ECONOMIC EVALUATION OF PROTECTION AGAINST FREEZES IN SATSUMA MANDARIN PRODUCTION 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. ___________________________________ Jeanne K. Lindsey Certificate of Approval: ___________________________ ________________________ Robert G. Nelson, Co-Chair Robert C. Ebel, Co-Chair Professor Associate Professor Agricultural Economics and Horticulture Rural Sociology ____________________________ ________________________ Patricia A. Duffy William A. Dozier, Jr. Profesor Profesor Agricultural Economics and Horticulture Rural Sociology ____________________________ George T. Flowers Interim Dean Graduate School ECONOMIC EVALUATION OF PROTECTION AGAINST FREEZES IN SATSUMA MANDARIN PRODUCTION Jeanne K. Lindsey 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 ECONOMIC EVALUATION OF PROTECTION AGAINST FREEZES IN SATSUMA MANDARIN PRODUCTION Jeanne K. Lindsey 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 VITA Jeanne (Williams) K. Lindsey, daughter of Thomas Sewell Williams and Patricia Caylor Williams, was born June 6, 1957 in New Orleans, Louisiana. She is a graduate of Saint Mary?s Dominican High School in New Orleans. She earned a Bachelor of Science degree and a Master of Science degree in Horticulture from Auburn University in August, 1979, and March, 1982, respectively. In 1998, she received the Charter Life Underwriter designation from The American College in Bryn Mawr, Pennsylvania and she is a past president of the Columbus, Georgia Chapter of the Society of Financial Service Professionals. Work experience includes fruit research as a Research Associate for Auburn University; landscape contracting, design, and nursery management as the owner of Classic Designs; and insurance and financial product sales as a representative for Prudential Financial. She is currently employed by the US Department of Agriculture in the Risk Management Agency as a Risk Management Specialist. She is married to John F. Lindsey, Jr., the son of Reverend John Frank Lindsey, Sr. and Juanita Fuller Lindsey, and has three children, Stacia Marie Knowles, Eric Daniel Lindsey, and Katherine Lee Lindsey. v DISSERTATION ABSTRACT ECONOMIC EVALUATION OF PROTECTION AGAINST FREEZES IN SATSUMA MANDARIN PRODUCTION Jeanne K. Lindsey Doctor of Philosophy, May 10, 2008 (M.S., Economics, Auburn University, 2005) (M.S., Horticulture, Auburn University, 1982) (B.S., Horticulture, Auburn University, 1979) 97 typed pages Directed by Robert G. Nelson and Robert C. Ebel This dissertation consists of three essays that evaluate the effect of freeze-risk reduction techniques on discounted net returns for Satsuma mandarin (Citrus unshiu Marc.) in the northern Gulf Coast region of the United States. In all studies, enterprise budgets are simulated over a 20-year investment horizon. Mean and distribution of net returns, and break-even prices are used to compare risk reduction methods. The first essay evaluates the effect of multi-peril crop insurance and freeze protection with micro- sprinkler irrigation on discounted net returns for one-acre grove units. Using weather data from the period 1948-2004, freeze occurrence probabilities for the Fairhope, Alabama area were calculated to be 14-percent for severe freeze and 11-percent for moderate freeze. Freeze protection in combination with crop insurance resulted in the vi highest mean and lowest variability in net return at market prices above break even. Increased yields and net returns due to freeze protection were attributed to the elimination of the need to replant after a severe freeze. Government subsidy for crop insurance premiums increased total discounted net returns; and indemnities lowered the distribution of negative net returns for the 20-year simulation period. In the second essay, micro- sprinkler and high tunnel technologies for freeze protection are compared to no protection for 10-acre Satsuma groves in the Fairhope, AL area. Micro-sprinkler technology eliminates the tree loss, but not crop loss, due to freezes. High tunnel technology eliminates the loss of either trees or crop for any freeze event. Relative to the high tunnel groves, average yield over the 20-year period was reduced by 25-percent for micro- sprinkler irrigated groves and 53-percent for unprotected groves. The high tunnel strategy was preferred to the micro-sprinkler protection only at market prices above $0.83 per pound. In the third essay, net returns for groves with micro-sprinkler and high tunnel technologies were compared to no protection at varying freeze probability levels. With severe freeze probability levels of 5-percent and greater, net returns for micro-sprinkler groves were greater than for unprotected groves. At a market price of $0.50 per pound, high tunnel groves had greater mean net returns than the micro-sprinkler technology only when total freeze events exceeded 50-percent. vii ACKNOWLEDGEMENTS I would like to thank my parents, Si and Pat Williams, for instilling a life-long love of learning in all of their children. Their support, love, and guidance through out my life have been so very important to me. I am thankful for the support of my committee members, Bob Nelson, Bob Ebel, and Patricia Duffy who always took the time to help and encourage me whenever I needed it. I would like to extend special thanks to Bill Dozier who has been a mentor and friend to me for many years, as well as one of my committee members. There are many dear friends and relatives, especially our friends at St. Mark in Columbus, GA, who have helped my family and me during the last five years. Special thanks, however, go to Jolly Roberts, Monica and Darrell Beck, Tricia Marshall, and Sandy and Jeff Krietemeyer for ?room and board? whenever I needed it, childcare, encouragement, and laughter. I would also like to thank my in-laws, John and Juanita Lindsey for their love and unwavering faith in me. Most of all, I am grateful to have a wonderful husband and children who have given me endless support and still love me in spite of everything. I thank God everyday for all the wonderful gifts and blessings He has bestowed on me. viii Style manual or journal used: Journal of Agricultural and Applied Economics Computer software used: Microsoft Word 2003, Microsoft Excel 2003, and Simetar ix TABLE OF CONTENTS LIST OF TABLES............................................................................................................. xi LIST OF FIGURES ......................................................................................................... xiii CHAPTER 1. EVALUATION OF NET RETURNS TO FREEZE PROTECTION AND CROP INSURANCE IN ALABAMA USING MONTE CARLO SIMULATION.....................................................................................................................1 1.1 Introduction...............................................................................................................1 1.2 Review of Literature .................................................................................................2 Satsuma Mandarin .................................................................................................2 Freeze Risk ............................................................................................................4 Risk Management and Evaluation Methods ..........................................................5 1.3 Data and Methodology..............................................................................................6 Production Budget .................................................................................................7 Crop Insurance Policies .........................................................................................8 Freeze Probability and Models ............................................................................10 1.4 Results and Discussion ...........................................................................................13 1.5 Conclusions.............................................................................................................18 CHAPTER 2. EVALUATION OF FREEZE PROTECTION TECHNOLOGIES FOR SATSUMA MANDARIN PRODUCTION IN ALABAMA WITH ENTERPRISE BUDGET SIMULATION .................................................................................................20 2.1 Introduction.............................................................................................................20 2.2 Review of Literature ...............................................................................................22 Satsuma Mandarin ...............................................................................................22 Freeze Protection .................................................................................................23 x 2.3 Data and Methodology............................................................................................26 Yield Data.............................................................................................................26 Yield Ratio............................................................................................................27 Production Costs...................................................................................................28 Freeze Risk ...........................................................................................................31 The Model.............................................................................................................32 2.4 Simulation Results and Discussion.........................................................................35 Simulations with Standard Values.......................................................................35 Stochastic Dominance..........................................................................................38 Equivalent Prices .................................................................................................42 High Tunnel Cost Analysis..................................................................................43 2.5 Conclusions.............................................................................................................48 CHAPTER 3. EFFECT OF LOCAL VARIATION IN FREEZE PROBABILITY ON NET RETURNS FROM THREE PROTECTION TECHNOLOGIES.............................52 3.1 Introduction.............................................................................................................52 3.2 Review of Literature ...............................................................................................54 Weather Data and Satsuma Cold Acclimation??? ...???.??.????54 Satsuma Production Areas....................................................................................56 3.3 Data and Methodology............................................................................................57 The Simulation Model .........................................................................................57 Model Variables...................................................................................................58 Freeze Probability Matrix ....................................................................................61 3.4 Results and Discussion ...........................................................................................62 Net Returns ..........................................................................................................62 A Decision Process ..............................................................................................68 3.5 Conclusions.............................................................................................................72 REFERENCES ..................................................................................................................74 APPENDIX A: SIMULATION VARIABLES FOR CHAPTER 1 ..................................79 APPENDIX B: SIMULATION VARIABLES FOR CHAPTERS 2 AND 3....................82 xi LIST OF TABLES Table 1.1 Values Used in Simulations that Change with Tree Age.....................................8 Table 1.2 Discounted 20-Year Net Returns for Satsuma under Varying Risk Management Scenarios ..........................................................................................14 Table 1.3 Percentage Change in Standard Deviation of Discounted 20-Year Net Returns when Compared to Base Scenario............................................................17 Table 2.1 Effect of Freeze Event on Yield of Simulated Satsuma Grove .........................34 Table 2.2 Twenty-Year Discounted Costs and Returns for 10-Acre Satsuma Grove in South Alabama with Different Freeze Protection Technologies .......................36 Table 2.3 Twenty-Year Key Output Variables from Simulations using Standard Parameters...............................................................................................................37 Table 2.4 Effects of Market Price and Absolute Risk Aversion Coefficient on Certainty Equivalents and Ranking for Three Satsuma Production Strategies with Negative Exponential Utility Function..........................................................42 Table 2.5 Simulated Discounted 20-Year Net Return to Management at Different Price Levels for Satsuma with Different Freeze Protection Technologies ? Fairhope, Alabama.................................................................................................45 Table 2.6 Equilibrium Prices between Freeze Protection Technologies with Varying Yield Ratios and High Tunnel Fixed Cost...............................................46 Table 2.7 Simulated Discounted 20-Year Net Return to Management at Different Price Levels for Satsuma with Changes in High Tunnel Variable Costs ..............47 Table 3.1 Array of Severe by Moderate Freeze Occurrence .............................................63 Table 3.2 Twenty-Year Discounted Net Returns for 10-Acre Unprotected Satsuma Grove with Varying Probabilities of Moderate and Severe Freeze Occurrence...64 xii Table 3.3 Twenty-Year Discounted Net Returns for 10-Acre Satsuma Grove with Micro-jet Freeze Protection and Varying Probabilities of Moderate and Severe Freeze Occurrence......................................................................................65 Table 3.4 Break-Even Prices for Simulations of Satsuma Groves with Different Levels of Freeze Protection and Varying Probabilities of Moderate and Severe Freeze Occurrence......................................................................................68 Table 3.5 Equivalent Prices for Simulations of Satsuma Groves with Different Levels of Freeze Protection and Varying Probabilities of Moderate and Severe Freeze Occurrence......................................................................................69 xiii LIST OF FIGURES Figure 1.1 Response of Discounted 20-Year Net Returns for Satsuma under Varying Risk Scenarios ? Fairhope, Alabama.......................................................15 Figure 1.2 Distribution of Negative Discounted 20-Year Net Returns for Satsuma under Varying Risk Scenarios ? Fairhope, Alabama.............................................17 Figure 2.1 Cumulative Distributions of 20-Year Discounted Net Returns to Management for Three Satsuma Freeze Protection Technologies at Market Price = $0.50/lb .........................................................................................38 Figure 2.2 Cumulative Distributions of 20-Year Discounted Net Returns to Management for Three Satsuma Freeze Protection Technologies at Market Price = $1.00/lb .........................................................................................39 Figure 2.3 Twenty-Year Discounted Net Returns to Management for Satsuma at Different Market Prices ? Fairhope, Alabama.......................................................43 Figure 3.1 Twenty-Year Discounted Net Returns for Satsuma Groves with 10-percent Probability of Severe Freeze and 20% Probability of Moderate Freeze....................................................................................................................66 Figure 3.2 Equivalent Price between Micro-jet and High Tunnel at Freeze Probability Levels of 5-percent to 60-percent .......................................................67 1 CHAPTER 1. EVALUATION OF NET RETURNS TO FREEZE PROTECTION AND CROP INSURANCE FOR SATSUMA MANDARIN USING MONTE CARLO SIMULATION 1.1 Introduction Specialty crops offer agricultural producers the opportunity to increase net returns per acre relative to traditional crops. Returns may increase because of price premiums associated with niche markets or lack of competition. In many investment areas, higher returns are often associated with increased risk or uncertainty; production of specialty crops is no exception. Methods chosen for risk mitigation depend on the type of risk. Evaluation of possible risk management techniques is important to enable producers to make good management decisions concerning which method, if any, to employ. Initial investment decisions may depend on the whether the risks can be economically managed. This is particularly important for perennial fruit crops that require significant investment and several years of growing time before positive returns are realized. For the present study, the specialty crop chosen is Satsuma mandarin, a type of citrus that is grown in the United States in the northern Gulf Coast area, from Texas to Florida, and in Arizona and California. The primary production risk facing this crop is the risk of freeze injury. This paper will use Monte Carlo simulation to evaluate several 2 possible risk management techniques for managing freeze risk in this crop in the Alabama area. 1.2 Review of Literature 1.2.1 Satsuma Mandarin Satsuma mandarin (Citrus unshiu Marc.) is one of the most cold-tolerant citrus species available for commercial production in the US. The fruit characteristics of this species have been studied by Ebel, et al. (2004) and include sweet flavor, ease of peeling, and seedlessness. Consumer preference surveys have demonstrated that these characteristics are desirable to potential consumers (Campbell, et al., 2004). Satsuma quality, both internal sugar-to-acid ratio and external orange color development, benefit from cool temperatures in the fall (Ebel, et al., 2004). Fruits mature between mid-October and mid-December. While fruit may be held on the tree longer than this, marketing is best done during the holiday season when the highest prices can be obtained. The Gulf Coast area of the United States is desirable for production because the warm temperate zone growing conditions allow for good tree growth and the relatively cool fall temperatures allow for good fruit quality development. The southern part of Alabama has a long history of Satsuma production (Ebel, et al., 2004), as they have been produced there since the early 1900?s. However, a once- viable citrus industry was decimated by killing freezes in the 1930?s and 1940?s (Winberg, 1948a, 1948b, 1948c). Recent developments in micro-sprinkler freeze protection have mitigated the risk to Satsuma due to freeze loss (Nesbitt, et al., 2000). This development and the absence of serious freeze events in the area since 1989 have 3 contributed to revived interest in commercial production of Satsuma in the lower- Alabama area (Ebel, et al., 2005). Based on historical levels of tree acclimation to cold in this region, the threshold for economically important injury is between 18 and 22 o F (-7 to -5 o C) (Ebel, et al., 2005; Nesbitt, et al., 2000; Nesbitt, et al., 2002). At 14 o F (-10 o C), stem dieback will occur and whole trees are susceptible to death if they are not fully hardened off. Temperatures below 12 o F (-11 o C) have historically resulted in tree death for unprotected trees. During a freeze event, micro-sprinklers placed within the tree canopy will protect the trunk and major scaffold branches through the release of the latent heat of fusion as the water spray freezes (Nesbitt, et al., 2000). This method of freeze protection decreases the severity of the freeze injury to the tree and prevents tree death. In the south Alabama region, freeze protected trees that experienced extensive injury to the canopy were able to return to full production the year following the freeze event (Nesbitt, et al., 2000). Because micro-sprinkler freeze protection does not extend to the outer canopy, freeze events that cause injury to leaves, or leaves and stems, will have the same effect on both protected and unprotected mature trees. Freeze events that could potentially kill unprotected trees, will have a lesser impact on freeze-protected trees. The protected trees will miss a year of production while canopy re-growth occurs, but the grove will not need to be replanted: (Bourgeois and Adams, 1987; Bourgeois, Adams, and Stipe, 1990; Nesbitt, et al., 2000). 