TDOC a Z TA245.7 8,1650 figzitfio » November 1990 Qf ' Optimal Cropping Strategies Considering Risk: Texas T ra ns -Pecos The Texas A ricultural Ex eriment Station, Charles J. Arntzen, Director, The Texas A& ‘Why System, College Station, Texas g p __.. Q n n n! REIMIHQIQQEW ' [Blank Page in Original Bulletin] * '5"; a Optimal Cropping Strategies Considering Risk: Texas Trans-Pecos John R. Ellis, Assistant Professor Agricultural Economics Washington State University Ronald D. Lacewell, Professor Agricultural Economics Texas A&M University Jaroy Moore, Professor Soil Physics Texas Agricultural Experiment Station - Pecos, Texas James Richardson, Professor Agricultural Economics Texas A&M University [Blank Page in Original Bulletin] V a», SMl-dE/Qgl 6- {D501 lglclOl NOV. Table of Contents Executive Summary .................................................................................................................................................... .. 5 Acknowledgements .................................................................................................................................................... .. 6 Introduction .................................................................................................................................................................. .. 7 Objectives .............................................................................................................................................................. .. 7 Study Area ............................................................................................................................................................ .. 7 Methodology ................................................................................................................................................................ .. 8 Risk Analysis ........................................................................................................................................................ .. 8 Crop Growth Model - EPIC ...................................................................................................................................... .. 10 Firm Simulation - FLIPSIM ...................................................................................................................................... .. 10 Research Organization .............................................................................................................................................. .. 10 Preliminary Analysis .................................................................................................................................................. .. 12 Alternative Crops and Irrigation Strategies ...................................................................................................... .. 12 Calibration of EPIC ............................................................................................................................................ .. 12 Simulated Yields .................................................................................................................................................. .. 13 Enterprise Budget Development ........................................................................................................................ .. 15 Whole-Farm Analysis ................................................................................................................................................ .. 17 Whole-Farm Scenarios ........................................................................................................................................ .. 17 FLIPSIM Assumptions ........................................................................................................................................ .. 18 Optimal Cropping Patterns ................................................................................................................................ .. 21 Firm Survival ...................................................................................................................................................... .. 22 Summary and Conclusions ................................... .; ................................................................................................. .. 24 Preliminary Risk Analysis Results .................................................................................................................... .. 24 Whole-Farm Results ............................................................................................................................................ .. 25 Conclusions .......................................................................................................................................................... .. 25 References .................................................................................................................................................................. .. 26 Appendix .................................................................................................................................................................... .. 27 [Blank Page in Original Bulletin] l 5*:- Executive Summary I-Iigh production costs and several years of poor Q output prices have placed most Texas Trans-Pecos ag- ricultural producers in poor financial health. Many are operating with small or negative net worth, prompting most commercial lenders to abandon the area. Cur- rently, surviving producers rely heavily on govem- ment farm programs including the commodity loan and target prices programs and operating loans from the Farmer's Home Administration. Reductions in government price and income support levels within the 1985 farm bill will likely lessen profit potential even further. _ This analysis examines alternative paths of adjust- ment for the remaining producers in the region. This study was designed to identify the mixture of crops or irrigation levels, given the high cost of irrigation water and declining government support, that would pro- vide the best chance of reducing current high debt loads and help ensure survival of the firm. Two simulation models were employed in an- Q swering the question posed above. Recent advances in biophysical modelling offer the opportunity to exam- ine a variety of production questions without resorting to expensive and time-consuming field trials. The EPIC (Erosion Productivity Impact Calculator) generalized crop growth model, originally developed by the U.S. Department of Agriculture (Williams et al.,1984a), was used to develop yield distributions for selected row crops and various irrigation schemes for cotton. EPIC is designed to reflect numerous aspects of the crop pro- duction process including weather, hydrology, sedi- mentation, nutrient cycling, plant growth, tillage, soil temperature, and irrigation effects. Detailed soil, yield, and historical weather information were combined to calibrate EPIC crop parameters for the Trans-Pecos. A second simulation model, known as FLIPSIM (Finn Level Policy Simulation Model) and developed by Richardson and Nixon, was also employed within this study. FLIPSIM is a FORTRAN-based simulation model, which may be used to reflect annual produc- tion, farm policy, marketing, financial management, and income tax considerations within a multi-year framework. FLIPSIM has an imbedded single year lin- \ ear and quadratic programming model, which can be used to formulate optimal annual cropping plans based upon changing price and yield expectations. FLIPSIM *may be used to track key financial variables including debt, net worth, annual income, government payments, and ending equity for up to l0 years. Conditions for continued economic survival can also be specified, thereby allowing one to calculate probabilities of firm survival by performing numerous multi-year simula- fions RESEARCH ORGANIZATION The importance of cotton within the region prompted an emphasis on irrigation levels for that crop. Twelve furrow- and 13 sprinkler-irrigation schemes for vari- ous levels of farm program cotton yield were exam- ined. These alternatives included furrow-irrigation schemes with a preplant irrigation and alternatively, schemes in which the seed are planted before any irrigation and then irrigated to get germination (post- plant irrigations only). Varying numbers of additional postplant irrigations may then be applied. Budgets developed for the various irrigation schemes, plus EPIC simulated yields, were used to generate distributions of whole-farm net returns as- suming an all cotton crop mix. Generalized stochastic dominance techniques (Meyer) were then employed to rank the various schemes. The latter step also entails eliciting utility of net returns points from several pro- ducers and estimating Pratt risk coefficients. Once de- termined, the dominant cotton irrigation schemes were combined with the remaining noncotton crops (barley, forage sorghum, red top cane, and grain sorghum) in several 3-year rotation schemes. EPIC simulations for these rotations revealed only small impacts upon cot- ton yields. Because there was very little effect of crop rotation upon cotton yield, plus the assumed short relevant time horizon of producers with high debt loads, the analysis was based on a single year quadratic programming model with FLIPSIM to generate crop- ping plans in each year of simulation. Strong rotational effects may have necessitated use of a multi-year plan- ning model. Conditions on a 1,600-acre representative farm with 720 acres of cotton base and 360 acres of small grain base were then simulated using FLIPSIM. Five possible cotton irrigation schemes plus four noncotton crops were included as cropping alternatives. Survival of the farm firm was liberally defined as maintaining a debt-to-equity ratio below 9O percent. Fifty, 5-year simulations for numerous scenarios concerning water availability, farm program yield, tenure arrangement, and starting debt were performed. Resulting probabili- ties of survival were calculated based upon the number of simulations that proceeded for the full 5-year period. Similar calculations were made concerning the proba- bility of realizing a 5 percent return on beginning eq- uity. RESULTS Stochastic dominance analysis of furrow-irrigated cotton schemes indicated a preference for the 10-acre- inch preplant schemes over the 8-acre-inch preplant schemes or the water-up schemes. Ranking of sprin- kler-irrigation schemes, consisting of two 3-acre-inch preplant irrigations plus varying postplant irrigations, showed a preference for more water per application as the base cotton yield increased. With an 800-pound base yield, 18 postplant acre-inches applied in 2-acre- inch increments was preferred. Application rates of 2.4 and 3.0 acre-inches dominated for the 1,000- and 1,200- pound base yield cases. These schemes applied 24 and 30 postplant acre-inches of water, respectively. Rank- ings for the sprinkler schemes indicated a preference for a number of applications rather than a given rate. Each of the top schemes has 9 or 10 postplant applica- tions in addition to the preplant irrigations. Whole-farm simulation results covered a variety of scenarios and indicated the importance of water availability in the area. The furrow-irrigation analysis was based on varying water availability conditions (four, six, or nine 700-gpm-wells) and conditions of medium and high starting debt. The medium debt scenario assumed that 50 percent (70 percent) of the value