1 A SHRIMP AQUACULTURE INVESTMENT BY GRAPEFRUIT PRODUCERS IN THE INDIAN RIVER PRODUCTION REGION OF FLORIDA: A RISK ASSESSEMENT By JENNIFER L. CLARK A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2007
2 2007 Jennifer L. Clark
3 To Christopher C. Ferraro for his friendship, love, and logic.
4 ACKNOWLEDGMENTS For it matters not how small the beginning may seem to be, what is once well done, is done forever.--Thoreau First and foremost, I thank God for the opportuni ty to pursue and complete this masters program. I am humbled and truly grateful for the opportunity. My gratitude extends to the University of Florida Department of Food a nd Resource Economics faculty, administration, and staff for providing a consummate learning environm ent in which to achieve excellence and meet the highest level of academic and personal goals. I am very proud to be a Florida Gator and graduate of the College of Agriculture and Li fe Sciences at the University of Florida. I especially thank Frank Sonny Williamson for his progressive vision and pioneering efforts to establish a shrimp aquaculture research facility at the University of Florida/ Indian River Research and Education Center located in Fort Pierce, Florida. Appreciation is also extended to Governor J.E. Jeb Bush and the Fl orida Department of Agriculture and Consumer Services (Charles Bronson, Commissioner a nd Sherman Wilhelm, Director) for providing resources to promote aquaculture research for agricu ltural producers in the State of Florida. This thesis research project could not have been comp leted without the efforts of these individuals. I extend deepest appreciation to my committee members who have assisted me in my research effort to complete a Master of Science degree. This project has been a labor of love, frustration, sweat, and hard work not to men tion many sleepless nights. I thank each of you for your advisement as we transformed this research idea into a tangible contri bution to the field of agribusiness and applied agricultural economics. Much gratitude is extended to my committee chair, Dr. Charles Adams, for considering that th e research idea for this project had merit when it was first presented to you. Your experience w ithin the aquaculture community and investment perspective was invaluable as we attempted to de fine the scope, of what was at times, a very
5 formidable challenge. Many thanks are extende d to Dr. Ferdinand Wirth for always finding time to discuss the real-world influe nces inherent in modeling a dyna mic, biological system. I am grateful for the opportunity to have worked with you. Sincere appreciation is expressed to Dr. P.J. van Blokland, for his patience and tenacity that lent valuable support to this project. It was his insight and decorous rhetoric that served to fortify the co llective rapport of this group. With warm regard, heartfelt appreciation is given to Dr. Rick Weldon for sharing his economic insight, research acumen, grounding in fluence, and mentoring support during the transformative progression from student to researcher. I thank my family and friends for their encouragement and patience in my academic pursuit. Each of you, in your own special way, has contributed to my success and achievement. I am grateful for your abounding love and support. Special thanks go to Chris Ferraro who allowe d me to lean on his strength during moments of mental and physical exhaustion and who se lflessly dedicated countless hours during my time of academic accomplishment. His effort was truly remarkable and, for me, has forged a bond of friendship and love that will forever be incomparable. Thank you Beaux always.
6 TABLE OF CONTENTS page ACKNOWLEDGMENTS...............................................................................................................4 LIST OF TABLES................................................................................................................. ..........9 LIST OF FIGURES................................................................................................................ .......10 ABSTRACT....................................................................................................................... ............11 CHAPTER 1 INTRODUCTION..................................................................................................................13 Preamble....................................................................................................................... ..........13 Problematic Situation.......................................................................................................... ....15 Researchable Problem........................................................................................................... .18 Research Objectives............................................................................................................ ....19 Hypotheses..................................................................................................................... .........20 2 SOME ECONOMIC IMPLICATIONS FOR ESTABLISHING A SHRIMP AQUACULTURE INDUSTRY IN FLORIDA......................................................................24 Introduction................................................................................................................... ..........24 Economic Context............................................................................................................... ...25 Value-Added Shrimp vs. Low-Cost Commodity Shrimp.......................................................26 Competition and Market Conduct..........................................................................................27 Differentiated Product......................................................................................................... ...28 Direct Marketing to Local Restaurants...................................................................................29 Consumer Demand for Locally-Grow n, Fresh (Never-Frozen) Shrimp................................30 Factor Resources Available for New Industry Creation.........................................................31 Generic and Collective Advertising Programs.......................................................................32 Summary........................................................................................................................ .........36 3 DEVELOPING A STOCHASTIC SPREAD SHEET MODEL TO ANALYZE THE POTENTIAL OF A SHRIMP AQUACULTURE INVESTMENT.......................................39 Introduction................................................................................................................... ..........39 Research Objective............................................................................................................. ....40 Simulation Model Methodology.............................................................................................41 Production System Methodology...........................................................................................42 Hypothetical Production System Design................................................................................43 Capital Cost................................................................................................................... ..44 Product Size................................................................................................................... ..45 Growth Function..............................................................................................................46 Survival Rates................................................................................................................. .48
7 Feed Costs..................................................................................................................... ..49 Inflation...................................................................................................................... .....49 Market Prices.................................................................................................................. .49 Revenue Calculation........................................................................................................50 Hurricane Probabilities....................................................................................................52 Key Output Variables (KOVs) of Financial Performance......................................................53 Scenario Overview.............................................................................................................. ....55 Scenarios 1 and 2Deterministic An alysis of Two Stocking Densities..........................55 Scenario 3 Probability of Positiv e NPV at a Higher Stocking Density........................55 Scenario 4 Higher Price Premium................................................................................56 Scenario 5 Reducing Initial Capital Costs....................................................................56 Scenario 6 Reducing the Discount Rate.......................................................................57 Scenario 7 Below Average Survival Rates...................................................................57 Scenario 8 Random Kill and Hurricane Events............................................................58 Deterministic Analysis Results...............................................................................................59 Baseline Operating Variables..........................................................................................59 Scenario 1 Stocking Density 80 Shrimp per m2...........................................................59 Scenario 2 Stocking Density 100 Shrimp per m2.........................................................60 Stochastic Scenario Framework.............................................................................................61 Stochastic Analysis Results....................................................................................................63 Scenario 3 Probability of Economi c Success at a Higher Stocking Density................63 Scenario 4 Higher Price Premium................................................................................64 Scenario 5 Reducing in itial capital costs......................................................................64 