4 1.2.2 Freeze Risk While Satsuma are the most cold tolerant among commercial citrus crops, the single greatest risk factor facing potential growers of this citrus in the northern Gulf Coast region of the southeastern US is the risk of tree injury or tree death due to freezing weather. The second greatest risk factor is crop loss due to freeze injury to leaves, flowers, or flower stems. The use of micro-sprinklers within a tree canopy has been found to be an effective method of reducing tree loss due to freeze for citrus (Ebel, et al., 2004). During a severe freeze event, this type of protection system would protect the citrus trees from dying, but it would not prevent the loss of the next season?s crop (Ebel, et al., 2005). A study of damaging freeze events in Baldwin County, Alabama during the period of October 1948 to March 2004 found that there were 8 years in which severe freeze events occurred and six additional years in which only moderate freeze events occurred (Ebel, et al., 2005). A severe freeze was classified as one that caused extensive tree injury or tree death, and a moderate freeze was classified as one that caused extensive leaf injury and some stem dieback to the extent that the next season?s fruit crop was destroyed. Based on this information, the long-term probability of severe freeze is 14-percent and moderate freeze is 11-percent. (Note: If both a severe and moderate freeze event occurred in the same growing period, only the severe freeze was counted for probability calculation purposes.) During the 1948-2004 period in this region, all freeze events occurred between the 12 th of December and the 9 th of March when mature fruit would not typically be present on the trees or most fruit would have been already harvested. 5 1.2.3 Risk Management and Evaluation Methods Producers have various methods for managing the risks they face. Production and management practices such as grove site location, fertilization practices, and the use of freeze protection systems can have an effect on production risk and variability of returns for a grove. Marketing strategies can affect price risk. Another method of risk mitigation used by many agricultural producers is Federal Crop Insurance. Market insurance is a risk transfer method that reduces the effects of economic loss on an insured?s net revenue but it does not change the probability of a loss occurring (self-protection or risk avoidance) or reduce the severity of a loss (self- insurance or risk reduction) (Ehrlich and Becker, 1972). Installation of a freeze protection system, as used in this study, is primarily a self-insurance measure; the system will not affect the probability of a freeze occurring but will reduce the amount of tree injury resulting from a severe freeze. The use of crop insurance by a producer is a market insurance method of risk management. Provision of crop insurance protection for specialty crops is currently a priority for the USDA Risk Management Agency (USDA, Risk Management Agency, 2004). A Satsuma crop insurance program for fruit or trees is not in place for south Alabama. If a named peril policy was available, however, it could mitigate the negative economic impact caused from a moderate or severe freeze by providing funds to replant or to help pay fixed and direct costs during years when trees or the crop is lost. (Note: disaster assistance is available through NAP ? Noninsured Crop Disaster Assistance Program ? for crops in counties without an insurance program. This program is administered 6 through the Farm Service Agency of USDA but it is not included in the present evaluations.) Various techniques have been used by researchers to evaluate risk and uncertainty in agricultural production including mean-variance models, linear programming, and simulation. A mean-variance linear programming model, considering changes in marginal benefits, was used by Featherstone and Moss (1990) to evaluate diversification opportunities for Florida citrus growers. MOTAD, a linear programming model that minimizes total absolute deviation from a mean rather than minimizing variance was used to evaluate optimal mixes of citrus types and planting density in Texas (Teague and Lee, 1988). When MOTAD was used in combination with simulation to evaluate production and marketing strategies on net farm income in Oklahoma, Mapp, et al. (1979), found that the simulation model was able to evaluate interactions of stochastic variables between years that were not possible with the MOTAD model. Simulation techniques were also used in combination with the Dixit-Pindyck model to evaluate investment behavior and sources of risk for grapefruit producers in Texas (Elmer, et al., 2001). Using these techniques, Elmer et al. determined that freeze risk was a greater source of uncertainty facing Texas grapefruit producers than market prices or expanded trade effects associated with NAFTA. 1.3 Data and Methodology Data collected for this study includes yield records, production costs and historical temperature records. Yield data for an ?Owari? Satsuma mandarin grove was collected by the Alabama Agricultural Experiment Station Gulf Coast Research and 7 Extension Center at Fairhope, Alabama. Yield data spans 16 years from initial planting in 1990 through crop harvest 2005-06. In the simulations, the planting density assumption is 116 trees per acre. Trees are assumed to have no yield for the first two growing seasons and reach a maximum yield of 400 lb/tree (23.2 tons/acre) by the ninth growing season (Table 1). 1.3.1 Production Budget Production costs were obtained from a Satsuma enterprise budget developed by Hinson and Boudreaux (2006) for Louisiana producers. Alabama producers are expected to have similar production methods and costs. The production budget includes costs for all labor and materials, as well as fixed costs of machinery and packing line equipment (Appendix Tables A2 and A3). Charges for a drip irrigation system are included in the simulation budget but were not in the Louisiana budget. Land is assumed to be an appreciable asset and there is no land charge included in the budget. Tree yields, and fixed and direct production costs vary by the age of the tree and are presented in Table 1. Yield related costs of fruit harvest, grading, and packaging are calculated based on yield level and the presence of a crop; the variable harvest costs used in the simulation are $6.40 per 40 lb bushel and the direct harvest costs are $211 to $215 per acre (Appendix Table A3). Freeze protection costs were developed from information supplied by the Gulf Coast Research Center and are presented in Appendix Table A3. The freeze protection system modeled is a micro-sprinkler irrigation system with one emitter per tree situated in the canopy at 5 feet above ground. Emitter delivery rate used is 30-gph. The study is 8 modeled using one 4? well system with a 60-gpm capacity for each acre. It is important to note that, during a freeze event, the systems must operate simultaneously and continuously for all acreage to be protected. The cost of freeze protection is $6,350 per acre covers a 4-inch well, a 60-gpm pump, and all below ground pipes; these costs are amortized at 6-percent over the 20-year period. In addition, above ground pipes and emitters have a cost of $185 and they are replaced every 4 years (amortized at 6-percent over each 4 year period), and there is a $25 per year maintenance charge. With these assumptions, the total cost of freeze protection charge is $632 per acre per year. Table 1. Values Used in Simulations That Change with Tree Age Varible Unit 123456789+ Yield lb/tree 0 0 70 120 190 250 350 350 400 Fixed & Direct Costs Production $/acre 2141 1157 1363 1693 1813 1813 1813 1813 1813 Harvest and Packing $/acre 0 0 211 211 215 215 215 215 215 Tree Policy Indemnity a $/acre 899 1634 2179 2451 2723 2723 2723 2723 2723 a Texas 2008 Citrus I tree policy with 65-percent coverage level. Sources: Hinson, et al., Louisiana Agricultural Experiment Station Info. Series No. 140, 2006. Alabama Agricultural Experiment Station Gulf Coast Research and Extension Center, Fairhope, Alabama. Tree Age - Growing Season 1.3.2 Crop Insurance Policies Two hypothetical crop insurance policies were modeled after existing policies and actuarial tables for citrus trees in Cameron County, Texas, and for mandarin fruit in Riverside County, California (USDA-RMA, 2007a, 2007b, and 2007c). Values for these policies that were used in calculations for the simulations are presented in Appendix Table A1. The tree policy used values from the Texas Citrus I policy. This policy has a fixed liability per acre with a graduated indemnity rate that reaches its maximum if loss 9 occurs during the fifth and subsequent growing seasons as presented in Table 1 and Appendix Table A1. The graduated indemnity schedule reflects an assumed decrease in risk of tree injury due to freeze as the trees mature. In the simulations, the grove was planted in March of the first year or in any re-plant year. The policy period is from November 21 to November 20 and there is never an insurance premium due for the tree policy in the initial year of the simulation because and there is no chance of a freeze event in the first year. The fruit crop insurance policy in the simulation uses values for mandarins from the Arizona-California Citrus policy and a 65-percent coverage level. In this policy, the grove cannot be insured until the sixth growing season. The liability per acre is more complicated to calculate for this policy as it depends on the past yield performance of the insured grove, termed ?actual production history? (APH). The APH is an average from the yield database containing a minimum of 4, building to 10, years. If there are less than 4 years in the database, a transitional-yield (T-yield) may be substituted for the missing yield; if there is a covered weather-related loss, a yield adjustment (YA) will substitute 60-percent of the T-yield for the lost yield in the database (USDA-RMA, 2006). In the liability calculation, the APH is multiplied by the coverage level and then by the price per unit (25-lb carton for this policy) assigned by RMA each year. With these assumptions, the liability calculation is: (1) Liability = APH x Price Election x Coverage Level. Premium rates for the tree and fruit crop insurance policies in Texas or California could not be used for these scenarios because of differences in freeze risk exposure 10 (Elmer, et al., 2001). The Risk Management Agency rate setting procedures are normally based on county/state indemnity experience for a particular crop (Schnapp, et al., 2000). Without any indemnity history, different methods need to be employed to determine a rate. In this study, base insurance premiums were calculated to produce a 1.00 premium to loss ratio based on the simulated loss experience with the given freeze probabilities. A catastrophic load was added to the base premium by dividing it by .88. The insured grove was assumed to have only one unit. With these assumptions, and the deduction of the appropriate government subsidy, a producer premium of $155 per acre was calculated for a 65-percent coverage level for the tree policy. For the fruit crop insurance policy, the total base premium rate of $0.313 was calculated using procedures described above. Premiums were charged to the appropriate simulation grove in the sixth and subsequent years. Producer premiums for each insured year in the simulation were calculated with the following equation: (2) Producer Premium = (Liability x base rate) ? subsidy; where the appropriate subsidy rate for the 65-percent coverage level in 59-percent of the calculated premium. 1.3.3 Freeze Probability and Models Daily min-max temperature data from 1948 through 2006 was obtained from the weather station located at the Gulf Coast Research Center. Economically important freezes were determined through a prediction formula developed by Ebel, et al. (2005) and compared to field observations for severity rating. These ratings were used to 11 calculate probabilities of economically damaging freeze occurrence in the Fairhope, Alabama area. Freeze severity ratings are: 1) Slight ? some injury to leaves, 2) Moderate ? extensive leaf injury and some stem dieback and 3) Severe ? widespread tree death. Only moderate and severe freeze events are considered economically important in this study. Based on this information, the simulations use the severe freeze probability of 14-percent and the moderate freeze probability of 11-percent. A hypothetical one-acre Satsuma grove was the unit of study. A Monte Carlo simulation with 100 iterations was performed using Excel 2003 for each of four scenarios for a 20-year period at each of seven different farm-level market prices ranging from $0.20 per pound to $0.50 per pound. The random event was the incidence of severe, moderate or no freeze. The generated random number matrix was consistent between scenarios. Net returns were calculated based on the costs and returns associated with the freeze event status and the tree age. The four scenarios were then evaluated based on discounted net returns totaled over the 20-year period and on the distribution of negative discounted 20-year net returns over the range of market prices. The scenarios evaluated were: 1) No freeze protection and no crop insurance (NP_NI), 2) Freeze protection and no crop insurance (P_NI), 3) No freeze protection plus crop insurance for tree loss (NP_I), and 4) Freeze protection plus crop insurance for fruit loss (P_I). Without freeze protection, trees were assumed to lose one crop year if a moderate freeze occurs, and were assumed to die and to be replanted if a severe freeze occurs. The number of times the grove will be replanted in the simulated 20-year production periods 12 is limited only by the severe freeze probability and the random draw of a severe freeze. The probability of a severe freeze occurring was set at 14-percent and the probability of a moderate freeze occurring is set at 11-percent based on historical data for the Fairhope, Alabama area. With freeze protection, trees were assumed to respond to both severe and moderate freezes with the loss of one crop year and no tree deaths ever occurred. A crop insurance policy insuring tree loss was considered more valuable to a producer when no freeze protection was present. A crop insurance policy insuring fruit production was considered more valuable to a producer when freeze protection was in place because there no tree loss occurred under this scenario. The discounted total net returns equation for the base scenario NP_NI is: 20 (3) NR d = ? [(PY j (f, t) ? C j (t) ? X j (y)) / (1+r) j ] j=1 where NR d = total discounted net returns, j = the simulation year, f = freeze event, t = tree age, P = market price for fruit, Y j (f, t) = yield in the jth year as a function of freeze event and tree age, C j (t)= fixed and direct costs in the jth year as a function of tree age, and X j (y) = variable costs as a function of yield in the jth year, and r = the discount rate. This equation is modified for the different scenarios as follows: 20 (4) P_NI: NR d = ? [(PY j (f(cp), t) ? C j (t) ? CP ? X j (y)) / (1+r) j ] j=1 20 (5) NP_I: NR d = ? [(PY j (f, t) + I j (f, t) ? C j (t) ? CI j ? X j (y)) / (1+r) j ] j=1 20 (6) P_I: NR d = ? [(PY j (f(cp), t) + I j (f) ? C j (t) ? CP ? CI j ? X j (y)) / (1+r) j ] j=1 13 where the terms described above are applicable and CP = the fixed cost of freeze protection, f(cp) = freeze event as a function of freeze protection, I j (f, t) = the insurance indemnity in the jth year as a function of freeze event and tree age, CI j = cost of crop insurance policy in the jth year, and I j (f) = insurance indemnity in the jth year as a function of freeze event. 1.4 Results and Discussion Discounted 20-year net returns for each of the scenarios are presented in Table 2 for market prices ranging from $.20 per pound to $.50 per pound. Breakeven prices of $.258 to $.291 per pound were calculated for the different scenarios with the lowest price being required by enterprises under the P_I scenario and the highest price being required under the NP_NI scenario. Below this cluster of breakeven prices, returns to freeze protection are less than returns to no freeze protection because of the large capital investment needed for the freeze protection system. Above the breakeven prices, however, the returns increase due to higher yields from protected groves. Total net returns over the 20-year period have a positive relationship to market price with the magnitude of the response being dependent on yield, ?NR/?P = Y(f,t) (from equations (3) and (5)) or ?NR/?P = Y(f(cp),t) (from equations (4) and (6)). Yield is a function of freeze event and tree age; the severity of the freeze event changes in response to freeze protection measures, but not to insurance protection. Total yield will increase with freeze protection because, in the event of a severe freeze, protected trees lose only one year of production whereas unprotected trees must be replanted and it will take two years before the grove starts to become productive again. 