of long (intermediate) term assets was still out- standing. Asset percentages outstanding in the high starting debt scenario were 60 and 80 percent, respec- tively, for the long and intermediate debt classifica- tions. Cropping pattern effects included a mixture of cotton irrigation schemes plus minor acreages of barley and forage sorghum in the four 700-gpm-well scenar- ios. Water resource limitations in this scenario resulted in a mixture of irrigation schemes, a conclusion not possible using stochastic dominance analysis alone. Greater water availability in the six- and nine-well scenarios prompted selection of a single cotton irriga- tion scheme and increased forage sorghum acreages. Barley production remained almost nonexistent be- cause of relative price disadvantage. Cropping pattern results for the 1,000-pound base, yield-sprinkler scenarios showed similar results. Lim- ited water supplies in the three-well (1,000 gpm/ well) scenario resulted in use of two irrigation schemes. In- creased water supplies in the five- and seven-well sce- narios lessened or removed the necessity of mixing cotton irrigation schemes. The apparent preference for a given number of irrigation applications indicated in the preliminary stochastic dominance analysis did not appear in the quadratic programming results. In some cases, schemes that ranked as low as third for a given base yield level had the highest crop acreage in the quadratic programming cropping plan. 6 _Firm survival and economic indicator results for the various scenarios reflected the importance of water availability as well as beginning debt. For furrow-irri- gated farms with four 700-gpm-wells, a 750-pound base cotton yield, and a medium starting debt position,the probability of survival was estimated as 100 percent. This value fell to 58 percent for producers in the high starting debt situation. Probabilities of success (realiz- ing a 5 percent return on beginning equity) were 10 and 2 percent for the two scenarios, respectively. Sprinkler- irrigation results were similar. Probabilities of success and survival generally increased with increased water supplies. Despite the high probabilities of survival reported above, producer net worth declined in all cases over the simulated 5-year period. Average de- clines, across the various furrow-irrigated scenarios examined, range from 13 to 94 percent with similar values applying for the sprinkler runs. CONCLUSIONS Results of this analysis indicate that highly lever- aged producers in the Texas Trans-Pecos will very likely not survive the reductions in target prices em- bodied in the 1985 and proposed for the 1990 farm bill. For those that attempt to do so, their chances are increased by use of more water-intensive, cotton irri ga- tion schemes (i.e., preplant schemes with 10-acre-inches vs. 8 acre-inches, or the heavier sprinkler schemes). Even with cotton target prices in the $0.70 to $0.80 range, the analysis predicts continued declines in farm net worth for medium- and highly leveraged produc- ers. Several consecutive years of above-average prices and yields would likely be required to remove the current debt load and assure a sound agricultural economy in Texas Trans-Pecos. ACKNOWLEDGEMENTS Special thanks go to David Bessler and J. Rod Martin who gave freely of their expertise concerning numer- ous facets of this study. Their contributions added greatly to the depth and scope of this effort. Numerous other individuals contributed valuable time and energy. James (Jimmy) Williams of the U.S. Department of Agriculture, Blacklands Research Cen- ter at Temple, spent many hours explaining various as- pects of the EPIC crop growth model. Dan Taylor and Alan Jones, also at Temple, gave freely of their time as well. Rich Patterson and Charles Stichler (Texas Agri- cultural Extension Service, Ft. Stockton) aided greatly in interviewing area producers and determining rele- J vant cropping alternatives. Special thanks go to the ‘E numerous farmers that participated. They each took time out from busy schedules. We hope that results presented here will, in some way, repay them for their time. Lastly, the authors express their deep gratitude to Ed Rister, John Stoll, and Wyatt Harman for their in- depth review and many helpful comments. Q ‘K Optimal Cropping Strategies Considering Risk: Texas Trans-Pecos John R. Ellis, Ronald D. Lacewell, Iaroy Moore, and James Richardson INTRODUCTION Over the last decade and a half, the Trans-Pecos region of Texas has experienced major changes in input prices, crop prices, and technology. Economic condi- tions vary considerably from those prevailing at the time of the last major analysis by Condra, and input recommendations made in the ECONOCOT program (Texas Agricultural Extension Service, 1977) are in need of reassessment. The majority of producers in the region are experiencing serious financial stress (Hoer- mann) and are attempting to survive in one of the more costly-production regions of the country (U.S. Depart- ment of Agriculture, 1985, 1964). As the health of the national farm economy has declined, producers have come to rely more heavily on the farm program for survival. The 1985 farm bill, however, mandates de- creased support for crops traditionally grown in the region (Knutson et al). Recent advances in biophysical simulation hold much promise in helping the agricultural economists examine input use decisions. Application of models, which accurately predict yield for varying irrigation and fertilizer levels under alternative weather condi- tions, permits one to examine production problems in depth while also considering the influence of farm programs. Application of such models allows a reason- able reflection of the multitude of factors that affect the agricultural producer’ s decision process. Some of those factors include resource constraints, financial condi- tion of the fann firm, government farm program fea- tures, and the goals and preferences of the producer. OBJECTIVES The overall objective of this study is to evaluate economic risk implications of alternative production strategies in the region. Irrigation water is a major limiting factor of production, prompting an emphasis in the analysis upon both optimal levels of application and timing under stochastic weather conditions. Spe- cific objectives of the research are: -1) To examine the risk implications of alternative cropping strategies upon per-acre net returns. 2) To develop whole-farm crop production plans when considering declining government price supports. 3) To determine probabilities of farm survival in a stochastic production/ price environment for the cropping pattern plans developed in objective two. STUDY AREA The Texas Trans-Pecos region is located in the western portion of the state and consists of over 18 million acres of rangeland, desert, desert mountains, and irrigated cropland (Fig. 1). A relatively sparse population, approximately 600,000 (Dallas Morning News), relies upon petroleum, agriculture, and tour- ism as the mainstays of the currently struggling econ- omy. Approximately 85 percent of the land area is farms and ranches, but a much smaller percentage is currently under cultivation (U.S. Department of Com- meroe, 1984). Major agricultural products include cattle, cotton, small grains, hay, vegetables, and cantaloupe. Receipts from crops marketed in 1984 totaled more than $87 million, or 2.4 percent of the state's total (Texas Crop and Livestock Reporting Service). Histori- cally, agricultural production in this region has cen- tered in two counties, Reeves and Pecos; and these two areas are the focus of the present study. Selected his- torical acreages for the 10 counties comprising the Trans-Pecos region and the two counties serving as the study area appear in Table 1. The value of 1984 crop production in Reeves and Pecos counties totaled $23. million. Groundwater supplies provide the majority of ir- rigation water in the region. Ninety-three percent of Texas High Plains New Mexico Presidio Brewster Figure 1. Texas Trans-Pecos Study Area. Table 1. Selected Historical Crop Acreages; Texas Trans-Pecos. Total Grain Forage Pecan Irrigated Year Cotton Sorghum Crops Alfalfa Wheat Other Grains Vegetables Acres“ Pecos County 1969 16,001 11,054 13,519 1,700 6,000 4,600 650 2,500 55,043 1974 10,053 5,890 14,494 4,480 5,200 8,938 1,000 800 51,795 1979 4,814 697 1,573 3,512 1,825 7,548 2,121 3,431 27,291 1984 11,116 1,235 1,863 6,096 1,571 2,246 1,879 2,612 31,231 Reeves County 1969 44,033 6,500 14,594 782 7,788 5,550 0 1 ,700 82,035 1974 40,070 5,320 12,722 4,620 1,800 10,793 0 1,180 78,180 1979 10,179 601 1,263 6,174 4,368 3,391 330 2,289 36,502 1984 13,065 218 1,672 3,435 157 4,965 362 504 27,061 Trans-Pecos” 1969 116,366 38,274 36,412 17,615 15,015 15,725 1,277 7,157 253,118 1974 106,282 20,687 33,124 41,75 7,665 28,016 3,295 6,499 252,636 1979 76,646 10,056 4,579 46,144 11,839 18,224 8,827 15,824 209,447 1984 60,180 9,996 6,305 38,642 13,355 8,326 9,249 11,048 162,391 1974, 1979, and 1984. Report 294, Austin, Texas. 1986. °Sum of individual acreages may not agree with total due to crops grown but not shown. “Includes irrigated acreages for Brewster, Culberson, El Paso, Hudspeth, Jeff Davis, Loving, Pecos, Presidio, Reeves, and Ward counties. Sources: (1 .) Texas Water Development Board, Inventories of Irrigation in Texas 1958, 1964, 1969, and 1974. Report 196, Austin, Texas. 1975. (2.) Texas Water Development Board, Surveys of Irrigation in Texas 1958, 1964, 1969, the irrigation water used in Pecos and Reeves counties in 1984 came from groundwater sources (Texas Water Development Board). Water quality varies across farms and, total dissolved solids average less than 2,000 ppm in the Coyanosa area and about 3,000 ppm in the Pecos pump area (Condra; Texas Water Development Board). Typical pump depths range from 200 to 500 feet (Henggeler). Development of large-scale irrigation from ground- water sources in the Trans-Pecos region began in the 1940's, peaking at 356,000 acres in 1964 (Table 2). De- clining cotton yields, changes in the government farm program, and vastly increased production costs re- duced acreages to 252,000 acres by 1974. Continued increases in energy costs as well as falling output prices reduced acreages further to 162,000 acres by 1984 (Texas Water Development Board, 1986). Many pro- ducers have been forced to the Farmer's Home Ad- ministration (FmHA) when seeking operating loans, and many have negative net worth (Hoermann). Both FmHA lending limits and farm program participation limitation restrictions have further reduced produc- tion options in the region. METHODOLOGY Production variability (risk) and the decision envi- ronment are major components of the production 8 problem under investigation. Two computer simula- tion models were employed, both of which incorporate some aspect of risk analysis. A desire to examine possible cotton irrigation strategies prompted use of the EPIC (Erosion Productivity Impact Calculator) generalized crop growth model. This model is capable of estimating yield distributions for a variety of crops and reflecting the agronomic benefits associated with various rotations. The heavy reliance of Trans-Pecos producers on government farm programs and a desire to reflect the decision environment faced by todays’ agricultural producer prompted use of the FLIPSIM (Firm Level Policy Simulation Model) computer model. Risk Analysis Price and yield variability are the two major sources of risk faced by agricultural producers. Government