 Regression Analysis of Base Model with Party Unity Substitution for 102nd to 108th Senate: Medical Malpractice Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores Year Elected 1.266 0.617 0.305 2.051 Current Term -0.950 3.164 -0.054 -0.300 Northeast -3.607 19.329 -0.050 -0.187 South -1.343 14.601 -0.022 -0.092 West -23.211 16.121 -0.343 -1.440 Diff DW Nom 24.135 50.583 0.064 0.477 Fed Spending 6.140 22.147 0.073 0.277 Per Cap Inc -0.001 0.002 -0.096 -0.413 African Amer. -59.213 54.903 -0.193 -1.079 Hispanic 26.211 60.175 0.091 0.436 Med Mal Crisis 21.126 15.366 0.293 1.375 Health 73.004 81.578 0.139 0.895 Lawyer -172.553 128.306 -0.287 -1.345 E-score 0.218 0.248 0.209 0.876 102nd Party Unity -0.241 0.090 -0.691 -2.693* N 102 R2 0.918 Adjusted R2 0.742 Constant 46.974 Year Elected 0.429 0.529 0.107 0.811 Current Term 6.275 2.568 0.317 2.444* Northeast 8.701 16.318 0.121 0.533 South 13.754 12.690 0.226 1.084 West -22.862 15.440 -0.338 -1.481 Diff DW Nom 27.490 38.660 0.071 0.711 Fed Spending 0.168 18.082 0.002 0.009 Per Cap Inc 0.000 0.001 0.055 0.359 African Amer. -87.170 35.256 -0.286 -2.473* Hispanic 36.405 46.035 0.127 0.791 Med Mal Crisis 6.661 10.734 0.092 0.621 Health 13.666 46.412 0.039 0.294 Lawyer -54.536 88.023 -0.104 -0.620 E-score -0.150 0.156 -0.143 -0.961 103rd Party Unity -0.344 0.056 -1.002 -6.143** N 102 R2 0.956 Adjusted R2 0.861 Constant 23.765 261 Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores Year Elected 0.120 0.229 0.054 0.525 Current Term -0.194 1.329 -0.015 -0.146 Northeast -4.378 6.897 -0.083 -0.635 South -8.616 6.888 -0.198 -1.251 West -10.825 6.206 -0.231 -1.744 Diff DW Nom 22.838 17.606 0.129 1.297 Fed Spending 7.081 10.950 0.089 0.647 Per Cap Inc 0.001 0.001 0.176 1.116 African Amer. -12.108 31.194 -0.056 -0.388 Hispanic -2.618 26.491 -0.010 -0.099 Med Mal Crisis 8.333 4.427 0.194 1.883 Health -116.009 42.862 -0.293 -2.707** Lawyer 38.018 32.747 0.117 1.161 E-score -0.097 0.128 -0.111 -0.762 104th Party Unity 0.020 0.036 0.085 0.541 N 103 R2 0.327 Adjusted R2 0.206 Constant 23.757 Year Elected 1.166 0.682 0.290 1.710 Current Term -3.801 4.625 -0.209 -0.822 Northeast 6.573 26.399 0.091 0.249 South 15.882 21.835 0.261 0.727 West -24.466 22.244 -0.361 -1.100 Diff DW Nom 30.144 61.915 0.081 0.487 Fed Spending 11.263 28.882 0.123 0.390 Per Cap Inc 0.000 0.002 0.058 0.247 African Amer. -99.307 67.615 -0.332 -1.469 Hispanic 30.445 54.065 0.109 0.563 Med Mal Crisis 2.184 19.803 0.030 0.110 Health 125.373 208.290 0.126 0.602 Lawyer -48.022 91.019 -0.112 -0.528 E-score -0.434 0.434 -0.255 -1.000 105th Party Unity -0.354 0.071 -1.021 -4.998** N 100 R2 0.897 Adjusted R2 0.676 Constant 34.061 262 Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores Year Elected 0.785 0.484 0.195 1.622 Current Term 6.606 2.541 0.333 2.599* Northeast 9.357 17.061 0.130 0.548 South 6.038 13.396 0.099 0.451 West -28.035 14.783 -0.414 -1.896 Diff DW Nom -15.476 69.314 -0.041 -0.223 Fed Spending 7.260 25.068 0.088 0.290 Per Cap Inc 0.000 0.001 0.028 0.110 African Amer. -54.317 51.987 -0.183 -1.045 Hispanic 15.021 36.362 0.055 0.413 Med Mal Crisis 6.245 13.536 0.087 0.461 Health 158.045 224.716 0.168 0.703 Lawyer -18.281 64.378 -0.042 -0.284 E-score -0.011 0.277 -0.009 -0.040 106th Party Unity -0.322 0.063 -0.981 -5.132** N 102 R2 0.951 Adjusted R2 0.846 Constant -6.207 Year Elected 0.362 0.206 0.073 1.755 Current Term -0.649 1.274 -0.021 -0.510 Northeast 2.725 6.957 0.021 0.392 South 6.780 6.492 0.064 1.044 West -5.795 6.192 -0.050 -0.936 Diff DW Nom 7.435 16.879 0.018 0.440 Fed Spending -1.394 9.391 -0.009 -0.148 Per Cap Inc -6.184E-05 0.001 -0.005 -0.079 African Amer. -73.315 29.359 -0.139 -2.497* Hispanic -16.521 25.989 -0.029 -0.636 Med Mal Crisis 6.168 4.505 0.059 1.369 Health 5.334 50.165 0.004 0.106 Lawyer -11.488 31.462 -0.016 -0.365 E-score -0.071 0.088 -0.038 -0.804 107th Party Unity -0.534 0.028 -0.953 -18.847** N 102 R2 0.896 Adjusted R2 0.876 Constant 58.575 263 Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores Year Elected 0.591 0.226 0.125 2.612** Current Term -4.217 1.366 -0.140 -3.087** Northeast -7.964 7.688 -0.060 -1.036 South 4.732 7.551 0.044 0.627 West -15.769 7.068 -0.138 -2.231 Diff DW Nom 11.851 17.896 0.030 0.662 Fed Spending 1.640 8.523 0.012 0.192 Per Cap Inc 0.001 0.001 0.044 0.721 African Amer. -117.649 32.873 -0.224 -3.579** Hispanic 35.644 26.156 0.066 1.363 Med Mal Crisis 10.682 5.089 0.101 2.099* Health -47.386 43.455 -0.050 -1.090 Lawyer -10.988 34.191 -0.015 -0.321 E-score 0.033 0.123 0.017 0.267 108th Party Unity -0.498 0.034 -0.921 -14.808** N 100 R2 0.874 Adjusted R2 0.849 Constant 45.038 * p < 0.05, ** p < 0.01 Party unity substitution produced statistically significant results for length of tenure in office and length of service within his or her current term. These results were not statistically significant in the Base Model. A positive correlation exists between length of time a senator has served in his or her current term in office and support for medical malpractice reform in the 103rd and 106th Senate, but the association is negative in the 108th Senate. A positive association indicates that the further into his or her current term in office the more likely the senator supports economically efficient policies. That the association is negative in the 108th Senate but a longer tenure of overall service is positive suggests differences in support for economic efficiency might result from the point which the senator has reached in his or her career. Controls for state economic conditions indicate that percent African American in a state is an indication of lower levels of support for medical malpractice reform. 264 Medical malpractice reform laws within a state are a statistically significant predictor of a senator supporting medical malpractice reform legislation for implementing federal policies. Coefficients of determination were higher after substituting party unity into the model, suggesting greater explanation for variation that exists in the model. Table 4.24 presents regression results for each Senate where statistically significant results are available after substituting ADA into the model. No statistically significant results were produced in the 102 nd Senate. Output for that Congress is not included in Table 4.25. ADA is a statistically significant measure of ideology in each remaining Congress (103 rd , 104 th , 105 th , 106 th , 107 th , and 108 th Senate) and is negatively correlated with medical malpractice reform as anticipated in Hypothesis 2a: Legislators with higher ADA scores vote in opposition to medical malpractice reform. Statistically significant values for ADA when the variable is substituted into the model offer a comparison to statistically significant output for E-score in the Senate Base Model. That E-score is not statistically significance when ADA is substituted into the model suggests a greater relative importance of liberal-conservative ideology as a predictor of behavior. 265 Table 4.24 Regression Analysis of Base Model with ADA Substitution for 103rd to 108th Senate: Medical Malpractice Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores Year Elected 0.033 0.920 0.008 0.036 Current Term 5.624 4.430 0.285 1.270 Northeast 7.543 27.137 0.096 0.278 South 4.762 21.039 0.077 0.226 West -27.072 26.639 -0.399 -1.016 Diff DW Nom 72.983 65.320 0.190 1.117 Fed Spending 11.858 31.523 0.133 0.376 Per Cap Inc 0.001 0.002 0.167 0.628 African Amer. -101.125 60.176 -0.333 -1.680 Hispanic 0.950 74.399 0.003 0.013 Med Mal Crisis 1.175 18.231 0.015 0.064 Health 30.526 86.100 0.087 0.355 Lawyer -42.044 155.798 -0.081 -0.270 E-score -0.115 0.255 -0.110 -0.452 103rd ADA -0.877 0.270 -0.971 -3.242** N 102 R2 0.897 Adjusted R2 0.641 Constant 41.744 Year Elected 0.166 0.236 0.074 0.704 Current Term -0.345 1.341 -0.027 -0.257 Northeast -4.658 6.916 -0.088 -0.673 South -8.489 6.942 -0.195 -1.223 West -11.607 6.312 -0.248 -1.839 Diff DW Nom 20.264 17.365 0.115 1.167 Fed Spending 5.418 10.970 0.068 0.494 Per Cap Inc 0.001 0.001 0.151 0.952 African Amer. -13.368 31.118 -0.062 -0.430 Hispanic -2.338 26.469 -0.009 -0.088 Med Mal Crisis 8.032 4.414 0.187 1.819 Health -121.603 41.182 -0.307 -2.953** Lawyer 32.526 33.237 0.100 0.979 E-score -0.048 0.127 -0.055 -0.381 104th ADA 0.076 0.082 0.144 0.928 N 103 R2 0.337 Adjusted R2 0.216 Constant 23.970 266 Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores Year Elected 0.959 0.883 0.238 1.086 Current Term -3.269 6.145 -0.179 -0.532 Northeast 6.111 35.727 0.085 0.171 South 12.739 28.514 0.209 0.447 West -22.531 29.158 -0.333 -0.773 Diff DW Nom 75.875 80.775 0.204 0.939 Fed Spending 14.055 38.176 0.154 0.368 Per Cap Inc 0.001 0.002 0.147 0.459 African Amer. -82.688 88.064 -0.276 -0.939 Hispanic 36.731 71.183 0.131 0.516 Med Mal Crisis 4.764 26.252 0.066 0.181 Health 172.012 272.855 0.172 0.630 Lawyer -82.292 118.315 -0.192 -0.696 E-score -0.455 0.589 -0.267 -0.772 105th ADA -0.789 0.230 -0.971 -3.428** N 100 R2 0.824 Adjusted R2 0.448 Constant 52.725 106th Year Elected 0.659 0.533 0.164 1.236 Current Term 7.189 2.781 0.363 2.585* Northeast 12.019 18.644 0.167 0.645 South 0.095 14.972 0.002 0.006 West -33.818 16.432 -0.499 -2.058 Diff DW Nom -15.825 76.132 -0.042 -0.208 Fed Spending 20.276 28.296 0.246 0.717 Per Cap Inc 0.001 0.002 0.128 0.452 African Amer. -26.440 58.024 -0.089 -0.456 Hispanic 10.215 39.760 0.038 0.257 Med Mal Crisis 5.656 14.830 0.078 0.381 Health 279.062 251.433 0.296 1.110 Lawyer -1.286 71.516 -0.003 -0.018 E-score 0.002 0.306 0.002 0.007 ADA -0.706 0.155 -1.000 -4.556** N 102 R2 0.941 Adjusted R2 0.815 Constant -10.620 267 Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores Year Elected 0.387 0.243 0.078 1.594 Current Term 0.147 1.498 0.005 0.098 Northeast 6.071 8.190 0.047 0.741 South 2.884 7.706 0.027 0.374 West -7.913 7.325 -0.068 -1.080 Diff DW Nom 10.358 19.908 0.025 0.520 Fed Spending 4.967 11.093 0.032 0.448 Per Cap Inc 9.013E-05 0.001 0.008 0.097 African Amer. -87.970 34.579 -0.167 -2.544* Hispanic -12.371 30.710 -0.021 -0.403 Med Mal Crisis 9.620 5.298 0.092 1.816 Health -6.120 59.143 -0.005 -0.103 Lawyer -21.158 37.106 -0.029 -0.570 E-score 0.002 0.103 0.001 0.017 107th ADA -1.163 0.076 -0.913 -15.272** N 102 R2 0.855 Adjusted R2 0.825 Constant 98.096 Year Elected 0.512 0.260 0.108 1.971 Current Term -3.348 1.568 -0.111 -2.136 Northeast 3.559 8.811 0.027 0.404 South 0.582 8.718 0.005 0.067 West -13.352 8.123 -0.117 -1.644 Diff DW Nom 10.918 20.600 0.028 0.530 Fed Spending 5.409 9.796 0.039 0.552 Per Cap Inc -5.731E-05 0.001 -0.005 -0.070 African Amer. -102.431 37.994 -0.195 -2.696** Hispanic 27.162 30.048 0.050 0.904 Med Mal Crisis 9.953 5.856 0.094 1.700 Health -86.887 49.661 -0.091 -1.750 Lawyer -10.741 39.436 -0.014 -0.272 E-score 0.065 0.143 0.034 0.455 108th ADA -1.171 0.097 -0.887 -12.107** N 100 R2 0.833 Adjusted R2 0.800 Constant 118.364 * p < 0.05, ** p < 0.01 Results for state economic conditions find an inverse relationship between percent African American in a state and a senators? support for medical malpractice reform in the ADA substitution. The greater the percentage of African Americans in a state the less likely a senator is to vote for economically efficient policies. Length of 268 service in current term in office is statistically significant in the 106 th Congress and indicates that senators further into their current term vote in higher numbers for medical malpractice reform. ACU is a statistically significant predictor of legislative voting when substituted into the model. Senators with higher ACU scores vote in support of medical malpractice reform in greater numbers. This positive association between ACU and an economically efficient policy is consistent with Hypothesis 3a: Legislators with higher ACU scores vote in support of medical malpractice reform. Compared to the base model substituting ACU into the model produces a much higher frequency that a senator?s time in office and point within his or her career is much more important to legislative voting records. The longer a senator has served and the longer the time since he or she was elected to the current term in office the greater the support for medical malpractice reform. The lone exception is the 108 th Congress where length of service in current term is inversely related to supporting economically efficient policies. Many of the patterns of statistical significance within the model mirror party unity substitution. This suggests a link to liberal-conservative decision-making as Republicans are unified within their party in support of medical malpractice legislation. E-score is not statistically significant in the model and raises questions about its use when a relatively strong measure of ideology, like ACU, is included in a model of legislative voting. 269 Table 4.25 Regression Analysis of Base Model with ACU Substitution for 103rd to 108th Senate: Medical Malpractice Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores Year Elected 0.408 0.542 0.101 0.753 Current Term 7.651 2.720 0.386 2.813* Northeast 7.883 16.712 0.109 0.472 South 0.615 13.161 0.010 0.047 West -32.028 16.535 -0.473 -1.937 Diff DW Nom 53.911 39.704 0.140 1.358 Fed Spending 14.661 19.403 0.166 0.756 Per Cap Inc 0.001 0.001 0.140 0.864 African Amer. -87.778 36.207 -0.288 -2.424* Hispanic 1.835 46.399 0.006 0.040 Med Mal Crisis 9.801 10.943 0.136 0.896 Health 40.132 50.651 0.114 0.792 Lawyer 19.836 98.714 0.038 0.201 E-score -0.051 0.156 -0.049 -0.324 103rd ACU 0.865 0.145 1.087 5.959** N 102 R2 0.954 Adjusted R2 0.854 Constant -62.787 Year Elected 0.071 0.227 0.032 0.314 Current Term -0.064 1.311 -0.005 -0.049 Northeast -4.365 6.740 -0.083 -0.648 South -7.577 6.815 -0.174 -1.112 West -10.608 6.099 -0.226 -1.739 Diff DW Nom 21.675 17.121 0.123 1.266 Fed Spending 4.973 10.855 0.062 0.458 Per Cap Inc 0.001 0.001 0.144 0.933 African Amer. -12.722 30.743 -0.059 -0.414 Hispanic -1.595 26.115 -0.006 -0.061 Med Mal Crisis 7.873 4.367 0.184 1.803 Health -109.260 41.062 -0.276 -2.661** Lawyer 32.143 32.205 0.099 0.998 E-score -0.001 0.124 -0.001 -0.009 104th ACU -0.130 0.079 -0.251 -1.662 N 103 R2 0.347 Adjusted R2 0.229 Constant 31.042 270 Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores Year Elected 0.959 0.911 0.238 1.052 Current Term -3.159 6.346 -0.173 -0.498 Northeast -1.382 35.699 -0.019 -0.039 South 1.083 28.621 0.018 0.038 West -18.220 29.721 -0.269 -0.613 Diff DW Nom 59.385 83.004 0.159 0.715 Fed Spending 15.536 39.512 0.170 0.393 Per Cap Inc 0.000 0.002 0.043 0.134 African Amer. -48.845 87.196 -0.163 -0.560 Hispanic 17.322 72.384 0.062 0.239 Med Mal Crisis 9.225 26.483 0.128 0.348 Health 175.764 281.221 0.176 0.625 Lawyer -80.184 121.904 -0.187 -0.658 E-score -0.306 0.586 -0.179 -0.522 105th ACU 0.773 0.237 0.884 3.266** N 100 R2 0.814 Adjusted R2 0.414 Constant -4.234 Year Elected 0.665 0.529 0.165 1.259 Current Term 7.169 2.757 0.362 2.600* Northeast 19.133 18.411 0.265 1.039 South 1.916 14.743 0.031 0.130 West -28.613 16.084 -0.423 -1.779 Diff DW Nom -27.105 76.158 -0.072 -0.356 Fed Spending 15.930 27.753 0.193 0.574 Per Cap Inc 0.000 0.002 0.058 0.212 African Amer. -22.608 57.700 -0.076 -0.392 Hispanic -0.273 39.292 -0.001 -0.007 Med Mal Crisis 3.772 14.695 0.052 0.257 Health 254.040 247.958 0.270 1.025 Lawyer -16.122 70.112 -0.037 -0.230 E-score 0.066 0.293 0.052 0.225 106th ACU 0.723 0.157 0.951 4.609** N 102 R2 0.942 Adjusted R2 0.818 Constant -66.620 271 Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores Year Elected 0.580 0.254 0.116 2.280* Current Term -0.279 1.568 -0.009 -0.178 Northeast 8.258 8.552 0.063 0.966 South 2.295 8.066 0.022 0.285 West -5.483 7.636 -0.047 -0.718 Diff DW Nom 19.016 20.830 0.046 0.913 Fed Spending -3.842 11.571 -0.025 -0.332 Per Cap Inc 0.000 0.001 -0.016 -0.197 African Amer. -77.004 36.179 -0.146 -2.128* Hispanic -21.677 32.020 -0.037 -0.677 Med Mal Crisis 9.799 5.537 0.094 1.770 Health -9.967 61.814 -0.008 -0.161 Lawyer -29.783 38.647 -0.041 -0.771 E-score 0.006 0.108 0.003 0.051 107th ACU 1.187 0.083 0.911 14.381** N 102 R2 0.842 Adjusted R2 0.812 Constant -8.270 Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores Year Elected 0.568 0.259 0.120 2.190* Current Term -4.290 1.567 -0.142 -2.738** Northeast -0.453 8.778 -0.003 -0.052 South 2.720 8.667 0.025 0.314 West -13.867 8.093 -0.121 -1.714 Diff DW Nom 28.113 20.598 0.072 1.365 Fed Spending 11.834 9.746 0.086 1.214 Per Cap Inc 0.001 0.001 0.079 1.121 African Amer. -125.372 37.652 -0.239 -3.330** Hispanic 36.083 30.042 0.067 1.201 Med Mal Crisis 13.741 5.858 0.130 2.346 Health -105.138 49.322 -0.110 -2.132 Lawyer -13.670 39.228 -0.018 -0.348 E-score 0.158 0.138 0.081 1.150 108th ACU 1.267 0.104 0.866 12.184** N 100 R2 0.834 Adjusted R2 0.802 Constant -50.727 * p < 0.05, ** p < 0.01 Substituting DW Nominate into the model produces statistically significant results for that variable in each Congress where variability exists in the model (102nd 272 through 108th Senate) with the only exception being the 102nd Senate, where DW Nominate is not statistically significant. In each Congress but the 105 th , the statistically significant results are positively correlated with the dependent variable, as anticipated from Hypothesis 4a: Legislators with higher DW Nominate scores vote in support of medical malpractice reform. In the 104 th Congress a directional change to a negative correlation indicates that those senators with higher DW Nominate scores do not support medical malpractice reform. Lower DW Nominate scores indicate senators with more liberal legislative positions support medical malpractice reform. That this directional change occurs in the 104 th Congress, when higher Health contributions are a statistically significant indicator of opposition to medical malpractice reform, the model suggests an effect between self interest and ideology is in play. Length of time in current term positively impacts senatorial support for medical malpractice reform, as a trend exists in DW Nominate substitution where those senators serving a longer tenure and who are further into their current term in office support medical malpractice reform in greater numbers. Senators from Western states vote in opposition to medical malpractice reform, while those in states with established medical malpractice laws support policies expanding those laws at the federal level. Higher levels of minorities within a states? population predict senators will not support medical malpractice reform. 273 Table 4.26 Regression Analysis of Base Model with DW Nominate Substitution for 102nd to 108th Senate: Medical Malpractice Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores Year Elected 1.601 0.677 0.386 2.364* Current Term -1.751 3.450 -0.100 -0.507 Northeast -0.416 20.982 -0.006 -0.020 South -11.079 15.220 -0.182 -0.728 West -29.726 17.464 -0.439 -1.702 Diff DW Nom 28.022 54.935 0.074 0.510 Fed Spending 12.271 24.126 0.146 0.509 Per Cap Inc -0.001 0.002 -0.083 -0.321 African Amer. -34.307 56.569 -0.112 -0.606 Hispanic 33.794 64.858 0.117 0.521 Med Mal Crisis 26.033 16.044 0.361 1.623 Health 107.883 84.974 0.205 1.270 Lawyer -236.993 132.007 -0.395 -1.795 E-score 0.275 0.264 0.264 1.043 102nd DW Nominate 54.454 23.860 0.655 2.282 N 102 R2 0.904 Adjusted R2 0.699 Constant 38.480 Year Elected 0.892 0.683 0.222 1.306 Current Term 8.746 3.410 0.442 2.565* Northeast 16.327 20.683 0.226 0.789 South -1.428 15.962 -0.023 -0.089 West -41.644 21.258 -0.615 -1.959 Diff DW Nom 42.331 47.782 0.110 0.886 Fed Spending 24.141 24.496 0.273 0.985 Per Cap Inc 0.001 0.002 0.186 0.914 African Amer. -74.017 43.060 -0.243 -1.719 Hispanic 19.142 56.407 0.067 0.339 Med Mal Crisis 9.403 13.214 0.130 0.712 Health 68.513 66.871 0.194 1.025 Lawyer 23.902 122.467 0.046 0.195 E-score -0.091 0.190 -0.087 -0.477** 103rd DW Nominate 97.358 20.647 1.191 4.715 N 102 R2 0.932 Adjusted R2 0.788 Constant -50.767 274 Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores Year Elected 0.043 0.226 0.020 0.192 Current Term -0.096 1.298 -0.007 -0.074 Northeast -3.447 6.706 -0.065 -0.514 South -6.468 6.800 -0.149 -0.951 West -9.697 6.073 -0.207 -1.597 Diff DW Nom 25.019 17.054 0.142 1.467 Fed Spending 4.674 10.728 0.059 0.436 Per Cap Inc 0.001 0.001 0.122 0.793 African Amer. -13.393 30.462 -0.062 -0.440 Hispanic -1.359 25.872 -0.005 -0.053 Med Mal Crisis 8.258 4.320 0.193 1.912 Health -108.712 40.424 -0.274 -2.689** Lawyer 28.919 31.989 0.089 0.904 E-score 0.038 0.124 0.044 0.309 104th DW Nominate -15.999 7.647 -0.321 -2.092* N 103 R2 0.359 Adjusted R2 0.243 Constant 25.214 Year Elected 1.504 0.846 0.374 1.778 Current Term -6.455 5.918 -0.354 -1.091 Northeast 17.058 33.341 0.237 0.512 South 8.879 25.411 0.146 0.349 West -35.938 27.337 -0.531 -1.315 Diff DW Nom 39.806 72.858 0.107 0.546 Fed Spending 24.093 35.079 0.264 0.687 Per Cap Inc 0.001 0.002 0.209 0.715 African Amer. -94.269 80.024 -0.315 -1.178 Hispanic 38.954 64.246 0.139 0.606 Med Mal Crisis -4.109 24.570 -0.057 -0.167 Health 194.891 246.790 0.195 0.790 Lawyer -72.927 106.905 -0.170 -0.682 E-score -0.553 0.537 -0.324 -1.029 105th DW Nominate 88.298 22.072 1.119 4.000** N 100 R2 0.857 Adjusted R2 0.550 Constant 3.997 275 Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores Year Elected 0.899 0.583 0.223 1.542 Current Term 9.208 3.102 0.465 2.968* Northeast 21.521 20.484 0.299 1.051 South -6.902 17.101 -0.113 -0.404 West -40.798 18.617 -0.602 -2.191 Diff DW Nom -43.027 86.373 -0.115 -0.498 Fed Spending 25.203 31.758 0.306 0.794 Per Cap Inc 0.001 0.002 0.223 0.688 African Amer. -10.748 65.109 -0.036 -0.165 Hispanic 19.435 44.109 0.072 0.441 Med Mal Crisis 2.415 16.346 0.033 0.148 Health 251.740 276.217 0.267 0.911 Lawyer -34.579 77.222 -0.079 -0.448 E-score -0.119 0.366 -0.094 -0.326 106th DW Nominate 88.838 22.327 1.147 3.979** N 102 R2 0.928 Adjusted R2 0.775 Constant -61.188 Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores Year Elected 0.485 0.288 0.097 1.683 Current Term -0.679 1.781 -0.022 -0.381 Northeast 3.646 9.718 0.028 0.375 South -4.084 9.280 -0.038 -0.440 West -10.965 8.728 -0.094 -1.256 Diff DW Nom 15.104 23.586 0.037 0.640 Fed Spending -1.031 13.114 -0.007 -0.079 Per Cap Inc -0.001 0.001 -0.049 -0.523 African Amer. -71.858 41.023 -0.136 -1.752 Hispanic -30.870 36.214 -0.053 -0.852 Med Mal Crisis 9.484 6.275 0.091 1.511 Health -40.906 70.049 -0.032 -0.584 Lawyer -38.494 43.863 -0.053 -0.878 E-score 0.092 0.121 0.049 0.764 107th DW Nominate 93.971 7.839 0.853 11.988** N 102 R2 0.797 Adjusted R2 0.758 Constant 66.322 276 Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores Year Elected 0.622 0.283 0.131 2.202* Current Term -4.843 1.709 -0.160 -2.834** Northeast -9.639 9.599 -0.073 -1.004 South -14.949 9.679 -0.140 -1.544 West -24.016 8.932 -0.210 -2.689 Diff DW Nom 27.440 22.371 0.070 1.227 Fed Spending 12.440 10.583 0.090 1.175 Per Cap Inc 0.001 0.001 0.053 0.699 African Amer. -99.430 41.154 -0.190 -2.416 Hispanic 42.501 32.787 0.078 1.296 Med Mal Crisis 16.601 6.402 0.157 2.593** Health -122.151 53.473 -0.128 -2.284 Lawyer -15.044 42.639 -0.020 -0.353 E-score 0.063 0.159 0.033 0.398 108th DW Nominate 101.312 9.480 0.896 10.686** N 100 R2 0.804 Adjusted R2 0.766 Constant 35.504 * p < 0.05, ** p < 0.01 E-score, a component of ideology, is statistically significant in most Congresses in the Base Model but is not when a relatively strong ideology variable ? DW Nominate ? is added. This suggests that the effect of liberal-conservative ideology is strong not only in nominal terms through ADA and ACU measures but also with weighted measures that capture a time element to ideology. Legislator Party was substituted into the model to measure if an effect exists on legislative voting surrounding political party control of the Senate. When testing Hypothesis 10a: Legislators from the minority party (Senate) are more likely than majority party legislators to support medical malpractice reform, legislator party is statistically significant in each Congress where variability exists in the model (102 nd through 108 th Congress) with the exception of the 104 th Congress. 