Scenario 6 Reducing the Discount Rate.......................................................................65 Scenario 7 Below Average Survival Rates...................................................................65 Scenario 8 Random Kill and Hurricane Event Probabilities Included.........................66 Summary........................................................................................................................ .........67 4 PROBABILISTIC FINANCIAL ANALYSIS FOR TWO AGRICULTURAL ENTERPRISES: APPLICATION FOR A GRAPEFRUIT & SHRIMP DIVERSIFICATION STRATEGY........................................................................................85 Introduction................................................................................................................... ..........85 Problem Setting................................................................................................................ ......86 Portfolio Selection Problem............................................................................................87 Stochastic Simulation Problem........................................................................................88 Model Development.............................................................................................................. .90 Shrimp Production System..............................................................................................90 Grapefruit Production System.........................................................................................92 Forecast of the Critical Variables....................................................................................92 Forecast of Stochastic Distributions................................................................................94 Expected Value Calculation............................................................................................95 Stochastic Probability Calculations.................................................................................97 Results........................................................................................................................ .............97 Summary........................................................................................................................ .........98
8 5 SUMMARY, CONCLUSIONS, AND IMPLICATIONS....................................................103 Conclusions.................................................................................................................... .......104 Implications................................................................................................................... .......107 Future Research Needs.........................................................................................................109 APPENDIX A HYPOTHETICAL SHRIMP PRODUCTION SYSTEM CONSTRUCTION COSTS.......110 B HYPOTHETICAL SHRIMP PRODUC TION SYSTEM REPLACEMENT SCHEDULE....................................................................................................................... ..115 LIST OF REFERENCES.............................................................................................................117 BIOGRAPHICAL SKETCH.......................................................................................................130
9 LIST OF TABLES Table page 2-1 Factors influencing market conduct within an industry.....................................................38 3-1 Capital costs for greenhouse..............................................................................................70 3-2 Capital costs for four 0.29-acre ponds...............................................................................70 3-3 Inflation ratesfor inputs used in the model for years 2006-2021.......................................71 3-4 Saffir-Simpson wind damage scale....................................................................................73 3-5 Storm event probability and damage description for St. Lucie County.............................73 3-6 Baseline variables for a five-acre shrimp farm investment...............................................74 3-7 Scenario 1 Net cash income for a 5-acre shrimp farm investment at 80 shrimp per m2 stocking density............................................................................................................75 3-8 Scenario 2 Net cash income for a 5-acre shrimp farm investment at 100 shrimp per m2 stocking density............................................................................................................77 3-9 Scenario 6 Probability values for calc ulated discount rates ranging from 3% to 7% and corresponding 50% NPV values................................................................................82 4-1 Calculated shrimp variable costs (V.C .) per pound (tail-weig ht) at two stocking densities, 80 shrimp/ m2 and 100 shrimp/ m2..................................................................100 4-2 Grapefruit variable costs (VC) per ac re, assuming 95 trees planted per acre..................101 4-3 Type of probability distributions defined for each stochastic variable............................101 A-1 Greenhouse Capital Cost Budget.....................................................................................110 A-2 Four 0.29-Acre Ponds Capital Cost Budget.....................................................................112 B-1 Greenhouse Component Es timated Useful Life..............................................................115 B-2 Pond Component Esti mated Useful Life.........................................................................116
10 LIST OF FIGURES Figure page 1-1 Illustration of citr us canker infection.................................................................................22 1-2 Map of the Indian River production region in Florida.......................................................22 1-3 Outline of steps in the risk management process...............................................................23 2-1 Wholesale and HRI shrimp prices 19932006..................................................................38 2-2 Examples of FDACS Florida Agricult ural Promotional Campaign material....................38 3-1 Flow chart of m odel component parts...............................................................................69 3-2 Low-salinity shrimp production system diagram..............................................................69 3-3 L. vannamei shrimp growout weight calculation...............................................................71 3-4 Shrimp imports (all product form s) by hemisphere for years 1989-2006. ......................72 3-5 Retail (HRI) historical shrimp prices January 1999-August 20 06 (31-35 tails/ lb)...........72 3-6 Scenario 3 CDF of net present value of net cash income with 100 shrimp per m2 stocking density............................................................................................................... ..79 3-7 Scenario 4 CDF of net present value of net cash income with $0.66 price premium......80 3-8 Scenario 5 CDF of net present valued of net cash income with 50% reduced capital costs.......................................................................................................................... ..........81 3-9 Scenario 6 CDF of net present value of net cash income at varying discount rates.......82 3-10 Scenario 7 CDF of net present value of net cash income with reduced shrimp survival at varying discount rates......................................................................................83 3-11 Scenario 8 CDF of net present value of net cash income with random kill and hurricane events at varying discount rates.........................................................................84 4-1 Geometric interpretation of cropping combinations for portfolio analysis.....................100 4-2 Prediction error illustration..............................................................................................101 4-3 Portfolio frontier for NPV of net cash income for three investment scenarios...............102 4-4 CDF of net present value of cash flows (NPV) for each enterprise.................................102
11 Abstract of Thesis Presen ted to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy A SHRIMP AQUACULTURE INVESTMENT BY GRAPEFRUIT PRODUCERS IN THE INDIAN RIVER PRODUCTION REGION OF FLORIDA: A RISK ASSESSEMENT By Jennifer L. Clark August 2007 Chair: Charles M. Adams Major: Food and Resource Economics Recent risky events including crop diseas es, weather events, and non-agricultural development have negatively impacted grapefru it production in Florida. Traditional risk management tools include insurance, c ontracting, government programs, and crop diversification. Existing culture technologies and high consumer demand for shrimp may make this species a feasible candida te as a crop diversification strategy for Florida farmers. The primary objective of this research project was to identify potential costs, benefits, and associated risks for Florida grap efruit producers interested in a low-salinity shrimp production investment. The economic success of this invest ment is determined by the positive probability of net present value (NPV) of net cash income over a 15-year planning horizon. Various management implications are examined via sc enario simulation includ ing different stocking densities, a higher local price premium, lower ca pital construction costs, lower survival rates, and the addition of random kill and hurricane events. The computer software, Simetar, used in this analysis operates as an Excel add-in to generate statistical probabili ties specific to the production assumptions programmed into a spreadsheet accounting model. Preliminary mode l output indicates that this hypothetical, lowsalinity shrimp investment is capital intensive (relative to grapefruit production) and a low
12 probability exists for a positive NPV of net cash inco me. In subsequent scenarios, an analysis of various management implications indicates a higher probability of economic success of the shrimp investment, although the pot ential for negative NPV values co ntinue to persist when the probabilities of random kill and hurricane events are considered.