farm programs mitigate a portion of the price risk, although marketing tools such as forward contracting or commodities futures contracts may also be used. Producers generally attempt to reduce production risk via their chosen production practices and risk manage- ment tools such as crop insurance. Within this analysis, two major risk analysis tech- niques were employed. A quadratic programming (QP) planning model within FLIPSIM determined the an- nual crop mix (including combination of cotton irriga- Table 2. Historical Irrigation Summary; Texas Trans-Pecos. Trans-Pecos Acres Acre-Feet ° Acre-Feet Irrigated Year Irrigated (on-farm use) (per acre) Wells Acres Pecos County 1958 1 17,413 345,266 2.94 636 0 1964 119,313 367,455 3.08 1,166 0 1969 55,043 201,748 3.66 912 0 1974 51,795 183,669 3.54 911 0 1979 27,291 94,462 3.46 915 3,097 1984 31 ,232 90,022 2.88 850 3,794 Reeves County 1958 96,000 358,568 3.83 850 0 1964 118,200 414,217 3.50 975 0 1969 82,035 334,392 4.08 1,010 640 1974 78,170 319,785 4.09 995 1,100 1979 36,502 127,469 3.49 975 1 1,370 1984 27,061 89,688 3.31 935 7,774 Total Trans-Pecos 1958 319,365 1,067,801 3.34 2,467 1,755 1964 356,185 1,101,237 3.09 3,273 2,507 1969 253,118 961,732 2.80 3,056 2,302 1974 252,636 932,108 3.69 3,107 4,442 1979 209,447 662,962 3.17 3,182 42,634 1984 162,391 544,563 3.35 3,022 22,897 °May include both surface and groundwater use. Texas 1986. Source: Texas Water Development Board. Surveys of Irrigation in Texas 1958, 1964, 1974, 1979 and 1984. Report 294, Austin, tion schemes). QP is a popular risk analysis tool (Freund; Anderson et al.) and is one of the many that rely on the premise that decision makers choose from among various alternatives by maximizing their expected well-being or utility. The technique assumes that utility is a quad- ratic function of expected returns. (1.1) U(R) = R'X = XXIX Where X = the vector of activities, R = expected mone- tary returns for each activity, Z = covariance matrix for net returns, and 7. = the Pratt risk aversion coefficient (Pratt). Lambda is used to reflect the relative weight of the variance and covariance of expected returns within the decision makers’ utility function. Equation (1.1) is usually maximized subject to constraints on water la- bor, land, and capital. Quadratic programming assumes either negative exponential or ‘quadratic utility. In many cases one does not know the specific utility functional form. For such cases or those cases in which there is limited infor- ‘ mation on producer attitudes toward risk, efficiency criteria may be used to select efficient subsets of invest- ment alternatives. The efficient subset contains the preferred choices of the individual whose preferences conform to the restrictions associated with the given ef- ficiency criterion. Numerous efficiency criteria exist (Anderson et al.), each with different restrictions on the underlying utility function. Stochastic dominance forms include the first, second, and third degree versions (FSD, SSD, and TSD) as well as stochastic dominance with respect to a function (SDWRF). Use of these criteria generally involve comparing the cumulative distribution func- tions of net returns for the various alternatives (e.g., irrigation schemes) under consideration. Attention is focused here on SDWRF, of which, FSD and SSD are special cases. SDWRF is the most dis- criminatory of the four versions noted (Meyer), yet requires greater information concerning the decision maker's preferences. SDWRF orders uncertain choices for decision makers whose absolute risk aversion func- tion MR) lies within a specified range. This function, which yields the so-called Pratt coefficient at a point, is expressed as (1.2) MR) — U" (R)/U'(R) where U(R) is the individual's utility of net returns function. The requirement of a specific range on MR) 9 allows the greater discriminatory power of SDWRF. Interested readers are referred to the reference section (Anderson et al.; Barry; Markowitz) for further details concerning use of these techniques. Crop Growth Method - EPIC EPIC is a generalized crop growth model devel- oped by the U.S. Department of Agriculture (Williams et al., 1984a and 1984b) and was originally designed to evaluate the impacts of altemative management schemes on both crop yield and soil erosion. The model consists of several major components designed to reflect weather, hydrology, sedimentation, nutrient cycling, plant growth, tillage, soil temperature, and irrigation. Major points favoring use of the EPIC model in- clude reflection of the impacts of alternative irrigation and fertilization schemes as well as that of varying crop rotations on yield over time. This latter facet is sorne- what unique since most crop growth models focus on a single crop in a single year, trading general applica- bility for supposed increased accuracy. Primary EPIC output variables of interest in this study include crop yield, plant water, and nitrogen stress, as well as esti- mates of wind and water erosion. Potential disadvan- tages include the required calibration effort, the restric- tion to a user-determined but pre-set management scheme (independent of weather conditions), and the data-intensive nature of the model. The flexibility of EPIC in estimating yields for several crops, as well as the ability to generate crop yield distributions over a large variety of management schemes, prompted its use within the current study. Extensive calibration efforts were required, both to gain familiarity with the model and to specify crop parameters for Texas Trans-Pecos and range of input variables pertinent to that region. A detailed descrip- tion of the applications of EPIC is presented by Ellis. Firm Simulation - FLIPSIM The Firm Level Policy Simulation Model (FLIP- SIM) is a FORTRAN-based model developed by Richardson and Nixon. A wide variety of agricultural policy options may be reflected using FLIPSIM, al- though the basic farm programs (Glaser; Knutson et al.) involving target prices, nonrecourse loans, and Findley loans (marketing loans in the case of cotton) are the main options emphasized here. In addition to pro- visions of the current (1985) farm bill, the model was used to reflect numerous other aspects of the decision environment facing the agricultural producer. These include equipment replacement, qualification for fi- nancing, annual case withdrawals for family living expenses, and income taxes under current tax laws. The FLIPSIM model operates recursively, simulat- ing the annual production, farm policy, marketing, fi- nancial management, and income tax aspects of a farm firm over a multiple-year planning horizon. Periods ranging from 1 to 10 years may be reflected with FLIPSIM tracking such key variables as debt, net worth, annual income, government payments, and ending eq- uity. Conditions for continued economic survival may 1O also be specified, resulting in the ability when making stochastic runs to estimate the probability of firm sur- vival over the chosen planning horizon. The annual quadratic programming planning model with FLIPSIM may be used to determine annual crop mix based on expected returns over variable costs. Expected net returns per acre are calculated by sub- tracting the expected nonlabor, noninterest variable production costs per acre from expected cash receipts. Government payments are included, if applicable, in expected cash receipts. Expected variable costs are the sum of per acre input costs inflated for annual increases in production costs. Research Organizations Several phases of analysis were required within this research effort. These steps are outlined in Figure 2. Step 1 entailed interviews with producers and agri- cultural experts in the region. Results of those inter- views were used to develop the cultural practices and irrigation alternatives to be evaluated by EPIC (step 2). Collection of detailed information concerning major soils in the area (Hallmark et al.) and weather data (U.S. Department of Commerce, 1966-1985) was also re- quired (step 3). Step 4 required collection of variety test data for cotton and expert opinions concerning the mean and range of yields for several other crops. Once obtained, the data noted above measured in calibrating EPIC for crop production in the Texas Trans- Pecos. Williams and Jones, two of the original develop- ers of EPIC, suggested allowable ranges for selected crop growth parameters. Their input was critical in the calibration effort for cotton, barley, grain sorghum, red top cane, and forage sorghum. Several agricultural producers were also inter- viewed in an effort to elicit information concerning their attitudes toward risk. The modified Ramsey tech- nique of elicitation (Ramaratnam; Lin et al.) was used to obtain data points relating utility to whole-farm net returns. Utility of net returns functions and corre- sponding Pratt risk coefficients (Pratt) were then esti- mated assuming negative exponential, quadratic, and Fourier functional forms. The latter form is flexible (Gallant), allowing reflection of a wider range of deci- sion behavior than the negative exponential or quad- ratic forms. The rest of step 6 entailed the development of crop enterprise budgets for various crops and water availability situations. Two major factors prompted consideration of only cotton at this juncture of the analysis. First, current acreages of other row crops are relatively minor. Sec- ond, Raskin and Cochran caution that Pratt risk coeffi- cients should be used within the context in which they were elicited. Utility is usually elicited for net returns at the whole-farm level, yet crop enterprise budgets are generated on an acreage basis. Raskin and Cochran suggest multiplying per-acre returns by farm size if a whole-farm Pratt range is known. For multi-product farms such scaling of returns requires some assump- tion of crop mix. Given the minor role of other row crops and the primary interest in cotton irrigation JL s1? F\ Y 6. utility function estimation, Pratt COGffiCiGIIiZ derivation, enterprise budget development 1. interview producers, ag »-—- extension and experiment station personnel (2. develop pertinent cultural practices and irrigation schemes k 3. obtain soil and ¢ weather data 4. variety test data or 'expert' opinion on range and mean of yields 5. simulate crop yields (EPIC) 9. price correlation data 7. combine budget information stochastic dominance, and Pratt risk coeflicients to yield subsets of risk effecient irrigation schemes 8. generate yield distributions for various cropping rotations, examine rotational effects (EPIC) 10. generate 250 yield observations for pertinent cotton irrigation schemes and remaining crops (EPIC) l multiple 5 year simulations to obtain crop mix and probabilities of farm firm survival. (FLIPSIM) Figure 2. Progression of research approach. calibration: cotton, barley, grain sorghum, red top cane, forage-so 11 schemes, only cotton was considered in generating whole-farm net returns to be subjected to ranking by stochastic dominance techniques. EPIC-simulated cot- ton yields obtained in step 5 were combined with budget information developed in step 6 to yield distri- butions of whole-farm net returns for the various cot- ton irrigation schemes under consideration. Stochastic dominance, with respect to function techniques using the previously estimated Pratt risk coefficients and the net returns distributions, were used to select subsets of cotton irrigation schemes to be included in subsequent whole-farm simulation analyses. The resulting