277 Table 4.27 Regression Analysis of Base Model with Legislator Party Substitution for 102nd to 108th Senate: Medical Malpractice Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores Year Elected 1.248 0.617 0.301 2.024 Current Term -0.763 3.158 -0.044 -0.242 Northeast -6.852 19.300 -0.095 -0.355 South 0.219 14.741 0.004 0.015 West -21.412 16.152 -0.316 -1.326 Diff DW Nom 11.117 50.232 0.029 0.221 Fed Spending 4.718 22.112 0.056 0.213 Per Cap Inc -0.001 0.002 -0.127 -0.561 African Amer. -53.898 54.092 -0.175 -0.996 Hispanic 24.901 60.122 0.086 0.414 Med Mal Crisis 21.925 15.221 0.304 1.440 Health 63.272 82.654 0.120 0.766 Lawyer -136.262 133.492 -0.227 -1.021 E-score 0.238 0.242 0.229 0.985 102nd Legislator Party -40.525 15.000 -0.676 -2.702* N 102 R2 0.918 Adjusted R2 0.743 Constant 71.277 Year Elected 0.394 0.511 0.098 0.770 Current Term 5.680 2.457 0.287 2.311 Northeast 5.923 15.714 0.082 0.377 South 16.654 12.313 0.273 1.353 West -17.510 14.629 -0.259 -1.197 Diff DW Nom 16.494 37.573 0.043 0.439 Fed Spending -7.299 17.211 -0.083 -0.424 Per Cap Inc 0.000 0.001 -0.020 -0.138 African Amer. -85.157 34.057 -0.280 -2.500* Hispanic 42.297 44.788 0.148 0.944 Med Mal Crisis 8.409 10.360 0.117 0.812 Health -5.040 43.066 -0.014 -0.117 Lawyer -88.644 81.900 -0.170 -1.082 E-score -0.152 0.151 -0.146 -1.008 103rd Legislator Party -56.087 8.797 -0.935 -6.376** N 102 R2 0.959 Adjusted R2 0.870 Constant 77.859 278 Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores Year Elected 0.129 0.229 0.058 0.563 Current Term -0.181 1.332 -0.014 -0.136 Northeast -4.884 6.938 -0.093 -0.704 South -8.580 6.904 -0.197 -1.243 West -11.128 6.209 -0.238 -1.792 Diff DW Nom 21.301 17.740 0.121 1.201 Fed Spending 7.715 10.958 0.097 0.704 Per Cap Inc 0.001 0.001 0.196 1.240 African Amer. -12.180 31.254 -0.056 -0.390 Hispanic -2.779 26.538 -0.011 -0.105 Med Mal Crisis 8.264 4.435 0.193 1.863 Health -123.066 43.106 -0.310 -2.855** Lawyer 41.209 32.737 0.127 1.259 E-score -0.146 0.126 -0.168 -1.164 104th Legislator Party -0.021 6.349 -0.001 -0.003 N 103 R2 0.325 Adjusted R2 0.203 Constant 23.501 Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores Year Elected 1.110 0.649 0.276 1.710 Current Term -2.745 4.322 -0.151 -0.635 Northeast 1.238 24.659 0.017 0.050 South 17.870 21.017 0.293 0.850 West -21.219 21.140 -0.313 -1.004 Diff DW Nom 14.777 59.652 0.040 0.248 Fed Spending 5.290 27.431 0.058 0.193 Per Cap Inc 2.045E-05 0.002 0.003 0.014 African Amer. -105.965 65.213 -0.354 -1.625 Hispanic 27.609 51.722 0.099 0.534 Med Mal Crisis 5.357 18.634 0.074 0.288 Health 57.006 199.658 0.057 0.286 Lawyer -26.912 87.714 -0.063 -0.307 E-score -0.334 0.406 -0.196 -0.823 105th Legislator Party 60.384 11.451 1.007 5.274** N 100 R2 0.905 Adjusted R2 0.703 Constant 15.083 279 Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores Year Elected 0.872 0.491 0.217 1.776 Current Term 6.468 2.588 0.327 2.499 Northeast 8.249 17.390 0.114 0.474 South 8.725 13.555 0.143 0.644 West -27.156 15.022 -0.401 -1.808 Diff DW Nom -18.996 70.752 -0.051 -0.268 Fed Spending 5.808 25.455 0.071 0.228 Per Cap Inc -1.601E-05 0.001 -0.003 -0.011 African Amer. -65.574 52.700 -0.221 -1.244 Hispanic 16.995 37.050 0.063 0.459 Med Mal Crisis 6.981 13.782 0.097 0.507 Health 134.724 228.124 0.143 0.591 Lawyer -15.819 65.631 -0.036 -0.241 E-score 0.037 0.276 0.029 0.133 106th Legislator Party 57.026 11.363 0.951 5.019** N 102 R2 0.949 Adjusted R2 0.840 Constant -32.183 Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores Year Elected 0.287 0.207 0.058 1.385 Current Term -1.007 1.283 -0.032 -0.785 Northeast -0.418 7.017 -0.003 -0.060 South 9.540 6.512 0.090 1.465 West -3.515 6.217 -0.030 -0.565 Diff DW Nom 8.633 16.982 0.021 0.508 Fed Spending -4.307 9.445 -0.028 -0.456 Per Cap Inc 0.000 0.001 -0.028 -0.408 African Amer. -64.400 29.570 -0.122 -2.178* Hispanic -26.161 26.085 -0.045 -1.003 Med Mal Crisis 4.661 4.541 0.045 1.026 Health 25.180 50.506 0.020 0.499 Lawyer -26.862 31.358 -0.037 -0.857 E-score 0-.044 0.088 -0.024 -0.500 107th Legislator Party -92.442 4.941 -0.926 -18.708** N 102 R2 0.894 Adjusted R2 0.874 Constant 116.442 280 Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores Year Elected 0.555 0.223 0.117 2.487* Current Term -4.163 1.349 -0.138 -3.087** Northeast -11.085 7.616 -0.084 -1.456 South 5.603 7.455 0.052 0.752 West -15.359 6.978 -0.134 -2.201* Diff DW Nom 12.326 17.677 0.032 0.697 Fed Spending 0.608 8.424 0.004 0.072 Per Cap Inc 0.000 0.001 0.036 0.607 African Amer. -113.205 32.495 -0.216 -3.484** Hispanic 28.482 25.759 0.053 1.106 Med Mal Crisis 9.280 5.021 0.088 1.848 Health -21.885 43.216 -0.023 -0.506 Lawyer -11.962 33.756 -0.016 -0.354 E-score 0.078 0.119 0.040 0.652 108th Legislator Party 90.156 5.988 0.899 15.057** N 100 R2 0.877 Adjusted R2 0.853 Constant 1.144 * p < 0.05, ** p < 0.01 Directional impact of the variable must be considered with party control of the Senate within each Congress. Referring to party divisions summarized in Table 4.1, Democrats controlled the Senate during in the 102 nd and 103 rd Congresses and Republicans controlled the Senate from the 104 th through the 106 th , and in the 108 th Congress. Control varied in the 107 th Congress. Considering directional movements of the variable, Legislator Party, finds that when Democrats are in the majority standardized coefficients are negative, indicating greater support for medical malpractice reform from Republicans, the party that does not control the Senate. Positive, statistically significant, directional impact of the variable in the 105 th through 108 th Congresses indicates that Republicans vote increasingly for medical malpractice reform regardless of party control. 281 Health is statistically significant and is negatively related to support for medical malpractice reform. Higher health contributions reduce support and are the only statistically significant in the 104 th Congress for Legislator Party substitution as well as the Base Model and each of the other substitutions, with the exception of DW Nominate substitution. That Health is statistically significant in the Congress that produced sweeping Republican control of both House and Senate shows a strong self-interest component affecting legislative voting when ideology and party control are not statistically significant in the model. Percent African American within a state consistently predicts reduced support for medical malpractice reform by senators from those states. In most Congresses coefficients of determination indicate that legislator party substitution explains the most variability in the model in comparison to the base model and each other substitution. Minimum wage Regression analysis for independent and control variables for the Base Model in the Senate are summarized in Table 4.28 for minimum wage as the dependent variable. E-score is statistically significant in the 100 th , 101 st , and 108 th Congresses and is a positive predictor of opposition to increasing the minimum wage. These results are consistent with expected results from Hypothesis 1b: Legislators with higher E-scores vote in opposition to increasing the minimum wage. 282 Table 4.28 Regression Analysis of Base Model for 100th to 108th Senate: Minimum Wage Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores Year Elected -0.162 0.567 -0.026 -0.285 Current Term 0.214 2.512 0.008 0.085 Northeast 0.908 14.864 0.008 0.061 South -20.504 14.693 -0.208 -1.395 West -8.080 13.787 -0.078 -0.586 Diff DW Nom -47.136 39.807 -0.116 -1.184 Fed Spending 35.490 19.283 0.237 1.840 Per Cap Inc 0.002 0.002 0.132 0.924 African Amer. 69.515 81.593 0.114 0.852 Hispanic -167.264 59.866 -0.336 -2.794** Minimum Wage -14.644 10.793 -0.163 -1.357 Business 10.912 14.036 0.086 0.777 Labor -16.326 15.515 -0.105 -1.052 100th E-score 1.001 0.164 0.657 6.114** N 101 R2 0.630 Adjusted R2 0.524 Constant -53.058 Year Elected 0.736 0.701 0.155 1.049 Current Term -0.954 2.890 -0.045 -0.330 Northeast -4.365 16.452 -0.044 -0.265 South 18.409 16.025 0.251 1.149 West 5.814 17.328 0.071 0.336 Diff DW Nom 79.812 40.997 0.258 1.947 Fed Spending 14.623 25.670 0.108 0.570 Per Cap Inc 0.003 0.002 0.270 1.386 African Amer. -47.180 75.199 -0.132 -0.627 Hispanic -11.310 65.213 -0.021 -0.173 Minimum Wage -12.181 11.037 -0.170 -1.104 Business -31.507 15.388 -0.302 -2.047* Labor -80.069 18.726 -0.599 -4.276** 101st E-score 0.573 0.233 0.387 2.461* N 101 R2 0.525 Adjusted R2 0.346 Constant -65.189 283 Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores Year Elected 1.095 1.091 0.189 1.003 Current Term 0.155 5.166 0.006 0.030 Northeast 5.714 27.190 0.057 0.210 South -2.621 22.093 -0.031 -0.119 West -3.904 25.380 -0.041 -0.154 Diff DW Nom -5.612 87.698 -0.011 -0.064 Fed Spending 19.196 35.636 0.163 0.539 Per Cap Inc -0.001 0.003 -0.106 -0.443 African Amer. -72.702 88.723 -0.169 -0.819 Hispanic -171.664 98.579 -0.426 -1.741 Minimum Wage -55.289 18.410 -0.664 -3.003* Business -43.494 92.445 -0.096 -0.470 Labor -397.108 219.548 -0.361 -1.809 102nd E-score 0.469 0.276 0.322 1.696 N 102 R2 0.879 Adjusted R2 0.667 Constant 65.155 Year Elected 0.239 1.270 0.043 0.188 Current Term 5.121 9.367 0.185 0.547 Northeast -15.713 30.452 -0.156 -0.516 South -5.622 26.015 -0.066 -0.216 West 23.832 32.616 0.252 0.731 Diff DW Nom 104.364 114.534 0.194 0.911 Fed Spending 37.183 50.344 0.301 0.739 Per Cap Inc 0.003 0.004 0.263 0.666 African Amer. -43.735 96.835 -0.103 -0.452 Hispanic -253.029 116.493 -0.631 -2.172 Minimum Wage -61.812 17.602 -0.743 -3.512** Business -169.516 107.053 -0.340 -1.583 Labor -323.734 398.686 -0.307 -0.812 103rd E-score 0.329 0.398 0.225 0.827 N 102 R2 0.824 Adjusted R2 0.515 Constant -40.875 284 Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores Year Elected -0.733 1.508 -0.130 -0.486 Current Term -1.999 8.127 -0.078 -0.246 Northeast -27.479 32.050 -0.273 -0.857 South 3.784 35.435 0.044 0.107 West 38.566 36.448 0.407 1.058 Diff DW Nom 181.560 141.044 0.349 1.287 Fed Spending -28.992 64.415 -0.227 -0.450 Per Cap Inc 0.000 0.004 0.040 0.101 African Amer. -141.639 152.110 -0.339 -0.931 Hispanic -162.013 116.008 -0.414 -1.397 Minimum Wage -73.969 24.567 -0.889 -3.011* Business -141.102 111.501 -0.331 -1.265 Labor 198.439 428.297 0.208 0.463 105th E-score 0.518 0.936 0.217 0.553 N 100 R2 0.766 Adjusted R2 0.357 Constant 123.620 Year Elected 0.123 1.051 0.022 0.117 Current Term -2.293 7.263 -0.083 -0.316 Northeast -9.786 29.635 -0.097 -0.330 South 18.059 27.326 0.212 0.661 West 25.770 31.732 0.272 0.812 Diff DW Nom 49.385 108.012 0.094 0.457 Fed Spending -22.326 47.015 -0.194 -0.475 Per Cap Inc 3.983E-05 0.003 0.005 0.016 African Amer. -60.020 92.334 -0.145 -0.650 Hispanic -127.368 92.550 -0.335 -1.376 Minimum Wage -48.458 20.976 -0.582 -2.310* Business -75.197 82.004 -0.217 -0.917 Labor 136.225 240.455 0.149 0.567 106th E-score 0.950 0.448 0.533 2.119 N 102 R2 0.839 Adjusted R2 0.557 Constant 45.318 285 Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores Year Elected 0.330 1.086 0.059 0.304 Current Term 3.222 4.552 0.132 0.708 Northeast -22.900 28.151 -0.227 -0.813 South -4.358 24.299 -0.051 -0.179 West 13.599 28.602 0.144 0.475 Diff DW Nom 81.658 106.632 0.156 0.766 Fed Spending 33.739 34.288 0.306 0.984 Per Cap Inc 0.002 0.002 0.326 1.068 African Amer. -105.479 95.001 -0.255 -1.110 Hispanic -230.576 85.202 -0.626 -2.706* Minimum Wage -81.089 19.739 -0.974 -4.108** Business -116.250 72.773 -0.329 -1.597 Labor -190.477 307.438 -0.147 -0.620 107th E-score 0.386 0.337 0.279 1.144 N 102 R2 0.836 Adjusted R2 0.550 Constant -35.667 Year Elected 1.564 1.106 0.278 1.415 Current Term 10.701 5.999 0.420 1.784 Northeast 1.264 23.243 0.013 0.054 South 37.881 25.906 0.445 1.462 West 27.807 30.645 0.294 0.907 Diff DW Nom -45.607 82.961 -0.087 -0.550 Fed Spending 8.642 26.442 0.087 0.327 Per Cap Inc 0.000 0.002 -0.029 -0.123 African Amer. -52.311 94.422 -0.127 -0.554 Hispanic -129.469 71.618 -0.350 -1.808 Minimum Wage -34.528 17.737 -0.415 -1.947 Business -112.304 66.913 -0.311 -1.678 Labor -143.254 192.799 -0.140 -0.743 108th E-score 1.264 0.432 0.875 2.926* N 100 R2 0.894 Adjusted R2 0.709 Constant -110.966 * p < 0.05, ** p < 0.01 In testing self-interest variables ? Business and Labor ? both are statistically significant in the 101 st Congress and each is negatively correlated with opposition to minimum wage increases. With the minimum wage dependent variable coded to reflect higher levels of economic efficiency, a vote in support of economic efficiency is a vote against increasing the minimum wage, and a vote that is economically inefficient is a 286 vote in support of increasing the minimum wage. An anticipated correlation between higher business contributions and opposition to minimum wage legislation was expressed in Hypothesis 5b: Legislators with higher business political contributions to total contributions vote in opposition to increasing the minimum wage. Senators receiving greater business contributions were found to support increasing the minimum wage. The relationship between Labor and support for minimum wage increases is expressed in Hypothesis 6b: Legislators with higher labor political contributions to total contributions vote in support of increasing the minimum wage. Directional movement between higher labor contributions and support for increasing the minimum wage was as hypothesized. Controls for Hispanic were statistically significant in the 100 th and 107 th Congresses. Higher levels of Hispanics in a state?s population are positively associated with a senator voting to increase the federal minimum wage. This directional impact differs from most Congresses in the House for either dependent variable, where the association between higher levels of minorities within the population was positively associated with support for economically efficient legislation. From the 102 nd to 106 th Congresses, inclusive, senators from states with minimum wage laws were found to support federal minimum wage legislation. No other independent or control variable was statistically significant during those Congresses. While support for minimum wage legislation is economically inefficient, support for federal minimum wage legislation when a state has a minimum wage law indicates a senator is weighing the relative 287 economic inefficiency of such vote to economic realties associated with capital flows to states with lower minimum wage laws. No other variables were statistically significant in the base model. Without statistical significance for current term, length of service and ideological differences between legislator and median party, Hypothesis 8b: The closer senators are to the end of their current term in office, the less likely they are to support increasing the minimum wage, Hypothesis 9b: The longer a legislator has served, the less likely he or she will support increasing the minimum wage, and Hypothesis 11b: The greater the division between the ideology of the legislator and the median ideology of the legislator?s party, the less likely the legislator supports increasing the federal minimum wage, respectively, are not supported. Coefficients of determination for the base model in the Senate are generally higher overall when applied to the minimum wage rather than medical malpractice. Like the analysis of the other models in the House and the medical malpractice model in the Senate, coefficients of determination are weaker in the base model relative to the substituted variables. This model is no exception. A weaker coefficient of determination implies the model does not explain variance as well as those models with higher coefficients of determination. When party unity is substituted into the base model, Party unity is statistically significant in each Congress in the model and predicts that Republican senators are more likely to oppose increasing the minimum wage, an economically inefficient position. This is consistent with the association expected in Hypothesis 7b: Republican legislators are less likely to vote for increasing the minimum wage more often than Democrats. The 288 negative sign of the standardized coefficient is based on coding in order to discern Republican from Democrat. Table 4.29 Regression Analysis of Base Model and Party Unity Substitution for 100th to 108th Senate: Minimum Wage Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores Year Elected -0.277 0.510 -0.045 -0.544 Current Term -1.530 2.305 -0.058 -0.664 Northeast -3.292 13.386 -0.029 -0.246 South -18.337 13.195 -0.186 -1.390 West -3.167 12.444 -0.031 -0.255 Diff DW Nom -71.007 36.325 -0.175 -1.955 Fed Spending 35.134 17.300 0.234 2.031* Per Cap Inc 0.001 0.002 0.084 0.650 African Amer. 71.481 73.200 0.117 0.977 Hispanic -150.482 53.909 -0.302 -2.791 Minimum Wage -11.467 9.723 -0.127 -1.179 Business 10.754 12.592 0.085 0.854 Labor 5.544 15.194 0.036 0.365 E-score 0.393 0.224 0.258 1.751 100th Party Unity -0.293 0.082 -0.524 -3.589** N 101 R2 0.708 Adjusted R2 0.617 Constant -6.581 Year Elected -0.103 0.433 -0.022 -0.238 Current Term 0.746 1.746 0.035 0.427 Northeast -9.651 9.888 -0.098 -0.976 South 9.545 9.672 0.130 0.987 West 5.984 10.393 0.073 0.576 Diff DW Nom 3.869 26.284 0.013 0.147 Fed Spending -6.109 15.603 -0.045 -0.392 Per Cap Inc 0.002 0.002 0.124 1.047 African Amer. 30.510 46.092 0.085 0.662 Hispanic -4.276 39.122 -0.008 -0.109 Minimum Wage -8.619 6.634 -0.120 -1.299 Business -9.387 9.618 -0.090 -0.976 Labor -11.635 14.007 -0.087 -0.831 E-score 0.236 0.146 0.159 1.620 101st Party Unity -0.355 0.043 -0.782 -8.177** N 101 R2 0.834 Adjusted R2 0.765 Constant -0.925 289 Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores Year Elected 0.619 0.666 0.107 0.930 Current Term -0.084 3.102 -0.003 -0.027 Northeast -4.217 16.519 -0.042 -0.255 South 5.788 13.436 0.068 0.431 West 2.963 15.336 0.031 0.193 Diff DW Nom -17.294 52.728 -0.033 -0.328 Fed Spending 6.052 21.655 0.051 0.279 Per Cap Inc 0.000 0.002 0.030 0.200 African Amer. -51.876 53.525 -0.121 -0.969 Hispanic -85.870 63.133 -0.213 -1.360 Minimum Wage -26.501 13.291 -0.318 -1.994 Business 1.597 56.684 0.004 0.028 Labor -180.628 143.006 -0.164 -1.263 E-score -0.085 0.218 -0.058 -0.388 102nd Party Unity -0.361 0.093 -0.740 -3.899** N 102 R2 0.962 Adjusted R2 0.880 Constant 43.852 Year Elected 0.065 0.611 0.012 0.107 Current Term 1.658 4.547 0.060 0.365 Northeast -8.396 14.693 -0.083 -0.571 South 7.923 12.758 0.093 0.621 West 13.058 15.800 0.138 0.826 Diff DW Nom -14.123 59.446 -0.026 -0.238 Fed Spending 7.977 24.810 0.065 0.321 Per Cap Inc 0.002 0.002 0.183 0.963 African Amer. -10.159 46.948 -0.024 -0.216 Hispanic -67.511 66.140 -0.168 -1.021 Minimum Wage -23.397 11.172 -0.281 -2.094 Business -75.986 54.407 -0.152 -1.397 Labor -176.304 193.539 -0.167 -0.911 E-score -0.007 0.201 -0.005 -0.034 103rd Party Unity -0.339 0.064 -0.707 -5.261** N 102 R2 0.964 Adjusted R2 0.888 Constant 10.776 290 Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores Year Elected 0.266 0.551 0.047 0.484 Current Term 2.116 3.715 0.087 0.570 Northeast -3.759 15.226 -0.037 -0.247 South 11.995 17.766 0.141 0.675 West 13.192 18.523 0.139 0.712 Diff DW Nom -43.512 62.653 -0.082 -0.694 Fed Spending -7.905 23.864 -0.061 -0.331 Per Cap Inc 0.001 0.002 0.056 0.342 African Amer. -39.464 55.807 -0.094 -0.707 Hispanic -48.478 65.934 -0.122 -0.735 Minimum Wage -25.492 14.732 -0.306 -1.730 Business -32.558 56.737 -0.080 -0.574 Labor -36.055 135.754 -0.040 -0.266 E-score 0.121 0.419 0.057 0.289 104th Party Unity -0.350 0.082 -0.745 -4.261** N 103 R2 0.958 Adjusted R2 0.869 Constant 36.233 Year Elected -0.028 0.661 -0.005 -0.043 Current Term -1.845 3.503 -0.072 -0.527 Northeast -2.706 14.418 -0.027 -0.188 South 6.839 15.283 0.080 0.448 West 9.550 16.437 0.101 0.581 Diff DW Nom -38.771 71.012 -0.074 -0.546 Fed Spending -21.140 27.797 -0.166 -0.761 Per Cap Inc -0.001 0.002 -0.064 -0.374 African Amer. -40.215 67.709 -0.096 -0.594 Hispanic -24.019 55.034 -0.061 -0.436 Minimum Wage -20.942 13.789 -0.252 -1.519 Business -23.581 51.895 -0.055 -0.454 Labor 160.025 184.730 0.168 0.866 E-score 0.123 0.409 0.052 0.302 105th Party Unity -0.401 0.067 -0.827 -6.005** N 100 R2 0.962 Adjusted R2 0.880 Constant 100.356 291 Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores Year Elected -0.033 0.546 -0.006 -0.060 Current Term -0.845 3.782 -0.031 -0.223 Northeast -18.413 15.489 -0.183 -1.189 South 6.389 14.394 0.075 0.444 West 8.916 16.847 0.094 0.529 Diff DW Nom -25.556 58.231 -0.049 -0.439 Fed Spending 1.565 24.914 0.014 0.063 Per Cap Inc 0.001 0.001 0.112 0.677 African Amer. -17.266 48.761 -0.042 -0.354 Hispanic -32.683 51.990 -0.086 -0.629 Minimum Wage -19.150 12.506 -0.230 -1.531 Business 16.089 46.680 0.046 0.345 Labor -90.771 133.600 -0.099 -0.679 E-score -0.397 0.366 -0.223 -1.085 106th Party Unity -0.437 0.092 -0.951 -4.764** N 102 R2 0.962 Adjusted R2 0.881 Constant 58.383 Year Elected 0.097 0.547 0.017 0.177 Current Term 1.898 2.297 0.078 0.826 Northeast -9.995 14.346 -0.099 -0.697 South 10.546 12.541 0.124 0.841 West 13.709 14.336 0.145 0.956 Diff DW Nom -19.184 57.148 -0.037 -0.336 Fed Spending -5.617 18.913 -0.051 -0.297 Per Cap Inc 0.001 0.001 0.128 0.813 African Amer. -34.985 49.673 -0.084 -0.704 Hispanic -51.783 55.772 -0.141 -0.928 Minimum Wage -27.896 14.553 -0.335 -1.917 Business -38.611 39.662 -0.109 -0.973 Labor -40.422 157.010 -0.031 -0.257 E-score -0.101 0.195 -0.073 -0.517 107th Party Unity -0.386 0.077 -0.830 -4.984** N 102 R2 0.964 Adjusted R2 0.887 Constant 42.033 292 Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores Year Elected 0.671 0.684 0.119 0.981 Current Term 5.955 3.701 0.234 1.609 Northeast -10.778 13.924 -0.107 -0.774 South 25.756 15.453 0.302 1.667 West 24.605 17.951 0.260 1.371 Diff DW Nom -51.317 48.568 -0.098 -1.057 Fed Spending -8.727 16.058 -0.088 -0.543 Per Cap Inc 9.996E-05 0.001 0.014 0.098 African Amer. -16.296 55.967 -0.040 -0.291 Hispanic -59.372 45.351 -0.161 -1.309 Minimum Wage -23.289 10.745 -0.280 -2.167 Business -59.569 41.270 -0.165 -1.443 Labor -61.711 114.610 -0.060 -0.538 E-score 0.361 0.337 0.250 1.069 108th Party Unity -0.284 0.070 -0.626 -4.045** N 100 R2 0.968 Adjusted R2 0.900 Constant -0.943 * p < 0.05, ** p < 0.01 E-score is no longer statistically significant when party unity is substituted and statistically significant controls also disappear for state minimum wage laws. Senators from states with a minimum wage higher than the current federal minimum wage were more likely to support passage of federal legislation in the base model. With party unity added to the model the effect of party is obvious as Republicans are unified around this legislation. Standardized coefficients are relatively strong for party unity in all Congresses in the model, but have somewhat weaker impacts on average than standardized coefficients for ADA and ACU. A statistically significant relationship exists in the 99th Congress between federal spending ratios in a state and support for a minimum wage increase. Senators from states with higher levels of federal spending going into their coffers measured in relation to tax revenue remitted to the federal government are more likely to oppose increasing the 293 minimum wage. Coefficients of determination indicate a relatively good fit for the model in explaining variance along the regression line. Substituting ADA into the model (Table 4.30) finds a strong basis for liberal- conservative ideology exists. In testing Hypothesis 2b: Legislators with higher ADA scores vote in support of increasing the federal minimum wage, ADA is statistically significant at p < 0.01 in each Congress and a positive predictor of support for increasing the minimum wage as expected. Perhaps more importantly, standardized coefficients show the per unit effect of applying the ADA independent variable to the dependent variable is relatively stronger for ADA than other variables in the model, including ACU. Impact of the variable is important because not only is the variable statistically significant and as a result less likely to be the result of chance, but it is also responsible for greater movement in the dependent variable on a per unit basis. ADA is the only statistically significant variable in the model for the 101 st through 108 th Congresses. In the 100 th Congress controls for state economic factors are statistically significant. Higher percentages of Hispanics in a population are positively related to support for a minimum wage increase. This effect has not changed from the base model. 