13 CHAPTER 1 INTRODUCTION Preamble Agricultural investments are subject to ma ny unforeseeable elements because of their biological nature and market de pendency. Many of these elements are beyond the producers control. While farm profit can be viewed as the reward for accepting risk, the risk to the agricultural producers future cash income can be significant and may increase over time (Hardaker, 2004). Although in this context risk can be associated with the possibili ty of greater reward, it is the necessary balance between risk exposure and risk tole rance that management seeks to meet strategic goals (Olsen, 2004). Production and price risks are traditionally recognized as the greatest sources of agricultural risk (Fumasi, 2005). Economists differ entiate between risk a nd uncertainty with the assumption that the risky events can be appr oximated via a known probability density function, whereas the probabilities of uncer tain events are simply unknown (Roberts et al., 2004). Welldefined probability estimates, specific to the firms financial performance, can provide a proactive methodology to analyze the risk parameters associated with an agricultural investment. Traditional risk management tools include crop insurance, government programs, contracting, hedging, options, fina ncial reserves, and choice of production activities (Miller et al., 2004). Additionally, crop divers ification is one activity that can reduce the firms exposure to financial risk. The success of a diversific ation strategy is highly de pendent on the correlation between commodity characteristics, including vari ety, planting date, input prices, market prices, environmental conditions, and changes in cons umer demand (Olsen, 2004). Reducing the dispersion in overall returns by selecting a combin ation of characteristics with low or negative correlations is the objective of a crop diversification st rategy (Hardaker, 2004). Consideration of
14 the variance of net incomes across alternatives, as the dispersion measurement, is one approach to defining the risk reduction achieved through crop diversification (Robinson, 1979). Crop selection in a diversification strategy is similar in nature to the portfolio selection problem in financial management. Typically, th e decision maker relies heavily on the portfolio selection hypothesis, traditionally used to evalua te financial instruments, which suggests that the investor does (or should) cons ider that expected return is a desirable thing and variance of return is an undesirable thing (Markowitz, 195 2). This hypothesis was extended to include farmers and assumes that agricu ltural investors are generally risk averse (Johnson, 1967). The general strategy of portfolio selec tion in the agricultural production context sets forth to calculate the probability distribution of net returns from farm enterprises, and select the efficient combination of these enterprises that minimize th e variance of expected net income. Stochastic simulation models, as a method for risk analysis, c ontinues to be widely us ed to calculate risk exposure via the probability of an events o ccurrence (Fumasi, 2005; Ri chardson et al., 2000; Zucker & Anderson, 1999; Griffin & Thacker, 1994; Medley et al., 1994; Gempesaw II et al., 1992; Karp et al., 1986; Sadeh et al., 1986; Hanson, et al., 19 85; Bailey & Richardson 1985). Using a stochastic simulation add-in for spreadsheet software such as Simetar (Richardson et al., 2006) in financial spreadsheet applications represents an approach developed to calculate the probability dist ribution of net returns to an ag ricultural investment. This approach can be used as a measure of the risk associated with a crop diversification strategy. The Simetar add-in module provides stochast ic functions not availa ble in the standard Microsoft Excel spreadsheet format. Utilizing this software, any financial spreadsheet model can be operated in a probabilistic mode by assi gning user-defined probability distributions to variables that are perceived to introduce risk into the proposed investment (Richardson, 1976).