cotton irrigation schemes were combined with the remaining noncotton crops in sev- eral 3-year rotation schemes deemed likely for use in the region. Potential agronomic benefits of, for ex- ample, a cotton-cotton-barley rotation were examined by obtaining crop yield distributions (by year in the rotation) using EPIC (step 8). Examination of rotational results revealed only small impacts upon cotton yields due to rotational effects, allowing use of a single-year QP planning model. Step 9 entailed collection of historical output price data for the region to calculate a correlation matrix among such prices. Stochastic price simulation within FLIPSIM employs a factored matrix derived from the price correlation matrix to reflect relative price rela- tionships among crops (Richardson and Condra; King). This matrix and mean projected output prices (Knut- son et al.) were used to generate stochastic output prices. EPIC was used once again in step 10 to obtain 250-yield estimates for the various cotton irrigation schemes and crops considered. Simulated yields were used within FLIPSIM such that each yield (cotton, barley, etc.) within a simulated year of farm firm activ- ity corresponded to a given EPIC-simulated weather year. Probabilities of farm firm survival for several water resource and starting financial situations were estimated in step 11 using FLIPSIM. Fifty, 5-year simu- lations were performed for each scenario. Interpreta- tion and summarization of FLIPSIM results took place in step 12. Preliminary Analysis Details concerning the required data and resulting analyses in determining per-acre risk impacts of vari- ous cropping strategies are discussed below. Alternative Crops and Irrigation Strategies Cotton continues to be the major crop in the Trans- Pecos, due both to a relative advantage in the govem- ment farm program as well as favorable weather con- ditions. Discussions with Texas Agricultural Experi- ment Station and Extension Service personnel aided in identifying potential alternative crops such as barley, grain sorghum, red top cane, and forage sorghum. Sprinkler-irrigated barley, accompanied with stocker cattle grazing, is grown as a winter crop. Spring barley is normally furrow-irri gated. Grain sorghum is usually grown in the region using furrow-irrigation practices, but was also included in the analysis as an alternative under sprinkler-irrigation. Red top cane is usually furrow-irrigated, and in some cases lucrative contracts for the harvested seed may be obtained. Both red top cane and forage sorghum hay are sold as roughage to local feedlots and dairies. The importance of cotton prompted an emphasis upon possible irrigation schemes for that crop, while ir- rigation for the remaining crops was assumed to be at levels obtained in interviews with regional experts. A limit of 1O cropping activities within the FLIPSIM quadratic programming planning model contributed to this convention. Cotton furrow-irrigation practices range from an 8- to 10-acre-inch preplant irrigation plus several postplant irrigations to "water-up" schemes where the seed is planted in dry soil and several postplant irrigations are applied. Sprinkler application rates generally vary from 1- to 3-acre-inches. Applica- tions less than 1-acre-inch are generally ineffective due to the low humidity and high summer temperatures. Conversely, center-pivot sprinkler systems may be- come mired in wet soil at application rates greater than 3-acre-inches. Descriptions and code names for the various cot- ton irrigation schemes are presented in Table 3. All sprinkler strategies assumed two 3-acre-inch water-up applications, followed by a varying number of addi- tional irrigations. The S1807 sprinkler scheme, for example, consists of the two 3-acre-inch water-up ap- plications plus 18-acre-inches applied in seven equal applications across the irrigation season. Furrow strate- gies include water-up schemes (POST3, POST6) and preplant schemes with an 8- or 10-acre-inch preplant application plus a varying number of postplant irriga- tions. A POST3 scheme consists of an 8-acre-inch wa- ter-up application plus two 5-acre-inch applications for a total of three postplant irrigations. A PP2-10 scheme consists of a 10-acre-inch preplant application plus two 5-acre-inch applications. The PP2-8 scheme is similar except that it entails an 8-acre-inch preplant irrigation. Calibration of EPIC Detailed information from cotton variety tests (Gannaway, Texas Agricultural Experiment Station, 1982, 1983, and 1985) and actual historical weather data (Moore, 1980-86) was used to adjust selected EPIC crop parameters until reasonable agreement between aver- age variety test results and predicted yields was ob- tained. Williams and Jones provided initial relevant ranges on selected parameters including harvest in- dex‘, pest factor, and runoff ratio. Percent differences between average variety test yields and EPIC pre- dicted yields range from -12.3 to +2.8 percent for ‘Harvest index is the fraction of aboveground biomass which is harvested. 12 cotton. Extensive regional variety test data for grain sorghum, barley, forage sorghum, and red top cane did not exist, prompting calibration using mean yield data elicited from experts in the area. Selected final EPIC crop growth parameters appear in Table A1 of the Appendix. Simulated Yields Prior to calculation of net return budgets, EPIC was employed to estimate average yield values for the various cotton irrigation schemes. Estimation of those yields was a two-step process. Nitrogen (N) fertiliza- tion levels were first calculated using the automatic fer- tilization option within EPIC. A stress level trigger of 5 percent was assumed, promoting the application of N whenever the EPIC N stress index rose above that value. Average uN and standard deviation 0N values for the amount of N applied were calculated using 25 EPIC runs. Nitrogen levels assumed for the final irriga- tion simulations were set at uN + 0N to ensure that N Table 3. Description of Alternative Cotton Irrigation Strategies; Texas Trans-Pecos. Total Acre-Inches Preplant Water-up Postpiant Applied Rate Number Rate Number Rate (in.) (in.) (in.) Sprinkler Schemes° S1806 24 2 3 6 3.0 S1807 24 2 3 7 2.6 S1809 24 2 3 9 2.0 S1812 24 2 3 12 1.5 S1818 24 2 3 18 1.0 S2408 30 2 3 8 3.0 S2410 30 2 3 10 2.4 S2412 30 2 3 16 2.0 S2416 30 2 3 10 1.5 S3010 36 2 3 12 3.0 S3012 36 2 3 12 2.5 S3015 36 2 3 20 2.0 S3020 36 2 3 20 1.5 Furrow Schemes” PP2-8 18 8 2 5.0 PP2-10 20 10 2 5.0 POST3 18 1 8 2 5.0 I PP3-8 23 8 3 5.0 PP3-10 25 10 3 5.0 POST4 23 1 8 3 5.0 PP4-8 28 8 4 5.0 PP4-10 30 10 4 5.0 POSTS 28 1 8 4 5.0 PP5-8 33 8 5 5.0 PP5-10 j 35 10 5 5.0 POST6 ‘ as 1 a s 5.0 °Sprinkler scheme mnemonics SXXYY imply two, 3-inch, water-up irrigations XX acre-inches applied in YY positions. “Furrow scheme mnemonics PPX-Y imply a preplant irrigation of Y acre-inches plus X postplant irrigation of 5-acre-inches. A POSTZ scheme implies an 8-acre-inch water-up plus Z-1, 5-acre-inch, postplant irrigations. 13 stress on the plant did not confound interpretation of irrigation results. Applied N levels, as well as sum- mary statistics for the final furrow and sprinkler simu- lations, appear in Table 4. Simulation-determined N levels were at or below levels common in the region. Furrow yield results corresponded fairly well with prior expectations in terms of absolute and relative magnitude. Yields for the schemes may be considered in subgroups of three. For example, the POST3, PP2-8, and PP2-10 schemes apply an almost identical amount of total water. Within that subgroup, and similar groups as well, average yield for the water-up scheme (POST3 in our example) lies somewhat between that for the 8- inch and 10-inch preplant schemes. Interviews with producers revealed some preference for the water-up schemes. Using maximum expected yield as a crite- rion, the 10-inch preplant scheme showed a higher yield than a corresponding postplant irrigation-only l4 Table 4. Simulated Cotton Yield Summary Data, Furrow- and Sprinkler-Irrigation; Texas Trans-Pecos. Mean Yield/Acre Std. Dev. c.v.° Skewness Min. Max. Nitrogen (lbs) (lbs) (lbs) (lbs) (lbs) Furrow Schemes“ POST3 525 94 .1790 ~ .8056 388 752 70 POST4 656 1 15 .1753 .51 77 478 896 94 POSTS 776 120 .1546 .4220 61 1 996 1 14 POST6 916 120 .1310 .4568 731 1151 139 PP2-8 505 - 93 .1842 .8137 377 718 82 PP2-10 555 97 .1 748 .7926 422 780 90 PP3-8 636 103 .1619 .7328 506 870 108 PP3-10 690 106 .1536 .7226 557 932 121 PP4-8 780 1 1 7 .1500 .7031 641 1050 133 PP4-10 840 118 .1405 .7158 701 1114 146 PP5-8 901 149 .1654 .041 1 664 1 151 152 PP5-10 966 146 .1511 -.0280 716 1213 166 Sprinkler Schemes° S1806 799 127 .1630 .3948 61 1 1022 140 S1807 790 129 .1633 .4423 615 1035 141 S1 809 794 124 .1 562 .5708 634 1 044 138 S1812 789 126 .1597 .7229 639 1074 134 S1818 759 122 .1607 .6128 591 1009 130 S2408 1018 130 .1277 .2565 810 1292 164 S2410 1020 126 .1235 .2796 821 1260 ‘ 166 S2412 1017 128 .1258 .4886 825 1271 165 S2416 991 135 .1362 .5450 810 1281 162 S3010 1233 87 .0706 -.1908 1056 1416 192 S3012 1232 89 .0722 -.2787 1045 1421 189 S3015 1218 91 .0747 -.0155 1035 1423 190 S3020 1205 100 .0830 .0151 1031 1421 190 °Coefficient of variation risk measure: (std. dev./mean). “Furrow scheme mnemonics PPX-Y imply a preplant irrigation of Y acre-inches plus X postplant irrigations of S-acre-inches. A POSTZ scheme implies an 8-acre-inch water-up plus Z-1, postplant irrigation of S-acre-inches. Yield values are based on 25 observations and assuming a 24% turnout (lint/harvested biomass) ratio. °Sprinkler scheme mnemonics SXXYY imply two 3-inch, water-up irrigations plus XX acre-inches applied in YY applications. Yield values are based on 25 observations and assuming a 24% turnout (lint/harvested biomass) ratio. Jl. strategy. Final conclusions on this point, however, may not be drawn until economic criteria are applied. Cost of water, resource availability (especially timing), and distribution of yield may alter such a conclusion. Similar results apply for sprinkler-irrigation, al- though average yields in this case exceeded actual yields. This applies especially for the water intensive S30 schemes? Two possible explanations may be in- ferred: 1) the S30 schemes lie outside in terms of total applied water, the data used to calibrate EPIC for sprinkler irrigations; and 2) the pest factor may not be constant. Calibration of EPIC for sprinkler cotton cov- ered irrigation applications ranging from 18 to 24 inches. and hail and insect damage may be increasing func- tions (percentage-wise) of yield or plant population. For the S18 schemes, average yield varied little for application rates between 1.5 inches (S1812) and 2.57 inches (S1807). A similar plateau of constant yields occurs for application rates between 2 and 3 inches. Application rates apparently make little difference in average yield when operating within these "windows" of optimality. As with the furrow case, constraints on water availability or distributions of yields could de- termine an alternative preferred strategy. Yield risk conclusions drawn from the data in Table 4 is mixed. Standard deviations of yield gener- ally increase with increased water application (i.e., going from a POST3 to POST4 scheme). Yield variance, however, decreases with increased water application when going from the S24 to the S30 sprinkler schemes. An alternative measure of yield risk is the coefficient of variation. This measure declines unambiguously as total applied water increases. Yield variance, as a