Rising Health Care Costs and the Two Price Market
The Impact of Third-Party Payers
Except where reference is made to the work of others, the work described in this thesis is
my own or was done in collaboration with my advisory committee. This thesis does not
include proprietary or classified information.
Joshua J. Robinson
Certificate of Approval:
Michael Stern
Assistant Professor
Economics
Richard O. Beil, Chair
Associate Professor
Economics
T. Randoph Beard
Professor
Economics
George T. Flowers
Interim Dean
Graduate School
Rising Health Care Costs and the Two Price Market
The Impact of Third-Party Payers
Joshua J. Robinson
A Thesis
Submitted to
the Graduate Faculty of
Auburn University
in Partial Fulfillment of the
Requirements for the
Degree of
Master of Science
Auburn, Alabama
December 17, 2007
Rising Health Care Costs and the Two Price Market
The Impact of Third-Party Payers
Joshua J. Robinson
Permission is granted to Auburn University to make copies of this thesis at its
discretion, upon the request of individuals or institutions and at
their expense. The author reserves all publication rights.
Signature of Author
Date of Graduation
iii
Vita
Joshua J. Robinson, son of James C. and Valerie Robinson, was born April 3, 1984
in Fairfield, AL. He married his wife, Rachelle, on August 9, 2003. He graduated Summa
Cum Laude from Huntingdon College on May 6, 2006 with a double major in Mathematics
and Business Administration with a concentration in Economics and Finance. In August
of 2006, Josh entered the graduate school at Auburn University. He has been accepted to
pursue his PhD in Economics at Emory University in Atlanta, GA.
iv
Thesis Abstract
Rising Health Care Costs and the Two Price Market
The Impact of Third-Party Payers
Joshua J. Robinson
Master of Science, December 17, 2007
(B.A. Huntingdon College?Montgomery, 2006)
34 Typed Pages
Directed by Richard O. Beil
Martin Feldstein wrote several important papers in the 1970s concerning the reciprocal
growth of health care inflation and insurance levels. This paper attempts to develop a
new theoretical model based on a two-price market that will explain the reciprocal nature
discovered by Feldstein, but is not dependent purely on moral hazard to explain increased
costs. A general two price model is developed, and then applied to the health care industry.
v
Acknowledgments
I would like to thank my wife, Rachelle, first and foremost for her love, support, and
patience. I would also like to thank my parents for there encouragement and support
throughout my academic career. Lastly, I would like to thanks my advisor, Richard Beil,
for his guidance and contributions to this project.
vi
Style manual or journal used Journal of Approximation Theory (together with the style
known as ?aums?). Bibliograpy follows the American Economic Review style.
Computer software used The document preparation package TEX (specifically LATEX)
together with the departmental style-file aums.sty. Graphics preparation done using
Grapher and Seashore.
vii
Table of Contents
List of Figures ix
1 Introduction 1
2 The Theory of a First Mover Monopsonist 6
3 Insurance as the First Mover Monopsonist 11
3.1 Insurance: Production and Profit . . . . . . . . . . . . . . . . . . . . . . . . 12
3.2 The Competitive Fringe . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
4 Results and Applications 16
4.1 Third Party Payers and Cost . . . . . . . . . . . . . . . . . . . . . . . . . . 16
4.2 Agency and Insurance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
5 Conclusion 22
Bibliography 25
viii
List of Figures
2.1 First Mover Monopsonist with a Competitive Fringe . . . . . . . . . . . . . . . . . 9
4.1 Continual Feedback Resulting from Third-Party Payer . . . . . . . . . . . . . . . . 17
ix
Chapter 1
Introduction
Rising costs in the health care industry have been a contentious economic and political
topic for number of years. In 2004, nearly 1.9 billion dollars, was spent on health care
in the United States, an increase of nearly 8 percent from the previous year. Per capita
expenditures on health care have increased from 148 dollars in 1960 to over 6000 dollars
in 2004. Moreover the health sector is absorbing an ever increasing percentage of GDP. In
1960 health care expenditures accounted for only 5 percent of GDP; current figure is about
16 percent (National Center Health Statistics 2006) .
Although we can track health care inflation quite well, there is little consensus as to
why these costs continue to accelerate. It is often the case that policy analysts and political
scientists blame greedy hospital administrators and physicians (Brewster and Stowers 2004).