14 Table 2. Discounted 20-Year Net Returns for Satsuma under Varying Risk Management Scenarios Market Price $/lb NP_NI P_NI NP_I P_I 0.20 -14,437 -18,603 -12,456 -14,783 (2,333) (1,350) (1,013) (1,689) 0.25 -6,525 -5,839 -4,544 -2,019 (5,900) (3,291) (4,233) (1,416) 0.30 1,387 6,925 3,368 10,745 (9,475) (5,233) (7,773) (2,950) 0.35 9,299 19,689 11,280 23,509 (13,052) (7,175) (11,336) (4,791) 0.40 17,212 32,453 19,192 36,273 (16,629) (9,117) (14,907) (6,689) 0.45 25,124 45,217 27,104 49,037 (20,207) (11,060) (18,481) (8,607) 0.50 33,036 57,980 35,016 61,801 (23,785) (13,002) (22,056) (10,534) Price Intercept 0.291 0.273 0.279 0.258 a NP=no freeze protection, P=freeze protection, NI=no crop insurance, I=crop insurance. Values in parentheses are standard deviations. Scenario a The observed price/yield relationships indicate that the groves with freeze protection will benefit more from an increasing market price situation than unprotected groves, which will have a lower 20-year yield. Slopes for the regression equations do not vary by crop insurance policy, but are 1.6 times greater under the freeze-protected scenario than under the unprotected scenario as shown in Figure 1. 15 -30,000 -20,000 -10,000 0 10,000 20,000 30,000 40,000 50,000 60,000 70,000 0.20 0.25 0.30 0.35 0.40 0.45 0.50 Farm Gate Price per Pound ($) 20 -Y ear D i sco u n t ed N e t R e turn ($) NP_NI P_NI NP_I P_I Figure 1. Response of Discounted 20-Year Net Returns to Market Price for Satsuma under Varying Risk Scenarios ? Fairhope, Alabama The two types of insurance policies cannot be directly compared because they have different indemnity schedules. However, they each have the same effect on total discounted net returns within protection scenarios with the tree policy resulting in an increased discounted net return of $1,980 over the 20-year period and the fruit policy resulting in an increase of $3,820 (Table 2). The returns to each crop insurance policy are consistent between market prices. The insurance policy net returns would be expected to be negative because the premiums were set to result in a .88 loss ratio; however, the positive net return reflects the government subsidy for premiums. Since insurance is usually a negative sum game due to charges in excess of actuarially fair premiums (i.e. administrative costs, shareholder profits, catastrophic charges, etc.), its real value as a risk management tool results from its effect on income variability. Both insurance policies used in the simulations reduced the standard deviation of 20-year net returns under each protection scenario. The percentage change 16 in standard deviations, when compared to the base scenario of no-protection/no- insurance, is presented in Table 3. When evaluated on a percentage basis, both the use of freeze protection and the use of crop insurance have a stabilizing effect on the variability of total discounted net returns. The effects from insurance decrease with increasing market price and increasing total net returns while the effects from freeze protection are relatively consistent across market price. The greatest decrease was seen in the scenario with the interaction of freeze protection and crop insurance. A graph of the distributions of negative discounted 20-year net returns is illustrated in Figure 2. All risk management technique scenarios resulted in a decrease in the distribution of negative net returns at market prices of $.30 per pound and above. Freeze protection results in the greatest reduction with crop insurance having a lesser effect. The no-intervention scenario does not reach 0-percent negative net returns distribution given the market prices used in this study. Risk management using crop insurance reduces the variability of returns, but it does not allow producers to benefit from increasing market prices. The results of this study indicate that under increasing prices producers would be have higher net returns with the use of risk management techniques that allow a return to increased production. At zero-profit and very competitive prices, crop insurance may be better for reducing risk because of lower capital investment. The combination of crop insurance and freeze protection results in the greatest net returns and the lowest income variability at market prices above breakeven price. Installing freeze protection systems is costly in these scenarios with the initial investment for the system being three times greater than the initial investment for planting the Satsuma grove. However, if historical freeze event probabilities are 17 indicative of future events, and market prices are higher than breakeven prices, then producers will benefit from freeze protection investment. Table 3. Percentage Change in Standard Deviation of Discounted 20-Year Net Returns when Compared to Base Scenario a Market Price $/lb P_NI NP_I P_I 0.20 -42.1 -56.6 -27.6 0.25 -44.2 -28.3 -76.0 0.30 -44.7 -18.0 -68.9 0.35 -45.0 -13.1 -63.3 0.40 -45.2 -10.4 -59.8 0.45 -45.3 -8.5 -57.4 0.50 -45.3 -7.3 -55.7 a Base scenario is no-protection/no-insurance. b NP=no freeze protection, P=freeze protection, NI=no crop insurance, I=crop insurance. Scenario b 0 20 40 60 80 100 120 0.20 0.25 0.30 0.35 0.40 0.45 0.50 Farm Gate Price per Pound ($) Pe r c e n t NP_NI P_NI NP_I P_I Figure 2. Distribution of Negative Discounted 20-Year Net Returns for Satsuma under Varying Risk Scenarios ? Fairhope, Alabama 18 1.5 Conclusions Freeze protection systems and crop insurance were the two risk reduction methods that were chosen for evaluation in this economic study. Each method has different effects on costs and returns because of grove responses to random freeze events. The objective of this empirical study was to evaluate the net returns to freeze protection and to crop insurance for Satsuma using a Monte Carlo simulation procedure. Information obtained from this study would be useful in the decision making process for current and potential Satsuma producers in the northern Gulf Coast region of the United States and to all enterprises facing decisions between the use of self-insurance measures and market insurance for risk management. The results of this study indicate that under increasing prices producers would have higher net returns with the use of risk management techniques that allow a return to increased production. The 20-year net returns are greater for all risk reduction procedures when market prices are above breakeven price, with returns being greater to freeze protection than to crop insurance. Freeze protection measures allow the Satsuma enterprise to benefit from increasing prices because of increased yield over the 20-year period. Returns to crop insurance are fixed and do not increase with market price, however, the use of crop insurance decreased the variability of returns with a reduced standard deviation in discounted net returns and lower distribution of negative discounted net returns for the 20-year period. The combination of crop insurance and freeze protection resulted in the greatest net returns and the lowest income variability at market prices above breakeven price. 19 Installing freeze protection systems was costly in these scenarios with the initial investment for the system being three times greater than the initial investment for planting the Satsuma grove. However, with the freeze probabilities used in this study and market prices higher than breakeven prices, producers could benefit from freeze protection investment. 20 CHAPTER 2. EVALUATION OF PROTECTION TECHNOLOGIES FOR SATSUMA MANDARIN PRODUCTION IN ALABAMA WITH ENTERPRISE BUDGET SIMULATION 2.1 Introduction In the Gulf Coast region of Alabama, as in many areas of the United States, development pressure on the conversion of farm land to residential and other urban- related uses is a major factor in rising land prices (Lubowski et al., 2006; USDA, NASS, 2007; Wiebe and Gollehon, 2006). Increasing the returns per acre may be an important factor in keeping land in agricultural production in these areas. Perennial tree crops generally return high profits per acre and offer an attractive alternative for growers who are willing to convert from traditional row crops. Higher returns, however, come at a cost as perennial crops generally have more intensive labor, management, and capital requirements than row crops and they may carry additional production risks. As with any agricultural crop, production of perennial tree crops involves risks from many different sources. Decision makers must make risk reduction choices that are both effective and economically feasible given the operation?s particular objectives, constraints, assets, and time horizons. In many cases, risk reduction may involve significant investment and the effects of stochastic variables on a multi-year operation complicate the decision process. 21 Ex ante analysis with simulation offers valuable information for decision makers before committing to significant initial investments (Purvis et al. 1995). Simulation is a decision making tool that allows for interaction between years that is not possible with standard linear programming models (Mapp et al., 1979). Simulation models can evaluate alternative strategies while incorporating risk to answer the positive question of what is the likely outcome (Richardson, 2004). Through the simulation of an enterprise budget, insight can be gained from the distribution of net returns, and other variables of interest, in addition to their expected values (Nelson et al., 2001). Satsuma mandarin orange is a perennial crop that could potentially provide a high value alternative for agricultural producers in the northern Gulf Coast region of the United States. In Alabama, this area encompasses Baldwin and Mobile counties where there is growing interest in reviving a once viable Satsuma mandarin industry (Ebel et al., 2005). Research conducted by Auburn University on freeze protection of Satsuma provided the catalyst for the present economic study. Protection of plants from freeze, the primary risk factor, can be provided at different protection and cost levels. The economic consequences of the trade-off between risk-reduction and increased production costs have not been previously evaluated for Satsuma in this area. This information would be useful for current and potential producers to make investment decisions relating to the establishment of Satsuma groves and/or the installation of freeze protection. The primary objectives of this study were to: 1) determine if the proposed freeze protection methods are economically feasible, 2) evaluate the effect of these freeze protection methods on the riskiness of net returns, and 3) conduct sensitivity analysis on key input variables for high tunnel technology. To achieve these objectives, enterprise 22 budgets for three hypothetical Satsuma groves in south Alabama, with a 20-year investment horizon, will be simulated. Stochastic dominance was used to compare simulations with standard inputs. In the sensitivity analysis, break-even prices and equivalent prices were also compared. This paper will proceed with a review of the literature, a section on methodology to describe the data and the model, followed by presentation and discussion of the simulation results, and concluding remarks. 2.2 Review of Literature 2.2.1 Satsuma Mandarin Satsuma mandarin (C. unshiu Marc.) is one of the most cold-tolerant of the commercial citrus species grown in the US (Hodgson, 1967). These citrus fruits have characteristics (sweetness, easiness of peeling, and seedlessness) which consumers find desirable (Campbell, et al. 2004). A thriving Satsuma mandarin industry existed in the Gulf Coast region of Alabama during the early 1900?s until a succession of freezes around 1940 decimated the groves and later freezes discouraged replanting (Ebel, et al., 2004, 2005). The re-development of a Satsuma industry in the Gulf Coast region of Alabama has been encouraged by a combination of factors. Technological developments in the use of micro-sprinkler irrigation within tree canopies have been successful in mitigating freeze risk in Satsuma in the Gulf Coast region (Bourgeois and Adams, 1987; Bourgeois et al., 1990; Nesbitt et al., 2000). Farmers in the counties of Baldwin and, to a lesser extent, Mobile, are facing increasing development pressure and increasing land prices (Lubowski et al., 2006; USDA, NASS, 2007; Wiebe and Gollehon, 2006). Maintaining 23 active farms will require the production of high value crops as cropland becomes scarcer and farms are reduced in size. Nationally, there is also increased interest on the part of consumers to buy locally grown produce (Gray, 2005). This trend may contribute to future grower interest in Satsuma production for localized sales. To facilitate the development of this industry, the USDA funded a multi- discipline research effort by Auburn University, in partnership with Louisiana State University, aimed at reducing production and marketing risk, and evaluating germplasm for potential use in breeding programs for Satsuma in Alabama and Louisiana. Current experimentation on freeze protection in Alabama includes the evaluation of high tunnels and the continued evaluation of micro-sprinklers for freeze protection. These two freeze protection methods are used as alternative strategies to the base plan with no protection in the simulation analysis of this paper. 2.2.2 Freeze Protection Satsuma is one of the most cold hardy of the commercial citrus species grown in the United States. Leaves are more sensitive to cold injury than the stems or trunks, and the minimum air temperature that causes injury is dependant upon the duration of the freeze event and the level of tree acclimation to cold (Yelenosky, 1991; Ebel, et al., 2005). Many environmental and biological factors affect cold acclimation; however, air temperature preceding the freeze event in citrus is the most important factor (Yelenosky, 1985, 1991, 1996). To acquire maximum cold hardiness, requires exposure of the tree to air temperatures ? 50 o F (10 o C) during the 500 hours (? 3 weeks) prior to the freeze event (Yelenosky, 1991). Based on historical levels of tree acclimation to cold in south 24 Alabama, the threshold for economically important injury occurs is between 22 and 18 o F (Ebel et al., 2005, Nesbitt et al., 2000; Nesbitt et al. 2002). This threshold represents the point where there is extensive leaf injury and some stem dieback such that the crop for the following harvest season is destroyed but the tree can recover and produce a normal crop the following year. At 14 o F (-10 o C), there is extensive injury to stems and the whole tree is susceptible to death if it is not acclimated to cold. Temperatures below 12 o F (-11 o C) have historically resulted in tree death for unprotected trees regardless of the level of cold acclimation. In the simulations for this study, a moderate freeze is considered to be one that causes the loss of the crop, but not the death of the tree. A severe freeze is one that causes injury to an unprotected tree to the point that the tree cannot recover. There have been many methods of freeze protection employed in the production of citrus with no one method being completely effective for all types of freeze events (Martsolf, 2000; Powell and Himelrick, 2007). Only two methods of freeze protection were modeled in the present economic study. The first method involved the use of micro-sprinkler irrigation, with the assumption that trees would not die in the event of a severe freeze but there would be a loss of fruit in the event of either a moderate or a severe freeze. The second method modeled was the use of high-tunnel technology with an assumption of full protection for both the trees and fruit in the event of either severe or moderate freezes. For the southern parts of Alabama and Louisiana, the use of micro-sprinkler irrigation within the tree canopies has been demonstrated to be effective in protecting tree trunks and major scaffold limbs during damaging freeze events (Bourgeois and Adams, 25 1987; Bourgeois et al., 1990; Nesbitt et al., 2000). This type of system does not protect the outer tree canopy, however, and the crop for the following growing season could be lost because leaves and stems are important for flower and fruit retention. In south Alabama, trees that experienced extensive injury to the outer canopy were able to return to normal production by the second growing season following the freeze event (Nesbitt et al., 2000). During an entire freeze event, the micro-sprinkler irrigation system must apply water continuously to all trees in the grove. The primary mechanism providing protection is the release of latent heat of fusion, as water crystallizes into ice (Powell and Himelrick, 2007). The system must also be independent of electricity, which may be unavailable during some freeze events. The system requirements for freeze protection are much greater than those needed for a normal irrigation regime which can use low volume emitters and zonal controls. The well and pump systems required for the freeze protection system represent a significant investment on the part of the producer. High tunnels are unheated, greenhouse-like structures with metal ribs covered with plastic. They are used extensively in Europe, Asia, and Israel to grow high-valued crops in areas with high population densities and constrained land and water resources (Orzolek, Lamont, and White, 2002). This technology was pioneered in the United States by Otho Wells, at the University of New Hampshire, and is now being tested for agricultural applications by several universities (Lamont, 2003). High tunnels are being used to extend growing seasons, increase quality, and reduce pesticide inputs. While relatively inexpensive when compared to traditional greenhouses, protecting plants with 26 high tunnels adds significantly to establishment costs and can only be justified for high- value crops. Auburn University currently has two high tunnel demonstrations with high- density plantings of Satsuma; one in the Gulf Coast region and one in the center of the state at the Chilton Area Horticultural Substation where there is an increased risk of freeze. The tunnels are 96 feet long and 24 feet wide and cover 30 trees each. The expectation for producing Satsuma under high tunnels is to totally eliminate fruit loss and tree death due to freeze. The high tunnels are only covered during the winter months since they are only used for freeze protection. White polyethylene plastic is used to prevent greenhouse effects associated with clear plastic that would cause plants to deacclimate and become less cold tolerant, or to thaw too quickly in the event of freezing temperatures. High tunnels are designed so that the sides can be raised for ventilation, but the top stays in place. Plants are covered from December, after the threat of hurricanes, until March or April after the threat of freeze has passed. Irrigation with micro-sprinklers under the canopies is provided and the irrigation system can add some additional heat in the event of severe freeze; however, no other heat source was added to the high tunnels in the demonstrations. 