294 Table 4.30 Regression Analysis of Base Model and ADA Substitution for 99th to 108th Senate: Minimum Wage Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores Year Elected -0.080 0.501 -0.013 -0.160 Current Term -0.842 2.285 -0.032 -0.369 Northeast 4.390 13.522 0.038 0.325 South -24.900 13.206 -0.253 -1.885 West -9.828 12.360 -0.095 -0.795 Diff DW Nom -49.571 36.635 -0.118 -1.353 Fed Spending 43.899 17.152 0.293 2.559* Per Cap Inc 0.002 0.002 0.120 0.946 African Amer. 44.405 73.137 0.073 0.607 Hispanic -172.515 52.935 -0.348 -3.259** Minimum Wage -4.788 9.901 -0.053 -0.484 Business 8.698 12.417 0.068 0.700 Labor -0.812 14.486 -0.005 -0.056 E-score 0.490 0.202 0.317 2.424* 100th ADA -0.719 0.181 -0.525 -3.966** N 101 R2 0.722 Adjusted R2 0.633 Constant 2.324 Year Elected 0.474 0.476 0.100 0.996 Current Term 1.048 1.976 0.049 0.531 Northeast -9.149 11.142 -0.092 -0.821 South 7.179 10.959 0.098 0.655 West -2.880 11.783 -0.035 -0.244 Diff DW Nom -1.864 30.265 -0.006 -0.062 Fed Spending -6.790 17.640 -0.050 -0.385 Per Cap Inc 0.002 0.002 0.129 0.968 African Amer. -54.660 50.836 -0.153 -1.075 Hispanic -12.979 44.075 -0.025 -0.294 Minimum Wage 4.004 7.840 0.056 0.511 Business -7.172 11.015 -0.069 -0.651 Labor -16.591 15.802 -0.124 -1.050 E-score 0.260 0.164 0.176 1.586 101st ADA -0.904 0.135 -0.806 -6.708** N 101 R2 0.789 Adjusted R2 0.701 Constant 33.214 295 Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores Year Elected 0.180 0.779 0.031 0.231 Current Term 1.018 3.457 0.042 0.294 Northeast -3.296 18.344 -0.033 -0.180 South 1.135 14.784 0.013 0.077 West -14.597 17.239 -0.154 -0.847 Diff DW Nom -22.026 58.722 -0.042 -0.375 Fed Spending 9.595 23.952 0.081 0.401 Per Cap Inc 0.001 0.002 0.068 0.405 African Amer. -67.576 59.216 -0.157 -1.141 Hispanic -44.809 76.110 -0.111 -0.589 Minimum Wage 1.847 21.176 0.022 0.087 Business -16.839 62.202 -0.037 -0.271 Labor -182.129 160.217 -0.165 -1.137 E-score -0.316 0.300 -0.217 -1.053 102nd ADA -1.245 0.376 -1.121 -3.312** N 102 R2 0.953 Adjusted R2 0.852 Constant 98.161 Year Elected -0.164 0.942 -0.028 -0.174 Current Term -6.619 8.354 -0.243 -0.792 Northeast -20.982 20.912 -0.194 -1.003 South -3.838 15.845 -0.045 -0.242 West 7.594 20.548 0.081 0.370 Diff DW Nom -27.520 77.304 -0.052 -0.356 Fed Spending -6.630 35.752 -0.054 -0.185 Per Cap Inc 0.000 0.004 -0.009 -0.030 African Amer. 9.841 60.435 0.023 0.163 Hispanic -29.767 90.810 -0.076 -0.328 Minimum Wage 3.478 19.790 0.042 0.176 Business 22.458 100.232 0.046 0.224 Labor 95.124 324.125 0.091 0.293 E-score 0.075 0.267 0.052 0.281 103rd ADA -1.190 0.305 -0.954 -3.899** N 102 R2 0.949 Adjusted R2 0.823 Constant 132.102 296 Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores Year Elected 1.020 0.506 0.181 2.015 Current Term 6.094 3.317 0.249 1.837 Northeast 7.399 12.745 0.073 0.581 South 3.331 14.927 0.039 0.223 West 10.651 15.324 0.112 0.695 Diff DW Nom -37.923 52.128 -0.071 -0.728 Fed Spending -19.179 20.450 -0.149 -0.938 Per Cap Inc 0.000 0.001 -0.013 -0.094 African Amer. -68.732 45.818 -0.163 -1.500 Hispanic -43.378 54.905 -0.109 -0.790 Minimum Wage -10.067 13.357 -0.121 -0.754 Business -10.040 47.670 -0.025 -0.211 Labor 54.543 112.404 0.060 0.485 E-score -0.223 0.384 -0.105 -0.581 104th ADA -1.224 0.226 -1.106 -5.417** N 103 R2 0.971 Adjusted R2 0.909 Constant 102.990 Year Elected -0.460 0.495 -0.082 -0.929 Current Term -0.828 2.663 -0.032 -0.311 Northeast 2.400 11.098 0.024 0.216 South 2.671 11.595 0.031 0.230 West 8.350 12.478 0.088 0.669 Diff DW Nom -13.588 51.886 -0.026 -0.262 Fed Spending -20.657 21.101 -0.162 -0.979 Per Cap Inc 0.000 0.001 0.012 0.093 African Amer. -18.050 51.987 -0.043 -0.347 Hispanic 15.351 43.650 0.039 0.352 Minimum Wage -9.307 11.241 -0.112 -0.828 Business -40.907 38.461 -0.096 -1.064 Labor 195.701 140.141 0.205 1.396 E-score 0.146 0.310 0.061 0.471 105th ADA -1.099 0.134 -0.968 -8.229** N 100 R2 0.978 Adjusted R2 0.931 Constant 131.602 297 Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores Year Elected -0.478 0.375 -0.085 -1.275 Current Term -0.531 2.545 -0.019 -0.209 Northeast -13.088 10.350 -0.130 -1.265 South 4.106 9.708 0.048 0.423 West 8.269 11.306 0.087 0.731 Diff DW Nom -2.014 38.283 -0.004 -0.053 Fed Spending 2.535 16.723 0.022 0.152 Per Cap Inc 0.001 0.001 0.110 0.987 African Amer. 3.145 33.257 0.008 0.095 Hispanic -19.656 35.222 -0.052 -0.558 Minimum Wage -16.242 8.441 -0.195 -1.924 Business 19.795 31.185 0.057 0.635 Labor -59.456 87.707 -0.065 -0.678 E-score -0.513 0.247 -0.288 -2.080 106th ADA -1.034 0.135 -1.043 -7.662** N 102 R2 0.983 Adjusted R2 0.946 Constant 115.967 Year Elected -0.086 0.577 -0.015 -0.150 Current Term 1.835 2.406 0.075 0.763 Northeast -5.874 15.204 -0.058 -0.386 South 5.825 12.928 0.068 0.451 West 8.867 15.036 0.094 0.590 Diff DW Nom -12.576 59.417 -0.024 -0.212 Fed Spending -2.606 19.578 -0.024 -0.133 Per Cap Inc 0.001 0.001 0.092 0.550 African Amer. -6.904 54.066 -0.017 -0.128 Hispanic -38.792 60.524 -0.105 -0.641 Minimum Wage -23.100 16.109 -0.278 -1.434 Business -43.198 41.216 -0.122 -1.048 Labor 13.089 166.976 0.010 0.078 E-score -0.073 0.202 -0.052 -0.359 107th ADA -0.920 0.196 -0.865 -4.699** N 102 R2 0.961 Adjusted R2 0.876 Constant 91.310 298 Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores Year Elected 0.463 0.646 0.082 0.716 Current Term 5.070 3.475 0.199 1.459 Northeast -13.087 12.962 -0.130 -1.010 South 16.514 14.784 0.194 1.117 West 18.616 16.692 0.197 1.115 Diff DW Nom -33.287 44.933 -0.064 -0.741 Fed Spending -1.502 14.471 -0.015 -0.104 Per Cap Inc 0.000 0.001 0.054 0.421 African Amer. -9.937 51.902 -0.024 -0.191 Hispanic -67.830 41.056 -0.183 -1.652 Minimum Wage -22.912 9.929 -0.275 -2.308* Business -46.076 39.037 -0.128 -1.180 Labor -61.915 105.777 -0.061 -0.585 E-score 0.273 0.321 0.189 0.850 108th ADA -0.694 0.154 -0.652 -4.514** N 100 R2 0.973 Adjusted R2 0.915 Constant 35.244 * p < 0.05, ** p < 0.01 E-score remains statistically significant in the 100 th Congress after substituting ADA into the base model, but it fails tests of statistical significance in each of the other Congresses. The directional impact between higher E-scores and opposition to minimum wage increases is positive. E-score barely fails a test of statistical significance in the 106th Congress, but its directional impact is negative for that Congress. That the statistical significance of E-score diminishes in the model when ADA is added suggests a stronger relative position for liberal-conservative ideology when measured against economic efficiency. As Table 4.31 indicates ACU when substituted into the model is statistically significant and positively correlated with opposition to minimum wage increases in each Congress. In testing the effect of ACU the positive relationship between ACU scores and opposition to policies that increase the federal minimum wage is consistent with 299 Hypothesis 3b: Legislators with higher ACU scores vote in opposition to increasing the federal minimum wage. ACU is a strong predictor of opposition to increasing the federal minimum wage, and as a measure of liberal-conservative ideology the variable compares closely with ADA as an indicator of legislative decision-making. In the Senate model the variable is statistically significant at the p < 0.01 level in each Congress except the 102nd where the level of significance is p < 0.05. ACU produces a relatively strong per unit impact on changes in the dependent variable, but compared to the ADA model, the per-unit impact for ACU is slightly weaker. E-score and percent Hispanic remain statistically significant in the 100 th Congress after substituting ACU into the model. Higher E-scores predict less support for increasing the minimum wage while higher percentages of Hispanics are associated with greater support for minimum wage increases. Senators from the South are more likely to support increasing the federal minimum wage, while those from states that receive more federal spending from the federal government in relation to tax revenue generated by that state are less likely to support increasing the minimum wage. The statistical significance of E-score is consistent with the hypothesized effect of the variable on the model, but produces relatively less per-unit impact when compared to ACU. That E- score fails tests of statistical significance in each of the other Congresses (102 nd through 108 th ) in the ACU substitution is an indication of the importance of liberal-conservative ideology in legislative decision-making. 300 Table 4.31 Regression Analysis of Base Model and ACU Substitution for 100th to 108th Senate: Minimum Wage Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t ? scores Year Elected -0.226 0.501 -0.037 -0.451 Current Term 0.488 2.220 0.019 0.220 Northeast -0.599 13.136 -0.005 -0.046 South -27.718 13.114 -0.282 -2.114* West -8.224 12.179 -0.079 -0.675 Diff DW Nom -53.907 35.207 -0.133 -1.531 Fed Spending 37.655 17.043 0.251 2.209* Per Cap Inc 0.002 0.002 0.118 0.931 African Amer. 22.627 73.098 0.037 0.310 Hispanic -202.011 53.648 -0.406 -3.765** Minimum Wage -9.460 9.629 -0.105 -0.982 Business 6.174 12.460 0.049 0.496 Labor -6.707 13.931 -0.043 -0.481 E-score 0.431 0.207 0.283 2.080* 100th ACU 0.689 0.179 0.526 3.847** N 101 R2 0.717 Adjusted R2 0.629 Constant -43.034 Year Elected 0.200 0.419 0.042 0.478 Current Term 0.395 1.714 0.019 0.230 Northeast 0.576 9.734 0.006 0.059 South 5.683 9.585 0.077 0.593 West 7.782 10.237 0.094 0.760 Diff DW Nom -38.070 28.010 -0.123 -1.359 Fed Spending -15.228 15.574 -0.112 -0.978 Per Cap Inc 0.001 0.002 0.062 0.526 African Amer. 7.318 44.887 0.020 0.163 Hispanic -7.875 38.516 -0.015 -0.204 Minimum Wage -2.242 6.626 -0.031 -0.338 Business -0.478 9.815 -0.005 -0.049 Labor -1.856 14.477 -0.014 -0.128 E-score 0.215 0.144 0.145 1.496 101st ACU 1.047 0.125 0.891 8.371** N 101 R2 0.839 Adjusted R2 0.772 Constant -31.036 301 Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores Year Elected 0.149 0.816 0.026 0.183 Current Term 2.682 3.676 0.110 0.730 Northeast -0.938 18.986 -0.009 -0.049 South -0.214 15.347 -0.003 -0.014 West -3.939 17.608 -0.042 -0.224 Diff DW Nom -65.537 63.837 -0.124 -1.027 Fed Spending 15.945 24.746 0.135 0.644 Per Cap Inc 0.001 0.002 0.102 0.570 African Amer. -37.152 62.612 -0.086 -0.593 Hispanic -46.699 79.378 -0.116 -0.588 Minimum Wage -2.136 21.373 -0.026 -0.100 Business -30.985 64.264 -0.069 -0.482 Labor -248.461 159.681 -0.226 -1.556 E-score -0.080 0.261 -0.055 -0.306 102nd ACU 1.067 0.344 0.924 3.102** N 102 R2 0.949 Adjusted R2 0.840 Constant -37.907 Year Elected -0.070 .542 -0.012 -0.130 Current Term 1.388 4.025 0.050 0.345 Northeast -7.873 12.997 -0.078 -0.606 South -0.091 11.086 -0.001 -0.008 West 8.019 14.092 0.085 0.569 Diff DW Nom -11.309 52.196 -0.021 -0.217 Fed Spending 12.330 21.764 0.100 0.567 Per Cap Inc 0.002 .002 0.188 1.121 African Amer. 4.957 41.891 0.012 0.118 Hispanic -67.159 58.075 -0.168 -1.156 Minimum Wage -12.158 11.040 -0.146 -1.101 Business -33.923 50.590 -0.068 -0.671 Labor -165.964 171.280 -0.157 -0.969 E-score 0.063 .174 0.043 0.362 103rd ACU 0.918 .150 0.825 6.112** N 102 R2 0.972 Adjusted R2 0.913 Constant -50.078 302 Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores Year Elected -0.108 0.431 -0.019 -0.251 Current Term 0.128 2.916 0.005 0.044 Northeast 0.255 12.203 0.003 0.021 South -1.288 14.532 -0.015 -0.089 West 10.712 14.776 0.113 0.725 Diff DW Nom -21.805 50.392 -0.041 -0.433 Fed Spending -3.763 18.843 -0.029 -0.200 Per Cap Inc 0.000 0.001 0.026 0.199 African Amer. 11.149 46.925 0.026 0.238 Hispanic -18.979 54.118 -0.048 -0.351 Minimum Wage -11.221 12.753 -0.135 -0.880 Business -10.471 45.944 -0.026 -0.228 Labor -52.005 109.139 -0.057 -0.477 E-score -0.031 0.345 -0.015 -0.091 104th ACU 0.956 0.169 0.912 5.665** N 103 R2 0.973 Adjusted R2 0.916 Constant 4.535 Year Elected -0.393 0.430 -0.070 -0.914 Current Term -0.981 2.309 -0.038 -0.425 Northeast -1.186 9.500 -0.012 -0.125 South -10.743 10.171 -0.126 -1.056 West 13.347 10.674 0.141 1.250 Diff DW Nom -38.502 46.125 -0.074 -0.835 Fed Spending -13.521 18.355 -0.106 -0.737 Per Cap Inc 0.000 0.001 -0.050 -0.451 African Amer. 38.843 47.087 0.093 0.825 Hispanic -4.937 36.765 -0.013 -0.134 Minimum Wage -5.301 9.986 -0.064 -0.531 Business -30.625 33.674 -0.072 -0.909 Labor 149.574 121.678 0.157 1.229 E-score 0.212 0.268 0.089 0.793 105th ACU 1.166 0.121 0.951 9.607** N 100 R2 0.984 Adjusted R2 0.948 Constant 34.148 303 Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores Year Elected -0.467 0.370 -0.083 -1.262 Current Term -0.422 2.518 -0.015 -0.168 Northeast -7.875 10.228 -0.078 -0.770 South 4.263 9.595 0.050 0.444 West 11.482 11.103 0.121 1.034 Diff DW Nom -27.356 38.559 -0.052 -0.709 Fed Spending 1.793 16.517 0.016 0.109 Per Cap Inc 0.001 0.001 0.086 0.780 African Amer. 11.573 33.169 0.028 0.349 Hispanic -26.477 34.480 -0.070 -0.768 Minimum Wage -12.544 8.591 -0.151 -1.460 Business 11.504 30.422 0.033 0.378 Labor -96.861 88.239 -0.106 -1.098 E-score -0.440 0.237 -0.247 -1.859 106th ACU 1.103 0.142 1.034 7.759** N 102 R2 0.983 Adjusted R2 0.947 Constant 14.563 Year Elected 0.091 0.527 0.016 0.172 Current Term 1.184 2.235 0.048 0.530 Northeast -1.211 14.230 -0.012 -0.085 South 7.918 11.980 0.093 0.661 West 9.114 13.854 0.096 0.658 Diff DW Nom -22.029 55.246 -0.042 -0.399 Fed Spending -15.475 19.071 -0.141 -0.811 Per Cap Inc 0.000 0.001 0.047 0.303 African Amer. 0.527 50.219 0.001 0.010 Hispanic -11.769 58.777 -0.032 -0.200 Minimum Wage -13.765 16.047 -0.165 -0.858 Business -31.374 38.758 -0.089 -0.809 Labor 49.434 155.573 0.038 0.318 E-score -0.088 0.187 -0.063 -0.470 107th ACU 1.062 0.203 0.991 5.218** N 102 R2 0.967 Adjusted R2 0.895 Constant 4.129 304 Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores Year Elected 0.578 0.526 0.103 1.099 Current Term 4.115 2.932 0.161 1.403 Northeast -0.455 10.450 -0.005 -0.044 South 24.485 11.876 0.287 2.062 West 23.407 13.794 0.247 1.697 Diff DW Nom -72.341 37.576 -0.139 -1.925 Fed Spending -10.935 12.368 -0.110 -0.884 Per Cap Inc 0.001 0.001 0.085 0.788 African Amer. -19.405 42.823 -0.047 -0.453 Hispanic -45.532 35.383 -0.123 -1.287 Minimum Wage -17.340 8.521 -0.208 -2.035 Business -77.135 30.695 -0.214 -2.513* Labor -5.394 89.944 -0.005 -0.060 E-score 0.417 0.244 0.288 1.704 108th ACU 0.889 0.156 0.739 5.711** N 100 R2 0.981 Adjusted R2 0.941 Constant -52.112 * p < 0.05, ** p < 0.01 Substituting ACU into the analysis produces a model that explains more variation than base model variables alone. For each Congress coefficients of determination increased when adding ACU. With values for coefficients of determination exceeding 90 percent for ACU and ADA in the 104th, 105th, 106th, and 108th Congresses, the risk of autocollinearity exists from including either of these independent variables in the model. With higher coefficients of determination for ACU and ADA in later Congresses an indication of a potential linear relationship between each of the two variables and support for a minimum wage increase, the lack of variability for each variable is consistent with a polarizing liberal-conservative philosophy. 305 Substituting DW Nominate into the model as a measure of ideology across time, a relative importance of liberal-conservative ideology continues to exits. As Table 4.32 indicates, DW Nominate is statistically significant and positively correlated with opposition to increasing the minimum wage in each Congress in the model with measurable results (100 th through 108 th Congresses). The directional impact of the variable indicates that senators who are more conservative vote in opposition to increasing the minimum wage in greater numbers. This positive relationship between higher DW Nominate scores and opposition to minimum wage policies was anticipated in Hypothesis 4b: Legislators with higher DW Nominate scores vote in opposition to increasing the minimum wage. Relatively strong standardized coefficients in the 105 th , 106 th , and 107 th Congresses indicate the per-unit effect of DW Nominate from its application to the model produces relatively greater change in the dependent variable. Differences in the senator?s ideology in relation to the median ideology for the senate are statistically significant in the model. The variable is not statistically significant in the base model or any of the other substitutions. The negative correlation between differences in DW Nominate and median party and opposition to increasing the minimum wage suggests that the more the senator diverges from median ideological positions of his or her political party the less likely the senator will support an economically efficient position. 306 Table 4.32 Regression Analysis of Base Model and DW Nominate Substitution for 100th to 108th Senate: Minimum Wage Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores Year Elected -0.193 0.504 -0.031 -0.383 Current Term -1.047 2.304 -0.040 -0.454 Northeast -3.656 13.693 -0.032 -0.267 South -25.585 13.310 -0.260 -1.922 West -10.244 12.443 -0.099 -0.823 Diff DW Nom -75.229 37.316 -0.180 -2.016* Fed Spending 45.667 17.328 0.305 2.635** Per Cap Inc 0.002 0.002 0.148 1.163 African Amer. 49.751 73.466 0.082 0.677 Hispanic -181.974 53.400 -0.367 -3.408** Minimum Wage -6.648 9.869 -0.073 -0.674 Business 9.622 12.483 0.076 0.771 Labor 0.036 14.646 0.000 0.002 E-score 0.403 0.221 0.261 1.828 100th DW Nominate 73.577 19.021 0.557 3.868** N 101 R2 0.719 Adjusted R2 0.629 Constant -27.526 Year Elected -0.299 0.386 -0.063 -0.776 Current Term 0.856 1.540 0.040 0.556 Northeast -0.737 8.713 -0.007 -0.085 South -4.613 8.798 -0.063 -0.524 West 1.970 9.177 0.024 0.215 Diff DW Nom -57.910 25.841 -0.187 -2.241* Fed Spending -8.789 13.790 -0.065 -0.637 Per Cap Inc 0.001 0.001 0.069 0.654 African Amer. 52.231 41.059 0.146 1.272 Hispanic -14.191 34.506 -0.027 -0.411 Minimum Wage -4.404 5.893 -0.061 -0.747 Business -4.840 8.584 -0.046 -0.564 Labor 7.612 13.346 0.057 0.570 E-score 0.133 0.131 0.089 1.011 101st DW Nominate 101.600 10.361 0.995 9.806** N 101 R2 0.871 Adjusted R2 0.817 Constant 27.349 307 Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t ? scores Year Elected 0.795 0.603 0.137 1.319 Current Term -0.511 2.839 -0.021 -0.180 Northeast -1.924 15.021 -0.019 -0.128 South -6.463 12.155 -0.076 -0.532 West -8.868 13.974 -0.094 -0.635 Diff DW Nom -29.781 48.437 -0.056 -0.615 Fed Spending 15.426 19.575 0.131 0.788 Per Cap Inc 0.001 0.002 0.126 0.889 African Amer. -4.187 51.095 -0.010 -0.082 Hispanic -55.193 60.168 -0.137 -0.917 Minimum Wage -12.784 13.944 -0.154 -0.917 Business 10.371 52.174 0.023 0.199 Labor -168.771 131.080 -0.153 -1.288 E-score -0.180 0.211 -0.123 -0.851 102nd DW Nominate 113.118 25.575 0.973 4.423** N 102 R2 0.968 Adjusted R2 0.900 Constant 11.716 Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t ? scores Year Elected 0.275 0.683 0.049 0.403 Current Term 1.257 5.108 0.045 0.246 Northeast -1.483 16.670 -0.015 -0.089 South -4.561 13.990 -0.054 -0.326 West 0.444 18.276 0.005 0.024 Diff DW Nom -32.784 68.575 -0.061 -0.478 Fed Spending 12.164 27.623 0.098 0.440 Per Cap Inc 0.002 0.002 0.149 0.698 African Amer. 29.718 54.516 0.070 0.545 Hispanic -44.670 77.612 -0.111 -0.576 Minimum Wage -10.822 14.675 -0.130 -0.737 Business -44.229 63.817 -0.089 -0.693 Labor -93.909 220.249 -0.089 -0.426 E-score 0.059 0.222 0.040 0.266 103rd DW Nominate 98.767 21.724 0.864 4.547** N 102 R2 0.955 Adjusted R2 0.860 Constant 0.174 308 Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores Year Elected 0.381 0.528 0.068 0.721 Current Term 3.206 3.582 0.131 0.895 Northeast -0.239 14.435 -0.002 -0.017 South -2.966 17.304 -0.035 -0.171 West 7.777 17.422 0.082 0.446 Diff DW Nom -58.620 59.558 -0.110 -0.984 Fed Spending 1.290 22.097 0.010 0.058 Per Cap Inc 0.002 0.002 0.155 1.004 African Amer. 12.280 55.905 0.029 0.220 Hispanic -30.301 63.556 -0.076 -0.477 Minimum Wage -13.715 14.989 -0.165 -0.915 Business -43.830 53.704 -0.107 -0.816 Labor -64.685 129.608 -0.071 -0.499 E-score -0.059 0.421 -0.028 -0.141 104th DW Nominate 109.837 24.007 0.978 4.575** N 103 R2 0.962 Adjusted R2 0.882 Constant 9.450 Year Elected -0.176 0.564 -0.031 -0.312 Current Term -3.532 3.018 -0.139 -1.171 Northeast 6.062 12.761 0.060 0.475 South -10.362 13.272 -0.122 -0.781 West -3.291 14.709 -0.035 -0.224 Diff DW Nom -92.835 64.777 -0.178 -1.433 Fed Spending -30.120 23.858 -0.236 -1.262 Per Cap Inc -0.001 0.001 -0.103 -0.704 African Amer. -20.816 58.808 -0.050 -0.354 Hispanic 30.791 50.700 0.079 0.607 Minimum Wage -7.153 13.030 -0.086 -0.549 Business -52.013 43.129 -0.122 -1.206 Labor 372.135 160.472 0.390 2.319* E-score 0.431 0.347 0.181 1.242 105th DW Nominate 114.942 16.045 1.042 7.164** N 100 R2 0.972 Adjusted R2 0.912 Constant 123.041 309 Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores Year Elected -0.092 0.515 -0.016 -0.179 Current Term 2.729 3.682 0.099 0.741 Northeast -7.874 14.491 -0.078 -0.543 South -13.744 14.718 -0.161 -0.934 West -15.489 17.461 -0.164 -0.887 Diff DW Nom -73.894 57.980 -0.141 -1.274 Fed Spending 19.542 24.380 0.170 0.802 Per Cap Inc 0.002 0.001 0.286 1.744 African Amer. 49.592 49.909 0.120 0.994 Hispanic 39.065 55.613 0.103 0.702 Minimum Wage 1.274 14.090 0.015 0.090 Business 12.733 43.576 0.037 0.292 Labor -126.792 128.173 -0.139 -0.989 E-score -0.818 0.407 -0.460 -2.008 106th DW Nominate 154.848 30.092 1.429 5.146** N 102 R2 0.966 Adjusted R2 0.894 Constant 13.002 Year Elected -0.232 0.421 -0.041 -0.551 Current Term 1.780 1.743 0.073 1.021 Northeast -3.760 11.053 -0.037 -0.340 South -0.526 9.256 -0.006 -0.057 West 7.818 10.908 0.083 0.717 Diff DW Nom -46.347 44.534 -0.088 -1.041 Fed Spending -7.677 14.335 -0.070 -0.536 Per Cap Inc 0.001 0.001 0.198 1.682 African Amer. 4.149 39.418 0.010 0.105 Hispanic -6.739 45.679 -0.018 -0.148 Minimum Wage -15.664 12.038 -0.188 -1.301 Business -80.112 28.157 -0.227 -2.845* Labor 28.637 121.084 0.022 0.237 E-score -0.215 0.155 -0.155 -1.389 107th DW Nominate 118.188 17.002 1.111 6.952** N 102 R2 0.979 Adjusted R2 0.935 Constant 49.129 310 Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores Year Elected 0.717 0.773 0.127 0.928 Current Term 3.511 4.505 0.138 0.779 Northeast 2.212 15.366 0.022 0.144 South 13.553 18.588 0.159 0.729 West 14.867 20.618 0.157 0.721 Diff DW Nom -78.637 55.708 -0.151 -1.412 Fed Spending -6.272 18.032 -0.063 -0.348 Per Cap Inc 0.001 0.001 0.087 0.542 African Amer. 5.076 64.701 0.012 0.078 Hispanic -26.962 56.301 -0.073 -0.479 Minimum Wage -11.463 13.582 -0.138 -0.844 Business -77.203 45.442 -0.214 -1.699 Labor -19.626 132.630 -0.019 -0.148 E-score 0.399 0.384 0.276 1.039 108th DW Nominate 83.937 24.957 0.803 3.363** N 100 R2 0.960 Adjusted R2 0.873 Constant -13.496 * p < 0.05, ** p < 0.01 Both measures of self-interest ? Business and Labor ? were statistically significant in the model in the 107 th and 105 th Congresses, respectively. Neither variable was statistically significant in the base model in those Congresses. Senators receiving higher business contributions were more likely to support increasing the minimum wage, while senators receiving higher labor contributions were more likely to oppose increasing the minimum wage. Ironically, the directional impact of each variable was opposite the anticipated correlation as hypothesized in the base model. That DW Nominate is statistically significant in each Congress suggests an ideology component over time could be affecting anticipated self-interest movements. Substituting legislator party into the base model for minimum wage (Table 4.33) finds relatively few statistically significant relationships change as a result of the substitution. Senators receiving higher labor contributions and representing states with 311 higher percentages of Hispanics in the population are statistically significant variables in the model and positively correlated to greater support for increasing the minimum wage. Legislators with higher E-scores are more likely to oppose minimum wage increases. Each of these associations and directional impacts hold for both the base model and legislator party substitution. Legislator party is statistically significant in each Congress with the exception of the 101 st Congress. In testing Hypothesis 10b: Legislators from the minority party (Senate) are less likely than majority party legislators to support increasing the federal minimum wage, the model finds no statistical difference in minority or majority party status in predicting support. Generally, Republicans oppose increasing the minimum wage and Democrats support increasing the minimum wage regardless of party control of the institution. In the 107 th Congress this association was not possible to measure with party control of the Senate varying throughout the Congress. State minimum wage laws remained statistically significant in the 102 nd , 103 rd , and 107 th Congresses after substituting legislator party into the model. Senators representing states with minimum wage laws higher than proposed federal minimum wage legislation support increasing the federal minimum wage. This association exists with both Democratic control and varying control of the Senate and is consistent with base model analysis. 