15 Examples of risky variables can include prices an d yields of the proposed crops being considered for diversification. Through an iterative Latin Hypercube sampli ng process, random values can be generated from the specified distributions fo r each stochastic variab le and mathematically manipulated throughout an accounting framework of a financial model to forecast enterprise net returns (Hardaker, 2004). By repeating this proc ess multiple times, a probability distribution for the net returns of the proposed i nvestment can be determined. Simple statistics can then be used to calculate the expected value and variance of the probability distribution for the investor to use to evaluate a proposed enterprise or combination of enterprises. It is in this manner that risk is transformed into a single value (Richardson, 1976). Problematic Situation Citrus canker ( Xanthomonas axonopodis pv citri ), a bacterial pathogen, has been a periodic management issue for Florida grapefruit producers since its presumed introduction from Japanese rootstock in 1910 (Sc hubert & Sun, 2003). Symptoms of citrus canker disease include lesions on the fruit, leaves, and stems (Figure 11) that immediately rest rict producer access to the fresh product market and ultimately reduce harv est yields during the life of the tree due to defoliation, premature fruit drop, di e-back of twigs and general de bilitation of the tree (Graham et al., 2004). Although citrus ca nker was officially declared eradi cated in the state of Florida in 1994, a re-introduction of the disease emerged thr ough a residential area of Dade County in 1995 and eventually contaminated commercial groves lo cated in the southwest and southeast areas of the state (Schubert & Sun, 2003). In response to this outbreak, the state quickly implemented a Citrus Canker Eradication Program (CCEP) with remediation protocols that required the destruction and burning of all trees within a 1,900-foot radius of a ny infected tree (Zansler et al., 2005). Cleared grove acreage was then required to be plowed and left fallow for tw o years before the acreage could be replanted in
16 citrus (White et al., 2006). Although the CCEP protocols were amended in January 2006, by this date seven million citrus trees on 80,000 acres had been burned (Connor, 2006). Compounding the financial implications of citr us canker, commercial grapefruit producers in Florida recently experienced unpreceden ted production risks during the 2004 hurricane season. Commercial citrus production in the stat e of Florida is divided into four major production regions: the Central Florida region, the Southwest region, the Southern region, and the Indian River region (Figure 1-2). The Indian River produc tion region, located along the southeast Atlantic coastline, refers to Brevard, Indian River, Martin, Pa lm Beach, and St. Lucie counties (Muraro, 2003) This region is known for producing a high quality grapefruit (Obreza & Collins, 2002) and accounts for 73% of the grapef ruit acreage in the state of Florida (FASS, 2004). During the hurricane season of 2004, f our hurricanes impacted the Indian River production region during August and September cau sing widespread tree damage and a loss of 70 % to 80% of projected grapefruit yields (M uraro & Hebb, 2005). Prior to the 2004 hurricane season, Florida produced over 40% of the world s fresh and processed grapefruit products (Spreen et al., 2006), but after the 2004 hurricane season Florida production was reduced by an estimated 68% (Brooks et al., 2005). In addition to marginalized returns from c itrus production, the recent surge in demand for non-agricultural acreage for development has led many producers to receive increasingly competitive price valuations for their land resour ces in this region. Th e appraised value of grapefruit groves in the Indi an River production region has in creased from $3,929 per acre in 2003 (Reynolds, 2003) to over $15,000 per acre in 2007 (Connelly, 2007). Due to the financial losses resulting from citrus canker disease, CCEP efforts, hurricane events, and the opportunity cost of appreciating land values, producers wi thin the Indian River production region are
17 considering whether or not to either quit farm ing and sell now or continue farming and sell later. While some in this region have chosen the former option, producers that continue to operate in this industry are seeking potential risk management strategies to mitigate the exposure to future financial loss. Alternative crops are being c onsidered as a management strategy by citrus producers to utilize acreage left fallow as a result of the CCEP. Crop diversification is one risk management strategy being considered that may have the po tential to mitigate production risks regarding disease and weather events associated with gr apefruit production. A speculative interest in small-scale shrimp farming as a diversificati on strategy is being cons idered as a separate production enterprise in additi on to current grapefruit producti on. Biological protocols have been developed for the production of the shrimp species, Litopenaeus vannamei in inland, freshwater ponds in Florida. The sub-tropical climate of the Indian River production region supports an extended growing season for L. vannamei with best growth rates be tween 23-30C and a temperature tolerance down to 15C and up to 33C (Wyban & Sweeny, 1991). The extended season represents a potential 10-mont h growing season utilizing a one-month greenhouse headstart facility that can result in a la rge (25-gram) fresh product. Ad ditionally, irrigation groundwater from the Floridan aquifer that is high in dissolv ed minerals, especially ch lorides, has the correct mineral balance to support fres h-water acclimation and culture of this species (Wirth et al., 2004). High consumer demand exists for shrimp and well-established production technologies suggest that this species may be a feasible candidate for crop diversification on the fallow acreage in the Indian River production region of Florida.
18 Researchable Problem Production protocols for farming shrimp in th e United States are va stly different from production protocols employed in other areas of the world where commercial production takes place in outdoor ponds that receive tidal flows of brackish to full-strength sea water (Samocha, 1998). The limited availability of and environm ental concern for coas tal acreage in the US renders the investment in a typical salt-wat er, flow-through, shrimp production facility as economically risky (Van Wyk et al., 1999). As a re sult, this risk has promoted the development of new technologies that utilize highly-mineralized (low-salinity ) well-water accessed via deepwater well systems (Larramore et al., 2001; Samocha et al., 1998, 2002; Van Wyk et al., 1999). This technology allows shrimp production to move inland, away from the high cost and environmentally sensitive coastal corridor. Within the grapefruit producing community of the Indian River production region, there has been considerable interest in shrimp aquaculture as a production enterprise. Permitted wellpoint access to the Floridan aquifer and available inland acreage for pond development encourages the investigation of a shrimp aquacult ure enterprise as a diversified portfolio option. Although grapefruit producers are anxious to ex plore this diversification option, limited economic data exists specific to the financia l feasibility of a small-scale shrimp farming enterprise in this area of south Florida. Despite the continued ex istence of several private shrimp farms within the United States, the profitabili ty and cash flows of a low-salinity shrimp aquaculture investment still remains indete rminate. While production strategies and technologies continue to be made publicly available, limited risk analyses have been published specific to the financial feasibility of low-salini ty pond shrimp operations in this region of South Florida. Analytical tools that exist for the shrimp-farming industry are typically 1) elaborated in discussion during conference proceedings (Brow dy, 2007; Duncan, 2007; Zazueta, 2007; Roy et