per- centage of mean, falls indicating less risk as total water applied increases. Enterprise Budget Development Extensiveinterviews with area producers revealed detailed information concerning equipment comple- ment size and composition. Tillage and irrigation op- erations were then summarized for the cotton irriga- tion schemes and remaining crops. Three farm sizes (960, 1,600, and 2,500 acres) were assumed to apply in the region based on producer interviews and U.S. Ag- ricultural Census data (U.S. Department of Commerce, 1984). The Microcomputer Budget Management Sys- tem, or MBMS, (Texas Agricultural Extension Service, 1986) was used to develop crop enterprise budgets for each cotton irrigation scheme and crop by farm size. In- dividual budgets were calculated assuming two well sizes (700 and 1,000 gpm) for furrow-irrigated farms and one well size (1,000 gpm) for farms using center- pivot sprinklers. Comparative per acre returns to land, manage- ment, and risk for the various cropping alternatives are summarized for the 960-acre farm equipment comple- ment in Table 5. Yields used in the budgeting process were the average of 25 individual EPIC simulations for each crop irrigation scheme. Assumed output prices and selected input prices appear in footnote a of Table 5 and in Table A3 of the Appendix. Expected returns (above variable and fixed costs) continue to favor cotton over the small grains. The summary also indi- cates greater returns for the more water-intensive cot- ton irrigation schemes. Red top cane returns also ap- pear attractive, assuming that contracts for sale of the Table 5. Summary of Projected Per Acre Returns to Land, Management, and Risk“; Texas Trans-Pecos. Furrow Sprinkler Crop/lrrlg. Crop/lrrig. Scheme” $lAcre 8cheme° $lAcre Cotton Cotton POST3 71.99 81806-81818 171.18 POST4 1 15.01 82408-82416 258.24 POSTS 152.31 83010-83020 337.62 PO8T6 200.63 PP2-8 55.66 PP3-8 104.89 PP4-8 1 51 .1 5 PP5-8 _ 186.31 PP2-10 75.09 PP3-10 127.03 PP4-10 177.81 PP5-10 217.20 Sorghum -26.89 Sorghum -16.80 Barley -53.21 Barley -61.34 Forage Sorghum 65.00 Forage Sorghum 76.36 Red Top Cane 265.96 "Crop enterprise budget returns calculated assuming the following prices: lint ($.55/lb), cotton ($105/ton), grain sor- ghum ($3.21/cwt), barley ($1.90/bu), forage/red hay ($64/ ton), and red top cane seed ($10.40/cwt). Furrow returns based on well costs for a 700 gpm well using 7.5/kwh electricity. Sprinkler costs based on a 1 ,000 gpm well yield. “Furrow scheme mnemonics PPX-Y imply a preplant irriga- tion of acres-inches plus X postplant irrigations of 5-acre- inches. A POSTZ scheme implies an 8-acre-inch water-up plus Z-1 postplant irrigations of 5-acre-inches. °8prinkler scheme mnemonics SXXYY imply two 3-inch water-up irrigations plus XX acre-inches applied in YY ap- plications. zThe S30 schemes have two 6-acre-inch water-up applications plus 30 additional acre-inches applied across the irrigation season. A key to irrigation scheme mnemonics appears in Table 3. 15 seed are available to alleviate price risk. Implications for the 960-acre farm were similiar for the 1,600 and 2,500 acre farms. Probability distributions of whole-farm net re- turns to management and risk were calculated using EPIC-generated yields (as summarized in Table 4) and the previously developed cotton budgets. A 79.4-cent per pound price and a $110 per ton cottonseed price were employed. The lint price included 55 cents for market or retums from the loan program, and 24.4 cents per pound for deficiency payments. Producers were assumed to have access to 960 acres with 432 acres of cotton base acreage. A 25-percent set aside re- quirement reduced the 432 acres to 324 permitted or allowed acres available for cotton production. Thus, producers were assumed to farm only a third of their actual acreage, a situation not uncommon in the area given the relatively high cost of inputs, low prices for small grains, and limited availability of operating capi- tal. Several farm program yield (FPY) levels were assumed to apply depending on the average yield for the irrigation scheme under consideration. Summary statistics for the returns distributions appear in Table 6. For the sprinkler-irrigated schemes, average net returns appeared to be maximized for schemes having 8 to 10 applications. For example, within the S18 alternatives, the S1809 scheme has the highest average net return with a value of $36,712. (Note: As one applies more water [e.g., moves from the S18 to S24 to S30 schemes] the coefficientof variation of net retums declines.) Greater water application levels lead to more stable returns. Generally, similar conclusions apply to the fur- row-irrigated schemes. As one applies more water in going from the POST3 to POST6 schemes the coeffi- cient of variation declines. This trend does not apply, however, for the two preplant (8- and 10-inch) schemes. The coefficient of variation of net returns is a minimum for the four postplant application rate, yet increased slightly when more water is applied in the five post- plant scheme. Once calculated, the net return distributions were ranked using stochastic dominance with respect to a function technique. This preliminary screening was Table 6. Summary of Whole-Farm Returns to Management and Risk“; Texas Trans-Pecos. Irrigation Farm Program Returns to Management and Risk ($) Scheme” Yield (lbs) Mean Min. Max. Std. Dev. c.v.° S1806 800 34,381 8,830 71,339 19,262 .5603 S1807 800 36,008 9,481 73,292 19,572 .5436 S1809 800 36,712 12,411 74,595 18,884 .5144 S1812 800 35,982 13,062 79,145 19,204 .5337 S1818 800 31,294 5,899 69,386 18,470 .5902 S2408 1,000 50,555 18,975 92,227 19,715 .3899 S2410 1,000 50,919 20,602 87,344 19,113 .3753 S2412 1,000 50,476 21,253 88,972 19,392 .3842 S2416 1,000 46,531 18,975 90,560 20,436 .4392 S3010 1,200 63,813 37,508 92,855 13,421 .2103 S3012 1,200 63,423 34,252 92,854 13,635 2150 S3015 1,200 61,730 34,252 92,854 13,943 .2259 S3020 1,200 61,105 34,578 93,831 15,210 .2489 PP2-8 550 12,560 -6,909 44,856 14,199 1.1306 PP3-8 660 27,310 7,580 62,927 15,657 .5733 PP4-8 800 33,938 12,71 1 74,894 1 7,794 5243 PP5-8 930 38,594 2,690 76,595 22,642 5864 PP2-10 550 15,992 -4,1 78 50,139 14,748 9222 PP3-10 660 31,245 10,984 67,899 16,133 .5164 PP4-10 800 38,496 17,413 80,195 17,867 .4641 PP5-10 930 A 43,991 6,029 81,562 22,161 5038 POST3 550 16,066 -4,796 50,530 14,340 8926 POST4 660 29,057 1,997 65,465 17,544 .6038 POSTS 800 34,540 9,471 68,074 18,198 .5269 POST6 930 41,748 13,567 77,379 18,276 .4378 °Coefficient of variation risk measure: (std. dev./mean). “Calculations based on a 960-acre cotton farm with 432 acres of cotton program acreage. Prices assumed include $0.55/lb lint, $110/ton cottonseed, 7.5/kwh electricity, and $24lacre land charge. bSprin kler scheme mne monics SXXYY imply 2-inch water-up irrig ations plus XX acre-inches applied in YY applications. Furrow scheme mnemonics PPX-Y imply a preplant irrigation of Y acre-inches plus X postplant lrrigatlons of 5-acre-inches. A POSTZ scheme implies an 8-acre-inch water-up plus Z-1 postplant irrig ations of 5-acre-inches. 16 ~L it, "\ used to reduce the number of cotton irrigation strate- gies considered. Assumed Pratt intervals included a risk neutral interval (-.0O001, .00001) and two risk averse intervals ([.01, .0002] and [.O01, .0021). Results portrayed in Table 7 indicate several points. First, rela- tive rankings were invariant across the assumed Pratt intervals and farm program yield values. For the fur- row schemes, more water is almost universally pre- ferred. Only for the lowest base yield (550 lbs) is the water-up POST3 scheme preferred to its more water- intensive PP2-10 counterpart. In all cases, the 10-inch preplant dominates the 8-inch preplant schemes. Sprinkler scheme rankings place a 2-inch applica- tion rate (S1809) first for the lowest base yield. As base yield increases, however, the preferred application rate increases. A 2.4-inch application rate dominates for the S24 schemes. Rankings for the sprinkler schemes may also indicate a preference for number of applica- tions, rather than a given rate. Each of the top schemes had 9 or 10 postplant applications. Once chosen, the subset of alternatives selected via stochastic domi- nance techniques must face the additional test of re- source feasibility. A producer may prefer the net re- turns distribution associated with a water-intensive scheme. Limited water or labor supplies or financial constraints may, however, preclude consideration of that particular scheme. Given this possible resource constraint, five irrigation schemes were chosen for three assumed farm program yields using each of the two types of irrigation. These cotton irrigation altema- tives (Table 8) were chosen by: 1) taking the three most preferred schemes with average yields near the as- sumed farm program yield, and; 2) combining those schemes with the two top-ranked schemes from the adjacent farm program yield classification(s). The al- tered QP planning model within FLIPSIM selected the optimal annual crop mix from among the five desig- nated cotton schemes plus the noncotton crop altema- tives. Whole-Farm Analysis The previously described research focuses pri- marily on acreage-based issues. Examination of the op- timal choice of cotton irrigation schemes and the poten- tial impact on farm-firm survival is a natural extension. Given the declines in government target prices, what are the expected impacts on a whole-farm bases? The FLIPSIM policy simulation model provided an excel- lent means of reflecting the numerous factors acting in such a situation. Whole-Farm Scenarios Producers in Texas Trans-Pecos face a variety of fi- nancial and resource constraints, and the performance of farm firms in the region is definitely affected by ac- crued debt. Starting debt was assumed to have two possible values. These were referred to as medium and high debt as shown in Table 9. Differing levels of outstanding intennediate and long-term debt charac- terize the different debt scenarios. Additional factors affecting farm-finn perform- ance were water availability and cotton farm program yield. Details concerning the number of wells, starting debt levels, and assumed farm program yield are shown in Tables 10 and 11 for the furrow and sprinkler analy- sis. Furrow-irrigated farms were assumed to use 700 gpm irrigation wells, while sprinkler systems had ac- cess to 1,000 gpm wells. Two possible tenure arrange- ments were also assumed: full ownership and a mixed own/ cash lease scenario. Farm program acreages for Table ‘I. Stochastic Dominance Rankings oi Aitemative Cotton irrigation Schemes“; Texas Trans-Pecos. Fann Program Yield (lbs) Ranking 550 660 800 930 1 000 1 200 Furrow Schemes 1 POST3 PP3-1 0 PP4-10 PP5-10 2 PP2-1 0 POST4 POSTS POSTS 3 PP2-8 PP3-8 PP4-8 PP5-8 Sprinkler Schemes 1 S1809 S2410 S3010 2 S1807 S2408 S3012 3 S1812 S2412 S3015 4 S1806 S241 6 S3020 5 S1818 °Pratt coefficient ranges oi (-.0O001, .00001), (.0001, .0002), and (.001, .002) were employed, although relative rankings did not vary across those ranges nor across the three farm sizes investigated. Relative rankings were invariant to assumed farm program yield as well. 17 Table 8. Cotton Irrigation Schemes Selected for Inclusion in Whole-Farm llllodel; Texas Trans-Pecos. Fam1 Program Yield (lbs) 550 750 800 930 1 000 1 200 Furrow Schemes Considered° POST3 POST4 POSTS POST4 POSTS POSTS PP2-8 PP3-10 PP4-10 PP2-1 0 PP4-8 PP5-8 PP3-10 PP4-10 PP5-10 Sprinkler Schemes Considered” S1 807 S2408 S301 0 S1809 S2410 S3012 S1812 S2412 