312 Table 4.33 Regression Analysis of Base Model and Legislator Party Substitution for 100th to 108th Senate: Minimum Wage Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores Year Elected -0.352 0.529 -0.057 -0.666 Current Term -1.148 2.372 -0.044 -0.484 Northeast -6.690 14.005 -0.058 -0.478 South -19.140 13.624 -0.194 -1.405 West -3.863 12.854 -0.037 -0.301 Diff DW Nom -61.786 37.211 -0.152 -1.660 Fed Spending 35.219 17.871 0.235 1.971 Per Cap Inc 0.002 0.002 0.092 0.693 African Amer. 75.044 75.638 0.123 0.992 Hispanic -148.794 55.819 -0.299 -2.666** Minimum Wage -14.778 10.003 -0.164 -1.477 Business 11.017 13.008 0.087 0.847 Labor 1.447 15.544 0.009 0.093 E-score 0.490 0.228 0.322 2.153* 100th Legislator Party -39.466 13.117 -0.433 -3.009** N 101 R2 0.689 Adjusted R2 0.591 Constant 5.814 Year Elected 0.721 0.710 0.152 1.015 Current Term -0.752 2.971 -0.035 -0.253 Northeast -4.182 16.652 -0.042 -0.251 South 17.162 16.538 0.234 1.038 West 6.397 17.598 0.078 0.363 Diff DW Nom 78.681 41.584 0.255 1.892 Fed Spending 12.887 26.365 0.095 0.489 Per Cap Inc 0.003 0.003 0.248 1.209 African Amer. -41.524 77.507 -0.116 -0.536 Hispanic -13.121 66.148 -0.025 -0.198 Minimum Wage -11.304 11.400 -0.158 -0.992 Business -29.008 16.885 -0.278 -1.718 Labor -79.311 19.049 -0.594 -4.164** E-score 0.560 0.238 0.378 2.351* 101st Legislator Party 3.668 9.593 0.051 0.382 N 101 R2 0.527 Adjusted R2 0.330 Constant -61.562 313 Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores Year Elected 0.800 0.704 0.138 1.136 Current Term -0.786 3.320 -0.032 -0.237 Northeast -4.430 17.652 -0.044 -0.251 South 6.112 14.366 0.072 0.425 West 4.786 16.443 0.051 0.291 Diff DW Nom -16.036 56.255 -0.030 -0.285 Fed Spending 7.275 23.075 0.062 0.315 Per Cap Inc 0.000 0.002 0.010 0.065 African Amer. -58.496 56.976 -0.136 -1.027 Hispanic -103.041 66.065 -0.256 -1.560 Minimum Wage -34.995 13.116 -0.420 -2.668* Business 3.083 60.667 0.007 0.051 Labor -185.474 152.852 -0.168 -1.213 E-score -0.021 0.225 -0.014 -0.092 102nd Legislator Party -53.409 15.109 -0.637 -3.535** N 102 R2 0.956 Adjusted R2 0.863 Constant 75.899 Year Elected 0.113 0.653 0.020 0.173 Current Term 2.107 4.850 0.076 0.435 Northeast -9.093 15.696 -0.090 -0.579 South 9.393 13.714 0.110 0.685 West 15.519 16.835 0.164 0.922 Diff DW Nom -9.301 63.338 -0.017 -0.147 Fed Spending 8.066 26.542 0.065 0.304 Per Cap Inc 0.002 0.002 0.175 0.861 African Amer. -20.910 49.943 -0.049 -0.419 Hispanic -81.025 69.605 -0.202 -1.164 Minimum Wage -29.239 11.275 -0.351 -2.593* Business -85.229 57.667 -0.171 -1.478 Labor -192.835 206.488 -0.183 -0.934 E-score 0.003 0.215 0.002 0.012 103rd Legislator Party -54.128 11.203 -0.645 -4.832** N 102 R2 0.959 Adjusted R2 0.872 Constant 40.097 314 Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores Year Elected 0.318 0.563 0.057 0.565 Current Term 2.406 3.793 0.098 0.634 Northeast -4.541 15.484 -0.045 -0.293 South 14.944 18.041 0.175 0.828 West 16.119 18.955 0.170 0.850 Diff DW Nom -47.397 63.687 -0.089 -0.744 Fed Spending -6.133 24.143 -0.047 -0.254 Per Cap Inc 0.001 0.002 0.068 0.409 African Amer. -53.235 56.270 -0.126 -0.946 Hispanic -62.522 66.305 -0.157 -0.943 Minimum Wage -30.385 14.635 -0.365 -2.076 Business -37.624 57.584 -0.092 -0.653 Labor -27.056 137.766 -0.030 -0.196 E-score 0.214 0.413 0.100 0.518 104th Legislator Party 56.737 13.621 0.676 4.166** N 103 R2 0.957 Adjusted R2 0.865 Constant -0.883 Year Elected 0.071 0.753 0.013 0.094 Current Term -2.268 3.971 -0.089 -0.571 Northeast -5.503 16.231 -0.055 -0.339 South 11.246 17.375 0.132 0.647 West 13.163 18.480 0.139 0.712 Diff DW Nom -26.071 79.847 -0.050 -0.327 Fed Spending -22.382 31.500 -0.175 -0.711 Per Cap Inc -0.001 0.002 -0.054 -0.283 African Amer. -67.504 75.705 -0.161 -0.892 Hispanic -42.505 61.251 -0.109 -0.694 Minimum Wage -29.165 14.826 -0.350 -1.967 Business -31.163 58.515 -0.073 -0.533 Labor 145.502 209.523 0.153 0.694 E-score 0.136 0.463 0.057 0.294 105th Legislator Party 62.777 12.193 0.748 5.149** N 100 R2 0.951 Adjusted R2 0.846 Constant 71.984 315 Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores Year Elected 0.149 0.623 0.027 0.240 Current Term -1.133 4.320 -0.041 -0.262 Northeast -17.409 17.691 -0.173 -0.984 South 9.394 16.363 0.110 0.574 West 11.269 19.182 0.119 0.587 Diff DW Nom -33.550 67.423 -0.064 -0.498 Fed Spending 0.237 28.474 0.002 0.008 Per Cap Inc 0.001 0.001 0.081 0.432 African Amer. -27.237 55.413 -0.066 -0.492 Hispanic -41.570 59.030 -0.109 -0.704 Minimum Wage -22.515 14.063 -0.270 -1.601 Business 17.510 53.989 0.051 0.324 Labor -62.942 151.275 -0.069 -0.416 E-score -0.258 0.404 -0.145 -0.637 106th Legislator Party 72.648 18.324 0.866 3.965** N 102 R2 0.950 Adjusted R2 0.844 Constant 18.868 Year Elected 0.155 0.593 0.028 0.262 Current Term 2.454 2.485 0.100 0.988 Northeast -15.275 15.429 -0.152 -0.990 South 10.541 13.650 0.124 0.772 West 16.715 15.596 0.177 1.072 Diff DW Nom -5.184 61.251 -0.010 -0.085 Fed Spending -1.647 20.288 -0.015 -0.081 Per Cap Inc 0.001 0.001 0.163 0.956 African Amer. -59.265 52.774 -0.143 -1.123 Hispanic -79.282 57.453 -0.215 -1.380 Minimum Wage -35.576 14.812 -0.427 -2.402* Business -45.793 42.663 -0.130 -1.073 Labor -74.275 169.477 -0.057 -0.438 E-score -0.061 0.209 -0.044 -0.291 107th Legislator Party -61.377 13.738 -0.732 -4.468** N 102 R2 0.958 Adjusted R2 0.866 Constant 61.781 316 Congress Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores Year Elected 0.915 0.633 0.163 1.445 Current Term 4.636 3.621 0.182 1.281 Northeast -0.199 12.944 -0.002 -0.015 South 31.443 14.498 0.369 2.169 West 23.987 17.083 0.253 1.404 Diff DW Nom -72.548 46.600 -0.139 -1.557 Fed Spending -10.653 15.378 -0.107 -0.693 Per Cap Inc -5.029E-05 0.001 -0.007 -0.052 African Amer. -47.978 52.574 -0.116 -0.913 Hispanic -57.679 43.169 -0.156 -1.336 Minimum Wage -20.773 10.371 -0.250 -2.003 Business -69.419 38.540 -0.192 -1.801 Labor -26.441 110.658 -0.026 -0.239 E-score 0.584 0.287 0.404 2.034 108th Legislator Party 50.218 11.578 0.599 4.337** N 100 R2 0.971 Adjusted R2 0.910 Constant -36.199 * p < 0.05, ** p < 0.01 Coefficients of determination across the legislator party substitution indicate that adding the variable to the model increased the explanatory effect of the variables in the model on changes in the dependent variable, minimum wage, in most Congresses. In the 101 st Congress adding legislator party decreased the explanatory effect of the model over base variables. Model Summary of Senate Analysis An analysis of the Senate finds ideology and political party as strong predictors of legislative behavior. Consistently throughout the model and across both dependent variables, a positive association exists between both conservatism and Republican support for economically efficient outcomes. These trends exist across all measures of ideology ? ADA, ACU, and DW Nominate ? and are not sensitive to party control of the 317 Senate. Evidence of increasing polarization between parties occurs in later Congresses as less variability surrounds conservative or liberal ideology and economic efficiency. With legislative ideology moving toward the extremes, E-score measures of a senator?s voting patterns appear relatively less important as a predictor of behavior. Within the Senate analysis distinctions exist between results produced when analyzing each dependent variable that must be noted. In the medical malpractice dependent variable E-score has a stronger impact in the base model but less impact after variable substitution into the model. Self-interest is a much more effective predictor of behavior in the House than in the Senate. This fact could be the result of the relative strength of liberal-conservative ideology and party in the Senate or indicate the effectiveness of lobbying efforts with the small House districts. State economic conditions and a senator?s tenure in his or her career are important control variables in the medical malpractice model but are much less effective in the minimum wage model. The highly correlated effect of the substituted variables ? party unity, ADA, ACU, DW Nominate, and legislator party ? is evident in the minimum wage model. When substituted into the model each measure of ideology or party consistently produces strong results for that variable. That the substituted variable was often the only variable statistically significant in the minimum wage model indicates not only the importance of ideology and party in predicting behavior, but also the relative effect of each variable measured against self-interest and E-score. Analyzing effects of base and substituted models for both dependent variables in the Senate found important similarities and differences with results from the same models in an analysis of the House. E-score is a relatively stronger predictor of behavior 318 in the House relative to the Senate across each dependent variable. That the House is a more homogenous institution and characterized to a larger extent through ideological extremes does not appear to affect a legislator?s support of economic efficiency principles. ACU is a relatively stronger predictor of behavior in the House, and ADA is a stronger predictor of behavior in the Senate. The strength of the variable is measured through standardized coefficients showing per unit effects associated with application of the variable on the model. Differences in a legislator?s ideology and the median ideology of the party exist between House and Senate. In the house larger differences between a legislator?s ideology and median party ideology are associated with greater support for economically efficient policies. In the Senate the relationship is reversed as senators who diverge from median ideology of their party are less likely to support economically efficient policies. This relationship suggests that the effect of party is not static across institutions. Legislative decision making supporting economically efficient policies in the House is consistent with the greater effect of E-score in that chamber as a predictor of behavior. Hypothesis Testing Summary In answering the research question of the extent that E-score represents a component of ideology beyond mere liberalism and conservatism that can be used as a predictor of legislative voting, hypotheses were developed to measure the impact of ideology, self-interest, and party in a multivariate model. 319 Hypotheses for both dependent variables ? medical malpractice and minimum wage ? were tested for the House and Senate and reported above. Confirmations of each hypothesis are summarized for House and Senate in Table 4.34. Table 4.34 Summary and Confirmation of Hypothesis Tested for Medical Malpractice and Minimum Wage Dependent Variables in House and Senate Medical Malpractice Confirmation Hypothesis House Senate H 1a: Legislators with higher E-scores vote in support of medical malpractice reform. Yes Yes H 2a: Legislators with higher ADA scores vote in opposition to medical malpractice reform. Yes Yes H 3a: Legislators with higher ACU scores vote in support of medical malpractice reform. Yes Yes H 4a: Legislators with higher DW Nominate scores vote in support of medical malpractice reform. Yes No H 5a: Legislators with higher health care political contributions to total contributions vote in support of malpractice reform. Yes No H 6a: Legislators with higher legal political contributions to total contributions vote in opposition to medical malpractice reform. Yes Yes H 7a: Republican legislators are likely to vote for malpractice reform more often than Democratic legislators. Yes Yes H 8a: The closer senators are to the end of their current term in office, the more likely they are to support malpractice reform. --- No H 9a: The longer a legislator has served, the more likely he or she supports medical malpractice reform. No No H 10a: Legislators from the minority party (House) are more likely than majority party legislators to support medical malpractice reform No No H 11a: The greater the ideological division between the legislator and the median ideology of the legislator?s party, the more likely the legislator supports medical malpractice reform. Yes No 320 Minimum Wage Hypothesis House Senate H 1b: Legislators with higher E-scores vote in opposition to increasing the minimum wage. Yes Yes H 2b: Legislators with higher ADA scores vote in support of increasing the federal minimum wage. Yes Yes H 3b: Legislators with higher ACU scores vote in opposition to increasing the federal minimum wage. Yes Yes H 4b: Legislators with higher DW Nominate scores vote in opposition to increasing the minimum wage. Yes Yes H 5b: Legislators with higher business political contributions to total contributions vote in opposition to increasing the minimum wage. Yes No H 6b: Legislators with higher labor political contributions to total contributions vote in support of increasing the minimum wage. Yes Yes H 7b: Republican legislators are less likely to vote for increasing the minimum wage more often than Democrats. Yes Yes H 8b: The closer senators are to the end of their current term in office the more likely they are to support malpractice reform. - - - No H 9b: The longer a legislator has served, the less likely he or she will support increasing the minimum wage. No No H 10b: Legislators from the minority party (House) are less likely than majority party legislators to support increasing the federal minimum wage. No No H 11b: The greater the division between the ideology of the legislator and the median ideology of the legislator?s party, the less likely the legislator supports increasing the federal minimum wage. Yes No In the House ideology, self-interest, and party are clear predictors of behavior in this model. As a component of ideology E-score was found to be a predictor of legislative voting positively correlated with support for economically efficient public policies. Measuring ideology across a liberal-conservative spectrum as a predictor of legislative voting is clearly confirmed in this model. Analysis using two distinct policy areas as dependent variables identifies conservatives as supporting economically efficient principles in higher numbers than liberals. Republicans support economically efficient policies regardless of party control of the institution. The ideology of the 321 representative is a stronger predictor of economically efficient behavior than median party ideology. As legislators diverge from median party ideology in casting roll call votes those legislators support economically efficient legislation in increasing numbers. House member decision-making is also shaped by political contributions that impact the self-interest of legislators. Political contributions are confirmed as an effective means of shaping a legislator?s voting preferences in support or opposition to a policy. Minority or majority status of the legislator?s party is not a factor in the model. In the Senate ideology remains a strong predictor of behavior but the impact of its effect changes. E-score and liberal-conservative ideology ? ADA and ACU ? are strong predictors of behavior in the Senate model. For medical malpractice dependent variable a liberal-conservative spectrum does not function as well in predicting directional impact of ideology on behavior over time. DW Nominate as a weighted measure for capturing changes in ideology over time does not produce a consistent directional impact to confirm Hypothesis 4a. The hypothesis is confirmed in the minimum wage model, suggesting that the utility of the DW Nominate variable in predicting ideology may be situational to the policy. Differences in economic efficiency from ideological divisions between a senator and his or her median party ideology are confirmed in this model. Regardless of party control, Republicans support economically efficient legislation in the Senate, while Democrats do not. Length of time in office and time in current term are not factors in support of economically efficient legislation in the Senate model. Self-interests affect legislative decision-making but are inconsistent in an analysis of the Senate. Neither self-interest hypothesis is confirmed in medical 322 malpractice dependent variable. Only labor contributions are effective in predicting behavior for minimum wage dependent variable. Considering the effects of change in party control on legislative behavior is important to this analysis of economic efficiency. Statistically significant differences were not found between minority-majority party control of Congress or ideological differences between legislator ideology and median party ideology when analyzing each Congress separately. Time Series Analysis Interrupted time series was used in measuring the impact of changes in political party control of the political institution (House and Senate) on support for medical malpractice reform and legislation for increasing the federal minimum wage within the Congresses of the study. The intent of using interrupted time series in the model is to evaluate the impact of changes in political party control on legislative decision-making, and if such decision-making impacts economic efficiency. In addition to an analysis of each dependent variable, separate analyses were regressed on time and two dummy variables for each ideology independent variable (ADA, ACU, E-score, and DW Nominate). The model initially identified three data points for making this analysis: the 1986 congressional election with the Democratic Party regaining control of the Senate, the 1994 election with Republicans sweeping both houses of Congress, and the 2000 election with closely divided, Republican controlled Congress and a Republican president. In order to measure changes at each data point several votes are needed before 323 and after the event. Too few votes were available for analyzing legislative decision- making from changes in party control before and after events in 1986 and 2000. The Republican sweep of both the Senate and the House in 1994 (104 th Congress) was adopted as the basis for making interrupted time series analysis. The model is based on Kellough (1990, p. 84) and is presented by the following equation: Y t = b 0 + b 1 X 1t + b 2 X 2t + e The key additions in an interrupted time series design are two dummy counting variables. One dummy variable X 1t is coded zero for observations before changes in institutional control in the 104 th Congress (e.g., a Democratic majority in the Senate is replaced by a Republican majority) and one for observations thereafter. It is used as an indicator of whether a change in behavior occurred in and around the event in question. The second dummy variable X 2t is coded zero for observations prior to the change in party control and one for the first year after the change in party control, two for the next, three for the next, and so forth. This variable is called a post counter. It is used in determining whether any change in pattern detected is long term or short term in duration. The dependent variable Y t represents mean scores for medical malpractice reform or in opposition to increasing the minimum wage. The latter is measured in terms of opposition to the legislation rather than in support of the legislation to standardize both dependent variables to positively reflect economic efficiency. Each Congress represents the time variable for years. 324 Analysis For each dependent variable ? medical malpractice and minimum wage ? in the House and Senate mean scores were gathered representing dependent variable legislation from that Congress or, if no dependent variable legislation was considered in that Congress, from a scoring model representing all dependent variable legislation over the study. Scores were separated into mean scores for Republicans and mean scores for Democrats to analyze the magnitude of differences between each political party. The model was tested for House and Senate in exploring if statistically significant changes were observed as a result of time or if changes in party control beginning with the 104 th Congress produced changes in behavior. Time and measures of changes in party control (DUM1 and DUM2) were regressed against each dependent variable ? mean medical malpractice scores and mean minimum wage scores ? and also against each measure of ideology (E-score, ACU, ADA, and DW Nominate) for Republicans and Democrats collectively and Republicans and Democrats separately for each Congress (99 th through 108 th ) in the study. House Mean scores for medical malpractice and minimum wage were analyzed as dependent variables in the interrupted time series model. A measure of time (HOUSE) and each counting variable (DUM1 and DUM2) were regressed against mean scores for each dependent variable. No statistically significant combinations existed between the effects of time before and after changes in party control and changes in either dependent variable, mean medical malpractice scores and mean minimum wage scores. 