19 al., 2005; Green, 2004; McMahon & Baca, 2001; Sa mocha et al., 2001; Van Wyk, 2000), 2) developed as commercially-mar ketable instruments (AQUAFarmerTM, 2004; Hanson & Posadas, 2004; Ernst et al., 2000; Griffin & Tr eece, 1996), 3) or are site and fa cility-specific (Wirth et al., 2004; Van Wyk et al., 2000). A need exists for an analytical risk assessm ent tool regarding the financial parameters of a site-specific shrimp aquaculture investment using low-salinity aquaculture production protocol s for Florida producers. Research Objectives Almost every production decision made by the farmer involves risk (Jolly, 1983); however, risk management can be processed through a systematic series of steps as outlined by Hardaker, (2004) (Figure 1-3). The first four step s of the risk management process are addressed in a three-essay format comprisi ng Chapters 2, 3, and 4. The objective of the analysis in Chapter 2 is to introduce and establish a strategic business c ontext (Step 1 in Figur e 1-3) via macro and micro economic principles of agri business within which to analyze the investment risk associated with a shrimp production enterprise. This includes an overview of the Frozen-shrimp commodity market Implications for a differentiated locally-gr own, fresh-shrimp product utilizing a directmarketing strategy to local restaurants Resource availability for a quaculture industry creation Potential benefits to be gained from cooperative advertising. The objective in Chapter 3 is to identify and structure the enterprise investment problem (Steps 2 and 3) via the construc tion of a stochastic spreadsheet model that captures the building and operating cash flows generated from a shri mp production investment. A database of historical market and feed prices is used to linearly forecast fu ture revenue and cost values for the shrimp production investment. Stochastic s imulation software incorporates a probability
20 distribution for each of the forecasted values to ca pture the variability (risk) of net cash income over a fifteen-year planning horiz on. Specific attention is given to the baseline probabilities of receiving a positive net present valu es (NPV), at various required rates of returns, over the planning horizon and subsequently incorporating various scenario specific probabilities. These scenarios include the addition of a value-added price premium, reduced capital (construction) cost requirements, non-optimal shrimp survival rates, and the impact of hurricane and random kill events on the financial position of the enterprise investment. The objective in the Chapter 4 (Step 4) is to use the deterministic and stochastic capabilities of the model to compare the expect ed returns and variability of returns for the grapefruit, shrimp, and combined grapefruit-shrimp enterprises. This step of the risk assessment process analyzes the risk and retu rns of investing in a shrimp ente rprise as a crop diversification option for Florida grapefruit producers. Steps 5, 6, and 7 are not included in this analysis and are assumed to be addressed by the individual producer who would utilize the findings of this study to consider his/her individual ri sk preferences. Chapter 5 summarizes the results and provides a discussion of the applicability of this model as an extension tool for pr ospective producers at the firm level to use for their own preliminary risk management process. Hypotheses Hypothesis 1: An investment into shrimp aquacultu re production in low-salinity ponds is capital intensive and may not be a profitable investment. Hypothesis 2: Probability of acceptable financial pe rformance (as measured by the net present value of net cash income) is depe ndent upon premium market prices, lowered capital construction costs, highe r survival rates, the discount rate calculation used, and the impact of uncertain variables incl uding weather and random kill events, Hypothesis 3: A shrimp aquaculture investment may not minimize the variance of returns associated with a diversification strategy for grapefruit producers in the Indian River production region
21 Hypothesis 4: By using historical data to estimat e future probabilities of financial performance, investment risk can be quantifie d and analyzed via stochastic simulation to reflect the implications of investment returns.
22 Figure 1-1. Illustration of citr us canker infection. Source: (University of Florida, Interdisciplinary Center fo r Biotechnology Research, 2007) Figure 1-2. Map of the Indian River production region in Flor ida. Source: (FASS, 2007) Indian River production region
23 1. Establish Context 2. Identify Important Risky Decision Problem 3. Structure the Problem 4. Analyze Options and Consequences 5. Evaluate and Decide 6. Implement and Manage 7. Monitor and Review 1. Establish Context 2. Identify Important Risky Decision Problem 3. Structure the Problem 4. Analyze Options and Consequences 5. Evaluate and Decide 6. Implement and Manage 7. Monitor and Review Figure 1-3. Outline of steps in the risk manageme nt process. Source: (Hardaker et al., 2004)
24 CHAPTER 2 SOME ECONOMIC IMPLICATIONS FOR ESTABLISHING A SHRIMP AQUACULTURE INDUSTRY IN FLORIDA Introduction Biological growth parameters and production te chnologies associated with a low-salinity shrimp aquaculture investment have been system atically explored and th ese results have been widely disseminated. However, determining th e financial position of a low-salinity shrimp aquaculture investment is less clear and is hi ghly dependent upon numerous combinations of factor inputs that are site and size specific for each investment decision being considered. Engineering a business strategy for this type of entrepreneurial investment may be facilitated by examining fundamental economic principles associ ated with creating competitive niche markets. Two elements originally associated with creat ing an entrepreneurial business strategy have been defined as the bringing together factor s of production (Say, 1803) and the combining of various input factors to carry out a new industry organizati on (Schumpeter, 1934). These elements of defining a business strategy are appl ied to the entrepreneurial development of a differentiated, locally-grown, fresh-shrimp industr y where the perceived value of a fresh product will be supported by consumer willingness to pay for a value-added product. A potential competitive advantage exists for investors in Flor ida who are able to define a business strategy by utilizing available consumer information and ex isting factors of producti on that include arable land, skilled labor, and state-sponsored advertis ing resources in the creation of a business development strategy. Prospective shrimp cultu rists in Florida should strive to utilize a combination of these resources. The strategic behavior of the firm, in the context of creating a bus iness strategy, clearly defines a firms operating environment and provides an investor with critical information to navigate the competitive business complex associat ed with industry creation (Carlton & Perloff,