2.3 Data and Methodology 2.3.1 Yield Data Data evaluated for this study included yield records, production costs, and historical temperature records. Due to a lack of commercial production records, 27 experimental yield data was used to develop an expected production curve by tree year. Yield data for an ?Owari? Satsuma mandarin grove was collected by the Alabama Agricultural Experiment Station Gulf Coast Research and Extension Center at Fairhope, Alabama from initial planting in 1990 through crop harvest 2006-07. While this data set is not ideal, as trees have been used in various experiments, information on the changes in yield by tree age can be estimated from the grove average yields and standard deviations. Trees are not allowed to have production for the first two growing seasons in order to maximize tree growth. Average production levels off around the 9 th growing season and may exhibit biennial bearing in subsequent years (Ebel, et al., 2004). For the simulations, yield was assumed to increase from the third through the ninth year at which time an average production of 400 pounds per tree is reached and used as the mature average yield per tree (Appendix A1). To account for possible yield variation in early years and possible biennial bearing in later years, yields were modeled with ? 25-percent variation using the GRKS distribution for Simetar? that was developed by Gray, Richardson, Klose, and Schumann (Richardson, 2004). This distribution allows for the specification of the minimum, midpoint, and maximum values and approximately 95- percent of the simulated observations will fall between the minimum and the maximum and 50-percent are less than the midpoint. In addition, 2.2-percent of the observations will fall below the minimum and 2.2-percent will be above the maximum. 2.3.2 Yield Ratio An important factor for comparing the different systems is the ?yield ratio?. The yield ratio is defined as the ratio of the expected yield of one tree on trifoliate rootstock 28 to the expected yield of one tree on dwarfing rootstock. The yield ratio can also be take to mean the number of trees on dwarf rootstock needed to equal the yield of one tree on conventional trifoliate rootstock. The trifoliate rootstock, ?Rubidoux? is used for both the non-protected and the micro-sprinkler freeze protected Satsuma groves with conventional plant spacing, and the dwarfing rootstock ?Flying Dragon? is used for the high tunnel freeze protected grove with high-density spacing. For the unprotected groves, a planting density of 116 trees per acre with Satsuma ?Owari? on ?Rubidoux? trifoliate orange rootstock was modeled. For the high-tunnel planting, a high-density tree spacing of 6 by 12 feet was used with 30 trees per tunnel with dimensions of 96 feet length and 24 feet width. To maintain trees within the confines of the high tunnels, Satsuma ?Owari? is grown on a dwarfing rootstock, ?Flying Dragon?. To compare the returns for the two different planting densities and plant sizes, the budgets were standardized on an equivalent yield basis. The ?yield-ratio? reflects the expected yield from a conventional tree in relation to the expected yield from high-density tree on dwarfing rootstock. Conversely, it will also reflect the number of high density trees needed to produce the same yield as one tree on conventional-spacing and rootstock. The implications of the yield ratio are that the higher the ratio (minimum is 1.0), the more high-tunnel plantings needed to equal an acre of conventional-spacing trees, and the greater the cost of high tunnel freeze protection. To determine an appropriate yield-ratio, Japanese research was consulted due to the lack of published yield data for the dwarfing rootstock, ?Flying Dragon?, in the United States. Based on studies in Japan, and given their cultivars and growing conditions, a yield-ratio of 2.0 was estimated and used as the standard in the simulations (Takahara et 29 al., 2001; Noda et al., 2001; Yonemoto et al., 2005). The accuracy of the translation of this yield-ratio to the ?Owari? cultivar under Alabama growing conditions will be determined in future years as the experimental groves mature. 2.3.3 Production Costs General production costs were obtained from a Satsuma enterprise budget developed by Hinson, Boudreaux, and Vaughn (2006) for Louisiana producers. The Louisiana budget utilizes the Mississippi State Budget Generator for computations (Mississippi State University). All variable costs for labor and materials, and fixed costs for machinery and packing line equipment are included and are detailed in Appendix Tables A2 and A3. Land was not included in the Louisiana budget because of the many options for owning or obtaining the resource and users were instructed to adapt the budget to their own situation. In the current study, land was assumed to be an appreciable asset with the discounted terminal value equal to the initial value and no land charges were included in the budget. Alabama producers were expected to have production methods and costs similar to Louisiana producers. The Louisiana budget did not include costs for irrigation; however, irrigation charges for a micro-sprinkler system were included in all simulation scenarios in the study. The quantity of variable inputs differed between production systems and was influenced by freeze occurrence and severity, tree age, and cost of freeze protection modeled (Appendix Table A2). Fixed costs were amortized based on the schedule in Appendix Table A3 at 6-percent with zero salvage value. Fixed costs associated with grove establishment (planting) were treated as one-time costs in the year they were realized. For the unprotected grove simulations, re- 30 establishment costs were realized any time that a severe freeze event occurred and the grove was replanted. The micro-jet and high tunnel simulations never had to re-establish groves. Freeze protection costs were developed from information supplied by the Gulf Coast Research Center at Fairhope, Alabama. The micro-sprinkler freeze-protection system was modeled with one emitter per tree situated in the canopy at 5 feet above ground. Emitter delivery rate was 30-gph. Each acre required one 4-inch well system with a 60-gpm capacity pump. The total fixed cost of this system included 1) a well, pump, and all below ground pipes with a cost of $6,350 per acre, amortized at 6-percent over the 20-year period, and 2) above ground pipes, tubing, and emitters that are replaced every 4 years with a cost of $185 per acre, amortized at 6-percent over each 4-year period. In addition, there was a one-time installation charge of 25 labor hours per acre and there was an annual $25 per acre maintenance charge. With these assumptions, the annual fixed cost was $607 and the direct/variable cost was $25 for a total cost of $632 per acre for freeze protection after installation (Appendix Tables A2 and A3). The high tunnel system fixed cost was $4,500 per tunnel and included construction materials for a 96 x 24 foot structure with two end walls and two layers of 20-year ground cloth for weed control. Forty labor-hours were required for assembly of each tunnel. The construction material cost was amortized at 6-percent over the 20-year period, but installation labor is given a one-time charge. Annual maintenance costs for each tunnel included a charge of $164 for the 6-mil white plastic with a two-year life, and 12 labor hours for replacement, maintenance, and removal of the plastic. With these assumptions, the annual fixed cost was $392 per tunnel and the direct/variable cost was 31 $279 per tunnel for a total cost of $671 per tunnel after installation (Appendix Tables A2 and A3). The number of tunnels with high-density plantings on ?Flying Dragon? rootstock needed to equal an acre of conventional plant spacing on ?Rubidoux? rootstock varied by the assumed yield-ratio. 2.3.4 Freeze Risk Severe and moderate freeze events occur often enough in the Gulf Coast region of Alabama to introduce significant uncertainty into the production of cold-sensitive crops. In a study of the effect of freeze on Satsuma in the Baldwin County area, Ebel et al. (2005) compared daily min-max temperature data to reported tree injury for the period October 1948 through March 2004. During this period there were 8 years in which severe freeze events occurred and six additional years in which only moderate freeze events occurred. A severe freeze was classified as one that caused extensive tree injury or tree death. A moderate freeze was classified as one that caused extensive leaf injury and some stem dieback to the extent that the next growing season?s fruit crop was destroyed, but trees were back to full production by the 2 nd harvest season after the freeze. Based on this information, the long-term probability of a severe freeze in the Fairhope, AL area is 14-percent, and that of a moderate freeze is 11-percent. It is important to note that these two types of freeze event are, for our purposes, mutually exclusive; if both a severe and moderate freeze event occurred in the same growing period, only the severe freeze was counted. During the period of 1948-2004 in the Fairhope region, all freeze events occurred between the 12 th of December and the 9 th of 32 March when mature fruit would not typically be present on the trees or harvest would be near completion. For simplification in the simulation models, all freezes are assumed to occur at the beginning of the period (in winter after harvest of the previous growing season crop) and reduce the yield, as appropriate for the protection system, for the current growing season. 2.3.5 The Model The basic unit of study was a hypothetical 10-acre Satsuma grove with a 20-year investment horizon. A Satsuma grove could potentially remain productive for more than the 20-year period, however, this time-period was chosen as the maximum time in which an investor could make meaningful comparisons between alternative scenarios. Using a 2.0 yield ratio between conventional and high-density plantings, 7.7 high tunnels were required to equal one acre of conventional production and the unit is referred to as ?acre- equivalents? in the budgets. Thus the 10-acre-equivalent for high tunnels would be 77 tunnels. It should be pointed out that a whole number of tunnels was always used and that this may cause a few more or less trees to be included in the high tunnel grove than in the alternative groves depending upon the yield ratio used. The actual land requirement for high tunnel production will depend upon the placement of the tunnels; however, since land charges were not included in the budgets, the unit of land is of no consequence in the simulations. Simulations were run for the hypothetical groves to compare three alternative production methods. These production methods will be identified in all further discussion as 1) ?Unprotected? for the grove with no freeze protection, 2) ?Micro-jet? for 33 the grove with micro-sprinkler irrigation freeze protection, and 3) ?High Tunnel? for the high-density grove protected with high tunnels. Analysis of a multi-year operation is complicated by: 1) risk probabilities in multiple years which may be independent or correlated, 2) the impact that decisions and occurrences in one year have on decisions and outcomes in future years, and 3) input costs and average yields that vary with the age of the plant. The simulations in this study approached these considerations in the following ways: 1) Freeze events were assumed to be independent across years. Each year of the operation, the occurrence of a freeze event followed a uniform (0,1) distribution with Latin Hypercube sampling (Inman, Davenport and Zeigler, 1980). 2) All freezes were assumed to occur at the beginning of a calendar year and affect the yield in the coming fall. For clarity, the effect of freeze events on yield is presented in Table 1. If no freeze occurred, all groves produced yields and incurred costs dependent on the tree age, and the tree age advanced another year. If a moderate freeze occurred, a) High Tunnel groves produced a yield based on tree age and advanced one year in tree age, b) Unprotected and Micro-jet produced no yields in the fall, regardless of tree age, and c) tree age for all groves advanced another year. If a severe freeze occurs, a) Unprotected groves were assumed to die and were re-planted in late spring, b) Micro-jet groves lived and advanced one year in tree age, produced no fruit in the fall but had yields the following year based on tree age and that year?s freeze occurrence, and c) High Tunnel groves produced a yield based on tree age and advanced one year in tree age. 34 Table 1. Effect of Freeze Event on Yield of Simulated Satsuma Grove Freeze Event Unprotected Micro-jet High Tunnel No Freeze no effect no effect no effect Moderate lose crop lose crop no effect Severe lose tree lose crop no effect Satsuma Grove 3) Direct and variable input costs of materials and labor are a factor of tree age and fruit yield; tree age and/or yield were affected by freeze event for unprotected and micro- sprinkler protected groves (Appendix Tables A1 and A2). No limit was placed on the number of times a grove was replanted and incurred re-establishment costs again following a severe freeze in the simulations. An enterprise budget was used as the basis for the model with the random values being freeze event and tree yield. The key output variables were total revenue, returns above variable costs, and net returns to management. The budget for each hypothetical grove was simulated for 1,000 iterations using Excel 2003? and the add-in program Simetar?. The simulations were run at the standard values for fruit price, yield ratio, and high tunnel cost, which were $.50 per pound, 2.0, and $4,500 per tunnel, respectively. Additional scenarios were also simulated for fruit prices ranging from $.25 to $1.00 per pound, for yield ratios ranging from 1.5 to 3.0, for high tunnel fixed costs ranging from $1,500 to $5,500 per tunnel, and for high tunnel variable costs ranging from 50 to 125- percent of the standard. The yield ratio and tunnel cost scenarios had no effect on the returns for the unprotected and the micro-sprinkler protected trees and were used for sensitivity analysis of the high tunnel production system. 35 2.4 Simulation Results and Discussion 2.4.1 Simulations with Standard Values Results for the baseline simulations using standard parameters are presented in Table 2. This table details the values used as the key parameters, and summarizes the costs, revenues, and returns that were discounted at 6-percent and totaled over the 20-year simulation period for each production scenario. Interim values, totaled at 5-year increments, for the Income Above Variable Costs and Net Return to Management variables are also presented in this table. The interim values give an indication of how quickly each strategy produced positive net returns to management. Descriptive statistics for the 20-year key output variables are presented in Table 3. Average total fruit production for the 20-year period was 7,080,202 pounds for the High Tunnel grove. The High Tunnel strategy was modeled to give total protection from freeze losses for both the trees and fruit and therefore the yield represents the maximum possible in the absence of freezes. In comparison, the average production from the Micro-jet and the Unprotected groves were reduced by 24.7-percent and 53.3-percent, respectively. Simulations were standardized through the yield-ratio to equate the groves on an equivalent yield basis. In the absence of freezes, all groves would be expected to have the same total production. Fruit yields for the Unprotected and Micro-jet groves would change only with variation in freeze probability or changes in the tree yield assumptions. 36 Table 2. Twenty-year Discounted Costs and Returns for 10-Acre Satsuma Grove in South Alabama with Different Freeze Protection Methods. KEY PARAMETERS: Time Conventional Freeze Probability: Severe: 0.11 Moderate: 0.14 Period Discount Rate 0.06 Conventional Trees/Acre 116 High-Density Trees/Tunnel 30 5 Year 16,790 29,132 -27,226 10 Year 164,010 283,297 257,435 Standard 15 Year 288,933 507,909 517,597 Price ($/lb) 0.50 0.25 0.50 0.75 1.00 20 Year 383,018 676,173 712,019 Yield Ratio (Conven./High Density) 2.0 1.5 2.0 2.5 3.0 Tunnel Fixed Cost ($/Tunnel) 4,500 2,500 3,500 4,500 5,500 Tunnel Variable Cost ($/Tunnel) 1.00 0.50 0.75 1.00 1.25 5 Year -18,591 -28,057 -225,651 Standard 1.5 2.0 2.5 3.0 10 Year 110,369 194,765 -48,318 Tunnels/Conventional Acre 7.7 5.8 7.7 9.7 11.6 15 Year 221,672 395,955 131,642 High-Density Tree/Acer Equiv 232 174 232 290 348 20 Year 305,575 546,718 266,132 Unit $/Unit Unprotected Micro-jet High Tunnel GROSS RECEIPTS Yield/10 Ac or Ac equiv. lb 3,305,089 5,332,031 7,080,202 Discounted Revenue ac/ac equiv $828,601 $1,317,314 $1,748,953 VARIABLE/DIRECT COSTS Pest/Disease/Weed Control ac/ac equiv. variable 43,777 55,044 40,064 All other material inputs ac/ac equiv. variable 26,495 30,507 30,507 Other Labor ac/ac equiv. variable 52,988 57,262 57,262 Pruning Labor hr 9.60 5,431 6,261 12,523 Specific System Maintenance ac/ac equiv. variable 0 3,303 238,072 Harvest Labor & Materials bushel 3.50 145,108 230,541 306,067 Other Harvest/Pack Costs ac/ac equiv. variable 151,689 229,021 305,362 Interest on Operating Capital ac/ac equiv. variable 20,211 29,067 47,018 ESTABLISHMENT COSTS Land Prep/Plants/Labor ac/ac equiv. variable 34,997 13,852 26,236 FIXED COSTS Equipment & Irrigation ac/ ac equiv 42,450 43,583 43,583 Freeze Protection or Tunnel ac/ ac equiv variable 0 72,020 376,068 Conventional High Tunnel Discounted Net Return to Management: + Unprotected + Micro-jet High Density Discounted Income Above Variable Costs: Scenarios - Evaluated Yield Ratio Min, Mean, Max .75, 1.0, 1.25 GRKS Yield Distribution 37 Table 3. Twenty-Year Key Output Variables a from Simulations using Standard Parameters b Total Revenues $828,601 $1,317,314 $1,748,953 Standard Deviation 360,827 190,916 55,237 Coefficient of Variation 44 14 3 Income Above Variable Costs $383,018 $676,173 $712,019 Standard Deviation 222,405 125,547 45,103 Coefficient of Variation 58 19 6 Net Return to Management $305,575 $546,718 $266,132 Standard Deviation 232,630 125,547 45,103 Coefficient of Variation 76 23 17 a Values are discounted at 6%. b Yield ratio = 2.0, high tunnel cost = $4,500/tunnel, and market price = $.50/lb. Unprotected High TunnelMicro-jet Producers in the Gulf Coast area have been able to sell all of the fruit they produced at prices ranging from $.30 to $.80 per pound. Based on conversations with industry specialists, the market price of $.50 was used as the average expected price and the standard for the simulations. At the market price of $.50 per pound, all production methods modeled had positive net returns after 20 years. Using either freeze protection method reduced the variability of net returns, but at this market price, mean returns for the Micro-jet grove were superior to the other two methods. The Unprotected grove produced 38-percent less fruit than the Micro-jet grove due to severe-freeze-induced tree loss, and the subsequent production lag following replanting. The Micro-jet grove lost fruit, but not trees, after severe freezes and returned to normal production the following year. Even though the High Tunnel grove had greater total fruit production than the Micro-jet grove, at the standard market price, yield ratio, and tunnel cost, the 20-year net returns to management are 51-percent less. The relatively higher initial investment and 38 the higher annual maintenance cost for the high tunnels are the primary costs contributing to reduced net returns for the High Tunnel grove in comparison with the Micro-jet grove. 