While this may be the case in some instances, it is unlikely to be a major factor. Martin
Feldstein (1971, 1973) did research on hospital cost inflation?which remains the largest
component of health care expenditures (BCBS 2007). He restricted his studies to not-
for-profit hospitals, which rules out greed a determinate, yet there were still massive cost
increases. Technological advances and the development of new drugs are also cited as
sources of cost increases (Brewster and Stowers 2004). However as both Feldstein (1971)
and Newhouse (1996) point out, factors like greed and technological advancement only
explain how costs increases. They do not explain why costs were able to increase or why
the trend of inflation continues. If we can understand the why, we may be able to check
this inflation in time. So what is the why?
1
Feldstein published a series of papers in the 1970s examining health care inflation. He
first showed that a permanent excess demand for physician services allows doctors a great
deal of control over the quantity of care provided and the price charged. Consequently, as
the patient?s ability to pay increased, physicians? fees increased (Feldstein 1970). In another
paper, he examined price increases in the not-for-profit hospital industry. He found that
increases in the demand for hospital care allowed hospitals to increase prices, which, due
to their non-profit nature, translated into increases in the quality and sophistication of the
care provided (Feldstein 1971). These factors were Feldstein?s ?how?, but why were patients
becoming increasingly able to pay physicians? And why was the demand for hospital care
increasing? Feldstein?s answer was third-party payers.
Feldstein (1973) discovered a statistical relationship between the percentage of health
care expenditures paid for by third-party payers and the price of care. It is an endogenous
relationship. An increase in insurance increases utilization of care, which, in turn, increases
price (assuming a fixed supply). On the other hand, Feldstein also found empirical evidence
to suggest that an increase in price increases the demand for insurance1. Additionally, he
found that Americans were over-insured by a significant degree, and there was room for
large welfare gains by reducing the level of insurance2.
1There are two factors at play here. First, the increase in price increases net health care expenditures,
which increases the demand for insurance. However, the increase in price lowers the quantity of medical care
demanded, which would reduce the demand for insurance. Feldstein found that the former effect dominated
the latter. In a later study (Feldstein 1977), he discovered that the quality change resulting from the price
increase causes and outward shift in the demand for medical care. This exacerbates the direct relationship
between price and insurance.
2This study was repeated by Feldman and Dowd (1991). They found price-elasticities that were lower
than Feldstein?s estimates, and used an estimate of absolute risk-aversion that was much higher than the
one used in Feldstein?s model. They found that Americans were not as over-insured as Feldstein estimated;
thus the welfare loss of excess insurance was smaller, but still significant.
2
In a more recent study, the growth of per capita medical spending in the US from 1960
to 1993 was examined empirically (Peden and Freeland 1998). They included explanatory
variables such as the coinsurance rate3, income, and technology. They found that coinsur-
ance levels were responsible for almost half of growth in per capita medical spending in the
34 year period. Moreover, from 1983 to 1993 the low coinsurance level accounted for 60 to
65 percent of the spending growth.
The question now becomes what is it about introducing a third-party payer into the
medical care market that causes such massive inflation? Many economists try to explain
this phenomenon with the notion of moral hazard. Because insurance effectively lowers the
price of medical care to almost zero at the time of utilization, insured individuals will likely
increase utilization of health services until their marginal utility of medical care is near zero
as well4(Feldstein 1971, Newhouse 1996). Additionally there is an incentive compatibility
problem (See Section 4.2) between the doctor and the insurance company. The insurer hires
a physician as its agent to provide health care to its customers, the patients, for as little cost
as possible. However, because his own profit is also part of the doctor?s utility function, there
is incentive to over-provide care. Schneider and Mathios (2006) showed both theoretically
and empirically that this was the case with traditional fee-for-service insurance?even when
monitoring efforts are employed. Feldstein (1970) found empirical evidence that physicians
actually raise fees when insurance levels increase. The over-utilization by both patient and
3Here defined as the percent of per capita spending not paid for by third-party payers (government or
private insurance) nor subsidized through tax breaks.
4That is, the additional utility gained by obtaining one more unit of health times the additional health
resulting from consuming one more unit of medical care. This is based on the derived demand for medical
care (see Santerre and Neun 2006). The consumer utilizes an optimal level of medical care when the marginal
utility of medical care divided by the price of medical care is equal to the marginal utility of all other goods
divided by their respective prices. Thus, assuming marginal utility is downward sloping, a lower price would
mean it would be optimal to lower the marginal utility by increasing consumption of the good.
3
doctor is often collectively referred to as moral hazard. Newhouse (1996) cites there is strong
evidence of moral hazard in the American health care system. In a randomized experiment
he conducted, patients with full coverage insurance spent about 40 percent more than those
with a high deductible.