25 2005). A business strategy is defined by six elemen ts: 1) the product-market scope in which the business is to compete, 2) the level of investme nt required, 3) the func tional area strategies needed to compete in the selected product market (e.g. product, form, price, distribution, etc.), 4) the strategic assets or competencies that underlie the strategy and provide a sustainable competitive advantage, 5) the allocation of re sources over the business units, and 6) the development of synergistic eff ects and the creation of value by having business units that support and complement each other (Aaker, 2001). The c ontext of each of th ese elements can be discussed relative to developi ng a locally-grown, fresh-shrimp aquaculture business strategy in Florida including Value-added shrimp versus low-cost commodity shrimp Competition and market conduct Differentiated product Direct-marketing to local restaurants Consumer demand for locally-grown, fresh (never-frozen) shrimp Factor resources available for new industry creation Generic and collective ad vertising programs Investors may be able to optimize profits by capitalizing on available consumer research data, restaurants as direct-market outlets, locally -available factors of production and publicly available advertising resources to create a ma rketing channel for local ly-grown, fresh (neverfrozen) shrimp produced in Florida. The discussion in this chapter focuses on the elements conducive to creating a potential competitive mark et advantage within the environment of a differentiation strategy as opposed to relegating market efforts toward a low-cost commodity approach (Porter, 1996). Economic Context The creation of a shrimp production industry in Fl orida is viewed as a risky investment due to undetermined financial parameters and incomple te market information that can be used to
26 generate a successful business strategy. Biol ogical protocols and production technologies for culturing low-salinity shrimp in Florida using the highly-mineralized water of the Floridan aquifer have been elaborated in a collaborativ e grant-funded commercial scale Penaeid shrimp demonstration project involving th e Florida Department of Agricu lture and Consumer Services and the University of Florida/ Indian River Re search and Education Center in Fort Pierce, Florida (Wirth et al., 2004). Market resear ch data has also been collected on consumer perceptions regarding fresh, lo cally-grown, shrimp purchasing behaviors (Wirth & Davis, 2001b). This chapter is organized to present some strategic elements that can be used to create value for a differentiation strategy marketing low-salinity shrimp cultured in Florida. Value-Added Shrimp vs. Low-Cost Commodity Shrimp World fisheries may have reached their carrying capacity between 1990 and 1997 as estimated by a flat, or in some areas declini ng, harvest yield (FAO, 1997). As a result, seafood production has been shifting from the traditional hunting and gathering aspect of wild-caught seafood to a more agrarian method of produc ing seafood products for consumption through aquaculture. The United States domestic shri mp industry supplies only 10% of US shrimp demand through wild-caught harves ts leaving over 90% of shri mp demand to be supplied by imported shrimp (NMFS, 2007). This trend has accel erated over the past decade as U.S. seafood consumption becomes more reliant upon fa rmed, rather than wild stocks. In 1998, 50% of shrimp imported into the U.S. domestic market were farm-raised and the global farmed-shrimp industry exported 305 million pounds of shrimp valued at $1.4 billion to the U.S. (Harvey, 1998). During the first six months of 2006, U.S. consumers demanded 525 million pounds of shrimp valued at $1.6 billion (Harvey, 2006). As a result of increasing consumer demand for shrimp, the number of farm ed-shrimp suppliers and total world supply of shrimp increased and the world price of head less, frozen shrimp decreased from $7.00 per pound
27 in 2000 to $2.60 in 2004 (Comtell, 2006). Howeve r, after January 2004, pr ice variability has narrowed within the wholesale and Hotel Restaura nt Institution (HRI) pr ice spreads, indicative of a maturing industry with production in over fifty countries (Figure 2-1). The HRI prices represent price data for Ecuadorian L. vannamei shrimp sized 31-35 tails per pound in a raw, headless, block-frozen form. This HRI price is an index price that characterizes the average retail price that a hotel, restaurant, or similar institution would pay a s eafood distributor for 3135 tails per pound count shrimp. HRI price data is collected from a cross-section sample of nationwide representative businesses via a w eekly nationwide telephone survey performed by Urner-Barry services. Competition and Market Conduct Market power within an industry is defined by the level of competition among participating businesses. The level of compe tition influences the conduct of firms operating in that industry. The structurec onductperformance (SCP) paradigm as first introduced by Mason (1939, 1949), is an analytical approach that us es inferences from microeconomic theory to determine the behavior of firms within an industrial organizational framework (Carlton & Perloff, 2005). The conduct of an industry is ge nerally classified by its pricing behavior and can range from low-bid price competition to highly m onopolistic pricing behavior which is in turn dependent upon market conditions consumer demand, and barriers to enter and exit the industry (Table 2-1). Structure and conduct together dete rmine the performance (i.e. the efficiency and degree of market integration, market price and marketing margins, accuracy and adequacy of information flows, etc.) of the marketing sy stem as a whole (Van Anrooy et al., 2006). The global shrimp industry has many s uppliers producing a relatively homogenous product, with slight differences in species, size, and form. As a result of a lack of market concentration (power), participants in a compe titive industry operate within a highly competitive
28 framework where the selling price level is bid down until only the most efficient producers can survive by making a normal profit. At this equili brium price, supply is sufficient to fill consumer demand which is addressed through a sophisticated network of domestic wholesalers. In this competitive structure, the producers only option is to become a price-taker for the homogenous good. One problem for small-scale diversified sh rimp farming operations in Florida is that higher unit-costs of production, relative to Asian and Latin Am erican producers who dominate the market, render a commodity-pricing structur e that is unprofitable for Florida producers. An alternative to a competitive price-taking market environment is the perceived-value pricing strategy that is associat ed with monopolistic competition. In this market structure a large number of producers can co-exist, but each sell s a product that is differentiated from the competition (Shaffner et al., 1998) such as dis tinguishing between the frozen commodity-shrimp market structure and a fresh (never-frozen), loca lly-grown shrimp market The benefit to the producer from a differentiated product strategy is the marginal increase in selling price associated with an increase in perceived product value by consumers. Rather than competing for world price in a highly competitive commodity market, focused product-differentiation efforts within a monopolistically-competitive environment could create a marginally profitable market position for small-scale shrimp aqu aculture producers in Florida. Differentiated Product Utility is defined as the satisfaction creat ed by the consumption of goods and services (Drummond & Goodwin, 2004). One approach on th is theory of consumer behavior is that rather than the goods or services being the direct objects of utility (satisfac tion), it is the product attributes and characteristics from which utility is derived (Lancaster, 1966). Consumer preferences for certain product ch aracteristics that will produce gr eater utility are then assumed to create value for the consumer. The primar y objective of a monopolisti cally-competitive firm