S3015 S2408 S1 809 S2408 S2410 S3010 S2410 “Furrow scheme mnemonics PPX-Y imply a preplant irrigation of Y acre-inches plus X postplant irrigations of 5-acre-inches. A POSTZ scheme implies an 8-acre-inch water-up plus Z-1 postplant irrigation of 5-acre-inches. “Sprinkler scheme mnemonics SXXYY imply two 3-inch water-up irrigations plus XX acre-inches applied in YY applications. Table 9. Beginning Debt Specifications; Texas Trans- Pecos“ °Developed from FmHA data. Debt levels are not inclusive of all producers in the Texas Trans-Pecos, but are gener- ally representative. “Debt ratios expressed are long- or intermediate-term debts divided by the value of long or intermediate-term assets. the 1,600-acre farm assume 720 acres for cotton and 360 acres of small grain base. Scenario names consist of five parts. The first letter defines furrow- or sprinkler- (F or S) irrigated scenar- ios. The second column of the name lists the number of wells for that scenario, with well numbers chosen to cover a reasonable range of water availability condi- tions. The third column conveys the land tenure speci- fication O for owner and L for the own / lease case. Owned land was valued at $300 per acre. 18 Cash lease costs vary from $40 to $90 per acre of allowed cotton acreage (Hoermann), prompting as- sumption of a cash lease cost of $60 per allowed cotton acre. The relative proportion of cotton acreage to total farmland reduced the $60 value to a net cost of $20.25 Debt Scenario” per leased acre. Medium or high starting debt levels are Medium mgh noted by M or H in column 5. The final designation (columns 5-8) is the assumed cotton farm program Outstanding long-term debt .50 .60 Yleld f“ Pounds of hflt _ _ Six broad scenarios were originally analyzed, three Outstanding intemediatwerm for furrow (500-lb, 7o50-lb, and 930—lb farm program debt _70 3o yield) and three similar scenarios for sprinkler. Only the 750-lb furrow and 1,000-lb sprinkler cases, how- ever, are examined here. Results for the remaining scenarios may be found in Ellis’ work. FLIPSIM Assumptions Numerous additional assumptions were required prior to simulating the activity of the 1,600 acre farm firm. A 5-year time horizon was chosen and expected target and market prices determined from Knutson et al., and projections from historical data (Table A2). Effective loan rates listed reflect the possible use of Findley loan provisions for small grains and the mar- keting loan for cotton. Projected cottonseed prices reflect the historical relationship between lint and seed prices. Expected forage hay prices reflect a trend line calculation from past observations. Grazing fees on barley were assumed to be $27/acre, inflated 3 percent per year. Red top cane was not considered in the whole-fariri analysis because of the specialized nature of seed contracts. Input prices (Table A3) were inflated 4 percent per year. b ~45! dozoo .2 22> EEmoEQ E22 - N 2 m _oo .5: u I .32 n 2 ._¢>¢_ 5% - 5. _8 .332 $23 ovw 53o 930m 8Q .mwmw_E>>o n ._ 53o u O .225. - m 6o .263 6 EwQESE - w Eu Emiccqm u m .305: u u. - P _ou 5056360 mEEmc ozmcmowu U_0_> 258M051 FCNL COZOU X X w 831.6» X X o ofizém X X v 855.... X X v 0525i X X m ofizomm X X m ofizofi X X m ommIOmm X X m ommIOmm X X w c.5102 X X m ommzomm X X o ofizow» X o ommzomm X X w ommIOom X X ¢ ofizovm X X ... 0310i X X w ommzow» X X w Qfizoi X < 850E =9: 622 ammo; :30 m=o>> oEmz E9: 522 0234 :30 n=o>> mEmz =2: dos. ammo; :30 m=o>> oEuz Boa E30 .02 ormcoow Eon E30 .oz ocmcoow San E30 .02 ocucoom 95cm... 95:3. 95:00. m2 o8 m2 9k ma: 0mm dooomécfl... mmxoh rwotmcoom co=mm_.:_ 325m .2 mu=w_._o~um._u;0 gum d» 2am» 19 22> £202.. chm... cozou douoo .2 20... 2.20.005 E8. - m 2 m _oo E9; n I 00E u s. ._0>0_ 5% - w .00 .0000_E3o n .. 63o u 0 .0.2.2 - m _oo 0:03 .0 .0082. - w .00 ..0_vE..Q0 - m .395. n u. - _. .8 Eo:c0>¢oo @582. o..0=0owo x . x 2 002x08 x x 2 ooo .208 x x 0 08.1000 x x m 000.1000 x x m 081000 x x 0 08:20am x x m o8 .2000 x x m 0002000 x x 0 0002.000 x x m 08.1000 x x 0 08:08 x x 0 08.2000 x x m ooo .2000 x x m 0002000 =9... 60s. 0000. :30 0=0>> 0E0z =2: 60s. 0000... :30 0=0>> 0E0z =2: 60s. 0000.. :30 0=0>> 0E0z 58 E30 .oz 2.2.2.0 58 E30 .oz ocucoom S00 E30 az 0:20am 0. 2.0... 0.2.0 h 0.2.0 ... 0n. 8w; 0n. 2s; 2.. 80 dooomécflh 00x0... ...0o_.0c0om 2.0.3.0.... .0_v.:....w .0. 0u=0_.080.0_.o 5.0“. . 5 050... 2O -\ The QP planning model within FLIPSIM included annual constraints for labor and numerous biweekly constraints limiting groundwater availability. Land constraints imposed farm program limits on cotton and small grain acreage as well as set aside require- ments. The $50,000 limit on deficiency payments was assumed ineffective. A fixed net return (over variable cost) covariance matrix based on estimated net returns for the 1982-1986 period was employed to reflect po- tential variation in net returns across alternative irriga- tion schemes and crops. Five observations were in- cluded in calculation of the covariance matrix due to limited historical price data and the assumption of a limited relevant historical period. Lambda, the Pratt coefficient of risk, was set at .00002 for the QP planning model. This value implies a slight aversion to risk, a result common in past risk studies (Anderson; King). The .00002 value is also rep- resentative of the values found in examining the Pratt coefficients derived from utility elicitation of several Trans-Pecos producers (Ellis). The negative exponen- tial, quadratic, and Fourier functional forms yielded fairly similar Pratt estimates for the range of net returns under consideration. Optimal Cropping Patterns Optimal crop mix varied significantly with water availability and FPY, although results only for the 750- pound furrow and 1,000-pound cotton farm program yield sprinkler scenarios appear here (Tables 12 and 13). Results for alternative levels of fann program yield appear in Ellis. Furrow acreage results for the 750- pound FPY scenario indicate significant impacts on crop mix and preferred irrigation schemes due to lim- ited water supplies. The impact of the assumed irriga- tion windows are most evident in the four-irrigation well scenario. Restricted water supplies and timing effects force the selection of the three separate irriga- tion schemes (POST4, PP3-10, and PP4-10) for cotton. The more water-intensive PP4-10 scheme, however, was the most preferred of the three. This preference apparently strengthened as target price declined. Over time, declining acreages for the less water intensive POST4 and PP3-10 strategies support this contention. Total cotton acreage declines as increasing input prices and declining target prices prompt putting more water on limited acreage. Barley production is positive at 161 acres in 1987, but continued expected weak prices for " small grains precluded production in later years. Small v'\ increases in forage hay prices resulted in small acre- ages in the latter years. Total cropped acreages also decline because of the absence of small grain produc- tion. . Greater water supplies in the six-well scenario result in selection of a single cotton irrigation strategy. All of the farm's allowed cotton acreage used the PP4- 10 irrigation plan. Timing aspects for irrigation appar- ently were no longer a limiting factor. Barley produc- tion continued in its weak posture, and the remainder of the increased water supplies were used for 128 acres of forage sorghum production in each of the 5 years considered. Total cropped acreages remained stable after 1988. Further increases in water supplies (9 wells) allow increased barley production in the first 2 years and a single year of grain sorghum production. Total cropped acres generally decline over time, but are highly de- pendent on the assumed trend in forage hay prices and inflationary pressure on input prices. Examination of results for the 550- and 930-pound cotton farm program yield scenarios (not shown) yield similar conclusions: 1) limited water supplies may force a mixture of irrigation schemes; 2) for the relative output prices assumed, small grain production pros- pects continue to be unfavorable; and 3) forage sor- ghum appears to be a viable production alternative, es- pecially if water supplies are adequate. Table 1 2. Optimal Crop Mix for Furrow Irrigation, 750- lb Base Cotton Yield“; Texas Trans-Pecos. Year Crop/Scheme“ 1987 1988 1989 1 990 1 991 (acres) (4 Wei/sf Cotton: POST4 117 94 94 30 30 PP3-10 88 65 65 O 0 PP4-10 335 381 381 445 440 Barley 161 0 O 0 0 Forage Sorghum 23 O 0 1 6 (6 Wells) Cotton: PP4-10 540 540 540 540 540 Barley 86 O 0 0 0 Forage Sorghum 128 128 128 128 128 (9 Wells) Cotton: p PP4-1 O 540 540 540 540 540 Barley 287 109 0 0 O Forage Sorghum 234 218 222 244 201 Sorghum 0 132 O 0 0 °Optimal cropping mix includes alternative irrigation strate- gies for cotton as well as acreages for noncotton crops. Base yield refers to the proven cotton yield tor government tarm programs. “Furrow scheme mnemonics PPX-Y imply a preplant irriga- tion of Y acre-inches plus X postplant irrigations of 5-acre- inches. A POSTZ scheme implies an 8-acre-inch water-up plus Z-1 postplant irrigation of 5-acre-inches. °Furrow irrigation scenarios assumed 700 gpm irrigation wells. 21 Table 13. Optimal Crop Mix for Sprinkler Irrigation, 1,000-lb Base Cotton Yield“, Texas Trans-Pecos. Year Crop/Scheme” 1987 1988 1989 1990 1991 (acres) (3 Wells)‘ Cotton: S2412 507 507 507 398 252 S3010 33 33 33 106 203 (5 Wells) Cotton: S2412 72 O 0 0 0 S3010 468 540 540 540 540 Forage Sorghum 102 78 78 78 78 (7 Wells) Cotton: S3010 540 540 540 540 540 Forage Sorghum 206 204 209 215 172 Sorghum 120 122 117 0 0 °Optimal cropping mix includes alternative irrigation strate- gies for cotton as well as acreages for noncotton crops. Base yield refers to the proven cotton yield for government farm programs. “Sprinkler scheme mnemonics SXXYY imply two 3-inch water-up irrigations plus XX acre-inches applied in YY ap- plications. °Sprinkler irrigation scenarios assumed 1,000 gpm irriga- tion wells. Results for the sprinkler scenarios are presented in Table 12. Limited water supplies in the three-well sce- nario resulted in a mixture of the S2412 and S3010 schemes. The apparent preference for a given number of irrigation applications indicated in the preliminary stochastic dominance analysis did not appear in the quadratic programming results. Total cropped acre- ages decline slightly over time. Limited water and the relative profitability of cot- ton resulted in production of only that crop in the three-well scenario. Cotton acreages were highest for the S2412 alternative, although stochastic dominance rankings (Table 7) for the 800-pound cotton base yield case placed the S2412 scheme third behind the S2410 and S2408 alternatives. Resource constraints resulted in selection of a significantly different irrigation scheme than that chosen with stochastic dominance. Increased water supplies in the five- and seven- well scenarios resulted in greater acreages of both forage sorghum and the more water-intensive cotton strategies similar to the furrow scenarios. The seven- well scenario had enough water to support grain sor- ghum production. Acreages for that crop declined, however, as target prices declined over time. Several observations arise from the preceding results. First, limited water availability resulted in a choice of schemes, some of which were deemed inferior when using stochastic dominance techniques. Second, as tar- get prices declined, producers had an incentive to shift cotton production to the more water-intensive