2.4.2 Stochastic Dominance The cumulative density functions (CDF) for the three standard simulations are illustrated in Figures 1 and 2, for market prices of $.50/lb and $1.00/lb, respectively. Under first-degree stochastic dominance, plan A would dominate plan B if F A (x) ? F B (x) for all levels of x (Harwood, et al., 1999). There is no clear dominance of one strategy 0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90 1.00 -$400,000 -$200,000 $0 $200,000 $400,000 $600,000 $800,000 $1,000,000 $1,200,000 20-Year Discounted Net Returns ($/Operation Unit) C u m u l a tiv e P r o b a b ili t y High Tunnel Micro-jet Unprotected Figure 1. Cumulative Distributions of 20-Year Discounted Net Returns to Management for Three Satsuma Production Strategies at Market Price = $.50 per Pound 39 0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90 1.00 -$500,000 $0 $500,000 $1,000,000 $1,500,000 $2,000,000 $2,500,000 $3,000,000 20-Year Discounted Net Returns ($/Operation Unit) C u m u la tiv e P r o b a b i l ity . Unprotected Micro-jet High Tunnel Figure 2. Cumulative Distributions of 20-Year Discounted Net Returns to Management For Three Satsuma Production Strategies at Market Price = $1.00 per Pound over the other in either price scenario because each CDF crosses another at some point. However, at $.50/lb market price, the Micro-jet strategy exhibits first degree stochastic dominance over the other two strategies except in the upper and lower tails of the distribution of returns. To rank the strategies, second-degree stochastic dominance procedures must be used and preference between the strategies would depend upon the decision maker?s utility function and risk aversion preference. Stochastic Dominance with Respect to a Function (SDRF) in Simetar? is a mathematically rigorous method of using the complete empirical distribution to rank scenarios with different risk strategies (Richardson, 2004). It relies on theory for the measurement of risk aversion developed by Pratt (1964) where decision makers have an expected utility function for money, u(x), that is increasing and twice differentiable. The absolute risk aversion coefficient (ARAC) is defined as r(x) = -u?(x)/u?(x). While Pratt limits the coefficient to describe risk adverse individuals, where u?(x) > 0 and u?(x) < 0, 40 the Simetar? program follows Meyer (1977) and allows for a grouping of decision makers into risk preference groups based on similar ARACs: Let U(r 1 (x), r 2 (x)) represent decision makers with preferences represented by r(x) over the range r 1 (x) ? r(x) ? r 2 (x) for all x. Using the above assumptions and distributions F and G bounded over the interval of 0 to 1, two risky alternatives, F(x) and G(x), with utility function u(x) are compared. Over the probability range of zero to one, F(x) is preferred to G(x) when: 1 1 (1) ? 0 u(x) d F(x) ? ? 0 u(x) d G(x). Rearrangement of equation (1) yields: 1 (2) ? 0 [G(x) ? F(x)] u?(x) dx ? 0. The SDRF program requires an assumption on the form of the utility function. Following Featherstone and Moss (1990), a negative exponential utility function was assumed such that: (3) U [W(x)] = -exp [-?W(x)], where wealth, W, is a function of net return, x, and ? is the Pratt absolute risk aversion coefficient. The negative exponential utility function assumes constant absolute risk aversion and increasing relative risk aversion. The stochastic dominance function also utilizes the utility function to calculate certainty equivalents (CE) coefficients to rank alternative strategies. The CE value is the net return required so that a decision maker with a given ARAC and utility function 41 would be indifferent between the investment and a no-risk investment. To calculate certainty equivalents from the negative exponential utility function, an assumption is also made that returns are distributed multivariate normal such that W(x) ~ N[?(x), ? 2 (x)]. Given these assumptions, Featherstone and Moss (1990) detail the derivation of the certainty equivalent formula from equation (3) by setting the inverse utility function to be equal to the expected utility. The resulting certainty equivalent formula is: (4) CE = W*(x) = ?(x) ? ?/2 ? 2 (x), where W*(x) is the certainty equivalent, ?(x) is the expected mean net return, ? 2 (x) is the variance, and ? is the Pratt ARAC. The SDRF program was run for the simulation distributions using an ARAC range of -0.1 ? r(x) ? +0.1 and the negative exponential utility function for three market price scenarios. The certainty equivalents and rankings of the different production strategies are presented in Table 4. The ranking preference, based on CEs changed as the absolute risk aversion coefficient changed from negative (risk loving) to positive (risk averse). With ARAC = 0 (risk neutral), the CE is equal to the mean of the net return to management with no consideration of the variance of the distribution. For a risk averse decision maker with the assumed utility function, the High Tunnel strategy is preferred in all market price scenarios; the lower variance in distribution of net returns to management for the High Tunnel strategy is a significant factor in this result at price levels of $.50 and $.75/lb. 42 2.4.3 Equivalent Prices The 20-year discounted net returns for all freeze protection strategies has a linear response to market price, as illustrated in Figure 3. Equivalent prices between the strategies, calculated from the response slopes, are $.253 between Unprotected and Micro-jet, $.521 between Unprotected and High Tunnel, and $.827 between Micro-jet and High Tunnel. The equivalent prices are market prices where the mean 20-year net returns to management are equal between the two strategies being compared. The large increase in the equivalent prices when alternative strategies are compared to the High Tunnel indicates that the expense of growing trees in high tunnels can only be justified if market prices are expected to exceed these equivalent prices, given the assumptions in the simulation and risk neutral preferences. Table 4. Effect of Market Price and Absolute Risk Aversion Coefficient on Certainty Equivalents and Ranking for Three Satsuma Production Strategies with Negative Exponential Utility Function CE b Rank CE b Rank CE b Rank Price = $.50/lb Unprotected 969,498 1 305,455 2 -189,748 3 Micro-jet 894,018 2 546,850 1 53,865 2 High Tunnel 402,665 3 266,192 3 138,863 1 Price =$.75/lb Unprotected 1,881,971 1 719,756 3 -162,619 3 Micro-jet 1,806,491 2 1,205,507 1 344,755 2 High Tunnel 1,360,695 3 1,140,669 2 935,385 1 Price = $1.00/lb Unprotected 2,794,444 1 1,134,057 3 -135,490 3 Micro-jet 2,718,963 2 1,864,163 2 635,644 2 High Tunnel 2,318,724 3 2,015,145 1 1,731,908 1 a Negative ARAC = Risk loving, 0 = Risk neutral, Positive ARAC = Risk averse. b CE = Certainty Equivalent. Absolute Risk Aversion Coefficient a -0.1 0 0.1 43 -1,000,000 -500,000 0 500,000 1,000,000 1,500,000 2,000,000 2,500,000 0.25 0.50 0.75 1.00 Market Price per Pound ($) N e t R e tu rn s ($ ) Unprotected Micro-Jet High Tunnel Figure 3. Twenty-Year Discounted Net Returns to Management for Satsuma at Different Different Market Prices, Fairhope, Alabama 2.4.4 High Tunnel Cost Analysis Total elimination of freeze risk for Satsuma production in the Gulf Coast region of Alabama would increase production efficiency and potentially benefit both producers and consumers. There is currently no oversupply of production and an opportunity for expanding sales exists. Under these conditions, a negative market price response would not be expected with increased production. However, the use of high tunnels to eliminate the risk of crop loss due to freeze may require higher average market prices than currently exist or lower production costs. Production costs for the High Tunnel strategy are affected by the yield ratio used to equate the high density planting to the conventional planting (yield from tree on standard rootstock in relation to yield from tree on dwarf rootstock), the fixed construction costs, and the annual maintenance costs. Simulations 44 were run with scenarios that varied these input values to evaluate their effect on net return. The simulation results from varying the high tunnel fixed costs and the yield ratio on 20-year net returns across a range of market prices are presented in Table 5. The effect of varying the initial tunnel construction cost on 20-year discounted net returns is a simple algebraic equation given that the discount rate and the amortization rate assumptions used in the simulations are both 6-percent. A one-dollar decrease in tunnel cost will result in a net return increase of one dollar per tunnel with the number of tunnels varying due to the yield ratio. Yield ratio had an inverse effect on net returns. Yield ratio is the ratio of the yield of a tree on conventional ?Rubidoux? rootstock to the yield of a tree on the dwarfing rootstock ?Flying Dragon?. As the yield ratio increased, there was a greater negative effect on the net returns due to the increased number of tunnels needed to equal conventional production. Each increase of .5 in the yield ratio resulted in approximately a 10-percent increase in the breakeven price for all high tunnel costs. Market price effects, for the different strategies, reflected changes in total fruit production and net returns exhibited a linear response to price. Each 25-cent increase in market price increased net returns by $414,301 for the Unprotected grove, $658,657 for the Micro-jet grove and $874,477 for the High Tunnel grove over the 20 year life of the project. 45 Table 5. Simulated Discounted a 20-Year Net Return to Management at Different Price Levels for Satsuma with Different Freeze Protection Methods - Fairhope, Alabama $1,500 $2,500 $3,500 $4,500 $5,500 Price/lb Unprotected Micro-Jet 0.25 -108,845 -11,807 -267,832 -325,827 -383,827 -441,809 -499,832 0.50 305,456 546,850 610,451 552,451 494,451 436,451 378,451 0.75 719,757 1,205,507 1,488,711 1,430,711 1,372,711 1,314,711 1,256,711 1.00 1,134,057 1,864,164 2,366,971 2,308,971 2,250,971 2,192,971 2,134,971 Price Intercept d 0.316 0.292 0.326 0.343 0.359 0.376 0.392 0.25 -377,308 -454,303 -531,303 -608,285 -685,303 0.50 497,192 420,192 343,192 266,192 189,192 0.75 1,371,669 1,294,669 1,217,669 1,140,669 1,063,669 1.00 2,246,145 2,169,145 2,092,145 2,015,145 1,938,145 Price Intercept d 0.358 0.380 0.402 0.424 0.446 0.25 -485,437 -582,418 -679,437 -776,437 -873,437 0.50 395,872 298,872 201,872 104,872 7,872 0.75 1,277,163 1,180,163 1,083,163 986,163 889,163 1.00 2,158,453 2,061,453 1,964,453 1,867,453 1,770,453 Price Intercept d 0.388 0.415 0.443 0.470 0.498 0.25 -594,433 -710,433 -826,433 -942,433 -1,058,433 0.50 283,849 167,849 51,849 -64,151 -180,151 0.75 1,162,113 1,046,113 930,113 814,113 698,113 1.00 2,040,377 1,924,377 1,808,377 1,692,377 1,576,377 Price Intercept d 0.419 0.452 0.485 0.518 0.551 a Discount rate is 6.0%. b Dollar cost per tunnel. c Yield ratio for Conventional:High Density (High Tunnel) plantings. d Break-even price/lb of fruit produced. Ratio c = 2.0 Ratio c = 1.5 High Tunnel Cost b Ratio c = 2.5 Ratio c = 3.0 An evaluation of the equivalent prices restates the relationship between the High Tunnel strategy and the other two strategies in terms of fruit market price (Table 6). Equivalent prices are the market price where the mean 20-year net returns are equal between the two strategies being compared and were calculated from the price response regression lines. Average market prices above the equivalent price indicate that net 46 returns are higher for the strategy that has the higher yield. With a yield ratio of 2.0 or less, the High Tunnel strategy would return more than the Unprotected strategy at market prices in the range of $.50 per pound, however, they would not return more than the Micro-jet strategy were equivalent prices are generally in excess of $.50 per pound. An advantage to using high tunnel technology would be the expectation of a crop in the event of either a severe or a moderate freeze; a risk adverse decision maker may consider the use of this technology for reasons other than achieving the greatest average net return. Horticultural research focusing on pruning methods, tree nutrition, and other factors affecting fruit yield may potentially impact the cost effectiveness of the high density planting. Table 6. Equilavent Prices between Freeze Protection Technologies with Varying Yield Ratios and High Tunnel Fixed Cost Unprotected 1 1 1 1 0.253 0.253 0.253 0.253 Mic-J 0.2530.2530.2530.253 1111 High Tunnel Fixed Cost: 1.50 2.00 2.50 3.00 1.50 2.00 2.50 3.00 HT - $1,500 0.335 0.396 0.454 0.511 0.428 0.558 0.680 0.801 HT - $2,500 0.367 0.437 0.506 0.573 0.494 0.648 0.791 0.933 HT - $3,500 0.398 0.479 0.558 0.636 0.561 0.737 0.901 1.066 HT - $4,500 0.429 0.521 0.611 0.698 0.627 0.827 1.012 1.198 HT - $5,500 0.460 0.563 0.663 0.761 0.693 0.916 1.123 1.330 Yield Ratio Yield Ratio Unprotected Micro-Jet Simulations that varied the annual maintenance costs for the High Tunnel strategy were also run (Table 7). A one-percent decrease in variable costs was found to increase 20-year discounted net returns by $2,492 at the standard yield ratio of 2.0. In comparison, a one-percent decrease in fixed construction cost resulted in an increase of $3,465 for 20-year discounted net returns. These changes in net return were static and 47 were not affected by market price of fruit. Reduction in either the initial construction costs or the annual variable costs were assumed to have no effect on the performance of the high tunnels. Cost reductions may be achieved through any number of ways including increased labor efficiency and volume discount purchases of materials. The effects of changing the variable and fixed costs on the break-even price for the high tunnel grove, at the standard yield ratio of 2.0, are also presented in Table 7. Table 7. Simulated Discounted a 20-Year Net Return to Management at Different Price Levels for Satsuma with Changes in High Tunnel Variable Costs - Fairhope, Alabama $1,500 $2,500 $3,500 $4,500 $5,500 Price/lb Unprotected Micro-Jet 0.25 -108,845 -11,807 -252,689 -329,689 -406,689 -483,712 -560,708 0.50 305,456 546,850 621,787 544,787 467,787 390,787 313,787 0.75 719,757 1,205,507 1,496,264 1,419,264 1,342,264 1,265,264 1,188,264 1.00 1,134,057 1,864,164 2,370,741 2,293,741 2,216,741 2,139,741 2,062,741 Price Intercept c 0.316 0.292 0.322 0.344 0.366 0.388 0.410 0.25 -315,005 -392,005 -469,010 -546,010 -623,005 0.50 559,490 482,490 405,490 328,490 251,490 0.75 1,433,966 1,356,966 1,279,966 1,202,966 1,125,966 1.00 2,308,443 2,231,443 2,154,443 2,077,443 2,000,443 Price Intercept c 0.340 0.362 0.384 0.406 0.428 0.25 -377,303 -454,303 -531,303 -608,285 -685,285 0.50 497,192 420,192 343,192 266,192 189,192 0.75 1,371,669 1,294,669 1,217,669 1,140,669 1,063,669 1.00 2,246,145 2,169,145 2,092,145 2,015,145 1,938,145 Price Intercept c 0.358 0.380 0.402 0.424 0.446 0.25 -439,605 -516,601 -593,601 -670,601 -747,601 0.50 434,894 357,894 280,894 203,894 126,894 0.75 1,309,371 1,232,371 1,155,371 1,078,371 1,001,371 1.00 2,183,848 2,106,848 2,029,848 1,952,848 1,875,848 Price Intercept c 0.376 0.400 0.420 0.441 0.464 a Discount rate is 6.0-percent. b Dollar cost per tunnel. c Dollar cost per tunnel. Note: The Yield Ratio for conventional:high density (High Tunnel) plantings is 2.0. Variable Cost = 1.25 * Standard High Tunnel Cost b Variable Cost = .50*Standard Variable Cost = Standard Variable Cost = .75 * Standard 48 Break-even prices at the 2.0 yield ratio range from $.322 to $.464 per pound with the input assumptions used in the simulations that varied high tunnel variable and fixed costs. Break-even prices across all yield ratios ranged from a low of $.326 to a high of $.551 per pound when only fixed costs were varied (Table 5). These prices indicate that it would economically feasible to produce Satsuma mandarins with high tunnel technology for freeze protection when market prices are above $.45 per pound at the standard yield ratio 2.0, and above $.55/lb at the 3.0 yield ratio. An important advantage of the high tunnels was the reduction in the variance of revenues and net return to management over the 20-year period (Table 3). However, because of the significant investment and maintenance costs for the high tunnels, other freeze protection methods may have greater average returns to management in the Gulf Coast area. The high tunnel economic evaluation developed for the Gulf Coast area may be preferred in other areas that have higher freeze risk as long as the tunnels offer sufficient protection from the minimum temperatures for the area of interest. It may be necessary to add supplemental heat to adequately protect trees with high tunnels in many areas; the analysis in this study will be relevant as long as the fixed and variable costs for the tunnel and additional heat source are maintained within the ranges evaluated. 2.5 Conclusions Three Satsuma mandarin groves with different freeze protection strategies were evaluated in this study through the simulation of their respective enterprise budgets over a 20-year investment horizon. The grove with no freeze protection (Unprotected) and the grove with micro-sprinkler freeze protection (Micro-jet) were modeled as 10-acre 49 enterprises with 116 trees to the acre. The higher-density planting with high tunnel freeze protection (High Tunnel) was equated to the other groves on a yield basis through a yield ratio which is the ratio of the expected yield from an ?Owari? Satsuma on ?Rubidoux? rootstock to the expected yield of ?Owari? Satsuma on the dwarfing rootstock, ?Flying Dragon?. With the standard yield ratio of 2.0, 77 high tunnels were needed to equal the expected production from the 10-acre enterprises. The groves were simulated using standard values of $4,500 high tunnel construction cost, a 2.0 yield ratio, and $.50 per pound wholesale market price for fruit. Total revenues, return above variable costs, and net return to management were key output variables that were discounted at 6 percent over the 20-year period. Using the standard input variables, all groves had positive discounted net returns to management after 20 years. At the $.50 per pound market price, the mean returns were highest for the Micro-jet strategy, but the returns for the High Tunnel strategy had the lowest variance. Preference between the strategies may depend upon the decision makers utility function and risk preferences. All strategies exhibited a linear response to market price. Changes in the market price resulted in the greatest change in 20-year net returns for the High Tunnel strategy, followed by the Micro-jet, and then the Unprotected. The ranking of net-return response to market price was due to the total fruit production for each strategy over the 20-year period. The use of high tunnels eliminated production loss due to tree or fruit injury in these simulations. The lowest total yields were attributed to the Unprotected strategy, which was subject to tree and/or fruit loss depending upon the freeze severity each year. 50 Total elimination of freeze risk in the Gulf Coast region of Alabama with the use of high tunnels necessitates a significant investment in high tunnel initial construction and annual maintenance costs. Maintaining, or improving, a 2.0 yield ratio between conventional plantings on ?Rubidoux? rootstock and high-density plantings on ?Flying Dragon? rootstock is a subject open for horticultural study. Yield ratio determines the number of tunnels needed to equal one acre of conventionally spaced production. As the yield ratio increases, more tunnels are needed to equal the potential yield of the trees with conventional plant spacing and net returns become more sensitive to the effect of tunnel costs. Sensitivity analysis varying the fixed and variable costs for the High Tunnel strategy showed that a one-percent decrease in high tunnel fixed cost resulted in a $3,465 increase in 20-year discounted net returns and a one-percent decrease in variable maintenance costs resulted in a lesser decrease of $2,492. Break-even prices across all yield ratios, fixed costs, and variable cost combinations ranged from a low of $.300 per pound to a high of $.551 per pound. The break-even prices indicate the market price at which it is economically feasible to produce Satsuma mandarins with high tunnel freeze protection under the different given assumptions. Freeze protection with high tunnels requires significant investment and maintenance costs for the high tunnels, and other freeze protection methods may have greater returns to management in the Gulf Coast area. The high tunnel economic evaluation developed for the Gulf Coast area may be used in any area with greater freeze risk as long as the tunnels offer sufficient protection from the minimum temperatures for the area of interest. Additional simulations based on this platform can be used to 51 determine the freeze probability conditions under which High Tunnel technology would be preferred to Micro-jet technology. 