29 is to provide greater value to consumers via di fferentiation (Porter, 1996 ). Market research specific to aquacultured shrimp indicates increase d marginal consumer u tility associated with market information regarding environmentally-sus tainable culture practi ces of aquacultured shrimp and corresponding country of origin labe ling (Kinnucan et al., 20 03; Wirth et al., 2007). Shrimp consumption in the U.S. takes place primarily in restaurants (Wirth & Davis, 2001a), making restaurants a prime outlet fo r direct-marketing efforts by producers. Additionally, utilizing a direct -marketing approach for a fresh (never-frozen), non-processed product further reduces the need for sophisticated processing facilities, thereby reducing unit costs of production. However, one requirement is that the perceived marginal value associated with this differentiated product be communicat ed throughout the marketing channel where it must ultimately be demanded by the consumer. This value might be communicated via the Florida Department of Agricultures marketi ng program for Florida seafood (FDACS, 2007) as discussed later in this chapter. Direct Marketing to Local Restaurants A conjoint analysis conducte d by Wirth and Davis (2001a) indi cates that product form is the restaurant buyers primary (60%) source of ut ility, followed by size (22 %), and price (14%). These findings indicate that an unprocessed (whole shrimp), direct-market approach could be a significant disadvantage for Florida farmers if processing activities are not incorporated into the production process. The findings indicate negative utility associated with a whole shrimp form. This could possibly be the result of an unfamiliar product form due to the ubiquitous standardized commodity shrimp inventory whic h is shipped headless in frozen blocks. Alternatively, the negative utility associated w ith whole shrimp could be associated with the unintended consequences of shifting the processi ng function from the produ cer to the restaurant buyer who would then be required to employ margin al labor for processing (i.e. head removal).
30 The latter disutility (due to increased labo r) might be negotiated via marketing efforts emphasizing negligible marginal labor employe d in de-heading shrimp during the currently employed de-shelling/ de-veining activity by restau rant employees. Marketing efforts could also convey additional product value a ssociated with the shrimp head s (for use in stocks, soups, alternative recipes, etc. ) thus allowing the producer to cons ider reversing the perceived negative utility. Consumer Demand for Locally-Grown, Fresh (Never-Frozen) Shrimp Consumer market research from academic institutions in Florida, Kentucky, Indiana, and Tennessee indicate that local seafood consumers ar e willing to pay a price premium for fresh (never-frozen) locally-grown seafood products (Wirth & Da vis, 2001b; WAS, 2007). In 2001, consumer research data was collected within a focus group setting in Florida in which 50% of participants indicated a willingness to pay a price premium of $1 more per pound, over and above the current market price of $7 per pound, for Florida farm-raised shrimp. Additionally, 88% of these participants indicated a willingn ess to pay a price premium of $1 more per pound, over and above a market price, for fresh, never-fr ozen shrimp, while 62% were willing to pay a price premium of $2 per pound for this fresh product (Davis & Wirth, 2001). Communicating the end-users (i.e., consumers ) perceived value for fresh shrimp products to the primary (restaurant) buyers could return a portion of the $1 or $2 local pr ice premium to be received by the producer. The total $1 or $2 price premium for locally-grown, fresh (never-frozen) Florida shrimp at the restaurant level is disaggregated as a portion of restaurant food cost and a portion of restaurant revenue. By using a direct-marke ting sales strategy to restaurants, the Florida producer is positioned to receive a portion of the marketing margin (the restaurants food cost percentage), which is typically received by a seafood wholesaler in the vertical food distribution system.
31 Typically, most full-service restaurants targ et a pre-determined food cost percentage between 28% to 32% of the menu price (Walker, 2002) as calculated in equation 2-1 (Powers, 1988). price Menu Percentage Cost Food Cost Food (2-2) The General Rule of Three is also an adag e that applies to food costs and restaurant profits that is simply stated as, the amount char ged for a food item (menu price) must be at least three times the total cost of the ingredient (Herbert, 1985). Using the Florida focus group consumer willingness to pay a price premium of $1 more per pound for a Florida-raised, fresh (neverfrozen) shrimp product, a restaurant ma nager could charge (and receive) a price premium by offering this differentiated product on the menu. The benefits to the producer are derived through the calculation of the retaurants food cost in equation 2-2; assuming a 33% food cost percentage and a $1 menu-price premium. The food cost is calculated to be $0.33 and assumed to be part of the marketing margin benefits tr ansferred directly to the Florida producer in a direct-market sales effort of fresh, never-f rozen shrimp. Communicating the value of consumers perception of fresh, Florida shrimp via a product differentiati on pricing strategy can effectively place the local farmer outside of th e frozen commodity-shrimp pricing structure. Factor Resources Available for New Industry Creation Effective assembly of factor resources in constructing a shrimp production facility, recruiting trained employees, in itiating production activities, and establishing a successful marketing channel is required prior to prof itably operating a low-salinity shrimp production enterprise. Efficiently managi ng expenses associated with pr imary factors of production is especially relevant for the perc eived high-capital, highrisk investment associated with shrimp aquaculture. Capital costs can easily become prohibitive in acquiri ng scarce land resources,