325 HOUSE, DUM1, and DUM2 were regressed against mean values for each ideology variable (E-score, ADA, ACU, and DW Nominate) in testing for the effect of time on changes in ideology. No statistically significant relationships were found in the ADA, ACU, and E-score analyses. Table 4.35 shows DW Nominate scores for all House members initially increased with the impact of changes in party control and continued increasing after the initial impact. The effect of time measured by HOUSE indicates a very small decline in DW Nominate scores over time, but the variable is not statistically significant. When Republican and Democrat DW Nominate scores are considered separately, no statistically significant relationships are found. This suggests that the overall effect of time and party control in the House does not affect Republicans and Democrats singularly, but rather has a collective impact on ideology. This is what one would expect given the fact that DW Nominate adjusts for ideological change over time. Table 4.35 Interrupted Time Series Analysis of Time and Changes in Party Control on Mean DW Nominate Scores for Republicans and Democrats in the House HOUSE Mean DW Nominate (Republicans and Democrats) Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores HOUSE -0.003 0.002 -0.140 -1.044 DUM1 0.089 0.010 0.833 8.675** DUM2 0.010 0.004 0.327 2.754* R squared 0.987 Adjusted R squared 0.980 Constant -0.044 Mean DW Nominate (Republicans) Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores HOUSE 0.017 0.012 0.586 1.362 DUM1 0.048 0.051 0.288 0.933 DUM2 0.004 0.018 0.089 0.232 R squared 0.865 Adjusted R squared 0.798 Constant 0.255 326 Mean DW Nominate (Democrats) Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores HOUSE -0.012 0.009 -0.692 -1.400 DUM1 -0.050 0.035 -0.501 -1.413 DUM2 0.008 0.012 0.280 0.638 R squared 0.823 Adjusted R squared 0.734 Constant -0.256 * p < 0.05, ** p < 0.01 Senate Regressing time and changes in party control on mean medical malpractice scores for Senate Republicans and Democrats combined (Table 4.36) finds no evidence of change from the passage of time alone (p = 1.00), but statistically significant changes occur when scores for medical malpractice are considered for Republicans and Democrats separately. Table 4.36 Interrupted Time Series Analysis of Time and Changes in Party Control on Mean Medical Malpractice Scores for Republicans and Democrats in the Senate SENATE Mean Medical Malpractice (Republicans and Democrats) Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores SENATE -1.432E-15 0.890 0.000 0.000 DUM1 -0.500 3.669 -0.075 -0.136 DUM2 1.500 1.258 0.815 1.192 R squared 0.568 Adjusted R squared 0.352 Constant 42.000 Mean Medical Malpractice (Republicans) Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores SENATE 6.263E-15 2.466 0.000 0.000 DUM1 -36.800 10.167 -1.369 -3.620* DUM2 12.000 3.487 1.610 3.441* R squared 0.798 Adjusted R squared 0.697 Constant 78.00 327 Mean Medical Malpractice (Democrats) Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores SENATE -2.326E-15 2.001 0.000 0.000 DUM1 34.500 8.251 1.297 4.181** DUM2 -12.300 2.830 -1.668 -4.346** R squared 0.864 Adjusted R squared 0.796 Constant 23.000 * p < 0.05, ** p < 0.01 For Republicans there is a change around the 104 th Congress indicating less Republican support for medical malpractice reform, but over the long term support increases over time. When measuring mean medical malpractice votes cast by Democrats, an initial change occurs with the change in party control. Democrats initially experience greater support for medical malpractice reform, which decreases over time. Time change itself is not statistically significant in the medical malpractice model for Republicans or Democrats and offers no support that time is responsible for changes. That Republican support initially drops and Democrat support initially increases reflects the pull of party on a senator?s ideology as party control changes. Republicans are influenced by Democrat control of the Senate to the extent that their behavior changes when Democrats are relegated to minority party status. Democrat senators experienced the opposite effect, as ascension to majority party status is associated with increased support for medical malpractice reform. In each scenario behavior reverts to long-term trends associated with support for medical malpractice reform by each party. Republican senators support increases over time after the initial decline and Democrat support declines after the initial increase. These movements in support are an indication that party and shifts in party control 328 influence senators and play a role in changes in the Senate, but senators gradually adjust to those changes over time. Testing if adherence to party principles is a factor in changes in support or if ideology is responsible for changes in legislative behavior, an interrupted time series analysis measuring time and party control was run against each ideology variable. None of the ACU and E-score equations showed statistically significant variables. For ADA (Table 4.37) the passage of time is related to increasing Senate scores over time. A change occurred around the Republican takeover with the 104 th Congress where scores declined. A gradual long-term decline continued but the results were not statistically significant. ADA scores for Republican senators experience an abrupt decline around the 104 th Congress. Republican ADA scores trend upward over time before the impact of changes in party control and after the initial decline, but the results are not statistically significant. ADA scores for Democrats are impacted by change in time and experience increases over the time period of the study. ADA scores for Democrats are not impacted by changes in party control, nor are their movements after the initial change in party control. Table 4.37 Interrupted Time Series Analysis of Time and Changes in Party Control on Mean ADA Scores for Republicans and Democrats in the Senate SENATE Mean ADA (Republicans and Democrats) Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores SENATE 2.400 0.898 1.756 2.672* DUM1 -12.600 3.703 -1.605 -3.403* DUM2 -0.400 1.270 -0.184 -0.315 R squared 0.686 Adjusted R squared 0.529 Constant 42.000 329 Mean ADA (Republicans) Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores SENATE 0.400 0.986 0.234 0.406 DUM1 -13.700 4.064 -1.396 -3.371* DUM2 1.300 1.394 0.477 0.933 R squared 0.758 Adjusted R squared 0.637 Constant 18.800 Mean ADA (Democrats) Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores SENATE 2.100 0.832 0.876 2.525* DUM1 6.400 3.429 0.465 1.866 DUM2 -1.500 1.176 -0.393 -1.275 R squared 0.912 Adjusted R squared 0.869 Constant 67.700 * p < 0.05, ** p < 0.01 Mean DW Nominate scores are impacted by time when Republican and Democrat combined Senate scores are included in the analysis. (See Table 4.38.) Given the method of calculation of the variable, a change in time for DW Nominate scores is expected, and an increase around the 104 th Congress is not surprising with Republicans ascending to majority status and generally having higher DW Nominate scores representing a more conservative ideology. That DW Nominate scores experience decreases over the time period of the study is surprising. Separating Republican and Democrat DW Nominate scores finds Republican scores increasing over time and Democrat scores decreasing. Initial changes in DW Nominate scores for each party were impacted by change in control of the senate to the Republicans with Republicans initially showing stronger conservatism through higher DW Nominate scores and Democrats initially showing stronger liberalism through declining DW Nominate scores. After the initial impact on DW Nominate scores for each party, continued effects were not observed in the model. 330 Table 4.38 Interrupted Time Series Analysis of Time and Changes in Party Control on Mean DW Nominate Scores for Republicans and Democrats in the Senate SENATE Mean DW Nominate (Republicans and Democrats) Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores SENATE -0.014 0.005 -1.023 -2.475* DUM1 0.101 0.022 1.329 4.482** DUM2 0.011 0.008 0.50 1.387 R squared 0.876 Adjusted R squared 0.814 Constant 0.001 Mean DW Nominate (Republicans) Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores SENATE 0.012 0.003 0.669 4.410** DUM1 0.032 0.011 0.300 2.756* DUM2 0.002 0.004 0.053 0.396 R squared 0.983 Adjusted R squared 0.975 Constant 0.266 Mean DW Nominate (Democrats) Independent and Control Variables Coefficient Estimates Standard Errors Standardized Coefficients t - scores SENATE -0.011 0.002 -0.975 -5.261** DUM1 -0.025 0.009 -0.393 -2.954* DUM2 0.007 0.003 0.388 2.356 R squared 0.975 Adjusted R squared 0.963 Constant -0.301 * p < 0.05, ** p < 0.01 With ADA and DW Nominate measuring different components of liberal- conservative ideology differences in results for the two variables were expected. Impact of change in time affects each measure of ideology, as a trend generally indicates gradual movement toward more liberal positions for Democrats and more conservative positions for Republicans. These effects were observed within ADA and DW Nominate analysis and analysis of medical malpractice dependent variable. Impact of change in party control frequently produced abrupt changes away from ideology principles, but such changes were not generally sustained. 331 The fact that E-scores were not impacted by a change in party control of either house suggests that legislators who support economically efficient policies are less affected by institutional effects within government. E-scores capture a component of ideology but in comparison with liberal-conservative measures are less aligned with party and therefore less affected by changes in party. As confirmed from the results, changes in E-score should not be a function of the legislative environment, but rather reflect support for policies that expand social benefit relative to cost. Chapter Five concludes the study and considers the effectiveness of E-score as a predictor of legislative behavior with policy applications beyond liberal-conservative ideology measures. 332 CHAPTER FIVE OVERALL CONCLUSIONS AND POLICY IMPLICATIONS The impetus for this research emanates from a need to better identify and measure policy implications of legislative voting. The focus of the research considered economic efficiency through an E-score as a variable that depicts social benefit maximization from legislative policy decisions. Traditional measures of ideology and self-interest identify characteristics of a legislator, but do not include the policy implications of legislation that he or she supports. This study considered two dependent variable policy areas: medical malpractice tort reform and federal minimum wage legislation. Each policy area was analyzed for members of the House of Representatives and Senate for the 99 th through 108 th Congress. Measures of legislative voting were analyzed not only as static measures that identified a legislator within each Congress, but also through the effect of changes in political and institutional variables over a period of years on legislative decision-making. In analyzing economic efficiency as a policy tool, this dissertation posed one primary research question: Does economic efficiency through an E-score function better than a traditional spectrum of liberal-conservative ideology in explaining the ideological position of a representative (House and / or Senate member), congressional activity, and public policy formulation? Strict use of liberal-conservative interest group ratings omits important strands of ideology and ideological change (Grafton & Permaloff, 2005a, p. 333 409). Ideology represents more than just relative liberalism and conservatism. Explaining behavior within these extremes fails to identify not only other facets of ideology, but also the importance of using ideology in analyzing economic consequences of political behavior. The study found that E-score is a useful tool in predicting behavior but liberal- conservative ideology and alignment with party are overall much stronger factors in legislative voting decisions. Liberal-conservative ideology and party each consistently produced stronger correlation coefficients and were statistically significant at higher levels of confidence in the model. Values are part of decision-making and economic efficiency is a value in that higher economic output begets higher standards of living, and greater social benefits. The study attempted to show that economic efficiency measured through an E-score more accurately captures why legislators support some bills and oppose others in an attempt to produce a public policy that increases net benefits to a constituency. The study included E-score in the vote models along with liberal-conservative measures of ideology, party, or self-interest variables. E-score is viewed here as a component of ideology, but how well ideology predicts behavior is a matter of perspective. E-score was often statistically significant in base model analysis that included self-interest and control variables but frequently failed tests of statistical significance when included with separate rotations of ideology and party variables. Higher coefficients of determination confirmed that liberal-conservative ideology and party variables explain more variance in the model than base model analysis alone. That party and ideology are strong predictors of behavior is not surprising (Kernell & 334 Jacobson, 2006, p. 462; Reichley, 1992, p. 414; Mullins, 1972, p. 509). The impact of their inclusion in vote models raises questions whether E-score adds a different perspective to values guiding legislative decisions beyond traditionally accepted measures of behavior. This chapter analyzes how the study answered the research question and justifies the E-score as a measure of ideology different from a liberal-conservative perspective. The inclusion of economic efficiency as a basis of policy analysis and usefulness of the E-score as a measure of economic efficiency are evaluated. Problems encountered with developing the E-score are discussed. Applications for E-score in public-private settings are explored. The chapter concludes with recommendations for using E-score in legislative vote models. Public Policy Development and Economic Efficiency A public policy is a chance to improve the world. Legislative decision-making involves setting goals for reaching policy outputs and outcomes. With specified goals, development of criteria and measures is possible. With legislative resources often scarce, policy outcomes that focus on providing the greatest social benefit measured against social cost is the heart of policy analysis debate. While Lasswell?s decisional processes expand democracy by increasing participation and benefit the public (Parsons, 1995, pp. 18-19), legislators also have an obligation to themselves and to their constituency to produce public policies that are not only necessary but also maximize social benefit outcomes within limited constraints. 335 For a public policy to produce outcomes that maximize net social benefits, values are part of the decision making process and political and administrative feasibility are included along with public perception of a problem and the necessary policy for correcting a problem. In the political arena values are often measured along a liberal- conservative spectrum but these designations describe characteristics of a legislator and not the outcome associated by his or her vote of support for or opposition to a public policy. E-score as an Ideological Tool This study argued that E-score is a measure of ideology that can be used to supplement traditional liberal-conservative tools, such as ADA, ACU, or DW Nominate scores. By focusing on a measure of economic efficiency to describe behavior through outcomes of voting decisions, using those decisions to predict support for policies that expand net benefits is possible. The study attempted to develop an E-score for measuring economic efficiency associated with legislative decision making and applied that score to public policy initiatives to test its effectiveness. Including liberal- conservative ideology variables, self-interest, and party variables in a multivariate analysis allowed for testing the effect of E-score in a model and measuring directional impact of its application. The study found that traditional measures of ideology ? ADA, ACU, and DW Nominate ? are highly correlated with each other but not with E-score. A positive correlation existed with E-score and ACU and DW Nominate and a negative correlation with E-score and ADA. These associations were relatively consistent throughout the 336 model. Party unity and legislator party scores were also highly correlated with ideology variables. Republicans were typically conservative and Democrats liberal. Extremes between ACU and ADA scores depict these relative differences in ideology as Republicans generally have higher ACU scores and Democrats higher ADA scores. Movements between Congresses for ADA, ACU, and DW Nominate scores experience relatively less variance than E-score. Less variance is an indication that liberal-conservative positions of legislators experience less change from Congress to Congress, while E-score values swing with much more variation. The purpose of incorporating more moderate values for E-score into a vote model was to alleviate extreme positions identified in using ADA, ACU, and DW Nominate scores to predict behavior. As a measure of economic efficiency, E-scores were not expected to measure extreme positions for ideology between each political party but rather capture a different component of ideology for Republicans and Democrats alike that is less extreme and more reliable. Building an E-score Model Traditional measures of liberal-conservative ideology have long histories in legislative vote models (Grafton & Permaloff, 2005b, p. 173; Levitt, 1996, Nelson & Silberberg, 1987; Kalt & Zupan, 1984, p.281; Friedrich, 1965). Vote selection criteria are well established for identifying votes that depict liberal and conservative ideologies across interest groups. ADA and ACU interest groups select issue areas where roll call votes are cast and support or opposition to the policy is measured in developing a score for each legislator. Legislators casting roll call votes in support of those policies with 337 which the interest group aligns were assigned higher scores from 0-100 and legislators casting roll call votes in opposition to those policies that the interest group supports were assigned lower scores from 0-100. DW Nominate scores represent liberal-conservative ideology but are built upon a model that captures changes in ideology over time as more and more votes are cast. With all three measures highly correlated, that DW Nominate scores include additional roll call votes that are different than roll call votes selected in ADA and ACU models supports each measure of liberal-conservative ideology as a readily available and accepted tool for describing legislative behavior. Measures of economic efficiency are not readily available and building a model required guidelines for standardizing vote selection. In developing legislative E-scores in this model criteria developed by Kennedy (2005, p. 60) and Stigler (1971) were used as a basis for vote selection and evaluation. Public policies for inclusion into an E-score were identified by those policies that either expand efficient policy output or did not produce injurious policies that involve greater social costs relative to social benefits. Public policy formulation that expands output and opportunities available to all must satisfy Pareto optimal conditions for inclusion into an E-score. With the intent of economic efficiency an expansion of output, which leads to greater social welfare, public policies must not hurt others in satisfying these principles. The criteria developed by Kennedy (2005) and Stigler (1971) for building an E-score and applied to analysis reported here consider the following economically inefficient: excise or direct monetary subsidies; regulations that limit competition; policies affecting consumption of goods that are either a substitute or complement; wage and price controls. 338 Using these categories a roll call vote in support or opposition to the selected legislation was evaluated. All legislation selected was identifiable as either enhancing economic efficiency or signaling a reduction in economic efficiency. Roll call votes were tabulated for each legislator voting. Roll call votes in support of economically efficient legislation were recorded as an economic efficiency-enhancing vote by that legislator; roll call votes in opposition to economically efficient legislation were recorded as an economically inefficient vote by that legislator. Roll call votes in support of economically inefficient legislation were recorded as an economically inefficient vote by that legislator; roll call votes by each legislator in opposition to economically inefficient legislation were recorded as economic efficiency enhancing. Legislation selected had to be unambiguous relative to the intent of the roll call vote. For example, when a final roll call vote on legislation contained multiple components the vote was not included because support or opposition to specific components in the legislation might not reflect support or opposition to the bill. The criteria considered not only the title of the legislation, but also the intent of the legislation. Understanding the intent of the legislation was crucial to deciding if a vote in support of the policy position was economically efficient or inefficient. For legislation where intent was unclear or multiple issue areas of the legislation created a bill with some components that were economically efficiency enhancing and some that were not, the legislation was omitted from inclusion into the E-score model. 