52 IV. EFFECT OF LOCAL VARIATION IN FREEZE PROBABILITY ON NET RETURNS FROM THREE PROTECTION TECHNOLOGIES 3.1 Introduction Reduction of risk in agriculture is a subject of much interest for producers and researchers alike and there is a large body of literature devoted to the subject. Risk, in its simplest term, refers to the possibility of experiencing a loss with a given probability of occurrence. In order to achieve effective risk reduction, either the severity of the loss or its probability must be reduced to the extent that the outcome is improved. Practices that reduce the severity of loss are termed ?loss reduction? or ?self insurance?, while practices that reduce the probability of loss are termed ?loss prevention? or ?self protection? (Briys and Schlesinger, 1990). It is not always possible, however, to know the distribution of a risk variable due to its random occurrence and this introduces uncertainty into the decision process (Knight, 1921). In Chapter 2, an economic evaluation of risk reduction methods for a Satsuma mandarin grove in the Fairhope, AL area, was presented. The methods evaluated were micro-jet sprinkler irrigation to prevent the loss of trees due to freeze, and high tunnels to prevent the loss of both trees and fruit due to freeze. Each of these methods is a form of self-insurance; they do not prevent the freeze event from occurring, but reduce the deleterious effects of the freeze event on profitability of the grove. Investment in 53 self-insurance has been shown to increase with an increase in the decision maker?s risk aversion (Hiebert, 1989, Dionne and Eeckhoudt, 1985, Ehrlich and Becker, 1972). With consideration of the level of risk aversion, both the expected net return and the associated distribution of returns are important to the decision maker. An assumption of risk neutral preference, however, simplifies the comparison of different self-insurance methods to an evaluation of the expected net return. A risk neutral individual would prefer the strategy that yields the highest expected net return. Satsuma mandarins require mild winters and cool autumn temperatures in allow for tree survival and optimum fruit quality (Ebel et al., 2004). In the United States these conditions are found in the northern Gulf Coast region states, and in certain areas of Arizona and California. In the Gulf Coast region, minimum winter temperatures may reach levels that cause tree injury. The evaluations in Chapter 2 were conducted using the probability of severe and moderate freeze events developed from weather data and observations of tree injury over a 56-year period for the Fairhope, AL area. Severe freeze was classified as one that caused extensive tree injury or tree death, and a moderate freeze was classified as one that caused extensive leaf injury and some stem dieback to the extent that the next season?s fruit crop was destroyed. While historical data is not an ideal predictor of future events, it is the best indication of the expected long-term freeze probability available in the absence of more accurate weather prediction models. The occurrence of weather events in any given location are considered acts of nature that cannot be directly influenced by the actions of a producer. However, an expectation of weather events, or probability of a weather event occurring, develops for different locations based on experience. A potential Satsuma producer in a location other 54 than the Fairhope, AL region would be expected to face different injurious freeze probabilities. Development of cultivars that are more cold tolerant than those used in current production or possible changes in global weather patterns could also change the expectation of freeze injury probability for a given location. It is not within the scope of this paper to determine how the changes would occur, but rather what would be the effect of different probabilities of severe or moderate freeze injury on the outcome of the simulation models. This information would be useful for decision makers facing uncertainty in future weather events. Net returns for Satsuma production under different freeze protection strategies would be expected to vary due to the occurrence of freeze events. This information would be useful to a decision maker facing uncertainty in future weather events or who has an expectation of freeze probabilities different from those used for the Fairhope, AL area. The objective of this study was to determine the effects of varying freeze probabilities on discounted net returns for hypothetical Satsuma groves that use different approaches to freeze protection. In the rest of this chapter, a review of the literature will be followed by a description of the methodology used, discussion of the results, and concluding remarks. 3.2 Review of Literature 3.2.1 Weather Data and Satsuma Cold Acclimation Long-term weather data, from 1948 to 2004, was matched to historical reports and research records of freeze injury on Satsuma in the Fairhope region to determine the probability of freeze occurrence and severity in this area (Ebel, et al., 2005). During 55 this period, no more than two freeze events occurred in any given winter season (December through March) and the duration was less than three days for all occurrences except one. The effect of critical temperatures on Satsuma plants will be dependent upon the plant?s level of acclimation to cold prior to the freeze event. The air temperature during the 500 hours (? 3 weeks) preceding the freeze event have been determined to be the most important factor affecting cold acclimation (Yelenosky,1985, 1991, 1996). Trees were found to acclimate to cold when the air temperatures were ? 50 o F (10 o C). Ebel, et al. (2005) developed a model to determine the expected sensitivity of Satsuma to cold injury that incorporated the level of tree cold acclimation prior to exposure to potentially injurious temperatures. Trees that were not fully acclimated experienced economically important injury at temperatures of 22 o F (-5.5 o C) and tree death at temperatures below 14 o F (-10 o C). When trees are fully acclimated they could withstand temperatures down to 18 o F (-7.7 o C) before experiencing economically important injury and tree death did not occur until temperatures reach 12 o F (-11.0 o C). Concerns about possible climate change, either from long-term natural weather patterns or human induced weather changes, are widespread; there are many interdisciplinary studies being conducted and models being developed to evaluate the impact of climate change (Goulder and Pizer, 2006; Reilly, et al., 2003; US Global Change Research Program, 2006). Easterling et al., (1999) reviewed the literature on recorded freeze data from 1766 though the 1990?s and the occurrence of freeze injury to citrus in Florida. The studies that were reviewed found an association between freeze injury and the strong positive mode of the Pacific-North American (PNA) circulation pattern and no association with the El Ni?o ? Southern Oscillation (ENSO). Katz, 56 Parlange, and Tebaldi (2003) evaluated the relationship of nine atmosphere-ocean circulation indices with min/max temperature and precipitation time-series data (1959- 1996) for the southeastern United States. They established an association of higher minimum and maximum winter temperatures and higher probability of precipitation when the Bermuda High was farther east than average. Long-term and short-term weather cycles appear to occur but are not yet predictable (US Global Change Research Group, 2006) and therefore are ignored in the current study. Nevertheless, if changes in long-term weather patterns result in warming trends, an increase in the minimum temperature that occurs in an area, could also decrease tree cold acclimation and result in more frequent, though less severe, tree injury. 3.2.2 Satsuma Production Areas Satsuma mandarin is one of the most cold-hardy of the commercially grown citrus; however minimum winter temperatures in the Gulf Coast region of the US may reach levels that cause injury to trees. Moving Satsuma production to areas with lower probability of freeze occurrence may have adverse effects on fruit quality. High air temperature during the final fruit maturation period of October through December promotes poor peel color development and may accelerate the decrease in acidity to the extent that flavor is less than ideal (Ebel et al., 2004). These quality features benefit from cool temperatures during the final fruit maturation. Producing Satsuma in areas that are further north could increase the probability of either severe or moderate freeze injury and require higher levels of freeze protection. The USDA Plant Hardiness Zone map may be useful to a potential producer to identify suitable production areas. 57 Fairhope, AL is located in USDA Plant Hardiness Zone 8b, with average annual minimum temperature range of 15 to 20 o F (-9.4 to -6.7 o C). Satsuma are expected to be hardy to (-10 o C) and should thrive in Zone 8b. Over the 56 year period, 1948-2004, there have been 11 years when the Gulf Coast Research Center at Fairhope, AL recorded minimum temperatures below this average with the absolute minimum recorded during this period being 5.2 o F (-14.9 o C). The hardiness zone map was published in 1960 and revised in 1965; it is drawn on average annual minimum temperatures, which, necessitates that there are occurrences of temperatures below this range. The weather data collected by the Gulf Coast Research Center indicates that annual minimum temperatures were below the Hardiness Zone Map an average of 20-percent of the time from 1948-2004. The USDA Plant Hardiness Zone map may give an indication of areas that are suitable for Satsuma production, but more detailed minimum temperature information is required to develop an appropriate freeze probability factor for a given area. 3.3 Methodology 3.3.1 The Simulation Model The models developed in Chapter 2 for Unprotected, Micro-jet, and High Tunnel groves were used for simulations with variations in freeze probabilities. Simetar?, an Excel add-in program, was used with Excel 2003 to simulate the performance of each grove over a 20-year period for 1,000 iterations. The primary output variable of interest was the accumulated discounted net returns to management: 58 (1) NR d = ? [ (PY j (f(?),t) ? C j (t) ? VC j (y)) / (1 + r) j ] where NR d = total discounted net returns over the 20-year period; j = the simulation year; P = market price per pound; Y = fruit yield as a function of tree age, t, and freeze event, f, that occurs with probability ?; C j (t) = fixed and direct costs in the jth year as a function of tree age; VC j = variable costs as a function of yield in the jth year; and r = the discount rate. The 20-year net return variable is linear in price. The simulations were conducted over a range of prices in order to calculate price response lines. Price response lines were used to determine break-even prices for each strategy and equivalent prices between the strategies for each simulation scenario. Break-even prices are equal to the price intercept from the price response line. Equivalent prices are the market price where the price response lines from two different strategies intersect: (2) EP ab = (PI a ? PI b ) / (S b ? S a ) where subscripts a and b refer to two different production strategies; EP is the equivalent price; and PI and S are the intercept and slope, respectively, of the applicable price response line. 3.3.2 Model Variables The basic unit of study was a hypothetical 10-acre Satsuma grove with a planting density of 116 trees per acre. There were three groves modeled: 1) one grove with no freeze protection which will be referred to as ?Unprotected?, 2) one with micro-jet sprinklers placed in the tree for freeze protection, referred to as ?Micro-jet?, and 3) one 59 grove protected by high tunnels, referred to as ?High Tunnel?. Trees in the High Tunnel grove were grown on the dwarfing rootstock, ?Flying Dragon?, and have a planting density of 6 feet in the row by 12 feet between rows so that each 96 x 24 foot high tunnel covers 30 trees. The dwarfing rootstock is desirable to more easily maintain tree growth within the confines of the high tunnels. The Unprotected and the Micro-jet groves on the conventional planting density were planted on ?Rubidoux? trifoliate orange rootstock. The groves with conventional planting density were equated to the high-density grove through equivalent yield and not through equivalent land area. Based on Japanese research (Takahara et al., 2001; Noda et al., 2001; Yonemoto et al., 2005), an assumption was made that it takes two trees on ?Flying Dragon? rootstock to produce the same yield as one tree on conventional ?Rubidoux? rootstock. This resulted in a 2.0 ratio between the yield of a conventionally grown tree and the high density tree on dwarfing rootstock. With a 2.0 yield ratio assumption, 7.7 high tunnels (231 trees) were needed to produce the same yield as one acre of trees on conventional rootstock and planting density. Thus, the 10-acre units for the Unprotected and the Micro-jet groves are assumed to have the equivalent yield potential of the High Tunnel grove with 77 high tunnel plantings. The Louisiana Satsuma production budget, developed by Hinson, Boudreaux, and Vaughn (2006), was used as the basis of the simulation model. It was assumed that production expenses for Louisiana producers would be similar for producers in other areas of the Gulf Coast region of the United States. Irrigation was not included in the Louisiana budget but was added to each of the simulated models, including the Micro-jet grove. The freeze protection system modeled for the Micro-jet grove was too large and expensive to operate on a regular basis to be efficient for irrigation needs. The cost of 60 establishing the groves and all variable and direct costs are realized in the year they occur. Fixed costs for machinery and irrigation are annual charges. All costs for the freeze protection technologies were obtained from the Alabama Agricultural Experiment Station Gulf Coast Research and Extension Center at Fairhope, Alabama. Fixed costs associated with freeze protection for the groves are amortized at 6- percent across their respective life expectancies. Fixed costs for the micro-jet freeze protection are $6,350 per acre for a well, pump, and all below ground pipes with a 20- year life expectancy, and $185 per acre for above ground parts with a 4-year life expectancy. Fixed costs for each high tunnel are $4,500 for the frame, end-walls, doors, hardware, and two layers of 20-year ground cloth. High tunnels are assumed to have a 20-year life expectancy. There are also significant variable costs associated with materials and labor to cover the tunnels with milky-white 6-mil polyurethane each year in December and to remove the covering after the danger of freeze. A yield curve based on tree age was developed from yield data collected on a Satsuma mandarin grove established in 1990 at the Gulf Coast Research and Extension Center in Fairhope, AL and is presented in Appendix 1.1. Trees were assumed to have no yield during the first two years of establishment and reach a mature average yield of 400 lb per tree by the ninth year after set out (Ebel, et al., 2004). A yield variation in the 25-percent range was observed among trees in the yield data collected by the Gulf Coast Research Center. The model used the GRKS distribution for Simetar? that was developed by Gray, Richardson, Klose, and Schumann (Richardson, 2004) to model a 61 25-percent variation from the average yield in any given year. This variation may be due to losses from sources other than freeze or it may be due to alternate bearing. An average price of $.50/lb was the assumed standard market price for all simulations. 