irriga- tion alternatives. Declining output prices, increasing input costs, and a desire to maintain farm program base acreage can be expected to prompt such a strat- egy. Interpretation of these results should be tem- pered, however, by recalling that farm program pay- ment limitations are assumed ineffective. If, however, program payment limitations are effective and the variable cost of production exceeded the loan rate or world market price, producer incentives would be to target production for their established farm program yield. Significant production above the FPY would have to be sold at a loss at market price or the loan rate. Additional acreage results of note include the con- tinued poor performance of small grains and the attrac- tiveness of forage sorghum as a cropping alternative given adequate water supplies. One might question whether total cropped acreages would decline as pro- jected. This question applies especially in the case of center-pivot sprinkler systems. In either case, it is likely that the land would no longer be cropped. This phe- nomenon has occurred commonly in the past (Tables 1 and 2). Sprinkler systems, although representing a large capital asset, would simply not be replaced. Reductions in sprinkler-irrigated acres in such instances might not proceed as rapidly as portrayed here, but would eventually occur nonetheless. Firm Survival FLIPSIM tracks numerous variables pertaining to the survival and relative financial health of the farm firm. Stochastic simulation results in a distribution of values for many of those variables. Discussion in this section focuses on mean values for selected measures of financial health as well as estimates of the probabili- ties of survival and success. Results pertaining to FLIPSIM simulations for the furrow-irrigated scenarios with a cotton farm program yield of 1,000 pounds per acre appear in Table 13. Probabilities of survival and success for the various scenarios are listed as well as starting net worth and debt. The average present value of ending net worth (assuming a 5 percent discount rate) is presented for comparison with its initial value. Average values for ending debt (nominal) as well as annual average cash receipts, net cash incomea, and government payments are presented in Table 14 as well. Under the liberal condition for firm survival (main- taining 10 percent equity), the majority of representa- 3Net cash income consists of total cash receipts less total cash expenses. Total cash expenses do not, however, include principal payments or family living expenses. 22 .22 620E. $0 m 052. 02221.8 02$ .32? Emwmin douoo .2 v6; 8.2005 E02 - .0 2 m _oo 50E n I .62: u _>_ ._m>m_ 5% . .18 .2362 38m 9.0 52o $28 Q09 mwmmiio n ._ .56 u O 6:52 - m 6o 22> B .382. - 0 _oo .6_v_ccqw u w £0.02 u m - F _8 E2Em>¢ou 05E? ocmcmomo N0 N0 F0 . F0 N0 N0 .00 .00 #0 ._.0 000. 5 mémcamm ._>o0 _m::c< 0.0 0.00 0.0T _..0 0.0- <2 0.0- 0.0? 0. F0- 0.0 000.5 0Eo0c_ £000 202 _m:cc< .6... Q00 .0v0 .000 . :2. .60. 00v 00¢ .0v0 .000 00040 2900mm 53 _m:cc< .000 4.0? .000 .000 .000 .000 .0 F0 .000 .000 .50 000. 5 50D 050cm .0 F0 .000 .0 F0 .0 F0 .8... .000 00v .000 00v .000 000. P0 50D 0cEE00m .00- .0 T .00- .00- .00. .0 T .00- .0 T . d0- .00- o». $0.550 E0905 N0 .000 .00 0E. .00 N00 .00 N00 NT .000 000.5 n53; ~02 E95 .>d N0? 600 N0? .000 .000 ._.00 .000 .000 600 .000 000. F0 _._to>> ~02 05.05000 .00 .00 .0 .0 F d? .00 .0 P .00 .0 .0 _. ...\¢ 0000000 2o gnmnoi .00 .00? .00 .02 .00 .00? .00 .00_. .00 .00? ..\o 622$ 2o a___nmpo.n_ 002.501 92.2.6“ 0022.21 0022.21 “.3109. 000E000 0001092 0320mm 00010.1 850E 2:3 0.. ammo-Z umocucoom A2002 ctr- E5“. co 50D 05230 0cm >=__..~__~>< 2203 0c_>._0> 0o $002M uwfiszwu .3. 030k A ~ A ~ 23 tive farms would still be in operation in 1991. Esti- mated probabilities of survival range from 56 percent for the four-well owned high-debt (F4OH750) scenario to 100 percent for all of the scenarios with so-called medium starting debt levels. Probabilities of success (earning a 5 percent return on beginning net worth), however, are much less likely. Among wholly owned scenarios, the six- and nine-well medium-debt scenar- ios have the greatest chance of success. Their chances of a reasonable return, however, were surpassed by the high probability of success (62 percent) experienced by the five-well lease / own medium-debt (F6LM750) sce- nario. Values for beginning and present value of average ending net worth portend continued erosion of pro- ducer equity with percent declines ranging from 13 to 94 percent. Firms in the medium starting debt classifi- cation fare somewhat better, yet still experience signifi- cant declines in net worth over the 5-year period. Several factors contribute to such declines. Increasing production costs, static land values, and reductions in government fann program target prices force produc- ers to borrow against owned land to continue produc- tion, reducing equity in the process. Outstanding debt values also increase. Government payments exceed net cash income in all cases, indicating a strong reliance on the price and income support programs. Average net cash income values are negative in most cases, even before extraction of living expenses and principal payments. In terms of resource scenarios, producers in the six-well medium-debt situation (F6OM750) appear to have a relative advantage over their four- and nine- well counterparts. Restricted water supplies in the four-well case (F4OM750) resulted in an average an- nual net cash income of $2,700 and ending debt of $341,000. Conversely, the F6OM750 operation has an average annual net cash income of $19,300 and ending debt of only $283,000. Additional water availability for the nine-well owner (F6OM750) results in essentially the same net cash income and $10,000 more average debt. Producers often view the water resource question in lélTTlS of how many acres to crop given a fixed water supply. The answer to such queries is a function of crop mix water requirements, output prices, and input prices. For the six 700-gpm well scenario, the 685 average cropped acres (Table 12) indicate a needed capacity of 6.1 gpm per cropped acre. This value increases slightly to 7.35 gpm per cropped acre for the none-well sce- nario. Net cash income values for the two scenarios (six- and nine-well) allow for land payments and re- pairs and maintenance on wells. Given the relative economic advantage of the six-well scenario with rela- tively good net cash income and lower ending debt, the optimal capacity ratio is likely around the 6.1 gpm per cropped acre value. Tenure arrangement also matters in the area as producers in the lease/ own scenarios fare better than their wholly owned counterparts. Probabilities of suc- cess and survival for the lease/ own situations exceed those for the owned scenarios. Declines in net worth 24 are less as well. Comparative results across tenure situations could change, however, if cash lease costs vary significantly from the effective rate of $20.25 per leased acre assumed here. Results for producers using sprinkler irrigation are summarized in Table 15. Probabilities of survival and success improve slightly over those occurring for the 750-pounds FPY furrow-irrigated case. Net worth continues to decline, even with the admittedly high simulated yields associated with sprinkler irrigation. Increased assets (i.e., sprinkler systems) result in greater starting debt levels than for furrow irrigation. Debt loads continue to increase, fueled by liberal loan poli- cies and growing demand accompanying falling target prices. Ending debt and average net cash income val- ues do, however, show a slight advantage for produc- ers in the five- or seven-well owned medium-debt classifications. SUMMARY AND CONCLUSIONS Provisions of the current government farm pro- gram are designed to reduce income support levels (i.e., target prices) during the 1986-1990 period. Im- pacts of reduced deficiency payments upon a high cost production area such as Texas Trans-Pecos could be dramatic. Agricultural producers in the region rely heavily on government farm program payments, and changes in those programs could imply major adjust- ments in production. Potential responses include al- teration of irrigation practices, changes in crop mix (possibly to nonprogram crops), and idling of acreages. Objectives of this study include an appraisal of po- tential cropping strategies at both the acre and whole- farm levels of analysis. Yield and price risk effects were accounted for explicitly due to the large role such fac- tors play in the ultimate success or failure of agricul- tural firms in the region. Preliminary Risk Analysis Results Traditional enterprise budgeting techniques dem- onstrated greater expected returns for more water-in- tensive cotton irrigation schemes. Forage sorghum and red top cane provided reasonable nonprogram crop- ping alternatives and small grains were deemed infe- rior due to insufficient market returns and farm pro- gram support. Use of stochastic dominance techniques with dis- tributions of whole-farm net returns for various cotton irrigation alternatives demonstrated a general prefer- ence for the more water-intensive schemes. Rankings for furrow irrigation generally identified in the 10-acre- inch preplant schemes as preferable to the water-up and 8-inch preplant schemes applying approximately the same amount of water. Only in the low farm program yield (550 lbs) case did the water-up scheme (POST3 in this case) dominate the 10-inch preplant alternative. Sprinkler rankings appeared to prefer a number of applications rather than a particular application rate. Preferred sprinkler schemes each had 9 or 10 postplant applications. Five cotton-irrigation schemes, chosen 19 if 1,000-lb Base Cotton Yield; Texas Trans-Pecos. '4 Table 15. Estimated Effects of Varying Water Availability and Starting Debt on Fann Firm Health, Sprinkler irrigation, Scenario“ Measure Unit S3OM1000 S3OH1000 S50M1000 S50H1 O00 S7OM1000 S7OH1000 n Probability of Survival % 100. 24. 100. 62. 100. 62. Probability of Success % 10. 4. 36. 8. 42. 14. Beginning Net Worth $1,000 378. 239. 379. 239. 381. 240. p.v. Ending Net Worth” $1,000 222. -48. 287. -48. 291. -49. (Percent Change) % -41 . -120. -24. -120. -24. -120. Beginning Debt $1,000 370. 510. 371. 511. 374. 515. Ending Debt $1,000 483. 710. 407. 656. 405. 657. Annual Cash Receipts $1,000 438. 439. 560. 560. 630. 629. Annual Net Cash income $1,000 -1.7 -24.9 18.4 -7.4 20. -5.9 Annual Govt. Payments $1,000 79. 82. 81. 82. 86. 