Can this really be the whole story, though? Schneider and Mathios (2006) showed
that moral hazard could be effectively held in check by using capitation, an agreement
between insurers and physicians in which physicians engage in supply-side cost sharing.
However, despite health maintenance organizations (HMO) and preferred provider orga-
nizations (PPO), which engage in a form of capitation, becoming the dominant trends in
insurance throughout the 1990s (Brewster and Stowers 2004), medical costs have continued
to rise and continue to swallow a larger and larger percent of GDP in the US.
In this paper, I will posit a new theoretical model for health care inflation. Like
Feldstein?s models (1970, 1971, 1973, 1977), the model presented here will center around the
reciprocating relationship between increasing insurance levels and price of care. However,
unlike Feldstein, my model is based on a bifurcated price market.
There is no such thing as a uniform price in the medical care industry. According
to a Pricewaterhouse Copper (PWC) (2005) report, large private and government insurers
negotiate with medical care providers to pay only a fraction of the list price of the procedure,
while uninsured individuals often pay list price. In fact, until 2004 hospitals could not offer
discounts to the uninsured without it affecting their Medicare reimbursement. The gap
between the list price and the negotiated fee can be extremely large; recent data gathered
about prices for appendectomy procedures (PWC 2005) showed that managed care providers
and government insurance programs typically pay less than a third of the hospital list price
4
on average. I will argue that the existence of a two price market is itself a source of perpetual
medical cost inflation.
The rest of this paper is organized as follows: in Chapter 2, I will posit the general
theoretical model for a two-price market. In Chapter 3, I will apply this model to the
medical care market. In Chapter 4, I will discuss the implications of the model developed.
And in Chapter 5, I will provide some concluding thoughts.
5
Chapter 2
The Theory of a First Mover Monopsonist
In this chapter the theoretical framework of a two-price market is developed1. In this
market there are two buying blocks: a large firm with significant monopsony power and
a competitive fringe of small buyers. In this model the mere existence of the large buyer
creates a market with two prices. In later sections, this model will be applied to the health
care industry, and I will argue that the two price market is a major source of health care
inflation.
Assume there exists a firm with significant monopsony power, and assume there is a
small fringe of buyers that behave as perfect competitors. Additionally, we will assume that
part of the dominant buyer?s monopsony power includes the ability to act as a first mover.
That is, all buying decisions made by the monopsonist occur before the buying decisions
of the fringe. The large buyer produces a good, X, using a variable input, q, and a fixed
input, f, according to the production function
X = X(q,f) (2.1)
Since the firm has monopsony power, it is not a price taker with respect to the variable
input. Thus the purchasing price, pd, for q is a function of the amount of the variable input
used.
pd = pd(q) (2.2)
1Blair and Harrison (1993) is used as a reference for the theory presented in this section
6
The function pd(q) comes from the supply curve for q. Thus we can assume a positive slope
or ?pd(q)?q ? 0.
The firm?s goal is to purchase inputs in such a way as to maximize profit, given by the
function
pi = rho1?X(q,f)?[pd(q)?q +??f] (2.3)
where rho1 is the output price2, ? is the fixed input price, and f is the amount of fixed input.
Since f is invariant in the short-run, the firm will maximize profit by adjusting the variable
inputs. The first order conditions are
?pi
?q = p?
?X
?q ?
bracketleftbigg
pd(q) +q? ?pd(q)?q
bracketrightbigg
= 0 (2.4)
Since ?X?q is the marginal product of the variable input,
parenleftBig
p? ?X?q
parenrightBig
is the marginal revenue
product?that is, the additional revenue generated by a one unit increase in q. Since the
dominant firm is buying all of its inputs at the same time, the purchase of an additional
unit of q will increase cost by pd(q) as well as by the additional amount that must be paid
for all other units of q as a result of ?pd(q)?q being positive. The sum of these two effects is
know as the marginal factor cost.
MFCq = pq(q) +q? ?pd(q)?q (2.5)
2Notice that if the firm has monopoly power as well as monopsony power that this value will not be
exogenous. Rather it will be a function of the quantity of X sold. This will be explored further in Section 3
7
The first order conditions of the firm show in order to maximize profit the quantity of
variable inputs must be chosen such that marginal revenue product equals the marginal
factor cost.
So far this is standard monopsony theory. Now we must consider the competitive fringe.