3.3.3 Freeze Probability Matrix There were two levels of freeze events that are economically important in the simulation models. Severe freeze was assumed to cause extensive injury or death of the tree, and moderate freeze is assumed to cause extensive leaf injury and some stem dieback to the extent that only the next season?s fruit crop would be destroyed. Trees that experience moderate freeze injury recovered and produced a normal crop the following year. Satsuma mandarin is considered hardy to 14 o F (-10 o C) if properly acclimated to cold and this is the threshold for severe freeze injury (Ebel et al., 2005). The threshold for moderate freeze injury, 18 to 22 o F (-7.7 to -5.5 o C) also depends upon adequate cold acclimation prior to the freeze event. The matrix of severe and moderate freeze probabilities was created from the array of severe freeze and moderate freeze probabilities with 5-percent intervals. The 5- percent interval has the added convenience of equaling 1.0 freeze difference when applied to the 20-year simulation investment horizon, i.e. 5, 10, and 15-percent probabilities equal 1, 2 and 3 freeze events, respectively, in a 20 year period (Table 1). The value of each element in the matrix is the result obtained with freeze probabilities for that particular column and row. The total probability of all freezes (severe and moderate) is found by adding the probabilities for the column and row. Since the zero-percent severe freeze column had no severe freezes, both the Unprotected and the Micro-jet 62 strategies would have the same total number and type of freezes. In the Micro-jet matrix table, the upper right triangle will be a mirror image of the lower left triangle as the effect of both severe and moderate freezes were assumed to be equal in the simulation model; the maximum number of freeze events possible in this matrix was 12 (60-percent total freezes) at the 30-percent severe, 30-percent moderate intersection. For the High Tunnel strategy, the simulation model treated all freezes the same and assumed that no injury occurred from any of the freeze events. All values in the freeze matrix were identical for the High Tunnel strategy and will be reported as a single value. Table 1. Array of Severe by Moderate Freeze Occurrence Moderate Freeze - % Probability 0 5 10 15 20 25 30 0 0, 0 5, 0 10, 0 15, 0 20, 0 25, 0 30, 0 5 0, 5 5, 5 10, 5 15, 5 20, 5 25, 5 30, 5 10 0, 10 5, 10 10, 10 15, 10 20, 10 25, 10 30, 10 15 0, 15 5, 15 10, 15 15, 15 20, 15 25, 15 30, 15 20 0, 20 5, 20 10, 20 15, 20 20, 20 25, 20 30, 20 25 0, 25 5, 25 10, 25 15, 25 20, 25 25, 25 30, 25 30 0, 30 5, 30 10, 30 15, 30 20, 30 25, 30 30, 30 Severe Freeze - Percent Probability 3.4 Results and Discussion 3.4.1 Net Returns The discounted 20-year net returns from the simulations of the Unprotected and Micro-jet protected groves are presented in Tables 2 and 3 for all freeze event combinations at three price levels. The values represent the expected return from a 10- acre grove for each scenario using a 6-percent discount factor. It should be noted that the 63 discussion of results for this study will be limited to expected returns and will not consider the distribution of returns. The net return calculation is slightly higher than a Net Present Value (NPV) calculation because the fixed expenses for equipment, Table 2. Twenty-year Discounted Net Returns a for 10-Acre Unprotected Satsuma Grove with Varying Probabilities of Moderate and Severe Freeze Occurrence Moderate Freeze % Probability 0 5 10 15 20 25 30 0 901,076 676,809 500,691 354,093 233,626 135,893 55,590 5 845,124 631,190 460,513 319,933 205,539 112,489 37,277 10 789,260 583,732 421,048 287,279 177,537 89,436 17,193 15 733,154 536,217 382,233 253,748 150,190 65,360 -3,762 20 677,609 489,745 343,004 220,531 121,751 40,414 -22,534 25 622,331 442,108 303,882 187,220 91,959 18,523 -42,708 30 566,189 396,182 264,259 151,772 65,889 -5,133 -62,811 0 1,779,338 1,380,943 1,068,212 808,297 594,649 421,492 279,843 5 1,679,364 1,299,116 996,149 746,686 543,810 379,014 246,377 10 1,579,521 1,214,180 925,203 687,734 493,240 337,133 209,837 15 1,479,315 1,129,030 855,372 627,140 443,724 293,541 171,717 20 1,379,951 1,045,667 784,911 567,436 392,404 248,345 137,505 25 1,280,988 960,421 714,553 507,456 338,624 208,598 100,814 30 1,180,729 878,015 643,444 443,674 291,479 165,752 64,210 0 2,657,599 2,085,065 1,635,734 1,262,501 955,673 707,072 504,172 5 2,513,603 1,967,042 1,531,785 1,173,439 882,082 645,540 455,478 10 2,369,782 1,844,629 1,429,358 1,088,189 808,943 584,829 402,481 15 2,225,476 1,721,842 1,328,511 1,000,979 737,258 521,722 347,196 20 2,082,294 1,601,590 1,226,819 914,342 663,058 456,276 297,543 25 1,939,644 1,478,735 1,125,224 827,692 585,288 398,673 244,336 30 1,795,269 1,359,848 1,022,629 735,577 517,070 336,637 191,231 a Discount rate is 6-percent and table values are dollars per 10-acre unit. Note: Returns for a 10-acre equivalent High Tunnel protected grove are $266,192, $1,140,669, and $2,015,145 for market prices of $0.50, $0.75, and $1.00 per pound, respectively. Market Price = $1.00/lb Severe Freeze - Percent Probability Market Price = $0.50/lb Market Price = $0.75/lb 64 irrigation, and freeze protection were amortized over the 20-year investment period according to assumed life expectancy and are not fully realized in the initial year incurred. The High Tunnel strategy net return values did not change in response to either severe or moderate freezes. The discounted net returns for the 10-acre equivalent High Tunnel grove are $266,192, $1,140,669, and $2,015,145 for market prices of $0.50, $0.75, and $1.00, respectively. In the absence of severe freezes, returns for both Unprotected and Micro-jet groves exhibit a linear response to increasing moderate freeze probability and the returns for Unprotected groves exceeded those for the Micro-jet grove at all moderate freeze probability levels by $75,481. This value is the discounted cost of installing and maintaining the micro-jet freeze protection for the 10-acre grove and is an unnecessary expense. However, at the 5-percent or greater probability levels of severe freeze, net returns for the Micro-jet grove exceeded the net returns for the Unprotected grove at all price levels. The loss due to a severe freeze in an unprotected grove is greater than the loss of the next season?s crop; it is also includes the cost of replacing the trees and the lost or reduced production during the re-establishment period. Within a given severe freeze probability level (greater than zero), the magnitude of the difference in net returns between the Unprotected and the Micro-jet groves exhibited an inverse response to moderate freeze probability level; the magnitude of this response also decreased as the probability of severe freeze increased. With an increase in moderate freeze events, there was an increased probability that the Unprotected grove would be in a re-establishment period and have a potential yield less than the Micro-jet grove for the simulation year; the potential difference increased with greater probabilities of severe freezes. 65 Table 3. Twenty-year Discounted Net Returns a for 10-Acre Satsuma Grove with Micro-jet Freeze Protection and Varying Probabilities of Moderate and Severe Freeze Occurrence Moderate Freeze % Probability 0 5 10 15 20 25 30 0 825,595 769,625 713,740 657,410 601,896 546,850 490,691 5 769,644 713,780 657,673 601,957 546,754 490,709 435,567 10 713,780 657,673 602,128 546,805 490,658 435,567 379,628 15 657,673 602,128 546,850 490,724 435,423 379,589 323,913 20 602,128 546,850 490,709 435,423 379,547 323,663 267,822 25 546,850 490,709 435,567 379,547 323,663 267,604 211,579 30 490,709 435,567 379,628 323,913 267,822 211,579 155,932 0 1,703,857 1,603,883 1,503,971 1,403,402 1,304,081 1,205,507 1,105,234 5 1,603,883 1,504,040 1,403,835 1,304,181 1,205,358 1,105,248 1,006,490 10 1,504,040 1,403,835 1,304,471 1,205,444 1,105,182 1,006,490 906,535 15 1,403,835 1,304,471 1,205,507 1,105,248 1,006,262 906,473 806,915 20 1,304,471 1,205,507 1,105,248 1,006,262 906,402 806,475 706,733 25 1,205,507 1,105,248 1,006,490 906,402 806,475 706,346 606,321 30 1,105,248 1,006,490 906,535 806,915 706,733 606,321 506,805 0 2,582,118 2,438,123 2,294,203 2,149,394 2,006,265 1,864,164 1,719,505 5 2,438,123 2,294,301 2,149,996 2,006,405 1,863,962 1,719,788 1,577,412 10 2,294,301 2,149,996 2,006,813 1,864,083 1,719,706 1,577,412 1,433,441 15 2,149,996 2,006,813 1,864,164 1,719,855 1,577,101 1,433,358 1,289,918 20 2,006,813 1,864,164 1,719,788 1,577,101 1,433,257 1,289,287 1,145,643 25 1,864,164 1,719,788 1,577,412 1,433,257 1,289,287 1,145,088 1,001,064 30 1,719,788 1,577,412 1,433,441 1,289,918 1,145,643 1,001,064 857,678 a Discount rate is 6-percent and table values are dollars per 10-acre unit. Note: Returns for a 10-acre equivalent High Tunnel protected grove are $266,192, $1,140,669, and $2,015,145 for market prices of $0.50, $0.75, and $1.00 per pound, respectively. Severe Freeze - Percent Probability Market Price = $0.50/lb Market Price = $0.75/lb Market Price = $1.00/lb It is notable that at market prices of $0.50 and higher, the discounted 20-year net returns were significantly positive for all strategies and freeze risk levels except for the highest freeze risk levels for the Unprotected grove. Producers facing severe freeze probability levels greater than 10-percent may do well to consider investment in freeze 66 protection. Whether investment in expensive high tunnels would yield greater net returns to management than the Micro-jet strategy depends on the interaction of expected market price and number of freeze events. The High Tunnel grove had higher net returns at $0.50/lb market price only when more than 10 total freeze events were expected over the 20-year period; as market price increased, however, the effect of greater total fruit yield for the High Tunnel grove decreased this turning point to 5 freeze occurrences at market prices of $0.75/lb and 3 freeze occurrences at market prices of $1.00/lb. -1,000,000 -500,000 0 500,000 1,000,000 1,500,000 2,000,000 2,500,000 0.25 0.50 0.75 1.00 Market Price per Pound ($) N e t R e tu r n s ( $ ) . Unprotected Micro-Jet High Tunnel Figure 1. Twenty-Year Discounted Net Returns for Satsuma Groves with 10-percent Probability of Severe freeze and 20-percent Probability of Moderate Freeze Net returns for all strategies and scenarios exhibit a linear response to market price as illustrated in the example of returns for Satsuma groves with the example of freeze probability levels of 10-percent severe and 20-percent moderate (Figure 1). This linear response allows for the calculation of break-even prices and comparative equivalent prices. Break-even prices occur at the market price where the 20-year net 67 returns equal zero and intersect the market price axis. The equivalent prices are found where two price response lines intersect and are calculated with Equation 2. If the market price is greater than a given equivalent price between two strategies, the strategy with the greatest total yield will have a higher net return. Total fruit yield is a function of freeze event and tree age (Equation 1) and given the assumptions used in the simulations for this study, total yield is highest for the High Tunnel grove. Total yield for the Micro-jet grove exceeds that for the Unprotected grove except in the absence of severe freezes where the yields are equal. Over the freeze probability range used in the simulations, the equivalent prices calculated for the Micro-jet and the High Tunnel technologies closely fit the equation: y = 12.582 x -0.8311 , as seen in Figure 2. y = 12.582 x -0.831 R 2 = 0.9939 $0.00 $0.50 $1.00 $1.50 $2.00 $2.50 $3.00 $3.50 $4.00 10 20 30 40 50 60 % Total Freeze Probability (Severe plus Moderate) E qui va l e nt P r i c e i n $ pe r P ound Figure 2. Equivalent Price between Micro-jet and High Tunnel at Freeze Probability Levels of 5-percent to 60-percent (above which High Tunnel has greater net returns) 68 Table 4. Break-even Prices for Simulations of Satsuma Groves with Different Levels of Freeze Protection and Varying Probabilities of Moderate and Severe Freeze Occurrence Moderate Freeze % Probability 0 5 10 15 20 25 30 0 0.243 0.260 0.279 0.305 0.338 0.381 0.438 5 0.247 0.264 0.285 0.312 0.348 0.394 0.455 10 0.250 0.269 0.291 0.321 0.359 0.410 0.478 15 0.254 0.274 0.298 0.330 0.372 0.428 0.505 20 0.259 0.280 0.306 0.341 0.388 0.451 0.535 25 0.264 0.287 0.315 0.354 0.407 0.476 0.574 30 0.269 0.294 0.326 0.370 0.427 0.508 0.624 0 0.265 0.269 0.274 0.280 0.286 0.292 0.300 5 0.269 0.274 0.280 0.286 0.292 0.300 0.301 10 0.274 0.280 0.286 0.292 0.300 0.309 0.320 15 0.280 0.286 0.292 0.300 0.309 0.320 0.332 20 0.286 0.282 0.300 0.309 0.320 0.332 0.347 25 0.292 0.300 0.309 0.320 0.332 0.347 0.366 30 0.300 0.309 0.320 0.332 0.347 0.366 0.389 0 0.424 0.424 0.424 0.424 0.424 0.424 0.424 5 0.424 0.424 0.424 0.424 0.424 0.424 0.424 10 0.424 0.424 0.424 0.424 0.424 0.424 0.424 15 0.424 0.424 0.424 0.424 0.424 0.424 0.424 20 0.424 0.424 0.424 0.424 0.424 0.424 0.424 25 0.424 0.424 0.424 0.424 0.424 0.424 0.424 30 0.424 0.424 0.424 0.424 0.424 0.424 0.424 High Tunnel Grove - Break-even Prices Severe Freeze - Percent Probability Unprotected Grove - Break-even Prices Micro-jet Grove - Break-even Prices 3.4.2 A Decision Process Break-even prices and equivalent prices for all freeze probability combinations are presented in Tables 4 and 5, respectively. The break-even prices in Table 4 condense the information presented in Tables 2 and 3 into a form that allows for an easier 69 comparison of the strategies. The simple decision rule is that at any given freeze probability level, the strategy with the lowest break-even price is the most efficient. This rule may work to evaluate the feasibility of a strategy under highly competitive prices; however, it does not consider the effect of increased yield potential of a strategy at market prices higher than the break-even price. Table 5. Equivalent Prices for Simulations of Satsuma Groves with Different Levels of Freeze Protection and Varying Probabilities of Moderate and Severe Freeze Occurrence Moderate Freeze % Probability 0 5 10 15 20 25 30 0 - 0.322 0.261 0.240 0.230 0.225 0.221 5 - 0.331 0.266 0.244 0.234 0.228 0.225 10 - 0.340 0.272 0.249 0.238 0.232 0.229 15 - 0.350 0.278 0.254 0.243 0.237 0.234 20 - 0.361 0.286 0.260 0.248 0.242 0.240 25 - 0.374 0.295 0.267 0.255 0.250 0.247 30 - 0.389 0.305 0.275 0.263 0.258 0.256 0 - 3.629 1.828 1.261 0.987 0.825 0.716 5 3.628 1.829 1.263 0.987 0.825 0.716 0.639 10 1.829 1.263 0.988 0.825 0.716 0.639 0.582 15 1.263 0.988 0.825 0.716 0.639 0.582 0.537 20 0.988 0.825 0.716 0.639 0.582 0.537 0.501 25 0.825 0.716 0.639 0.582 0.537 0.501 0.472 30 0.716 0.639 0.582 0.537 0.501 0.472 0.447 0 - 1.103 0.691 0.552 0.484 0.445 0.419 5 4.097 0.942 0.643 0.530 0.472 0.437 0.414 10 2.053 0.825 0.605 0.511 0.460 0.429 0.409 15 1.410 0.740 0.572 0.494 0.450 0.422 0.403 20 1.098 0.675 0.544 0.478 0.440 0.415 0.399 25 0.913 0.623 0.520 0.464 0.431 0.410 0.394 30 0.789 0.583 0.499 0.451 0.423 0.404 0.390 Severe Freeze - Percent Probability Unprotected to Micro-jet Equivalent Prices Micro-jet to High Tunnel Equivalent Prices Unprotected to High Tunnel Equivalent Prices 70 The equivalent price table (Table 5) will allow for the determination of the strategy that yields the highest net return at a given freeze probability combination and market price. The decision rule is to choose the strategy with the highest total crop yield if the market price is greater than the equivalent price. This process assumes that the yield relationship between strategies is Unprotected < Micro-jet < High Tunnel. Care must be taken, however, not to fall below the break-even price. A combination use of the tables would avoid this. To aid in future development of a computerized decision tool, a decision tree was developed that utilizes information from the break-even table and the equivalent price table. The following proposed decision process requires an assumption of the freeze probability combination and an average market price on the part of the user: 1) Is market price = equivalent price for Unprotected vs. Micro-jet? A) Yes ? net returns are equal. B) No ? go to 2) 2) Is market price > equivalent price for Unprotected vs. Micro-jet? A) No ? Is the market price > break-even price for Unprotected? a) No ? stop, net return will be negative. b) Yes ? Unprotected will have the highest expected return. B) Yes ? Is the market price ? equivalent price for Micro-jet vs. High Tunnel? a) No ? Is the market price > break-even price for Micro-jet? (1) No ? stop, net return will be negative. (2) Yes ? Micro-jet will have the highest expected return. b) Yes ? Is the market price = break-even price for High Tunnel? 71 (1) Yes ? net returns are equal for Micro-jet and High Tunnel. (2) No ? Is the market price > break-even price for High Tunnel? (a) No ? stop, net return will be negative. (b) Yes ? High Tunnel will have the highest expected return. C) Don?t know (Severe freeze probability = 0) Is market price ? equivalent price for Micro-jet vs. High Tunnel? a) No ? Is the market price > break-even price for Unprotected? (1) No ? stop, net return will be negative. (2) Yes ? Unprotected will have the highest expected return. b) Yes ? Is the market price > equivalent price for Micro-jet vs. High Tunnel? (1) Yes ? High Tunnel will have the highest expected return. (2) No? Unprotected will have the highest expected return. This decision process will always choose the strategy with the highest expected return. Comparison of the strategies will result in the same conclusions as comparing the net returns tables but has the added advantage of showing the market-price break point where one strategy will have higher returns than the other. The decision process may appear cumbersome and a decision maker could look at the tables and come to the same conclusion. If underlying assumptions for the simulations were changed, however, a new set of break-even and equivalent price tables would be produced. Studying many sets of tables would become tedious. The above process would be useful in the programming of a decision tool that allows for changes in underlying cost and yield variables. This decision process was developed for evaluating the Satsuma strategies that were simulated 72 for this study; however, it could easily be adapted to the evaluation of other risk reduction strategies in other crops. 