339 Comparing E-score to liberal-conservative model development In order to better understand steps surrounding vote identification and selection in the E-score model a comparison to ADA and ACU model development is necessary. Differences exist in legislation used in developing an E-score model from that legislation used for liberal-conservative measures of ideology. The criteria for selection of votes are more restrictive for E-score model relative to liberal-conservative models. The number of bills debated and roll call votes cast are comparatively fewer when using more restrictive criteria for selection. Appropriate roll call votes for inclusion in the model in each Congress may be limited. Fewer numbers of votes that are considered in building a model increases the probability of any one vote skewing the results for computing the E-score. Liberal-conservative models consider a broader array of legislation where liberal and conservative principles can be defined. Selected legislation is not restricted to economic output and is more plentiful in each Congress. Perhaps the biggest difference between building liberal-conservative models and an E-score model involves the subjectivity in choosing legislation that meets criteria developed by Stigler (1971) and Kennedy (2005). Legislation considered for vote models can be interpreted differently. With the goal of economic efficiency creating public policies that enhance net social benefits, to the extent that the outcome of the policy is interpreted to produce these results subjectively associated with anticipated outcomes varies. For example, public policies that promote free trade are within the criteria adopted by Stigler (1971) and Kennedy (2005), but free trade involves removing supports for local firms. Free trade is a culmination of economic theory principles that espouse the virtues of an absence of regulation (Stigler, 1971) but without local supports 340 domestic firms may suffer in the short term. How one interprets these effects is subjective and greater subjectivity in selection of votes decreases the accuracy of a vote model to predict changes in behavior. Table 5.1 summarizes comparisons between developing and using liberal-conservative ideology models and an E-score model with relative advantages of each provided. Table 5.1 Summary of Advantages in E-score Vote Model Relative to Liberal- Conservative Vote Models Advantages of Liberal-Conservative Models Advantages of E-score Model Use well established Economic efficiency measures outcomes as opposed to legislator characteristics Vote models function well Economic efficiency is a long standing and an important component of policy analysis Score calculations done by others and easily attainable Ability to analyze economic consequences of political behavior Vote selection may be less subjective than E-score; E-score must often deal with conflicting interpretation of economically efficient policies Less highly correlated with political party Number of roll call votes needed for model development is much lower Less highly correlated with liberal- conservative measures and liberal- conservative measures are highly correlated with each other Measures experience less variation between congresses Added to voting models containing liberal- conservative measures, the resulting models have added explanatory power Economic Efficiency and Vote Models With liberal-conservative ideology consistently having a stronger impact on predicting behavior than E-score in this study, the role of economic efficiency in other vote models must be addressed. That E-score functioned well in base model analyses measured against each dependent variable in this study but less frequently met tests of 341 statistical significance when liberal-conservative ideology and party variables were introduced in House and Senate does not preclude E-score from a role in other vote models. Base model results produced by E-score consistently were equal to or exceeded the impact of self-interest variables in the study. With self-interest a strong component of human behavior (Sears & Funk, 1990; Sen, 1990, p. 29; Buchanan & Tullock, 1962; Downs, 1957, pp. 6-7) the statistical significance of E-score in the model and expected directional impact of its application are arguments for its inclusion in the model to predict legislative behavior. Grafton and Permaloff (2005a, pp. 408-409) cite research that finds few references between ideology and public policy formulation. Including a measure of ideology, such as economic efficiency, to bridge the gap between liberal-conservative positions and why legislators support public policies offers an opportunity to expand the use of ideology in analyzing the need for public policies and the expectation from policy implementation. E-score is a tool for analyzing public policies that is arguably an extension of policy analysis theory for recognizing different paradigms of ideology and identifying weaknesses in those models (Danziger, 1995, pp. 443-444). Analytical perspectives offered from measuring which policies are expected to produce greater social benefit relative to social cost and assigning a score that immediately identifies each legislator?s voting record offers comparisons among legislators, parties, and ideologies not possible from a liberal-conservative spectrum alone. 342 E-score as a Predictor of Behavior in House and Senate The study found economic efficiency to vary relative to its impact in House and Senate. In the Senate the impact of party and liberal-conservative ideology appeared to have stronger predictive qualities than E-score. In the Senate base models for each dependent variable E-score functioned well with self-interest but lost all statistical significance in the medical malpractice model and most statistical significance across Congresses in the minimum wage model, when analyzed with party and ideology variables rotated into the model. This suggests that type of policy and self-interest that are related to economic conditions may affect support for economic efficiency and impact E-scores. In the Senate E-score appears to function better in absolute terms but does not produce statistically significant results when measured relative to party control or ideology over time. Through interrupted time series analysis measuring changes in party control from Democrat to Republican after the 1994 congressional election, the study analyzed if E-score varied with liberal-conservative ideology or predicted a different direction for legislative behavior. E-scores were not statistically significant in either chamber for either dependent variable. E-score was less reliable in analyzing impact of changes in party control in the House. With DW Nominate scores the only variable producing statistically significant results in the House over time an argument that liberal-conservative ideology produces stronger, more reliable predictive ability strengthens. Vandoren?s (1990) argument that pooling of data through time series analysis is necessary for understanding how congressional behavior is a function of policy 343 dimensions appears to hold for these findings. The virtues of E-score were expected to reflect statistically significant associations with a legislators? support for economically efficient legislation. Instead, changes in party control initially disturb liberal- conservative perspectives, but legislators over time are generally less affected by majority-minority status and revert to traditional liberal-conservative ideology, which is closely aligned with party. Mean ADA and ACU scores fluctuated between each measure of ideology more often in the Senate and varied within the period of the study (see Figures 4.1 and 4.2). The effect of liberal-conservative ideology appears to show a different perspective when considered in the Senate relative to the House. E-scores experience a pattern in the Senate also dissimilar to the House. House E-scores increased across the study and Senate E-score experienced more variance. Usefulness of E-score This study finds that E-score is a predictor of legislative behavior in both the House and Senate, but its impact on vote models is relatively less robust than the impact of liberal-conservative ideology and party variables. That E-score is statistically significant in the model and predicts behavior with a consistent directional impact is an argument for thinking of E-score not only as a variable that captures policy dimensions of legislative ideology, but also as a tool for expanding policy discussion. Kennedy (2005) built the E-score model and analyzed its application over two Congresses. The impact of the model and political application of the measure was extended in this study. Kennedy?s (2005) model focused on a legislator?s preference for economic efficiency 344 (pp. 45-56) but was not intended to produce results for measuring the extent that economic efficiency is a component of ideology with applications beyond a liberal- conservative spectrum when analyzed in a multivariate equation with self-interest and party variables controlled by economic and geographic conditions over time. Rhetoric is important in policy discussions at various levels ? between legislators and constituents and media and the public ? that a tool for assessing if a vote in support or opposition to a policy produces a result that can be measured by its social effects is important in capturing legislative voting impact. Introducing a measure of economic efficiency to disparate groups encourages dialogue and makes comparisons between policy positions quantifiable and expands legislative transparency. E-scores could assist in most components of policy analysis. Tying increased dialogue to economic differences in policy positions and measuring legislative behavior through an E-score produces a more open political process where cost-benefit comparisons are made not among liberal-conservative ideologies of legislators, but rather through the production of each legislator in managing public resources and maximizing net social benefits. Cost-benefit comparisons are conducted at many levels of government, the private sector, and in government-business partnerships. To the extent that a legislative E-score has application linking these relationships opens opportunity to use E-score in multiple settings. E-score and Public Policy The Kennedy (2005) study was a primary instrument for this study on economic efficiency and the use of E-score as a predictor of behavior, and is a basis for developing 345 E-score applications useful in studies of political behavior. A key component of the Kennedy model was a derivation of an E-score from standards explaining economically efficient and inefficient public policies. Why legislators support economically efficient policies and why shirking exists when public interests are not consistent with legislative ideology were examined in the Kennedy analysis. The analysis found E-score comparable with other vote models and provided an effective interest group rating alternative. Kennedy?s model made important distinctions between interest group ratings of ideology and E-score that were examined in this dissertation, forming a basis for comparison of E-score to traditional ideology and developing applications for its use in the policy process. Kennedy?s analysis of perceived subjectivity of interest groups in selectively choosing votes for analysis of their respective ideological positions raises questions of subjectivity associated with vote selection in not only an E-score model, but also any model where the meaning and intent of a vote is analyzed. An important difference in traditional interest group rating and E-score that Kennedy (2005) raises is not subjectivity in identifying and selecting scores for analysis, but rather how economic efficiency is a objective criterion and traditional measures of ideology are subjective criteria (p. 35). As a measurable component, E-score represents an unconditionally different aspect to ideology that has a multitude of public policy uses. Comparisons to traditional measures of ideology ? ADA, ACU, and DW Nominate ? are warranted in order to establish if E-score is a statistically significant contributor to vote models and to evaluate impact of its contribution relative to traditional measures. E- score can coexist with traditional measures of ideology in that ideology is complex and 346 should not be considered only in s single dimension. Policy processes offer opportunity to apply the standards that economic efficiency represents to improving policy understanding and discussion. Economic Efficiency and Discourse Opening paths encouraging dialogue in the political process leads to an expansion of democracy that Lasswell envisioned (as cited in Parsons, 1995, pp. 340- 342). Political actors, policy entrepreneurs, leaders of private and public firms, and constituents, to name a few, all benefit from increased discourse that identifies current economic and political issues that need policy consideration. Sabatier (1991) found that demand for increased discourse is a natural extension of feedback from socioeconomic factors within the political science discipline, where links between public policy objectives are increasingly tied to levels of income, education and unemployment levels. Recognizing greater emphasis between social, economic, and political problems within a society and governmental policy decisions, measuring policy effects in economic terms is important in understanding positive and negative impact of policy decisions. Economic efficiency occupies a key role in measuring these policy effects and, through an E-score, identifying trends in policy support or opposition. Flow of ideas in the political process is important for understanding conditions for political change. Persuasive discourse offers the principles for enlarging the role of political discussion where ideas and political persuasion occupy more prominent roles (White, 1994, p. 508). Examining the process could include ways of promoting equity but an increased role of management and scarcity of resources invites measuring cause 347 and effect through benefits relative to costs. Incremental adjustments are the norm in policy change (Linblom, 1959), but promoting new approaches to solving problems entails looking to the process of change and how incremental changes not only produce better policies for addressing policy issues, but also for maximizing the benefits of those policies to constituents and overall to the public. The role that economic efficiency plays in opening dialogue in the policy process provides a bridge between process and interests. While an E-score as a measure of legislative economic efficiency is not concerned with method or procedure, what processes or changes are needed to reach an objective and what outcome is desired naturally fits with the logic of an E-score. Party and liberal-conservative ideology are both strong influences on vote models in predicting legislative behavior, but E-score has a role in separating policy debate into development of policy preferences and bringing together policymakers, analysts, and voters in devising measurable standards for generating policy outputs in reaching desired policy outcomes. Application to Policy Analysis E-score has applications beyond expansion of dialogue in the political process. Using economic efficiency as a measure of benefit-cost analysis has the potential to extend beyond mere evaluative application. Sabatier (1991) finds that the policy process opens opportunities for linking political behavior to clear normative standards for good policy. Applying standards to desired policy outcomes involves linking client specifics with an objective. Economic efficiency is a logical standard that is a positive measure 348 for developing policy strategies, where a Pareto optimal distribution of resources benefits an intended party without unintentionally, negatively impacting another party. Policy solutions are not linear and are often a result of conflicting players inundating decision makers with requests that test rationality. Shifting political issues place demands on legislators that are difficult to measure with liberal-conservative ideology alone. Kingdon?s policy soup analogy (Parsons, 1995, pp. 192-194) is appropriate in describing the environment in which legislators must process information in weighing rational self interests associated with reelections (Downs, 1957) to an ideological base. E-score provides an objective measure for political contributors, fellow members of Congress, and constituents to evaluate relatively each legislator on the merits of support for economically efficient policies. While legislators are rational their legislative decisions are too often a matter of interpretation along liberal-conservative principles. An objective, numeric measure of each legislator produces a standard by which policy formulation can proceed. Heclo?s finding of a loosely organized policy process defined through ideas and policy experts (as cited in White, 1994, p. 515) is an argument for a standard from which policy makers develop problem solving goals and ideas are promoted and shaped. The difficulty in identifying and selecting roll call votes used in developing an E- score is a weakness of the measure. E-scores are based on a narrow set of criteria and roll call votes may or may not be available for each Congress on legislation that meets the criteria. Liberal-conservative measures are readily available and accessible, while E- scores must be explicitly defined and votes gathered. Votes are subjective to the extent that an economically efficient or inefficient policy is a matter of interpretation. E-score 349 is intended as a measure of social benefit maximization, but benefit-cost ratios may be incorrectly interpreted to support narrow benefits to a few. Ironically, interpretation of economically efficient or inefficient policies is defined along a liberal-conservative spectrum and party alliance. An argument for the usefulness of an E-score lends itself to framing of political issues. Preferences are constantly interpreted and ideas and coalitions change within Congresses and between Congresses. Schattschneider provides a compelling argument for management of scope and extent of conflict, where pressure groups, parties and institutions all seek to frame an issue around a cause (as cited in Parsons, 1995, p.126). E-score offers a measure with which these disparate groups can formulate an argument in support of their cause. With framing setting the direction of the agenda, understanding which policies are economically efficient and which are not is important but impractical to a casual observer. A measure of economic efficiency standardizes arguments in support or opposition to a policy on expected outcomes to a society and alleviates barriers to understanding and judging good and bad public policies. Table 5.3 summarizes an application of E-score in legislative vote models and in public and private sector setting outside of the U.S. Congress. 350 Table 5.2 Application Summary of E-score Model in Public Policy Development and Analysis Used in legislative vote models 1. Supplement measures of liberal-conservative ideology with a different dimension of ideology 2. Establish objective criterion for measuring behavior 3. Establish congressional baseline for comparing anticipated policy outcomes and effectiveness 4. Feedback mechanism for legislator commitment to public resource management Used as a measure of economic efficiency 1. Facilitate public policy discourse 2. Extend policy analysis objectives through the use of an economic efficiency measure in expanding public policy net social benefits 3. Introduce a standardized measure for analyzing public resource commitments 4. Measure decision making at state and local government levels 5. Comparison of public policy solutions to private sector needs Conclusion E-score is a statistically significant predictor of behavior in a multivariate analysis of medical malpractice reform and minimum wage legislation. The variable proved valuable in explaining behavior when evaluated with self-interest in a vote model. E-score was less effective in predicting behavior when tested with liberal- conservative ideology positions and party added to the analysis. The relative inability of E-score to predict behavior under these conditions suggests that legislative behavior is driven by a close alignment of liberal-conservative ideology and party principles. Self- interest variables function very well with E-score, but their effect is relatively weak compared to liberal-conservative ideology and party and their directional impact less reliable. 351 Applications of an objective, standardized measure of behavior are potentially numerous in each level of government and public-private relationships. With the policy process too often disjointed incrementalism, Baybrooke and Lindblom?s (as cited in Parsons, 1995, pp. 286-287) approach to the role of policy analysis offers a logical place for E-score in improving the policy formulation process. The use of an E-score is a practical application to long-term planning strategies through assignment of a number that is a reflection of policy outcomes. A quick glance tells stakeholders how a decision maker votes in support of policy and is a tool for gauging the effectiveness of a decision maker to produce desirable outcomes. With economic efficiency the maximization of net social benefits of a public policy, linking outcomes and legislative performance is necessary for improving policy formulation at all levels of government where accountability issues increasingly act as a tool for maximizing legislative resources. Cost, both expected and unexpected, associated with public policy development and formulation force decision makers to increasingly weigh policy options according to costs and benefits of those outcomes. 352 REFERENCES Abramowitz, A. I., & Saunders, K. L. (1998). Ideological realignment in the U.S. electorate. Journal of Politics, 60(3), 634-652. Acemoglu, D., & Robinson, J. A. (2001). Inefficient redistribution. American Political Science Review, 95(3), 649-661. Ainsworth, S. H. (1997). The role of legislators in the determination of interest group influence. Legislative Studies Quarterly, 22(4), 517-533. Aldrich, J. H. & Rohde, D. W. (2005). Congressional committees in a partisan era. In L. C. Dodd and B. I. Oppenheimer (Eds.), Congress reconsidered, (8th ed.), (pp. 249-270). Washington, DC: CQ Press. Alesina, A., & Rosenthal, H. (1989). Partisan cycles in congressional elections and the macroeconomy. American Political Science Review, 83, 373-398. Alesina, A., & Rosenthal, H. (1995). Partisan politics, divided government, and the economy. Cambridge, England: Cambridge University Press. Ansolabehere, S., & Snyder, J. M., Jr. (2000). Money and office: The source of the incumbency advantage in congressional campaign finance. In D. W. Brady, J. F. Cogan, & M. P. Fiorina (Eds.), Continuity and change in House elections (pp. 65-86). Stanford, CA: Stanford University Press. Apter, D. E. (1964). Introduction. In D. E. Apter (Ed.), Ideology and discontent (pp. 1- 46). New York: Free Press. 353 Arrow, K. J., (1963). Social choice and individual values. New York: John Wiley. Atlas, C. M., Gilligan, T. W., Hendershott, R. J., & Zupan, M. A. (1995). Slicing the federal government net spending pie: Who wins, who loses, and why. American Economic Review, 85(3), 624-629. Austen-Smith, D. (1993). Information and influence: Lobbying for agendas and votes. American Journal of Political Science, 37(3), 799-833. Baker v. Carr, 369 U.S. 186 (1962). Barry, B., & Hardin, R. (1982). Individual actions and collective consequences. In B. Barry and R. Hardin (Eds.), Rational man and irrational society? (pp. 19-37). Beverly Hills, CA: Sage. Bennett, C. F. (1979). Analyzing impacts of extension program. Washington, DC: U.S. Department of Agriculture. Bennett, C. F. (1982). Reflective appraisal of programs. Ithaca, NY: Cornell University Media Services. Berkowitz, E., & McQuaid, K. (1978). Businessman and bureaucrat: The evolution of the American social welfare system. Journal of Economic History, 38(1), 120- 142. Bethell, T. (1979). The changing fashions of liberalism. Public Opinion, 2, 41-46. Boadway, R. (1976). Integrating equity and efficiency is applied welfare economics. Quarterly Journal of Economics, 90(4), 541-556. Boozer, B.B., Jr., Westley, C., & Landry, R. J., III. (2007). Medical malpractice reform policy: Why economic efficiency matters. Manuscript submitted for publication. 