Since the dominant firm is allowed to make its buying decisions first, the competitive fringe
faces a new vertical (price) axis set at the quantity, qd, the dominant firm has chosen. That
is, the price the competitive fringe will pay is determined by an altered inverse demand
function.
pf = p(q +qd) (2.6)
Where pf is the price the fringe will pay, q is the quantity the fringe will buy, and qd is the
quantity the dominant firm has chosen to purchase. Since the price of q is an increasing
function of the amount of the input purchased, this will result in the competitive fringe
paying more of each unit ofq than the dominant firm. These results are displayed graphically
in Figure 2.1.
The green marginal revenue product (MRP) curve is the monopsonist?s derived demand
for the input good; it represents the additional revenue gained from employing each addition
input. The red marginal factor cost (MFC) curve is the incurred cost of employing each
additional input. The monopsonist chooses a quantity of the input good such that its profit
is maximized. As suggested by the first order conditions, this occurs when the MRP curve
intersects the MFC curve at QDom.. The suppliers are willing to sell a quantity of QDom.
at a price of PDom., which is the price the monopsonist pays. Once the monopsonist has
made its purchasing decisions, the competitive fringe enters the market?their demand is
represented by the blue curve labeled D. The monopsonist?s ability to act as a first mover
8
Figure 2.1: First Mover Monopsonist with a Competitive Fringe
creates a new vertical ?axis? for the competitive fringe; graphically shifting the demand
curve outward3. Since the fringe is perfectly competitive, it has no negotiating power.
Thus the price paid by the fringe is determined by its demand curve?s intersection with the
supply curve, which occurs at PFringe. The result is the monopsonist paying one price, and
the competitive fringe paying another, higher price.
In the next chapter, this general model will be adapted to the health care market.
Large third-party payers, like the government and large insurance corporations, will be the
3Although graphically the demand curve is shifting outward, the monopsonist has no direct effect on the
demand of the competitive fringe. Effectively what is occurring is the supply available to the fringe is being
reduced by the monopsonist?s purchases.
9
monopsonist. Those individuals not represented by third-party payers will be the com-
petitive fringe. It will be shown that the unique dependent relationship between these two
buying blocks creates much more dramatic results than those of the general model presented
above.
10
Chapter 3
Insurance as the First Mover Monopsonist
Now let us consider how the model presented in the previous chapter applies to the
health care industry. For consumers, health care is a variable input that they use to produce
personal health. However, in the United States only 13 percent of health care expenditures
came directly out of pocket from the consumer. Most consumers contract the purchase of
health care out to third-party payers in the form of purchasing health insurance. According
to a 2005 US Census Bureau estimate, 68 percent of all Americans were covered by some
form of private insurance. Many consumers? health care expenditures are covered by the
federal and state governments in the form of public health insurance, like Medicare or
Medicaid. The US Census Bureau?s estimates say that a little more than 27 percent of
Americans are covered by some form of publicly funded insurance. According to the Blue
Cross Blue Shield (BCBS)Medical Cost Reference Guide (2007), private insurance pays for
about 35 percent of national health care expenditures, and publicly funded insurance pays
for about 46 percent of the health care bill.
By anyone?s standard the federal government, who accounts for nearly half of all expen-
ditures in the medical care industry, has significant monosony power. Blue cross Blue Shield,
a private corporation, insures one third of Americans?the Alabama branch of BCBS covers
over three quarters of the state?s residence. Additionally, the prevailing trend in private
insurance is to use HMO and PPO style reimbursement. These aggressive cost manage-
ment efforts allow private insurer to pay reimbursements similar to those of Medicare and
11
Medicaid (PWC 2005). Hence it is realistic to assume that third-party payers in the health
care industry have significant monopsony power.
Here we will assume a representative medical insurer is our monopsonist, and the some
20 percent of medical care paid for with private funds is our competitive fringe.
3.1 Insurance: Production and Profit
The insurance company produces medical care coverage. That is, it allows consumers
to pre-pay a fixed amount, called a premium, for certain expected medical expenses during
a specific amount of time. The insurer takes on the risk of the actual amount of medical
expenditures its client will have, which are financed by collection of premiums based on
expected medical expenses. The consumer bases his expected medical expenses on the fee
he would pay the doctor per unit of care in the event of illness, F, the units of medical
care he expects to consume, and the probability of contracting the ailment that would
make consuming this care necessary. Additionally, the consumer is risk averse with respect
to income, which means his utility function is strictly increasing with respect to income
(?U?I ? 0), but increases at a decreasing rate (?2U?I2 ? 0). Thus, the consumer does not pay
the ?fair? amount for insurance1 . Since the insurance company will charge the maximum
amount possible, it will charge the consumer a loading factor, L, equal to the consumers
risk premium, which is the ratio between the maximum premium the consumer is willing to
pay and the fair premium2. Thus, the insurer collects revenues according to the following
1The actuarially fair price for insurance equals the total expenditure on care times the probability that
care will be necessary. Thus if p is the price of a unit of care, q is the quantity of care necessary for the
ailment, and ? is the probability of that ailment occurring, then p?q?? is the fair actuarial price for insurance.