3.5 Conclusions The purpose of the present study was to evaluate the effect of varying the severe and moderate freeze probabilities on discounted net returns for Satsuma groves with different levels of freeze protection. This information would aid potential producers in evaluating the feasibility of producing Satsuma mandarins in areas with freeze probabilities that vary from those for the Fairhope, AL area. The 20-year discounted net returns were calculated over an array of severe and moderate freeze probability combinations at 5-percent increments ranging from zero to 30-percent. The net returns were determined for Satsuma groves with three different levels of freeze protection. Only, in the absence of severe freezes do returns for the Unprotected grove exceed returns for the Micro-jet grove. When the probability of severe freeze increases to 5-percent and above, net returns are greater for the Micro-jet grove than the Unprotected Grove at all moderate freeze probabilities and market prices evaluated in the study. The 20-year total expected fruit yield is a function of freeze event and tree age and increases as the level of freeze protection increases with Unprotected < Micro-jet < High Tunnel. Increasing market prices results in a greater rate of return to the High Tunnel strategy than to the other strategies due to greater total yield over the 20-year period. Net returns are greater for the High Tunnel grove than the Micro-jet grove when total freeze events exceed 10 in the 20-year period at $0.50/lb market price; this also 73 occurs when there are more than 5 freeze occurrences at market prices of $0.75/lb and more than 3 freeze occurrences at market prices of $1.00/lb. Net returns for all strategies and scenarios exhibit a linear response to market price and this relationship is used to calculate break-even prices for each scenario and equivalent prices between the strategies. Break-even prices occur at the market price where the 20-year net returns equal zero and intersect the market price axis. The equivalent price is the market price where two price response lines intersect and have equal net returns. Evaluation of equivalent prices is a simple method of identifying the strategy with the highest net returns and has the added advantage of identifying the market price at which one strategy will return more than the other will. A decision process based on the break-even and equivalent prices was proposed for future use in programming a computer-based decision tool. 74 REFERENCES Bourgeois, W.J. and A.J. Adams. ?Low-Volume Scaffold Branch Irrigation for Citrus Freeze Protection.? HortScience 22(1987):48-50. Bourgeois, W.J., A.J. Adams, and D.R. Stipe. ?Effectiveness of Scaffold Branch Irrigation for Freeze Protection of Louisiana Citrus during 1989 Freeze.? Proc. Florida State Horticulture Society 103(1990):62-64. Briys, E. and H. Schlesinger. ?Risk Aversion and the Propensities for Self-Insurance and Self-Protection. Southern Economic Journal 57(1990):458-67. Campbell, B.L., R.G. Nelson, R.C. Ebel, W.A. Dozier, J.L. Adrian, and B.R. Hockema. ?Fruit Quality Characteristics that Affect Consumer Preferences for Satsuma Mandarins.? HortScience 39(2004):1664-69. Dionne, G., and L. Eeckhoudt. ?Self-Insurance, Self-Protection and Increased Risk Aversion.? Economic Letters 17(1985):39-42. Easterling, D.R., H.F. Diaz, A.V. Douglas, W.D. Hogg, K.E. Kunkel, J.C. Rogers, and J.F.Wilkinson. ?Long-Term Observations for Monitoring Extremes in the Americas.? Climatic Change 42(1999):285-308. Ebel, R.C., B.L. Campbell, M.L. Nesbitt, W.A. Dozier, J.K. Lindsey, and B.S. Wilkins. ?A Temperature Index Model to Estimate Long-Term Freeze-Risk of Satsuma Mandarins Grown on the Northern Coast of the Gulf of Mexico. Journal of the American Society of Horticultural Science 130(2005):500-07. Ebel, R.C., W.A. Dozier, B. Hockema, F.M. Woods, R. Thomas, and B.S. Wilkins. ?Fruit Quality of Satsuma Mandarin Grown on the Northern Coast of the Gulf of Mexico.? HortScience 39(2004):979-82. Ehrlich, I. and G.S. Becker. ?Market Insurance, Self-insurance and Self-protection.? Journal of Political Economy 80(1972):623-48. Elmer, N.A., A.P. Thurow, J.L. Johnson, and C.P. Rosson, III. ?An Ex Ante Assessment of Investments in Texas Grapefruit Under Uncertainty.? Journal of Agricultural and Applied Economics 33(2001):391-401. 75 Featherstone, A.M., and C.B. Moss. ?Quantifying Gains to Risk Diversification using Certainty Equivalence in a Mean-Variance Model: an Application to Florida Citrus.? Southern Journal of Agricultural Economics 45(December 1990):191-98. Freund, R.J. ?The Introduction of Risk into a Programming Model.? Econometrica 24(1956):253-61. Goulder, L.H., and W.A. Pizer. ?The Economics of Climate Change.? National Bureau of Economic Research, Working Paper 11923, 2006. Gray, T.W. ?Local-based, Alternative Marketing Strategy could Help Save More Small Farms.? Rural Cooperatives 72(May/June 2005). Internet site: http://www.rurdev.usda.gov/rbs/may05/local.htm (Accessed September 18, 2007). Harwood, J.L., R. Heifner, K. Coble, J. Perry, and A. Somwaru. Managing Risk in Farming: Concepts, Research, and Analysis. Economic Research Service, USDA, Agricultural Economic Report No. 774, 1999. Hiebert, L.D. ?Optimal Loss Reduction and Risk Aversion.? Journal of Risk and Insurance (June 1989):300-06. Hinson, R.A., J.E. Boudreaux, and A. Vaughn. Projected Costs of Establishing and Operating a Citrus Grove. Louisiana Agricultural Experiment Station Information Series No. 140, 2006. Hodgson, R.W. ?Horticultural Varieties of Citrus.? The Citrus Industry, Volume I: History, World Distribution, Botany, and Varieties. W. Reuther, H.J. Webber, and L.D. Batchelor, eds. Riverside, California: University of California, Rev. Ed. 1967. Internet site: http://lib.ucr.edu/agnic/webber/Vol1/Vol1TP.html (Accessed August 10, 2007). Inman, R.L., J.M. Davenport, and D.K. Zeigler. ?Latin Hypercube Sampling: a Program User?s Guide.? Technical Report SAND 79-1473, Sandia Laboratories, 1980. Katz, R. W., M. B. Parlange, and C. Tebaldi. ?Stochastic Modeling of the Effects of Large-Scale Circulation on Daily Weather in the Southeastern U.S.? Climate Change 60(2003):189-216. Knight, F.H. Risk, Uncertainty, and Profit. Boston, MA: Houghton Mifflin Co., 1921. Lamont, Jr., W.J., M.D. Orzolek, E.J. Holcomb, K. Demchak, E. Burkhart, L. White, and B. Dye. ?Production System for Horticultural Crops Grown in the Penn State High Tunnel.? HortTechnology 13(April-June 2003):358-62. 76 Lubowski, R.N., M. Vesterby, S. Bucholtz, A. Baez, and M.J. Roberts. Major Uses of Land in the United States, 2002. Washington D.C.: U.S. Department of Agriculture/Economic Research Service, Pub. No. EIB-14, May 2006. Mapp, Jr., H.P., M.L. Hardin, O.L. Walker, and T. Persaud. ?Analysis of Risk Management Strategies for Agricultural Producers.? American Journal of Agricultural Economics 61(1979):1071-77. Martsolf, J.D. ?Diversity of Frost Protection Methodology.? Proceedings of the Florida State Horticulture Society 113(2000):133-37. Mississippi State University ? Department of Agricultural Economics. Mississippi State Budget Generator. Internet site: http://www.agecon.msstate.edu/laughlin/msbg.php (Accessed August 10, 2007). Meyer, J. ?Choice Among Distributions.? Journal of Economic Theory 14(1977):326-36. Nelson, R.G., S.A. Duarte, and H.P. Masser. ?Financial risk Analysis of Three Aeration Regimes in Catfish Cage Culture.? Aquaculture Economics and Management 5(2001):171-77. Nesbitt, M.L., N.R. McDaniel, R.C. Ebel, W.A. Dozier, and D.G. Himelrick. ?Performance of Satsuma Mandarin Protected from Freezing Temperatures by Microsprinkler Irrigation.? HortScience 35(2000):856-59. Nesbitt, M.L., R.C. Ebel, D. Findley, B. Wilkins, F. Woods, and D.G. Himelrick. ?Assays to Assess Freeze Injury of Satsuma Mandarin.? HortScience 37(2002):871-77. Noda, K., H. Okuda, T. Kihara, I. Iwagaki, and K. Kawase. ?Effects of Rootstocks on Tree Growth and Fruit Quality in Very Early Ripening Satsuma Mandarin ?Yamakawa?.? Journal of the Japanese Society for Horticultural Science 70(2001):78-82. Orzolek, M.D., W.J. Lamont, and L. White. ?Promising Horticultural Crops for Production in High Tunnels in the Mid-Atlantic area of the United States.? Acta Horticulturae 633(2004):453-58. Powell, A.A., and D.G. Himelrick. Principles of Freeze Protection for Fruit Crops. Alabama Cooperative Extension System. Internet site: http://www.aces.edu/dept/peaches/freeze.html (Accessed 7/18/2007). Pratt, J.W. ?Risk Aversion in the Small and in the Large.? Econometrica 32(1964):22-36. 77 Purvis, A., W.G. Boggess, C.B. Moss, and J. Holt. ?Technology Adoption Decisions under Irreversibility and Uncertainty: An Ex Ante Approach.? American Journal of Agricultural Economics 77 (August 1995):541-51. Reilly, J., F. Tubiello, B. McCarl, D. Abler, R. Darwin, K. Fuglie, S. Hollinger, C. Izaurralde, S. Jagtap, J. Jones, L. Mearns, D. Ojima, E. Paul, K. Paustian, S. Riha, N. Rosenberg, and C. Rosenzweig. ?U.S. Agriculture and Cimate Change: New Results.? Climatic Change 57(2003): 43-69. Richardson, J.W. Simulation for Applied risk Management. Dept. of Agr. Econ., Texas A&M University, 2004. Schnapp, F. F., J.L. Driscoll, R.P. Zacharias, and G.R. Josephson. ?Ratemaking Considerations for Multiple Peril Crop Insurance.? CAS Forum, Winter 2000. Internet site: www.casact.org/pubs/forum. Takahara, T., T. Ogata, H. Fujisawa, and N. Muramatsu. ?Effect of Rootstocks on Tree Growth, Yield and Fruit Quality of ?Shirakawa? Satsuma Mandarin (Citrus unshiu Marc.).? Bulletin of the National Institute of Fruit Tree Science 35(2002):99-107. Tauer, L. ?Target MOTAD.? American Agricultural Economics Association 65(1983):606-10. Teague, P.W., and J.G. Lee. ?Risk Efficient Perennial Crop Selection: A MOTAD Approach to Citrus Production.? Southern Journal of Agricultural Economics (December 1988):145-52. U.S. Department of Agriculture?Agricultural Research Service. USDA Plant Hardiness Zone Map. Misc. Publication No. 1475, 1965. U.S. Department of Agriculture?National Agricultural Statistics Service. 2006 Alabama Agricultural Statistics. Montgomery, AL: USDA, NASS, Alabama Field Office, Bulletin No. 48, 2007. U.S. Department of Agriculture?Risk Management Agency. 2007 Crop Insurance Handbook. FCIC 18010-01, 2006. http://www.rma.usda.gov/handbooks/18000. U.S. Department of Agriculture?Risk Management Agency. 2008 Arizona-California Citrus Crop Provisions. 2007a. ttp://www.rma.usda.gov/policies/2008policy.html U.S. Department of Agriculture?Risk Management Agency. 2008 Texas Citrus Tree Crop Provisions. 2007b. http://www.rma.usda.gov/policies/2008policy.html U.S. Department of Agriculture?Risk Management Agency. Actuarial Documents. 2007c. http://www.rma.usda.gov/tools/adb.html 78 U.S. Department of Agriculture?Risk Management Agency. Report on Specialty Crop Insurance. May 2004. http://www.rma.usda.gov/pubs/2004/specialtycrop.pdf US Global Change Research Group. ?Chapter 4: Climate Variability and Change? in Change Science Program Strategic Plan. January 24, 2006. Internet site: http://www.usgcrp.gov (Accessed October 27, 2006). Wiebe, K. and N. Gollehon, eds. Agricultural Resources and Environmental Indicators, 2006 Edition. Washington D.C.: U.S. Department of Agriculture/Economic Research Service, Pub. No. EIB-16, July 2006. Winberg, O.F.E. ?The Satsuma.? Fairhope Courier Daily periodical published in Fairhope, AL, October 7, 1948a. Winberg, O.F.E. ?The Satsuma.? Fairhope Courier Daily periodical published in Fairhope, AL, October 21, 1948b. Winberg, O.F.E. ?The Satsuma.? Fairhope Courier Daily periodical published in Fairhope, AL, October 28, 1948c. Yelenosky, G. ?Cold Hardiness in Citrus.? Horticulture Review 7(1985):201-38. Yelenosky, G. ?Mini-review: Responses and Adaptations of Citrus Trees to Environmental Stresses.? Israel Journal of Botany 40(1991):239-50. Yelensosky, G. ?An Overview of Florida Citrus Freeze Survival.? Proceedings of the Florida State Horticultural Society 109(1996):118-23. Yonemoto, Y., T. Takahara, H. Okuda, and T. Ogata. ?Effects of ?Karatachi?, Common Trifoliate Orange (Poncirus trifoliata (l.) Raf.) and ?Hiryu?, Flying Dragon Trifoliate Orange (P. trifoliata var. monstrosa) Rootstocks on Tree Growth, Yield and Fruit Qualities in Young Tree of New Citrus Cultivars ?Amakusa? and ?Amaka?.? Horticultural Research (Japan) 4(2005):81-4. 79 APPENDIX A: SIMULATION VARIABLES FOR CHAPTER 1 Table A1. Information for Tree and Fruit Crop Insurance Policies Used in Simulations for Chapter 1 Rate or Item Unit 1 2 3 4 5 6+ Texas Citrus I Tree Policy a : Coverage level assumption 65% Maximum Reference Amount $4,190 Coverage Level x Reference $2,723 Liability by Tree Year percent 33.0 60.0 80.0 90.0 100.0 100.0 Liability by Tree Year dollar 899 1,634 2,178 2,451 2,723 2,723 Premium dollar 155 155 155 155 155 155 Arizona-California Citrus Policy b : Mandarins: Coverage level assumption 65% Price Election (per 25 lb carton) $5.70 T-Yield (25 lb carton per acre) 430 Yield Adjustment (t-yield x .60) 258 Total Premium $0.313 0.00 0.00 0.00 0.00 0.00 Formula c Government Premium Subsidy Rate 59% Liability per acre 0.00 0.00 0.00 0.00 0.00 Formula c a 2008 Policy with fixed coverage per acre of trees. b 2008 Policy with coverage per acre based on actual production history (APH). c Premium = Liability x base rate; Liability = APH x Coverage Level x Price Election. Sources: Hinson, et al., Louisiana Agri. Expt. Sta. Info. Series No. 140, 2006. USDA-RMA, 2008 Policy and Actuarial Documents, www.rma.usda.gov. Tree Leaf Year Value per Acre 80 Table A2. Fixed and Direct Costs (Excluding Harvest Costs) per Acre of Satsuma Used in Simulations for Chapter 1 Item Unit 1 2 3 4 5+ Rate 1 2 3 4 5+ Fertilizer 13-13-13 cwt 2.00 6.00 3.00 8.00 10.00 15.50 31.00 93.00 46.50 124.00 155.00 Amm Nitrate (34%) cwt 1.00 1.20 1.50 16.00 16.00 19.20 24.00 Fungicide dollar 18.94 0 172.53 214.30 228.79 1.00 18.94 172.53 214.30 228.79 Herbicide dollar 40.00 219.68 126.00 136.00 136.00 1.00 40.00 219.68 126.00 136.00 136.00 Insecticide dollar 32.30 40.73 69.94 192.16 190.82 1.00 32.30 40.73 69.94 192.16 190.82 Trees each 116.00 10.00 8.00 928.00 80.00 Labor Mark Rows hour 4.00 9.60 38.40 Plant hour 20.00 3.00 9.60 192.00 28.80 Nutrients hour 5.00 8.00 7.00 9.60 48.00 76.80 67.20 Prune hour 4.50 4.00 5.00 1.50 6.50 9.60 43.20 38.40 48.00 14.40 62.40 Strip Fruit hour 1.00 9.60 9.60 Scout hour 5.00 6.00 18.00 20.00 9.60 48.00 57.60 172.80 192.00 Operator hour 16.84 10.40 13.16 11.79 11.45 15.30 257.65 159.12 201.35 180.39 175.19 Diesel Fuel gal 39.21 18.79 24.19 29.30 28.53 2.23 87.44 41.90 53.94 65.34 63.62 Gasoline gal 0.90 1.80 2.10 2.63 2.37 4.73 5.52 Repair & Maint dollar 60.54 37.37 45.51 56.02 54.47 1.00 60.54 37.37 45.51 56.02 54.47 Interest on Op Capital dollar 87.49 40.78 58.87 83.61 99.02 1.00 87.49 40.78 58.87 83.61 99.02 Fixed Implements dollar 38.64 23.13 29.48 73.92 72.37 1.00 38.64 23.13 29.48 73.92 72.37 Tractor dollar 47.45 22.89 29.89 36.43 35.40 1.00 47.45 22.89 29.89 36.43 35.40 Self-Propelled dollar 4.65 9.30 10.85 0.00 0.00 1.00 4.65 9.30 10.85 0.00 0.00 Packing Line dollar 141.02 141.02 141.02 1.00 141.02 141.02 141.02 Total Direct and Fixed 1958.07 974.24 1180.20 1509.59 1630.10 a For each tree leaf year: Cost of Item per Acre = Quantity of Unit per Acre x Rate. Source: Hinson, et al., Projected Costs of Establishing and Operating a Citrus Grove. Louisiana Agri. Expt. Sta. Info. Series No. 140, 2006. Tree Leaf Year Tree Leaf Year Cost of Item per Acre a Quantity of Unit per Acre 81 Table A3. Fixed and Direct Costs per Acre of Satsuma and Variables Used in Simulations for Chapter 1 Item Unit 1 2 3 4 5+ Rate 1 2 3 4 5+ Direct Costs: Harvest Aid Field Box each 10.0 10.0 10.0 12.00 120.00 120.00 120.00 Harvest Container each 10.0 10.0 10.0 2.00 20.00 20.00 20.00 Electricity - Pack Line kwh 175.0 175.0 210.0 0.12 21.00 21.00 25.20 Repair & Maint - Pack Line dollar 50.00 50.00 50.00 1.00 50.00 50.00 50.00 _____ _____ _____ _____ _____ Total Harvest Direct Costs 0.00 0.00 211.00 211.00 215.20 Variable Costs b : Harvest Labor - per Bushel 2.25 Grading Labor - per Bushel 2.90 Marketing Box - Bushel 1.25 Total Variable Cost per Bushel 6.40 (Note: Bushel = 40 lb) Micro-Sprinkler Freeze Protection: Life Amoritization Rate 6% Well, pump, pipes 6,000 20 yr 523.08 523.08 523.08 523.08 523.08 Below ground pipes 350 20 yr 30.51 30.51 30.51 30.51 30.51 Tubing, emitters 185 4 yr 53.39 53.39 53.39 53.39 53.39 Annual Maintenance $25 25.00 25.00 25.00 25.00 25.00 a For each tree leaf year: Cost of Item per Acre = Quantity of Unit per Acre x Rate. b Harvest costs are not incurred if a freeze occurs in the simulation. Sources: Hinson, et al., Projected Costs of Establishing and Operating a Citrus Grove. Louisiana Agri. Expt. Sta. Info. Series No. 140, 2006. Alabama Agricultural Experiment Station Gulf Coast Research and Extension Center at Fairhope, Alabama. Quantity of Unit per Acre Cost of Item per Acre a Tree Leaf Year Tree Leaf Year 82 APPENDIX B: SIMULATION VARIABLES FOR CHAPTERS 2 AND 3 Table B1. Yield and Establishment Cost Variables used in Simulations for Chapter 2 and Chapter 3 Strategy a Unit 1 2 3 4 5 6 7 8 9+ Rate Revenue Yield/Tree - Conventional UP, MJ lb/tree 0 0 70 120 190 250 350 350 400 $0.25 Yield/Tree - Conventional UP, MJ bu/ac 0 0 203 348 551 725 1015 1015 1160 $10.00 Yield/Tree - High Density HT lb/tree 0 0 35 60 95 125 175 175 200 $0.25 Yield/Tree - High Density HT bu/ac equiv 0 0 202 347 549 722 1011 1011 1155 $10.00 (Note: 1 bu = 40 lb) Establishment Cost Land Preparation All $/ac, ac equiv 100 1.00 Conventional Spacing: Plants UP, MJ no./ac 116 12 $8.00 Labor - layout & plant UP, MJ hour/ac 24 3 $9.60 Labor - Strip fruit UP, MJ hour/ac 1 $9.60 High Density Planting: Plants/ac equivalent b HT no./ac equiv 231 23 $8.00 Labor - layout & plant HT hour/ac equiv 48 6 $9.60 Labor - Strip fruit HT hour/ac equiv 2 $9.60 a UP = Unprotected grove, MJ = Micro-jet grove, HT = High Tunnel grove. b Plants/ac equivalent changes with yield ratio: 1.5 = 174 plants, 2.0 = 231 plants, 2.5 = 291 plants, and 3.0 = 348 plants. Sources: Hinson, et al., Louisiana Agri. Expt. Sta. Info. Series No. 140, 2006. Alabama Agricultural Experiment Station Gulf Coast Research and Extension Center at Fairhope, Alabama. Leaf Year 83 Table B2. Direct and Variable Cost Variables used in Simulations for Chapter 2 and Chapter 3 Strategy a Unit 1 2 3 4 5 6 7 8 9+ Rate Direct Costs Pest/Disease/Weed UP, MJ $/acre 91.24 260.41 368.47 542.46 555.61 555.61 555.61 555.61 555.61 1.00 Pest/Disease - High Tunnel HT $/acre equiv 91.24 40.32 242.47 406.46 419.61 419.61 419.61 419.61 419.61 1.00 High Tunnel Plastic HT $/tunnel 164 164 164 164 164 164 164 164 164 7.70 Fertilize (13-13-13) All cwt/ac, ac equiv 2 6 4 9.2 11.5 11.5 11.5 11.5 11.5 $15.50 Fuel (diesel & gas) All gal/ac, ac equiv 40 20.6 26.3 29.3 28.5 28.5 28.5 28.5 28.5 $2.23 Repair/maintenance All $/ac, ac equiv 60.00 38.00 45.00 55.00 55.00 55.00 55.00 55.00 55.00 1.00 Operator Labor All hour/ac, ac equiv 16.8 10.4 13.2 11.8 11.45 11.45 11.45 11.45 11.45 $15.30 Labor: Prune - Conventional UP, MJ hour/ac 4.5 4 5 1.5 6.5 6.5 6.5 6.5 6.5 $9.60 Prune - High Density HT hour/ac equiv 9 8 10 3 13 13 13 13 13 $9.60 Fertilize All hour/ac, ac equiv 5 8 7 7 7 7 7 7 7 $9.60 Scouting All hour/ac, ac equiv 5 6 18 20 20 20 20 20 $9.60 Micro-jet Maintenance MJ hour/ac 3 3 3 3 3 3 3 3 3 $9.60 High Tunnel Maint HT hour/ac equiv 2 12 12 12 12 12 12 12 12 $9.60 Irrigation Maint All hour/ac, ac equiv 10 10 10 10 10 10 10 10 10 $9.60 Harvest Variable Costs Field Boxes All no./ac, ac equiv 50 10 10 10 10 10 10 $12.00 Harvest Labor All bu/ac, ac equiv 203 348 551 725 1015 1015 1160 $2.25 Pack - Box All bu/ac, ac equiv 203 348 551 725 1015 1015 1160 $1.25 Grad & Pack - Labor All hour/ac, ac equiv 68 116 184 242 338 338 387 $9.60 Pack line electricity All kwh/ac, ac equiv 175 175 200 259 338 338 387 $0.12 Pack line repair/maint All $/ac, ac equiv 50 50 50 50 50 50 50 1.00 Interest on Operating Cap All $/ac $0.048 a UP = Unprotected grove, MJ = Micro-jet grove, HT = High Tunnel grove. Sources: Hinson, et al., Louisiana Agri. Expt. Sta. Info. Series No. 140, 2006. Alabama Agricultural Experiment Station Gulf Coast Research and Extension Center at Fairhope, Alabama. Leaf Year 84 Table B3. Fixed Cost Variables used in Simulations for Chapter 2 and Chapter 3 Amortization Strategy a Unit 1 2 3 4+ Rate Cost Life Factor ($) (Year) (6%) Fixed Costs Tractors & equipment All $/ac, ac equiv 90.75 55.00 70.00 110.00 1.00 Pack line All $/ac, ac equiv 191.58 191.58 191.58 191.58 1.00 1,410 10 0.13587 Irrigation All $/ac, ac equiv 87.18 87.18 87.18 87.18 1.00 1,000 20 0.08718 Freeze Protection Well, pump, pipes MJ $/acre 553.59 553.59 553.59 553.59 1.00 6,350 20 0.08718 Tubing, emitters MJ $/acre 53.39 53.39 53.39 53.39 1.00 185 4 0.28859 Installation Labor MJ $/acre 25 $9.60 High Tunnel Structure HT $/tunnel 392.33 392.33 392.33 392.33 1.00 4,500 20 0.08718 Installation Labor HT hour/ac equiv 40 $9.60 a UP = Unprotected grove, MJ = Micro-jet grove, HT = High Tunnel grove. Sources: Hinson, et al., Louisiana Agri. Expt. Sta. Info. Series No. 140, 2006. Alabama Agricultural Experiment Station Gulf Coast Research and Extension Center at Fairhope, Alabama. Leaf Year