354 Bozeman, B. (2002). Public-value failure: When efficient markets may not do. Public Administration Review, 62(2), 145-161. Brady, D. W. (1978). Critical elections, congressional parties, and clusters of policy changes. British Journal of Political Science, 8(1), 79-99. Brady, D. W., Donofrio, R., & Fiorina, M. P. (2000). The nationalization of electoral forces revisited. In D. W. Brady, J. F. Cogan, & M. P. Fiorina (Eds.), Continuity and change in House elections (pp. 130-148). Stanford, CA: Stanford University Press. Brown, C., Gilroy, C., & Kohen, A. (1982). The effect of the minimum wage on employment and unemployment. Journal of Economic Literature, 20(2), 487- 528. Brunell, T. L., Koetzle, W., Dinardo, J., Grofman, B., & Feld, S. L. (1999). The R2 = .93: Where then do they differ? Comparing liberal and conservative interest group ratings. Legislative Studies Quarterly, 24(1), 87-101. Buchanan, J. M. (1972). Toward analysis of closed behavioral systems. In J. M. Buchanan and R. D. Tollison (Eds.), Theory of public choice (pp.11-23). Ann Arbor MI: University of Michigan Press. Buchanan, J. M., & Tullock, G. (1962). The calculus of consent. Ann Arbor, MI: University of Michigan Press. Bullock, C. S., III, & Brady, D. W. (1983). Party, constituency, and roll-call voting in the U.S. Senate. Legislative Studies Quarterly, 8(1), 29-43. 355 Burden, B. C., Caldeira, G. A., & Groseclose, T. (2000). Measuring the ideologies of U.S. senators: The song remains the same. Legislative Studies Quarterly, 25(2), 237-258. Burden, B. C., & Kimball, D. C. (2002). Why Americans split their tickets: Campaigns, competition, and divided government. Ann Arbor: University of Michigan Press. Browning, E. K., & Zupan, M. A. (2002). Microeconomics: Theory & applications, (7 th ed.). New York: John Wiley & Sons, Inc. Canon, D. T. (1989). The institutionalization of leadership in the U.S. Congress. Legislative Studies Quarterly, 14(3), 415-443. Cantor, D. M., & Herrnson, P. S. (1997). Party campaign activity and party unity in the U.S. House of Representatives. Legislative Studies Quarterly, 22(3), 393-415. Caro, R. (2002). Master of the Senate. New York: Knopf. Carsey, T. M., & Layman, G. C. (2004). Policy balancing and preferences for party control of government. Political Research Quarterly, 57(4), 541-550. Cater, D. (1964). Power in Washington: A critical look at today?s struggle to govern in the nation?s capital. New York: Vintage Books. Champagne, R. A. (1983). Conditions for realignment in the U. S. Senate, or what makes the steamroller start? Legislative Studies Quarterly, 8(2), 231-249. Clausen, A. R., & Cheney, R. B. (1970). A comparative analysis of Senate-House voting on economic and welfare policy, 1953-1964. American Political Science Review, 64, 138-152. 356 Coate, S. (2000). An efficiency approach to evaluation of policy changes. Economic Journal, 110(463), 437-455. Coleman, J. J. (1999). Unified government, divided government, and party responsiveness. American Political Science Review, 93(4), 821-835. Converse, P. E. (1964). The nature of belief systems in mass publics. In D. E. Apter (Ed.), Ideology and discontent (pp. 206-261). New York: Free Press. Cooper, J. (2005). From congressional dominance to presidential preeminence: Power and politics in late Nineteenth-Century America and today. In L. C. Dodd and B. I. Oppenheimer (Eds.), Congress reconsidered, (8th ed.), (pp. 363-393). Washington, DC: CQ Press. Cooper, J., & Brady, D. W. (1981). Institutional context and leadership style: The House from Cannon to Rayburn. American Political Science Review, 75(2), 411-425. Danziger, M. (1995). Policy analysis postmodernized: Some political and pedagogical ramifications. Policy Studies Journal, 23(3), 435-451. Deckard, B., & Stanley, J. (1974). Party decomposition and region: The House of Representatives, 1945-1970. Western Political Quarterly, 27, 249-264. deLeon, P. (1995). Democratic values and the policy sciences. American Journal of Political Science, 39(4), 886-905. Dodd, L. C. (2005). Re-envisioning Congress: Theoretical perspectives on congressional change ? 2004. In L. C. Dodd and B. I. Oppenheimer (Eds.), Congress reconsidered, (8th ed.), (pp. 411-445). Washington, DC: CQ Press. 357 Dodd, L. C., & Oppenheimer, B. I. (2005). A decade of republican control: The House of Representatives, 1995-2005. In L. C. Dodd and B. I. Oppenheimer (Eds.), Congress reconsidered, (8th ed.), (pp. 23-54). Washington, DC: CQ Press. Downs, A. (1957). An economic theory of democracy. New York: Harper and Row. Downs, A. (1998). The public interest: Its meaning in a democracy. In A. Downs (Ed.), Political theory and public choice (pp. 49-83). Eheltenham, England: Edward Elgar. Drucker, H. M. (1974). The political uses of ideology. New York: Harper and Row. Edwards, G. C., Barrett, A., & Peake, J. (1997). The legislative impact of divided government. American Journal of Political Science, 41(2), 545-563. Erikson, R. S. (1990). Roll calls, reputations, and representation in the U.S. Senate. Legislative Studies Quarterly, 15(4), 623-642. Erikson, R. S., & Wright, G. C. (2005). Voters, candidates, and issues in congressional elections. In L. C. Dodd and B. I. Oppenheimer (Eds.), Congress reconsidered, (8th ed.), (pp. 77-106). Washington, DC: CQ Press. Etzioni, A. (1990). Policy implications of socio-economics. Policy Studies Review, 9(3), 445-454. Fiorina, M. P. (1974). Representatives, roll calls, and constituencies. Lexington, MA: Lexington Books. Fiorina, M. P. (1994). Divided government in the American states: A byproduct of legislative professionalism. American Political Science Review, 88, 304-316. Fiorina, M. P. (1996). Divided government, (2 nd ed.). Needham Heights, MA: Allyn & Bacon. 358 Fiorina, M. P. (2005). Keystone reconsidered. In L. C. Dodd and B. I. Oppenheimer (Eds.), Congress reconsidered, (8th ed.), (pp. 159-179). Washington, DC: CQ Press. Fisher, L. (1998). The politics of shared power: Congress and the executive, (4 th ed.). College Station: Texas A&M University Press. Flanigan, W. H., & Zingale, N. H. (2002). Political behavior of the American electorate (10 th ed.). Washington, DC: CQ Press. Flower, L. L. (1982). How interest groups select issues for rating voting records of members of the U.S. Congress. Legislative Studies Quarterly, 7(3), 401-413. Frantzich, S. (1979). Who makes our laws? The legislative effectiveness of members of the U.S. Congress. Legislative Studies Quarterly, 4(3), 409-428. Freeman, R. B. (1996). The minimum wage as a redistributive tool. Economic Journal, 106(436), 639-649. Friedrich, C. J. (1965). Ideology in politics: A theoretical comment. Slavic Review, 24(4), 612-616. Fry, B. (1989). Mastering public administration: From Max Weber to Dwight Waldo. Chatham, NJ: Chatham House Publishers. Frymer, P. (1994). Ideological consensus within divided party government. Political Science Quarterly, 109(2), 287-311. Grafton, C., & Permaloff, A. (2005a). Liberal and conservative dissensus in areas of domestic public policy other than business and economics. Policy Sciences, 38(1), 45-67. 359 Grafton, C., & Permaloff, A. (2005b). The behavioral study of political ideology and public policy formulation. In C. Grafton and A. Permaloff (Eds.), The behavioral study of political ideology and public policy formulation (pp. 173- 192). Lanham, MD: University Press of America. Grofman, B., Griffin, R., & Berry, E. (1995). House members who become senators: Learning from a natural experiment in representation. Legislative Studies Quarterly, 20(4), 513-529. Groseclose, T., Levitt, S. D., & Snyder, Jr., J. M. (1999). Comparing interest group scores across time and chambers: Adjusted ADA scores for the U.S. Congress. American Political Science Review, 93(1), 33-50. Gujarati, D. N. (1988). Basic econometrics (2 nd ed.). New York: McGraw-Hill Publishing Company. Hager, G. L., & Talbert, J. C. (2000). Look for the party label: Party influences on voting in the U.S. House. Legislative Studies Quarterly, 25(1), 75-99. Hall, R. L., & Evans, C. L. (1990). The power of subcommittees. Journal of Politics, 52(2), 335-355. Hamilton, L. H. (2004). How Congress works and why you should care. Bloomington: Indiana University Press. Hammond, P. J. (1976). Equity, Arrow?s conditions, and Rawls? difference principle. Econometrica, 44(4), 793-804. Haynie, K. L. (2005). African Americans and the new politics of inclusion: A representative dilemma? In L. C. Dodd and B. I. Oppenheimer (Eds.), Congress reconsidered, (8th ed.), (pp. 395-409). Washington, DC: CQ Press. 360 Heinz, J. P., Laumann, E. O., Salisbury, R. H., & Nelson, R. L. (1990). Inner circles or hollow cores: Elite networks in national policy systems. Journal of Politics, 52(2), 356-390. Hibbing, J. R., & Tiritilli, E. (2000). Public disapproval of Congress can be dangerous to majority party candidates: The case of 1994. In D. W. Brady, J. F. Cogan, and M. P. Fiorina (Eds.), Continuity and change in House elections (pp. 110-129). Stanford, CA: Stanford University Press. Hird, J. A. (1991). The political economy of pork: Project selection at the U. S. Army Corps of Engineers. American Political Science Review, 85(2), 429-456. Hojnacki, M., & Kimball, D. C. (1998). Organized interests and whom to lobby in Congress. American Political Science Review, 92(4), 775-790. Holyoke, T. T. (2003). Choosing battlegrounds: Interest group lobbying across multiple venues. Political Research Quarterly, 56(3), 325-336. Hoover, K. R. (2003). Economics as ideology: Keynes, Laski, Hayek, and the creation of contemporary politics. Lanham, MD: Rowman & Littlefield. Ingberman, D., & Villani, J. (1993). An institutional theory of divided government and party polarization. American Journal of Political Science, 37, 429-471. Jacoby, W. G. (1986). Levels of conceptualization and reliance on the liberal- conservative continuum. Journal of Politics, 48(2), 423-432. Jackson, J. E., & Kingdon, J. W. (1992). Ideology, interest group scores, and legislative votes. American Journal of Political Science, 36(3), 805-823. Jones, B. D. (2001). Politics and the architecture of choice: Bounded rationality and governance. Chicago: University of Chicago Press. 361 Kalt, J. P., & Zupan, M. A. (1984). Capture and ideology in the economic theory of politics. American Economic Review, 74(3), 279-300. Kau, J. B., Keenan, D., & Rubin, P. H. (1982). A general equilibrium model of Congressional voting. Quarterly Journal of Economics, 97(2), 271-293. Kellough, J. E. (1990). Federal agencies and affirmative action for blacks and women. Social Science Quarterly, 71(1), 83-91. Kennedy, M. (2005). The determinants of legislators? preferences for economic efficiency. Dissertation Abstracts International, 65(12). (UMI No. 3157289) Kernell, S., & Jacobson, G. C. (2006). The logic of American politics. Washington, DC: CQ Press. Klosko, G., Muller, E. N., & Opp, K. D. (1987). Rebellious collective action revisited. American Political Science Review, 81(2), 557-564. Knuckey, J. (2005). Ideological alignment and partisan change in the American south, 1972-1996. In C. Grafton and A. Permaloff (Eds.), The behavioral study of political ideology and public policy formulation (pp. 43-61). Lanham, MD: University Press of America. Koford, K. (1989). Dimensions in congressional voting. American Political Science Review, 83, 949-975. Kollman, K. (1997). Inviting friends to lobby: Interest groups, ideological bias, and congressional committees. American Journal of Political Science, 41(2), 519- 544. 362 Krauss, M.I. (2003). Tort reform, Cato handbook for Congress: Policy recommendations for the 107th Congress. Retrieved October 15, 2005, from http://www.cato.org/pubs/handbook/hb107/hb107-33.pdf. Krauss, M.I., & Levy, R.A. (2004). Can tort reform and federalism coexist? Policy Analysis, 514, 1-31. Lacy, D., & Niou, E.M.S. (1998). Elections in double member district with nonseparable preferences. Journal of Theoretical Politics, 10, 89-110. Lacy, D., & Paolino, P. (1998). Downsian voting and separation of powers. American Journal of Political Science, 42, 1180-1199. LeGrand, J. (1990). Equity versus efficiency: The elusive tradeoff. Ethics, 100(3), 554- 568. LaPalombara, J. (1966). Decline of ideology: A dissent and interpretation. American Political Science Review, 60(1), 5-16. Lawrence, D. G., & Fleisher, R. (1987). Puzzles and confusion: Political realignment in the 1980s. Political Science Quarterly, 102(1), 79-92. Lee, D. S. (1999). Wage inequality in the United States during the 1980s: Rising dispersion or falling minimum wage? Quarterly Journal of Economics, 114(3), 977-1023. Leyden, K. M., & Borrelli, S. A. (1990). Party contributions and party unity: Can loyalty be bought? Western Political Quarterly, 43(2), 343-365. Levitt, S. D. (1996). How do senators vote? Disentangling the role of voter preferences, party affiliation, and senator ideology. American Economic Review, 86, 425- 441. 363 Lindblom, C. (1959). The science of muddling through. Public Administration Review, 19, 79-88. Maass, A. (1966). Benefit-cost analysis: Its relevance to public investment decisions. The Quarterly Journal of Economics, 80(2), 208-226. Mansbridge, J. J. (1990). Expanding the range of formal modeling. In J. J. Mansbridge (Ed.), Beyond self-interest (pp. 254-263). Chicago: University of Chicago Press. Mayhew, D. R. (1991). Divided we govern: Party control, lawmaking, and investigations, 1946-1990. New Haven, CT: Yale University Press. McArthur, J., & Marks, S. V. (1988). Constituent interest versus legislator ideology: The role of political opportunity cost. Economic Inquiry, 26, 461-470. McFarland, A. S. (1991). Interest groups and political time: Cycles in America. British Journal of Political Science, 21(3), 257-284. McRae, D., Jr. (1954). Some underlying variables in legislative roll call votes. Public Opinion Quarterly, 18(2), 191-196. Mebane, W. R., Jr. (2000). Coordination, moderation, and institutional balancing in American presidential and House elections. American Political Science Review, 94, 37-58. Menefee-Libey, D. (1991). Divided government as scapegoat. PS: Political Science and Politics, 24(4), 643-646. Minar, D. W. (1961). Ideology and political behavior. Midwest Journal of Political Science, 5, 317-331. Moe, T. M. (1984). The new economics of organization. American Journal of Political Science, 28(4), 739-777. 364 Monroe, K. M. (1991). The theory of rational action: Its origins and usefulness for political science. In K. M. Monroe (Ed.), The economic approach to politics (pp. 1-31). New York: HarperCollins. Moscardelli, V. G., Haspel, M., & Wike, R. S. (1998). Party building through campaign finance reform: Conditional party government in the 104 th Congress. Journal of Politics, 60(3), 691-704. Mullins, W. A. (1972). On the concept of ideology in political science. American Political Science Review, 66(2), 498-510. Nelson, D., & Silberberg, E. (1987). Ideology and legislator shirking. Economic Inquiry, 25, 15-25. Oliver, T. R. (1991). Health care market reform in Congress: The uncertain path from proposal to policy. Political Science Quarterly, 103(3), 453-477. Olson, M. (1965). Introduction. In M. Olson (Ed.), The logic of collective action; public goods and the theory of groups (pp. 1-3) Cambridge, MA : Harvard University Press. Ostrom, V. (1989). The intellectual crisis in American public administration. Tuscaloosa: University of Alabama Press. Ostrom, V., & Ostrom, E. (1971). Public choice: A different approach to the study of public administration. Public Administration Review, 13, 203-216. Parsons, D.W. (1995). Public policy: An introduction to the theory and practice of policy analysis. Aldershot, UK: Elgar. Patton, M.Q. (1997). Utilization focused evaluation: The new century text (3 rd ed.). Thousands Oaks, CA: Sage Publications. 365 Peltzman, S. (1985). An economic interpretation of the history of congressional voting in the twentieth century. American Economic Review, 75, 656-675. Polsby, N. W. (1968). The institutionalization of the U.S. House of Representatives. American Political Science Review, 62(1), 144-168. Poole, K. (1981. Dimensions of interest group evaluation of the U.S. Senate, 1969-1978. American Journal of Political Science, 25, 49-67. Poole, K. T. (1988). Recent developments in analytical models of voting in the U.S. Congress. Legislative Studies Quarterly, 13(1), 117-133. Poole, K., & Daniels, S. (1985). Ideology, party and voting in the U.S. Congress, 1959- 1980. American Political Science Review, 79, 373-399. Poole, K., & Rosenthal, H. (1985). A spatial model of legislative roll call analysis. American Journal of Political Science, 29, 357-394. Poole, K. T., & Rosenthal, H. (1997). Are legislators ideologues or agents of constituents? European Economic Review, 40(3-5), 707-717. Poole, K. T. & Rosenthal, H. (2001). D-Nominate after 10 years: A comparative update to Congress: A political-economic history of roll call voting. Legislative Studies Quarterly, 26(1), 5-29. Preston, N. S. (1969). The Senate institution. New York: Van Nostrand Reinhold Company. Putterman, L., Roemer, J. E., & Silvestre, J. (1998). Does egalitarianism have a future? Journal of Economic Literature, 36(2), 861-902. 366 Quirk, P. J. (1990). Deregulation and the politics of ideas in Congress. In J. J. Mansbridge (Ed.), Beyond self-interest (pp. 183-199). Chicago: University of Chicago Press Reichley, A. J. (1981). Conservatives in an age of change. Washington, DC: Brookings Institution. Reichley, J. (1992). The life of the parties: A history of American political parties. New York: Free Press. Reynolds v. Sims, 377 U.S. 533 (1964). Richardson, L. E., & Munger, M. E. (1990). Shirking, representation, and congressional behavior: Voting on the 1983 amendments to the Social Security Act. Public Choice, 67, 11-33. Riley, R. L. (1995). Party government and the contract with America. PS: Political Science and Politics, 28(4), 703-707. Roberts, J. M., & Smith, S. S. (2003). Procedural context, party strategy, and conditional party voting in the U.S. House of Representatives, 1971-2000. American Journal of Political Science, 47(2), 305-317. Rothenberg, L. S., & Sanders, M. S. (2000a). Severing the electoral connection: Shirking in the contemporary Congress. American Journal of Political Science, 44(2), 316-325. Rothenberg, L. S., & Sanders, M. S. (2000b). Lame duck politics: Impending departure and votes on impeachment. Political Research Quarterly, 53(3), 523-536. Rubin, P.H. (1995). Fundamental reform of tort law. Regulation, 4(26-33). Retrieved March 2, 2005, from http://www.cato.org/pubs/regulation/reg18v4b.html. 367 Sabatier, P. A. (1991). Political science and public policy. PS: Political Science and Politics, 24(2), 144-147. Schatteschneider, E. E. (1960). The semi-sovereign people: A realist?s view of democracy in America. New York: Holt, Rinehart, & Winston. Scheve, K., & Tomz, M. (1999). Electoral surprise and mid term the loss in U.S. congressional elections. British Journal of Political Science, 29, 507-521. Schickler, E. (2000). Institutional change in the House of Representatives, 1867-1998: A test of partisan and ideological power balance models. American Political Science Review, 94(2), 269-288. Schickler, E., & Pearson, K. (2005). The House leadership in an era of partisan warfare. In L. C. Dodd and B. I. Oppenheimer (Eds.), Congress reconsidered, (8th ed.), (pp. 207-225). Washington, DC: CQ Press. Schneider, J. E. (1979). Ideological coalitions in Congress. Westport, CT: Greenwood Press. Scholz, J. T., & Wood, B. D. (1999). Efficiency, equity, and politics: Democratic controls over the tax collector. American Journal of Political Science, 43(4), 1166-1188. Schuessler, A. A. (2000). A logic of expressive choice. Princeton, NJ: Princeton University Press. Schwab, L. M. (1988). The myth of conservative shift in American politics: A research note. Western Political Quarterly, 41(2), 817-823. 368 Sears, D. O., & Funk, C. L. (1990). Self-interest in Americans? political opinions. In J. J. Mandsbridge (Ed.), Beyond self-interest (pp. 147-170). Chicago: University of Chicago Press. Segal, J. A., Cameron, C. M., & Cover, A. D. (1992). A spatial model of roll call voting: Senators, constituents, presidents, and interest groups in Supreme Court confirmations. American Journal of Political Science, 36, 96-121. Seltzer, A. J. (1995). The political economy of the Fair Labor Standards Act of 1938. Journal of Political Economy, 103(6), 1302-1342. Sen, A. K. (1990). Rational fools: A critique of the behavorial foundations of economc theory. In J. J. Mansbridge (Ed.), Beyond self-interest (pp. 25-43). Chicago: University of Chicago Press. Shaffer, W. R. (1989). Rating the performance of the ADA in the U.S. Congress. Western Political Quarterly, 42(1), 33-51. Shipan, C. R., & Lowry, W. R. (2001). Environmental policy and party divergence in Congress. Political Research Quarterly, 54(2), 245-263. Silberman, J. I., & Durden, G. C. (1976). Determining legislative preferences on the minimum wage: An economic approach. Journal of Political Economy, 84(2), 317-330. Simon, H. A. (1996). The sciences of the artificial, (3 rd ed.). Cambridge, MA: MIT Press. Sinclair, B. D. (1977). Determinants of aggregate party cohesion in the U.S. House of Representatives, 1901-1956. Legislative Studies Quarterly, 2(2), 155-175. 369 Sinclair, B. (2000). Unorthodox lawmaking: New legislative processes in the U. S. Congress, (2 nd ed.). Washington, DC: CQ Press. Sinclair, B. (2005). The new world of U. S. senators. In L. C. Dodd and B. I. Oppenheimer (Eds.), Congress reconsidered, (8th ed.), (pp. 1-22). Washington, DC: CQ Press. Smith, C. E., Brown, R. D., Bruce, J. M., & Overby, L. M. (1999). Policy balancing and voting for Congress in the 1996 national election. American Journal of Political Science, 41, 1042-1056. Smith, S. S., & Gamm, G. (2005). The dynamics of party government in Congress. In L. C. Dodd and B. I. Oppenheimer (Eds.), Congress reconsidered, (8th ed.), (pp. 181-205). Washington, DC: CQ Press. Snyder, Jr., J. M., & Groseclose, T. (2000). Estimating party influence in Congressional roll-call voting. American Journal of Political Science, 44(2), 187-205. Sobel, R. S. (1999). Theory and evidence on the political economy of the minimum wage. Journal of Political Economy, 107(4), 761-785. Stailey, A.D. (2004). The health act?s same old story, different Congress dilemma: Overhauling the health act and unifying Congress as a remedy for tort reform. Tulsa Law Review, 40, 187-217. Stein, R. M. (1990). Economic voting for governor and U.S. senator: The electoral consequences of federalism. Journal of Politics, 52(1), 29-53. Stigler, G. J. (1965). The politics of political economists. In G. J. Stigler (Ed.), Essays in the history of economics (pp. 51-65). Chicago: University of Chicago Press. 370 Stigler, G. J. (1971). The theory of economic regulation. Bell Journal of Economics and Management Science, 2(1), 3-21. Stigler, G. J. (1975). The citizen and the state: Essays on regulation. Chicago: University of Chicago Press. Stokey, E., & Zeckhauser, R. (1978). A primer for policy analysis. New York: Norton. Strattman, T. (2000). Congressional voting over legislative careers: Shifting positions and changing constraints. American Political Science Review, 94(3), 665-676. Tarrance, V. L., Jr., & DeVries, W. (1998). Checked and balanced: How ticket splitters are shaping the new balance of power in American politics. Grand Rapids, MI: Eerdmans. Thurber, J. A. (1991). Representation, accountability, and efficiency in divided party control of government. PS: Political Science and Politics, 24(4), 653-657. Truman, D. (1960). The governmental process. New York: Alfred A. Knopf. Tuckel, P. (1983). Length of incumbency and reelection chances of U.S. senators. Legislative Studies Quarterly, 8(2), 283-288. Tullock, G. (1983). Economics of income redistribution. Boston, MA: Kluwer-Nijhoff Publishing. Tullock, G. (2002). People are people: The elements of public choice. In G. Tullock, A. Seldon, & G. L. Brady (Eds), Government failure: A primer in public choice. Washington, DC: Cato Institute. Tullock, G., Seldon, A., & Brady, G. L. (2002). Government failure: A primer in public choice. Washington, DC: Cato Institute. 