2It is important to remember here that the consumer is aware of the loading factor, and, because he is
risk averse, would still rather buy the insurance than do without it. The more risk averse an individual is
the higher the loading factor he is willing to accept.
12
equation
R = L???F(x)?x (3.1)
Where ? is the probability of incurring an ailment, and x is the expected units of medical
care needed. The expected fee, F, is a decreasing function of x?that is, ?F(x)?x < 0. This is
because the pricing schedule for the insurance premiums is inherited from the consumers?
demand for medical care. The value F reflects the price the consumer would pay for medical
care without insurance.
The insurance company has a pricing advantage over the consumer. Because of its
monopsony power, the insurer is able to negotiate lower fees for service. Additionally, the
actual quantity of medical care purchased on behalf of the consumer is most likely different
than the expected quantity, x. The bifurcated pricing is represented in the profit function.
pi = L???F(x)?x?N(q(x))?q(x) (3.2)
Where q(x), the actual quantity of medical care purchased by the insurer on behalf of its
customer, is an increasing function of the amount of expected care, x, purchased by the
consumer. N is the negotiated price paid by the insurer to the doctor; it is an increasing
function of the actual quantity of care purchased.
Because contract terms for medical care are made before the customer uses his insur-
ance, the insurance company has no direct effect on the actual quantity of care it purchases.
It can only influence q(x) by varying the amount of expected care it will insure. Thus the
insurer will maximize its profit function with respect to x3. The first order conditions are
3Some may be unsure about why I am using expected quantities instead of actual utilization. Besides
having the benefit of simplifying the problem, there are several rational reasons to do this. As has been
13
given below.
?pi
?x = L???
bracketleftbigg
F(x) + ?F?x ?x
bracketrightbigg
? ?q?x
bracketleftbigg?N
?q ?q(x) +N(q(x))
bracketrightbigg
= 0 (3.3)
In the first part of this expression, F(x) is the additional revenue gain by selling an addi-
tional unit of expected medical coverage. Since the function, F, comes from the consumer
demand, ?F?x is negative. Thus
parenleftBig?F
?x ?x
parenrightBig
is the revenue loss on the other units from selling
the additional unit of coverage. The sum of these values multiplied by the probability of
medical incident and the loading factor is the insurer?s marginal revenue. In contrast to
the generic bifurcated price model, the insurance company?s marginal product is assumed
to be constant4. Thus in order to insure a downward sloping ?demand? function for the
insurance company we must assume
parenleftBig
2? ?F?x
parenrightBig
>
parenleftBig?2F
?x2 ?x
parenrightBig
5.
Likewise, in the second part of the equation, N(q(x)) represents the addition cost from
purchasing an additional unit of medical care, and
parenleftBig?N
?q ?q(x)
parenrightBig
is the increased cost of all
other units resulting from the additional purchase. The sum of these values multiplied by
the change in q as a result of selling one more unit of x gives the insurer?s marginal cost. The
optimality condition requires the insurers marginal revenues to equal its marginal costs.
mentioned several times in this paper, the majority of insurers have switched or are switching to a managed
care platform. Under this format, insurers contract with medical service providers prior to utilization and
negotiate a fee schedule. Since this is done before services are rendered, the quantity consideration in
the price negotiations must be estimated based on the number of insured?that is, an expected quantity.
Additionally this also justifies the insurer?s first-mover status.
4This decision was made to simplify the problem. In reality, the insurer most likely has production
restraints in its ability to process large numbers of claims as well as inheriting production constraints from
the medical providers. However, this complicates the model severely, and a downward sloping demand curve
is easily attained by assuming the consumer has a downward sloping demand for medical care.
5If the marginal revenue product curve is to be downward sloping, then we need ?MRP
?x = L ?
?
bracketleftBig
2?F?x + ?2F?x2 x
bracketrightBig
< 0. Since ?F?x < 0 we must assume 2?F?x > ?2F?x2 x if we are to avoid making any as-
sumption about the second derivative of F that may limit its functional form and the generality of the
model.