87. col 2 - number of wells. “Present values (p.v.) calculated using a 5% interest rate. °Scenario naming convention: col 1 - F = furrow, S = sprinkler. col 3 - tenure, O = own, L = own/lease, (960 acres own, 640 acres lease). col 4 - debt level, M = med, H = high. col 5 to 8 - farm program yield. from among the top-ranked schemes at or near a par- ticular farm program yield, were combined with non- cotton crops for consideration in the whole-farm simu- lation analysis. Whole-Farm Results Optimal cropping patterns identified via applica- tion of the quadratic programming model within FLIP- SIM provided insight into irrigation management. Multiple irrigation schemes were chosen when water supplies were limited, indicating that timing of appli- cation becomes more important under those condi- tions. Furthermore, restricted water availability re- sulted in the selection of non-optimal, in terms of stochastic dominance rankings, irrigation schemes. Inclusion of resource feasibility constraints not ac- counted for in stochastic dominance analysis resulted in a different set 0f preferred activities, and in general, the more water-intensive alternatives were preferred. T he 10-inch preplant and water-up furrow-irrigation schemes were chosen over the 8-inch preplant strate- gies. Low water supplies resulted in a mixture of 2- and 3-inch sprinkler application rates, but moved toward the 3-inch rate with more adequate water supplies. De- c~clining target prices and relatively high production costs resulted in negligible barley and grain sorghum production, while forage sorghum production became -\profitable once cotton base acreage water require- ments were met. Significant firm survival results include the small chances of economic success and moderate to high chances of survival. The "quality" or length of that survival may be questionable given the analysis’ indi- cated erosion of net worth, negative net cash income, and increased levels of debt. Few of the scenarios examined offered a reasonable hope of reducing the significant debt levels many Trans-Pecos producers now hold. Results indicate that producers with moderate starting debt have a greater chance of surviving on- coming reductions in farm program income support (target price) levels. Producers with mixed own/ cash lease operations also may have a greater change of survival. In all cases, producers in the region will con- tinue to depend heavily upon government farm pro- grams. Resources related results indicate that returns will be maximized if water and land are combined in ratios of 6 or 7 gpm per cropped acre. Even these optimal rates will not, however, offset the adverse effects of high starting debt levels as evidenced by declining net worth in all but the most optimistic scenarios. CONCLUSIONS Producers in the Trans-Pecos will continue to struggle given the assumptions made in this analysis. Current projected prices (market and government) simply will not allow producers to overcome current debt loads. Significantly higher market prices over an extended period could aid greatly in that effort, but do not appear likely. 25 REFERENCES Anderson, I.R. "Simulation: Methodology and Application in Agricultural Economics." Review of Marketing and Agricultural Economics. 42(1974):3-55. Barry, P.I. and D.R. Willman. "A Risk Programming Analy- sis of Forward Contracting with Credit Constraints." Amer. I. Agric. Econ. 53(1976):62-70. Condra, G.D. "An Economic Feasibility Study 0f Irrigated Crop Production in the Pecos Valley of Texas." Texas Water Resources Institute TR-101, Texas Agricultural Experiment Station, 1979. Dallas Morning News. Texas Almanac and Industrial Guide, 1986 and 1987. Dallas: A.H. Belo Corporation, 1985. Ellis, I.R., R.D. Lacewell, and D.R. Reneau. "Estimated Eco- nomic Impact From Adoption of Water-Related Agri- cultural Technology." West. I. Agric. Econ. 10(1985):307- 21. Ellis, I.R. "Risk Efficient Cropping Strategies and Farm Sur- vival: Texas Trans Pecos" Ph.D. dissertation, Texas A&M University, 2987. Freund, R.I. "The Introduction of Risk Into a Prgramming Model." Econometrica 24(1956):253-63. Gallant, A.R. "The Fourier Flexible Form." Amer. I. Agric. Econ. 66(1984):204-O8. Gannaway, I. Cotton Improvement Specialist, Texas Agri- cultural Research and Extension Center, personal com- munication. Lubock, Texas, October 1986. Glaser, L.K. Provisions of the Food Security Act of 1985. U.S. Dept. of Agriculture, Economics Research Service, Ag- ricultural Information Bulletin No. 498, 1986. Hallmark, A.M., L.T. West, L.P. Wilding, and L.R. Drees. "Characterization Data for Selected Texas Soils," Texas Agricultural Experiment Station MP-1583, 1986. Henggeler, I. Unpublished irrigation plant efficiency tests, Pecos County, 1975-1982. Texas Agricultural Extension Service, Ft. Stockton, Texas. Hoermann, I. Director, Farmers‘ Home Administration Of- fice, personal communication. Ft. Stockton, Texas, 1985. Hoermann, I. Director, Farmers’ Home Administration Of- fice, personal communication. Ft. Stockton, Texas, 1985. King, R.P. "Operational Techniques for Applied Decision Analysis Under Uncertainty." Ph.D. dissertation, Michi- gan State University, 1979. Knutson, R.D., E.G. Smith, I.W. Richardson, I.B. Penson, Ir., D. W. Hughes, M.S. Paggi, RD. Yonkers, and D. T. Chen. Policy Alternatives for Modifying the 1985 Farm Bill. B-1561, Texas Agricultural Experiment Station, College Station, Texas, 1987. Lin, Wm., G.W. Dean, and C.V. Moore. "An Empirical Test of Utility vs. Profit Maximization in Agricultural Produc- tion." Amer. I. Agric. Econ. 56(1974):497-508. Markowitz, H.M. Portifolio Selection-Efficient Diversifica- tion of Investments. New York: Iohn Wiley and Sons, 1959. Meyer, I. ”Choice Among Distribution.” I. Econ. Theory 13(1977):325-36. Moore, I. "Historical Weather Data for Texas Agricultural Experiment Station. Pecos, Texas, 1980-86." Unpub- lished. 26 Pratt, I.W. ”Risk Aversion in the Small and the Large." Econometrica 32(1959):122-36. Ramaratnam, S.S. "Texas Coastal Bend Grain Sorghum Pro-f ducer’ s Fertilizer Decisions Under Uncertainty." Ph.D. dissertation, Texas A&M University, 1985. Raskin, R. and M.I. Cochran. ‘Interpretations and Transfor- mations of Scale for the Pratt-Arrow Absolute Risk Aversion Coefficient: Implications for-‘jGeneralized Sto- I. chastic Dominance.” West I. Agric. Econ. 11(1986):204-&J 10. Richardson, I.W. and C.I. Nixon. "Description of FLIPSIM V: A General Firm Level Policy Simulation Model". B- 1528. Texas Agricultural Experiment Station, College Station, Texas, 1986. Richardson, I.W. and G.D. Condra. "Farm Size Evaluation in the El Paso Valley: A Survival / Success Approach." Amer. I. Agric. Econ. 63(1981):430-37. Texas Agricultural Experiment Station. "Cotton Variety Tests in the Trans-Pecos Area of Texas, 1981 ." TR-82, El Paso and Pecos, Texas, 1982. Texas Agricultural Experiment Station. ”Cotton Variety Tests in the Trans-Pecos Area of Texas, 1982.” TR-83-1, El Paso and Pecos,Texas, 1983.Texas Agricultural Experiment Station. ”Cotton Variety Tests in the Trans-Pecos Area of Texas, 1984." Lubbock and Halfway, Texas, 1985. Texas Agricultural Extension Service. ”Drip Irrigated Cotton Symposium.” Midland, Texas, February 18-19, 1986. Texas Agricultural Extension Service. ”Texas ECONOCOT System." Texas A&M University System, Ft. Stockton, Texas, 1977. Texas Crop and Livestock Reporting Service. ”1984 Texasw Agricultural Cash Receipts and Price Statistics." Austin, Texas, 1985. Texas Water Development Board. ”Surveys of Irrigation in Texas, 1958, 1964, 1969, and 1984." Report No. 294. Austin, Texas, 1986. U. S. Department of Agriculture. ”Economic Indicators of the Farm Sector, Costs of Production, 1984." EDIFS 4-1, Econ. Res. Serv., 1985. U. S. Department of Agriculture. ”Costs of Producing Up- land Cotton in the United States." Econ. Res. Serv., Agric. Econ. Rept. 99, Washington, DC, 1964. U.S. Department of Commerce. ”1982 Census of Agriculture, Texas State and County Data." Vol. 1, Part 43, U.S. Govt. Printing Office, Washington, DC, 1984. U. S. Department of Commerce. "Climatological Data, Texas." U.S. Govt. Printing Office, Washington, DC, 1966-1985. Williams, I.R., C.A. Iones, and P.T. Dyke. "A Modelling Approach to Determining the Relationship Between Erosion and Soil Productivity." Trans. ASAE (1984a):129- 44. Williams, I.R., I.W. Putnam and P.T. Dyke. ”Assessing the Effects of Soil Erosion with EPIC." Proc. Natl. Symp. Erosion and Soil Prod., New Orleans, LA, December 1984b. Williams, S.R. and C.A. Iones. Texas Agricultural Experi- ment Station, personal communication. Temple, Texas, 1985 and 1986. v8 APPENDIX Table A1. Selected EPlC° Crop Parameters; Texas Trans-Pecos. Grain Red Top Forage Barley Cotton Sorghum Cane Sorghum Biomass/energy 30.0 20.0 40.0 28.0 32.0 Harvest index .42 .55 .33 .34 .33 Optimal temp. for plant growth 15.0 27.5 27.5 27.5 27.5 Min. temp. for plant growth 0.0 12.0 10.0 10.0 10.0 Max. leaf area index 8.0 5.0 5.0 5.0 5.0 Fraction of growing season when leaf area starts declining .80 .85 .80 .72 .90 Leaf area development parameter 1 15.01 15.01 15.01 15.01 15.01 Leaf area development parameter 2 50.95 45.95 50.95 50.95 50.95 Leaf area decline rate factor 2.0 2.0 .5 2.0 2.0 Biomass/energy decline rate factor 10.0 2.0 2.0 10.0 10.0 Aluminum tolerance 2.0 3.0 2.0 2.0 2.0 Maximum crop height 1.2 1.0 1.5 2.5 2.5 Maximum root depth 2.0 2.0 2.0 1.5 1.5 Harvest efficiency .95 .90 .95 .95 .95 Pest factor .95 .80 .95 .95 .95 Fraction water in yield .12 .01 .14 .11 .14 Nitrogen in plant at 0. growth .0600 .0580 .0440 .0440 .0440 Nitrogen in plant at .5 growth .0231 .0192 .0164 .0164 .0164 Nitrogen in plant at 1. growth .0134 .0177 .0128 .0128 .0128 P in plant at 0. growth .0084 .0081 .0060 .0060 .0060 P in plant at .5 growth .0032 .0027 .0022 .0022 .0022 P in plant at 1. growth .0019 .0025 .0018 .0018 .0018 Wind erosion factor (standing live) 3.39 1.138 .657 3.39 3.39 Wind erosion factor (standing dead) 3.39 .603 .657 3.39 3.39 Wind erosion factor (flat residue) 1.61 .332 .320 1.61 1.61 Crop category 5. 4. 4. 4. 4. Potential heat units 2056. 2400. 1918. 1918. 1205. °EPlC (Erosion Productivity Impact Calculator) refers to a generalized crop growth simulation model developed by the U.S. Department of Agriculture (Williams et al. 1984b). 27 Table A2. Prolected Market, Target, and Adjusted Loan Prices; Texas Trans-Pecos, 1 987-1991. Table A3. Selected 1986 Input Prices, Cro Calculations; Texas Trans-Pecos. p Enterprise Budget] \ . Commodity 1987 1988 1989 1990 1991 Cotton (cents/lb) Market 53.32 59.16 59.40 67.90 67.90 Target 79.40 77.00 74.50 72.90 72.90 Adjusted Loan 41.80 40.00 40.00 40.00 40.00 Cottonseed ($/ton) Market 85.31 94.66 95.04 108.64 108.64 Barley ($/ton) Market 1.79 1.84 1.92 2.08 2.08 Target 2.60 2.55 2.47 2.38 2.38 Adjusted Loan 1.49 1.42 1.34 1.27 1.27 Grain Sorghum ($/cwt) Market 3.31 3.41 3.55 3.83 3.83 Target 5.14 5.04 4.89 4.66 4.66 Adjusted Loan 3.14 2.98 2.83 2.68 2.68 Forage Hay ($/ton) Market 61.20 63.20 66.00 71.60 71.60 Grazing ($/acre) Market 27.00 27.81 28.64 29.50 30.39 Sources: Knutson et al., 1987 and projections from histori- cal data. 28 Cost Item Unit ($/unit) Electricity kwh .075 Gasoline gal i .90 Hired Labor (repair/maintenance) hr 6.00 Q Hired Labor (irrigation) hr 3.00 Insurance Rate (% of mkt. value) % 1.00 Interest Rate (intermediate borrow) % 10.00 Interest Rate (intermediate equity) % 10.00 Interest Rate (oper. capital borrow) % 10.50 Interest Rate (oper. capital equity) % 10.50 Interest Rate (positive cash flow) % 5.25 Interest Rate (investment capital) % - 7.50 Lube Multiplier none .10 Owner Labor (repair/maintenance) hr 6.00 Owner Labor (irrigation) hr 6.00 Personal Property Tax % .50 Nitrogen Fertilizer (NH3) lb .16 Nitrogen Fertilizer (dry) lb .28 Nitrogen Fertilizer (liquid) lb .28 Hail Insurance (cotton) $100 valuation 11.00 Herbicide (cotton) appl 6.00 Herbicide (sorghum) appl 5.50 Seed (cotton) lb .55 u Seed (barley) lb .15 “a Seed (grain sorghum) lb .80 Seed (forage sorghum) lb .44 Seed (red top cane) lb .30 i» U‘ ‘we [Blank Page in Original Bulletin] " 3., 4-, .4 = A; -, C? l\ Mention of a trademark or a proprietary product does not constitute a guarantee or warranty of the product by The Texas "l Agricultural Experiment Station and does not imply its approval to the exclusion of other products that also may be suitable. All programs and information of The Texas Agricultural Experiment Station are available to everyone without regard to race, color, religion, sex, age, handicap, or national origin. 1M—1 1 -90