371 Uslaner, E. (1999). The movers and the shirkers: Representatives and ideologues in the Senate. Ann Arbor: University of Michigan Press. Vandoren, P. M. (1990). Can we learn the causes of congressional decisions from roll call data? Legislative Studies Quarterly, 15(3), 311-340. Van Dyke, V. (1995). Ideology and political choice. Chatham, NJ: Chatham House Publishers. Vogel, D. (1980). The public interest movement and American reform tradition. Political Science Quarterly, 95(4), 607-627. Wachs, M. (1989). U.S. transit subsidy policy: In need of reform. Science, 244(4912), 1545-1549. Waldo, D. (1947). The administrative state: A study of the political theory of American public administration. New York: Ronald. Waldo, D. (1965). The administrative state revisited. Public Administration Review, 25(1), 5-12. Waldo, D. (1980). The enterprise of public administration. Novato, CA: Chandler & Sharp. Waterman, R. W. (1990). Institutional realignment: The composition of the U.S. Congress. Western Political Quarterly, 43(1), 81-92. Wattier, M.J., & Tatalovich, R. (2005). Senate voting on abortion legislation over two decades: Testing a reconstructed partisan variable. In C. Grafton and A. Permaloff (Eds.). The behavioral study of political ideology and public policy formulation (pp. 129-148). Lanham, MD: University Press of America. 372 Weimer, D.L., & Vining, A.R. (2005). Policy analysis: Concepts and practice, (4 th ed.). Upper Saddle River, NJ: Prentice Hall. Wesberry v. Sanders, 379 U.S. 1 (1964). White, L. G. (1994). Policy analysis as discourse. Journal of Policy Analysis and Management, 13(3), 506-525. White, Jr., R. D. (1999). More than an analytical tool: Examining the ideological role of efficiency. Public Productivity and Management Review, 23(1), 8-23. Wilcox, C., & Clausen, A. (1991). The rationality of roll call voting reconsidered. Legislative Studies Quarterly, 16(3), 393-406. Winston, C. (1993). Economic deregulation: Days of reckoning for microeconomists. Journal of Economic Literature, XXI, 1263-1289. 373 APPENDICES 374 APPENDIX A Medical Malpractice Legislation Selected As Dependent Variable Table A1. House Action to Amend HR 3174 Bill Title and Synopsis Date of Vote Vote (Yeas-Nays) Military Medical Malpractice Allow active duty armed services members to sue the government for medical and dental malpractice at government facilities. A nay vote is economic efficiency enhancing. 10/7/1985 R 104 ? 71 D 213 ? 19 Adopted 317-90 Table A2. House Action to Amend HR 162 Bill Title and Synopsis Date of Vote Vote (Yeas-Nays) High-Risk Occupational Disease Notification/ Liability Amendment to extend protection from medical malpractice lawsuits to individuals and physicians in high-risk occupations. A yea vote is economic efficiency enhancing. 10/15/1987 R 158 ? 11 D 58 ? 156 Adopted 216-197 Table A3. House Action to Amend HR 956 Bill Title and Synopsis Date of Vote Vote (Yeas-Nays) Product Liability / Medical Malpractice Cap Amendment to cap non-economic pain and suffering damages in all health care liability cases at $250,000 A yea vote is economic efficiency enhancing. 3/9/1995 R 204 ? 21 D 43 ? 149 I 0 ? 1 Adopted 247-171 375 Table A4. House Action on HR 956 Bill Title and Synopsis Date of Vote Vote (Yeas-Nays) Product Liability Cap punitive damages in all civil cases and limit those damages to cases where the plaintiff establishes the defendant intended to cause harm; prohibit product liability for products manufactured and sold more than 15 years ago; revise doctrine of joint and several liability in civil cases; bar compensatory damages if alcohol or drug use is determined to be the primary cause of injury; exempt from punitive damages the makers of drugs or medical devices approved by the Food and Drug Administration; and cap jury awards at $250,000 for non-economic factors in medical malpractice cases. A yea vote is economic efficiency enhancing. 3/10/1995 R 220 - 6 D 45 ? 154 I 0 ? 1 Adopted 265-161. Table A5. House Action on HR 4600 Bill Title and Synopsis Date of Vote Vote (Yeas-Nays) Medical Malpractice Awards/ Passage Passage of bill capping punitive damages that a plaintiff can receive in a medical malpractice case to the greater of $250,000 or double economic damages, with limits on attorneys? contingency fees. Malpractice suits must be filed within three years of injury or one year of its discovery, whichever is earlier, and no punitive damages can be assessed against drug and medical device manufacturers if their products are approved by the Food and Drug Administration or generally considered to be safe. A yea vote is economic efficiency enhancing 9/26/2002 R 203 - 15 D 14 ? 187 I 0 ? 1 Adopted 217?203 376 Table A6. Senate Motion to Table Amendment to HR 956 Bill Title and Synopsis Date of Vote Vote (Yeas-Nays) Product Liability Overhaul/ $500,000 Non-Economic Award Motion to table an amendment limiting non- economic damages for pain and suffering in medical malpractice suits to $500,000. A nay vote is economic efficiency enhancing. 5/02/1995 R 13 ? 41 D 43 ? 3 Motion carried 56?44. Table A7. Senate Motion to Table Amendment to HR 956 Bill Title and Synopsis Date of Vote Vote (Yeas-Nays) Product Liability Overhaul/ Obstetric Services Motion to table an amendment requiring ?clear and convincing? evidence in medical malpractice cases involving labor or delivery of a baby if the physician had not provided prenatal care. A nay vote is economic efficiency enhancing. 5/02/1995 R 10 ? 44 D 29 ? 17 Motion rejected 39 ? 61. Table A8. Senate Vote to Invoke Cloture on HR 956 Bill Title and Synopsis Date of Vote Vote (Yeas-Nays) Product Liability Overhaul/ Cloture Motion to invoke cloture on an amendment limiting non-economic damages for pain and suffering in medical malpractice suits to $500,000. A yea vote for cloture is economic efficiency enhancing. 5/04/1995 R 45 ? 9 D 2 ? 43 Motion failed to garner 60 votes required to invoke cloture and was rejected 47 ? 52. 377 Table A9. Senate Motion to Table Amendment to S 1052 Bill Title and Synopsis Date of Vote Vote (Yeas-Nays) Patient?s Rights/Malpractice Liability Motion to table an amendment exempting health care professionals who provide pro bono medical services to uninsured, indigent individuals from any malpractice liability. A nay vote is economic efficiency enhancing. 6/29/2001 R 2 - 45 D 49 ? 1 I 1 ? 0 Motion carried 52?46. Table A10. Senate Vote to Invoke Cloture on S 1052 Bill Title and Synopsis Date of Vote Vote (Yeas-Nays) Medical Malpractice/Cloture Motion invoking cloture on consideration of a bill capping damage awards in medical malpractice lawsuits against obstetricians and gynecologists. A yea vote for cloture is economic efficiency enhancing. 2/24/2004 R 47 - 3 D 1 ? 41 I 0 ? 1 Motion failed to garner 60 votes required to invoke cloture and was rejected 48 ? 45. 378 APPENDIX B Minimum Wage Legislation Selected as Dependent Variable Table B1. House Action on HR 2 Bill Title and Synopsis Date of Vote Vote (Yeas-Nays) Minimum Wage Increase / Passage Passage of legislation to increase the federal minimum wage from $3.35 an hour to $4.55 an hour over three years and provide a two- month training wage of 85 percent of the minimum for employees who have never held a job. A nay vote is economic efficiency enhancing. 3/23/1989 R 23 - 147 D 225 ? 24 Adopted 248 ? 171. Table B2. House Action to Override Veto of HR 2 Bill Title and Synopsis Date of Vote Vote (Yeas-Nays) Minimum Wage Increase / Veto Override Veto override of HR 2, a bill raising the federal minimum wage from $3.35 an hour to $4.55 an hour over three years and providing a 60 day training wage equal to 85 percent of the minimum for workers who have not worked a total of 60 days. A nay vote is economic efficiency enhancing. 6/14/1989 R 21 - 150 D 226 ? 28 Attempt failed to garner two-thirds majority necessary to override presidential veto and was rejected 247-178. 379 Table B3. House Action on HR 2710 Bill Title and Synopsis Date of Vote Vote (Yeas-Nays) Minimum Wage Increase / Passage Passage of a bill to increase the federal minimum wage from $3.35 an hour to $4.25 an hour over two years and provide a temporary training wage of 85 percent of the minimum for employees aged 16 to 19 years old. A nay vote is economic efficiency enhancing. 11/01/1989 R 135 - 35 D 247 ? 2 Adopted 382 - 37 Table B4. House Action to Amend HR 1227 Bill Title and Synopsis Date of Vote Vote (Yeas-Nays) Employee Commuting / Minimum Wage Increase Amendment to increase the federal minimum wage by 90 cents per hour over two years, thereby raising the minimum wage from $4.25 to $4.75 on July 1, 1996, and to $5.15 per hour on July 1, 1997. A nay vote is economic efficiency enhancing. 5/23/1996 R 77 - 156 D 188 ? 6 I 1 - 0 Adopted 266 - 162 Table B5. House Action to Amend HR 3846 Bill Title and Synopsis Date of Vote Vote (Yeas-Nays) Minimum Wage / Two ? Year Increase Amendment to increase the minimum wage by $1 over two years. A nay vote is economic efficiency enhancing. 3/09/2000 R 42 - 173 D 203 - 5 I 1 - 1 Adopted 246 - 179 380 Table B6. House Action on HR 3846 Bill Title and Synopsis Date of Vote Vote (Yeas-Nays) Minimum Wage / Continued Consideration Continue consideration of minimum wage bill despite point of order that bill constitutes an unfunded mandate. A nay vote is economic efficiency enhancing. 3/09/2000 R 70 - 139 D 199 - 5 I 1 - 1 Adopted 270 - 145 Table B7. Senate Vote to Invoke Cloture on S 837 Bill Title and Synopsis Date of Vote Vote (Yeas-Nays) Minimum Wage Restoration / Cloture Motion to invoke cloture on consideration of a bill raising the federal minimum wage to $4.55 an hour over three years. A nay vote is economic efficiency enhancing. 9/23/1988 R 8 - 32 D 48 - 3 Motion failed to garner 60 votes required to invoke cloture and was rejected 56 ? 35. Table B8. Senate Motion to Amend S 4 Bill Title and Synopsis Date of Vote Vote (Yeas-Nays) Minimum Wage Increase / Graham Amendment Motion to raise minimum wage from $3.35 an hour to $4.55 an hour by September 30, 1991, and to provide a training wage equal to 85 percent of the minimum for new employees who have not worked a total of 60 days and at least 30 consecutive days with one employer. No more than 25 percent of an employer?s workers could be on the training wage at the time, and the training wage provision would expire in September 1992. A nay vote is economic efficiency enhancing. 4/11/1989 R 8 - 37 D 53 - 2 Adopted 61 - 39 381 Table B9. Senate Action on HR 2 Bill Title and Synopsis Date of Vote Vote (Yeas-Nays) Minimum Wage Increase / Passage Passage of the bill to increase the minimum wage from $3.35 an hour to $4.55 an hour over three years and provide a training wage of 85 percent of the minimum for workers with minimal work experience. A nay vote is economic efficiency enhancing. 4/12/1989 R 10 - 35 D 52 - 2 Adopted 62 - 37 Table B10. Senate Action on HR 2710 Bill Title and Synopsis Date of Vote Vote (Yeas-Nays) Minimum Wage Increase / Passage Passage of a bill for increasing the minimum wage from $3.35 an hour to $4.25 an hour over two years and providing for a temporary training wage of 85 percent of the minimum wage for employees aged 16 to 19 years. A nay vote is economic efficiency enhancing. 11/08/1989 R 36 - 8 D 53 - 0 Adopted 89 - 8 382 APPENDIX C Roll Call Votes Selected by House and Senate in Compiling an E-score for Legislators Table C1. House Votes in 99 th Congress Legislation Bill Number Narrative Economic efficiency Textile Import Quotas (rule) HR 1562 Limit textile / shoe imports Nay Textile Import Quotas HR 1562 Passage imposing quotas (1986) Nay Omnibus Trade Bill HR 4800 Strike sections increasing duties Yea South Africa Sanctions HR 4868 Imposing economic sanctions Nay Textile Import Quotas HR 1562 Passage imposing quotas (1985) Nay Omnibus Trade Bill (Roth amend) HR 4800 Strike easing restrictions exports Yea Omnibus Trade Bill (passage) HR 4800 Increased trade restrictions Nay Table C2. Senate Votes in 99 th Congress Legislation Bill Number Narrative Economic efficiency Textile Import Quotas HR 1562 Import quotas Nay Textile Import Quotas (Thurmond amend) HR 1562 Table adding quotas Yea South Africa Sanctions S2701 Table striking a ban on imports Nay South Africa Sanctions (passage) HR 4868 Sanctions banning imports Nay South Africa Sanctions (passage over veto) HR 4868 Passage imposing sanctions Nay 383 Table C3. House Votes in 100 th Congress Legislation Bill Number Narrative Economic efficiency Textile / Apparel Trade Act (rule) HR 1154 Limit textile / apparel imports Nay Textile / Apparel Trade Act (passage) HR 1154 Limit textile / apparel imports Nay Transportation Appropriations / Airlines HR 4794 Eliminate subsidy small community Yea Textile / Apparel Trade Act (veto override) HR 1154 Passage over veto limiting exports Nay Table C4. Senate Votes in 100 th Congress Legislation Bill Number Narrative Economic efficiency Omnibus Trade Bill / Non- agricultural S 1420 Foreign subsidies Yea Omnibus Trade Bill / Newsprint Tariff S 1420 Table amendment exempting tariff Yea Omnibus Trade Bill (veto override) HR 3 Rejected: unfair trade / import damage Nay Retail Competition S 430 Vertical price fixing illegal Yea Textile Import Quotas / Footwear S 2662 Table amendment striking footwear quota Nay Textile Import Quotas / Profitability S 2662 Table amendment suspending quotas Nay 384 Table C5. House Votes in 101 st Congress Legislation Bill Number Narrative Economic efficiency Dairy Price Supports HR 3950 Increase price support Nay Textile Trade Act (rule) HR 4328 Limit growth of textile imports Nay Textile Trade (veto override) HR 4328 Rejected: Limit growth of textile imports Nay Textile Trade (concur in Senate) HR 4328 Limit textile imports; establish quotas Nay Hungary Most Favored Nation (rule) HR 1594 Adoption rule extend MFN 3 years Yea Hungary Most Favored Nation (passage) HR 1594 Extend MFN 3 years Yea Hungary Most Favored Nation (passage) HR 1594 Suspend rule extend MFN 5 years Yea Export Administration Act / Telecommunications HR 4653 Strike provisions easing export restrictions Nay Table C6. Senate Votes in 101 st Congress Legislation Bill Number Narrative Economic efficiency Gas Price Decontrol (passage) HR 1722 Eliminate price and non- price controls natural gas Yea Gas Price Decontrol / Price Escalator HR 1722 Table price escalator natural gas clauses Yea Honey Price Supports S 2830 Price support for next 4 years Nay Miscellaneous Tariffs HR 1594 Suspend duty ulcer treating drug Yea Sugar Price Supports S 2830 Table extending sugar price supports Yea Textile Trade Act / Consumer Costs HR 4328 Table reporting increased consumer costs Nay Textile Trade Act / GATT HR 4328 Table amendment supporting GATT Nay Textile Trade Act (passage) HR 4328 Limit textile imports Nay 385 Table C7. House Votes in 102 nd Congress Legislation Bill Number Narrative Economic efficiency Miscellaneous Tariff (rule) HR 4318 Increase tariff on minivans Nay Miscellaneous Tariff (passage) HR 4318 Increase tariff on minivans Nay Cable TV Regulation / Conference Report S 12 Cap cable rates / FCC set rates Nay Cable TV Regulation (passage) S 12 Stronger FCC setting rates Nay Cable TV Regulation (veto override) S 12 FCC sets cable rates Nay Striker Replacement (passage) HR 5 Prohibit employers hiring replacements Nay Table C8. Senate Votes in 102 nd Congress Legislation Bill Number Narrative Economic efficiency Central American Free Market Policies S 100 U.S. assistance to promote free trade Yea Cable TV Regulation / Conference Report S 12 Cap cable rates, increase FCC role Nay Cable TV Regulation (passage) S 12 Cable regulation Nay Cable TV Regulation (veto override) S 12 Passed increased FCC authority Nay National Energy Policy (cloture) S 1220 Limit debate on CAF? standards Yea 386 Table C9. House Votes in 103 rd Congress Legislation Bill Number Narrative Economic efficiency 1872 Mining Law (passage) HR 322 Increase environmental regulations / royalty Nay NAFTA Implementation (rule) HR 3450 Waive points of order and approve Yea NAFTA Implementation (passage) HR 3450 Approve NAFTA Yea Interstate Commerce Commission HR 2750 Eliminate ICC Yea GATT Implementation (rule) HR 5110 Rule for House floor vote to implement Yea GATT Fast Track Extension (passage) HR 1876 Negotiate strengthening GATT Yea GATT (passage) HR 5110 Implement GATT, reduce tariffs Yea China MFN Executive Order HR 4590 Amendment codifying granting MFN Yea China MFN HR 4590 Amendment denying MFN Nay Amtrak Appropriations HR 2750 Cut funding to Amtrak Yea Table C10. Senate Votes in 103 rd Congress Legislation Bill Number Narrative Economic efficiency Transportation Appropriations / ICC HR 2750 Table eliminating ICC funding Nay GATT / Budget Waiver HR 5110 Negotiate to implement GATT Yea GATT Fast Track Extension (passage) HR 1876 GATT accord under fast track Yea GATT (passage) HR 5110 Implement GATT, reducing tariffs Yea Agriculture Market Promotion HR 2493 Table amendment eliminating funding Nay 387 Table C11. House Votes in 104 th Congress Legislation Bill Number Narrative Economic efficiency Regulatory Moratorium / Small Business HR 450 Extend moratorium on small business regulations Yea Regulatory Moratorium / Telemarketing HR 450 Exempt from moratorium telemarketing Nay Regulatory Moratorium / Competitiveness HR 450 Exempt regulations benefiting U.S. firms Nay Telecommunications (rule) HR 1555 Remove telecommunication regulations Yea Telecommunications / Conference Report S 652 Promote competition and deregulation Yea Telecommunications (passage) HR 1555 Promote competition and deregulation Yea Table C12. Senate Votes in 104 th Congress Legislation Bill Number Narrative Economic efficiency Telecommunications / Barriers to Entry S 652 Strike authority pre-empting local regulations Nay Telecommunications / Conference Report S 652 Promote competition and deregulate Yea Telecommunications (passage) S 652 Promote competition and deregulate Yea Regulatory Overhaul S 343 Amendment increasing cost- benefit analysis threshold Nay Repeal Alaska Oil Import Ban (passage) S 395 Lift ban on export crude oil Yea Independent Regulatory Agency S 1 Table consideration of bills administered by independent regulatory agency Yea 388 Table C13. House Votes in 105 th Congress Legislation Bill Number Narrative Economic efficiency Steel Imports (passage) In HR Increase in steel imports is problem Nay Sub-Saharan Africa Trade (rule) HR 1432 New trade policies with these countries Yea Sub-Saharan Africa Trade (passage) HR 1432 Duty free trade policies Yea Amtrak subsidies / Labor Protection HR 2247 Limit labor protection; increase contracting Yea Caribbean / Central American Trade HR 2644 Suspend rules pass duty-free trade Yea Normal Trade Relations China HJ Res 121 Denying normal trade relations (MFN) Nay Fast-Track Authority (passage) HR 2621 Expedited implementation trade Yea Table C14. Senate Votes in 105 th Congress Legislation Bill Number Narrative Economic efficiency Tobacco Restrictions S 1415 Increase tobacco regulations Nay Tobacco Restrictions (cloture) S 1415 Cloture on bill increasing restrictions Nay Tobacco Restrictions / Remove Provision S 1415 Table striking provisions increasing tobacco taxes Nay Ethanol Tax Break S 1173 Remove extending ethanol tax break Nay Economic Sanctions S 2159 Table requiring 45 days notice implement sanctions; future sanctions end 2 years Nay 389 Table C15. House Votes in 106 th Congress Legislation Bill Number Narrative Economic efficiency Sub-Saharan Africa Trade HR 434 Grant duty free status Yea Steel Imports (passage) HR 975 Impose quotas, tariff surcharges Nay Regulatory Cost-Benefit Analysis (passage) HR 1074 OMB to make annual cost- benefit analysis assessing impact federal regulations Yea OSHA Ergonomics Regulations (passage) HR 987 Restrict issuing new ergonomic rules Yea Disapprove Normal Trade Relations with China (passage) HJ Res 57 Reject extension normal trade relations Nay Africa, Caribbean Trade (rule) HR 434 Extend tariff benefits to those nations Yea China Trade (passage) HR 4444 Make normal trade relations permanent Yea Table C16. Senate Votes in 106 th Congress Legislation Bill Number Narrative Economic efficiency Steel Import Quotas (cloture) HR 975 Limit debate on imposing quotas Yea Democratic Emergency Farm Aid S 1233 Table extending farm aid Yea Compromise Emergency Farm Aid S 1233 Table extending emergency farm aid Yea China Trade / Import Relief HR 4444 Market disruption import relief Nay Food / Medicine Sanctions S 1233 Table ending sanctions Nay Fuel Efficiency Standards HR 2084 Study raising CAF? standards Nay Steel, Oil, Gas Loan Guarantee (cloture) HR 1664 Limit debate on loan guarantees Nay Steel, Oil, Gas Loan Guarantee (passage) HR 1664 Establish loan guarantees Nay China Normal Trade Relations / Discharge SJ Res 27 Reverse extending normal trade relations Nay 390 Table C17. House Votes in 107 th Congress Legislation Bill Number Narrative Economic efficiency Farm Bill / Sugar Subsidy HR 2646 Reduce loan rates for raw sugar cane Nay Vietnam Trade (passage) HJ Res 101 Disapprove normal trade relations Nay Trade Promotion Authority (conference) HR 3009 Extend duty free status Columbia, Peru, Bolivia, and Ecuador Yea Trade Promotion Authority / Rule HR 3009 Adopt rule (H Res 509) for floor consideration of trade promotion and duty free status Yea Trade Promotion Authority / Rule HR 3005 Adopt rule (H Res 306) expediting floor consideration of trade negotiations between executive branch and foreign government Yea Trade Promotion Authority / Passage HR 3005 Expedited negotiations between executive branch and foreign governments promoting trade Yea Table C18. Senate Votes in 107 th Congress Legislation Bill Number Narrative Economic efficiency Farm Bill / Sugar Program S 1731 Table phasing out of sugar support Nay Andean Trade / Tariff Reduction HR 3009 Table prohibiting reduction in tariffs Yea Andean Trade / Motion to Proceed HR 3009 Motion to proceed to bill extending duty free status Yea Andean Trade / Passage HR 3009 Extend duty free status to products from Columbia, Peru, Bolivia, and Ecuador Yea Vietnam Trade / Passage HJRES 51 Normal trade relations with Vietnam Yea 391 Table C19. House Votes in the 108 th Congress Legislation Bill Number Narrative Economic efficiency U.S. ? Chile Trade (passage) HR 2738 Reduce tariffs and trade barriers Yea U.S. ? Australia Trade (rule) HR 4759 Floor consideration reducing barriers Yea U.S. ? Australia Trade (passage) HR 4759 Reduce tariffs and trade barriers Yea Miscellaneous Tariff Reductions HR 1047 Reduce / eliminate tariff on 300 chemicals Yea U.S. ? Morocco Trade (rule) HR 4842 Floor consideration reducing barriers Yea U.S. ? Morocco Trade (passage) HR 4842 Reduce U.S. ? Moroccan trade barriers Yea Gasoline Price Reduction HR 4545 Waivers from fuel additive requirement Yea Table C20. Senate Votes in the 108 th Congress Legislation Bill Number Narrative Economic efficiency U.S. ? Morocco Trade (passage) S 2677 Extend duty free to most products and reduce tariffs to other Morocco products Yea U.S. ? Australia Trade (passage) HR 4759 Extend duty free access and reduce tariffs Yea Fiscal 2005 Defense Authorization / U.S Foreign Subsidiaries S 2400 Extend restrictions on transactions of U.S. companies that do business with countries that sponsor terrorism to foreign subsidiaries where the U.S. firm owns 50 percent or more of company Nay Miscellaneous Tariff and Trade (cloture) HR 1047 Motion to invoke cloture on conference report on a bill suspending duties Yea