14
3.2 The Competitive Fringe
In this version of the model, the competitive fringe represents all medical care pur-
chases not made by way of insurance. This category of buyers defines a competitive fringe
extremely well. There is a large number of small buyers who are price takers, and the group
as a whole makes up less than 20 percent of all medical care purchases. The only additional
assumption that will be made about this group for the medical market model besides the
assumptions necessary to make them perfect competitors is that there medical consumption
plans are realized. That is, because the fringe of buyers is not purchasing medical care in
advance of actual usage, the model assumes the expected medical care purchases of the
fringe equal the actual medical purchases (x = q). The fringes inverse demand equation is
as follows
F(x) = F(x+xd) (3.4)
Where F is the price the fringe pays, x is the quantity of medical care purchased,
and xd is the expected quantity of medical care purchased by the third-party payer. As in
the general model, the existence of the monopsonist shifts the demand curve of the fringe
outward, increasing the price they pay.
The next chapter will summarize the major results of the bifurcated price model applied
to the medical industry. Using these results, this paper will explain both how and why
third-party payers cause inflation in the medical care market.
15
Chapter 4
Results and Applications
4.1 Third Party Payers and Cost
The major distinguishing feature of the bifurcated price model applied to the medical
insurance industry is that the price paid by the competitive fringe is a component of the
demand of the monopsonist. This small difference has a significant impact on the results of
the model. Remember that in the general two price model the monopsonist can affect the
price paid by the fringe (by increasing or decreasing its own demand), but the fringe has no
impact on the price paid by the monopsonist. The fact that the fringe price is an argument
in the monopsonist?s demand equation means not only the demand of the competitive fringe
influences the demand of the monopsonist, but also that the model is potentially unstable.
That is, if the demand for either the monopsonist or the competitive fringe increases for any
reason, cost and demand could increase indefinitely. For example, if people become more
risk averse and purchase more insurance, the insurer?s demand for expected medical care
will shift out. This will cause the quantity of expected medical care the insurance company
demands to increase. Because the quantity demanded by the insurer is the ?axis? for the
competitive fringe, the demand of the fringe will also shift out. The result is that the price
paid by both insurer and fringe competitor increases. Additionally, because an increase
in the fringe price will increase the demand for insurance, and thus, the monopsonist?s
demand for medical care, the insurer?s demand curve will shift out further causing the
fringe?s demand curve to shift outward as well. It is possible that there may not be a new
equilibrium; this process of repeated feedback could continue indefinitely. Thus, as far as
16
Figure 4.1: Continual Feedback Resulting from Third-Party Payer
the this model is concerned, any increase in demand from either the monopsonist or the
competitive fringe may cause perpetual price increases. The results are shown graphically
in Figure 4.1.
The green MRP curves represent the third-party payer?s derived demand for medical
care. The blue D curves represent the demand of the competitive fringe. Since the supply
of medical providers is assumed to be fixed, neither the S curve nor the MFC curve shift.
As the insurer?s demand for care increases the vertical ?axis? for the competitive fringe
shifts outward, which shift the demand curve of the fringe outward. This shift causes the
price paid by the competitive fringe to increase, which, in turn, increases the demand for
insurance, and thus further increases the insurer?s demand for medical care. A shift from
MRP1 to MRP2 and from D1 to D2 causes the prices paid by both monopsonist and
17
competitive fringe to increase. Additionally, in this graph the difference between the prices
has increased; this has further implications discussed in Sections ?? and 4.2. Notice that
no new equilibrium is made.
If this model were to ever be used in an empirical application, a dynamic model would
be more appropriate and informative. Depending on the functional form of the demand
curves, supply curves, and marginal factor cost curve, the model may find a new equilibrium
or it may be explosive. Either scenario would provide significant insight into the future
of health care. Feldstein (1977) explored the dynamics of his inflation model using data
from 1958 to 1973. He found that the reciprocating increases of insurance and price were
dynamically explosive for nearly half of the time period explored (1958-1965). He explained
the stabilization to be the result of quality increasing to a point where consumers were less
sensitive to a quality change. It would be interesting to see an empirical test with a two
price model to gauge the impact of recognizing a different market structure.
4.2 Agency and Insurance
Principle-agent theory models economics situations in which there is asymmetrical
information. The principle, who is not fully informed, hires and agent, who is fully informed,
to perform a service. Because the agent?s knowledge gives him a comparative advantage in
performing the task for which he is hired, the exchange should be pareto optimal. However,
the principles lack of knowledge creates an incentive compatibility problem. Since the agent
is trying to maximize his own independent utility function, his incentive is to induce demand
for the service he provides. This is a conundrum for the principle, who must try and design
a reward system that aligns the agent?s incentives with his own and keep the exchange
18
attractive to the agent. The difficulty of this task greatly depends size of the disparity of
information between the principle and the agent.
Traditionally economists have modeled the relationship between an consumer and his
insurance company as one where the consumer is the agent. It is believed that after an
individual purchases insurance, they will engage in riskier activities. This increases the in-
dividual?s probability of incurring a loss, which shifts the demand for medical care outward
and makes it more difficult for the insurance company to accurately manage risk. Addition-
ally, insurance effectively reduces the price of medical care to zero (or a small co-payment).
Because of this, individuals will now go see a doctor to treat illnesses that they may have
dealt with with home remedies (i.e colds, allergies, minor sprains, etc.). This problem is re-
ferred to as moral hazard. It is, therefore, up to the insurance company to realign incentives
by way of co-payments or not fully insuring the individual.
Considering the results of the model used in this paper, we will consider the problem
from a different angle. Consider that the consumer is largely unaware of the negotiated fee
paid by his insurer, and bases his purchasing decisions on the fee he would be charged with-
out insurance. This is an informational disparity between the consumer and his insurance
company. In this instance, however, the consumer is the principle and the insurer is the
agent.
Using the ability to negotiate for lower fees from physicians and hospitals, the insurance
company indirectly induces demand from the consumer. The consumer purchases insurance
based on expenditures (price times quantity) he would likely make on health care during
the coverage period. However, because the consumer is calculating this figure using a price
19
which is higher than the one the insurer is paying, he will spend too much on his insurance?
even if he predicts the same quantity of usage. Some may argue here that this is not induced
demand because the consumer could not have negotiated this price on his own. While it
may be true that individual consumers cannot negotiate price, the mere existence of a third-
party payer artificially inflates the ?market? price the consumer observes. Additionally, as
PWCs Health Research Institute (2005) reports the incentive for hospitals and physicians
is to charge an extremely high list price to make up for losses from powerful managed care
payers, like government programs and large insurers. Moreover, standard monopsony theory
tells us that any savings gained by the monopsonist are not passed on in output price, and
if the monopsonist also has monopoly power, output price will actually increase as a result
of the firm negotiating efforts?all savings will be retained as profits (Blair and Harrison
1993). In essence, the insurer creates its own demand by simply being a large buyer with
the ability to negotiate before other buyers.
Usually when a principle-agent problem exists, the research generally focuses around
designing a contract to realign the incentives of the principle and agent. However in this
case, this may be impossible. Recall that the loading factor is the observed profit of the
insurer that the consumer allows because of his risk aversion. A smaller loading factor
means that the degree of risk aversion necessary to make insurance attractive goes down.
Notice that from the consumer perspective
lim
q(x)?x
L???F(x)?x
F(q(x))?q(x) = L (4.1)
20
but from the insurer?s perspective
lim
q(x)?x
L???F(x)?x
N(q(x))?q(x) = L?
F(x)
N(x) > L (4.2)
That is, the consumer observes a much lower loading factor than actually is being
charged; so bifurcated pricing actually makes insurance seem like a ?better deal? to the
consumer. Thus the consumer is completely unaware that he is purchasing too much in-
surance; he simply believes he can not do without it. Feldstein (1973) made a similar
analysis referring to the ?price of insurance?. In his paper the price was equal to the ratio
of premium cost to expected benefits. Notice that if we consider expected benefits equal
to the actuarially fair price for insurance, the this ratio is exactly what I refer to as the
loading factor, L. Feldstein sites that the lower the cost benefit ratio (L), the higher the
consumer?s optimal level of insurance. Thus because the bifurcated-price market lowers
the loading factor (cost-benefit ratio), people are induced to buy more complete insurance.
This accelerates the reciprocating feedback between increases in insurance levels and price.
21
Chapter 5
Conclusion
Since the 1970s, there has been strong research to suggest that the proliferation of
insurance is the source of medical cost inflation. Martin Feldstein?s research provides strong
empirical evidence to support this fact. However, much of the ?how? part of this analysis
is blamed on problems of moral hazard. In this paper, it has been shown that this simple
assumption may not be adequate explanation of health-care inflation. To my knowledge,
there has been no research concerning the impact that having two market prices can have
on the market. However its role is significant.
Feldstein (1973, 1977) discovered a dynamic relationship between price and the level
of insurance. Being insured increases the demand for medical care through various forms
of moral hazard. The increase in demand results in higher prices. When prices increase,
the consumer?s expected expenditures increase, which further increases the demand for
insurance. He found that this feedback could be explosive or stable (and showed that it
had been both) depending on the consumer response to market conditions.
A similar relationship can be see from the model presented in this paper. Insurers
ability to negotiate prices before the competitive fringe effectively shifts price axis of the
fringe out to the quantity purchased by the insurer. This has the same effect of reducing
supply for the fringe, which now faces and artificially high price. Those individuals who
purchase insurance base their expenditure decisions on the price they would pay without
insurance (their other alternative), which causes them to over purchase. This results in over-
utilization, which increases the price the fringe will pay. The increase in the transparent
22
price will further increase the demand for insurance. This will shift the fringe ?axis? out
even further, and the fringe price will increase again. The process of continual feedback
could potentially go on without end.
Additionally by considering this problem in a two-price market, the analysis of this
paper reveals a new method of insurance inducement: the loading factor or cost-benefit
ratio. The ability to negotiate fees allows insurers to present a much lower cost-benefit ratio
to the consumer. This means that the level of risk aversion necessary to make additional
insurance optimal also falls. That is the existence of a two price market induces more people
to purchase insurance and induces insured people to purchase more complete coverage.
While there is little recognition of the two-price problem in the world of economics, it is
recognized on a national level. Part of the Bush Administration?s plan to fight health care
inflation involves increased transparency about negotiated fees on the part of government
funded and private insurers to allow everyone the same access to affordable care. Moreover
PWC?s Health Research Institute recognizes the cost shifting from large insurers to those
without coverage resulting from a multiple-price market as a major roadblock in providing
charitable insurance for the underprivileged. According to the Seattle Weekly, medical care
providers are also responding. Groups like Qliance, a boutique medical practice in Seattle,
WA, are providing primary care to working class and uninsured individuals at a fraction
of the typical costs; and they are doing so by not accepting any form of insurance. They
claim that the savings in employee hours and paperwork from not having to file insurance
claims along with a carefully altered business plan makes this alternative form of medical
service viable.
23
Further research in the area needs to focus on more empirical analysis of this model.
Specifically a dynamic price adjustment model stemming from the two-price formulation
may provide new insight as to the true nature of health cost inflation.
24
Bibliography
[1] Blair, Roger D. and Jeffery L. Harrison, Monopsony: Antitrust Law and Economics,
Princeton 1993.
[2] Brewster, Lawrence G. and Genie N.L. Stowers, The Public Agenda, Ontario, 2004.
[3] Feldman, Roger and Bryan Dowd, ?A New Estimate of the Welfare Loss os Excess
Health Insurance?, AER, Mar. 1991, 81, pp.297-301
[4] Feldstein, Martin S, ?The Rising Price of Physician?s Services?, The Review of Eco-
nomics and Statistics, May 1970, 52, pp. 121-133.
[5] . , ?Hospital Cost Inflation: A Study in Nonprofit Price Dynamics?, Dec. 1971, 61,
pp. 853-872.
[6] . ,?The Welfare Loss of Excess Health Insurance?, The Journal of Political Econ-
omy, Mar.-Apr. 1973, 81, pp.251-280.
[7] . , ?Quality Change and the Demand for Hospital care?, Econometrica, Oct. 1977,
45, pp. 1681-1702.
[8] Newhouse, Joseph P, ?Health Reform in the United States?, The Economic Journal,
Nov. 1996, 106, pp. 1713-1724.
[9] Peden, Edgar A. and Mark S. Freeland, ?Insurance Effects on US Medical Spending
(1960-1993)?, Health Economics, 1998, 7, pp. 671-687.
[10] Santerre, Rexford E. and Stephen P. Neun, Health Economics: Theories, Insights, and
Industry Studies, 4th ed, Mason 2007.
[11] Schneider, Helen and Alan Mathios, ?Principle Agent Theory and Health Care Utiliza-
tion?, Economic Inquiry, Jul. 2006, 44, pp. 229-241.
[12] BCBS Association, 2007 Medical Costs Reference Guide, 2007.
[13] National Center for Health Statistics, Health, United States, 2006 With Chartbook on
Trends in the Health of Americans , Hyattsville 2006.
[14] Pricewaterhouse Coppers? Health Research Institute, Acts of Charity: Charity Care
Strategies for Hospitals in a Changing Landscape, 2005.
25