32 while marginal firm profitability can be easily er oded using unskilled labor to manage a complex water quality environment. Competitive advantages may exist for shrimp aqu aculture producers in Florida, specifically regarding primary factors of production that may be available for these pro ducers. For example, Florida citrus producers have access (owned or leased) to scarce production acreage. Additionally, highly skilled labor and management are available through degree and certification programs jointly offered in the area of St. Lucie County, Florida through the Indian River Community College and Harbor Branch Oceanogra phic Institution. The use of skilled labor has the potential to minimize up-front costs of producti on typically associated with learning a new technology. A familiar maxim heard within the indus try is, You really arent an aquaculturist until youve killed a million animals. Advancing th e learning curve through skilled training in a teaching lab can reduce the probability of expe riencing an expensive catastrophic crop failure arising from technical inexperi ence. Skilled and experienced la bor reduces potential production risks through optimized feed regimes and effi cient labor resources, as well as serving to minimize water quality problems that can lead to disease issues or pond die-offs. Inefficient management can quickly result in sub-optimal financial performance due to reduced survival rates that result in low production yi elds. The availability of highvalue factor resources such as land and a skilled labor force by prospective shr imp growers in Florida has the potential to minimize the probability of a ne gative net present value of net cash income for the shrimp aquaculture investor. Generic and Collective Advertising Programs Generic advertising campaigns are government-sanctioned programs, funded through cooperative producer (supplier) monies and used to promote non-branded, homogeneous products (Ward, 2006). The use of generic promotions is becoming increasingly more important
33 to aquaculture producers as a mark eting tool (Kinnucan et al., 2003) The objective of this type of advertising strategy is to increase overall consumer demand for a product by extolling desirable attributes that are comm on to the product that consumer s desire but may not be readily apparent in the product during the search proce ss. Although a low-salinity shrimp aquaculture production facility in Florida se lling 25-gram (whole-weight) shri mp is limited to one production harvest per year, a generic advertising program could promote locally-grown shrimp as a specialty (rather than ye ar-round) product thus transferring valu e-added benefits associated with specialty products to local consumers. Increa sing the overall consumer demand for locallygrown Florida-farmed shrimp through a gene ric advertising program should additionally facilitate a demand for this product by local Flor ida restaurants that ar e involved in the supply chain and who directly benefit from the increased consumer demand. The Florida Department of Agriculture and Consumer Servic es (FDACS) Florida Agriculture Promotional Campaign (FAPC) is a state-sponsored identification and promotional program. The program is designed to increase sales of all Florida agriculture products by helping consumers easily identify agricultural products grown in Flor ida (FDACS, 2007). The use of this state-sponsored me mbership program by Florida shrimp producers could provide an advertising outlet to in creased consumer awareness for fres h, locally-grown shrimp products by using existing promotional material (Figure 2-2). The FAPC began in 1990 and provides produc er benefits through the use of FAPC marketing logos on producer promotional mate rials. The program generates state-wide advertising exposure from the generic Fresh from Florida advertising lo go via print, billboard, radio, and television outlets. The rational for a generic marketing strate gy specifically designed for Florida farm-raised shrimp may be to high light the local culture, fresh form, and quality
34 attributes of a fresh, whole-shrimp product. The FDACS implied quality standards represent credence attributes of safety and quality char acteristics presented with the approval of a government agency (Wirth et al., 2007) and de sired by consumers (Mabiso et al., 2005). The increase in consumer demand for cred ence attributes (attributes that are not discernable even after consuming the product) reflects the U.S. governments efforts in the 1990s to provide consumers with additional market information rather than to increase process or performance regulations (Caswell & Mojdus zka, 1996). Results from a 2001 University of Florida consumer survey indicated that 62% of respondents agr eed that it is important to know the state or country in which shrimp are harves ted prior to purchase (Wirth & Davis, 2001b). Consumer research data disseminating from the 2004 University of Florida Aquaculture Demonstration Project in Ft. Pi erce, Florida focused on shrimp product attributes, specifically regarding consumer utility associated with purch asing decisions (Wirth et al., 2007). These findings indicate positive shrimp product utility as sociated with country-of-origin labels (COOL) and a reduction of shrimp product utility associated with a lack of COOL or COOL associated with another country. By adopting an informational protocol, Flor ida shrimp producers could benefit from a state-wide generic promotion e ffort specific to product attribut es of Florida-farmed shrimp. Promoting Florida-farmed shrimp as antibioticfree could differentiate Florida-farmed shrimp attributes from the imported commodity-shrim p contamination issues regarding banned antibiotics, such as chloramphenicol. Or, an alternative/ additional strategy could promote production protocols of Florida-farmed shri mp as environmentally -responsible culture technologies that are located inland and away from sensitive coastal areas.
35 Although the worldwide trend for chlorampheni col application is decreasing (Collette, 2006) credence attribute perceptions regarding food safety issues continue to be important to consumers (Erdem & Swait, 1998). Food-safe ty assurances can communicate product differentiation (Allshouse et al., 2003) and as domestic demand for shrimp continues to be supplied by both domestic and intern ational seafood markets, quali ty assurances may lead to increased levels of seafood safety being cons idered a product credence / search attribute. Additionally, public opin ion has been negatively influenced by large-scale coastal-area shrimp farms that have converted substantia l mangrove resource areas into shrimp production ponds. Ponds that may be up to 200 acres in si ze potentially contribut e to salinization of surrounding lands and groundwater sources, as well as to the pollution of coastal waters from pond effluents (Lewis et al., 2002) A survey conducted by Johnst on et al., (2001), suggests that environmentally responsible cult ure practices are important to shrimp consumers who prefer products that do not negatively impact the environm ent. This research indicates that shrimp consumers are willing to pay a price premiu m for environmentally responsible products. Ecolabeling is a method used to signal to cons umers that efforts have taken place to avoid or reduce the environmental consequences of fish or shrimp production (Kinnucan, 2003) and this labeling has important welfare effects for cons umers who are concerned with environmental stewardship related to agricultural products (Wessells et al., 1999). Ecolabeling has been accepted by the General Agreement on Tarrifs and Trade (GATT) as a method of product differentiation (Wessels et al., 2001) which can be relevant for the consumer due to the homogenous nature of fresh and previously froz en shrimp products. If Florida producers can sufficiently promote product differentiation through ecolabeling (i.e ., communicating that Florida shrimp are produced in a non-destructive environmental manner), the enhanced marginal
36 utility associated with Florida shrimp products may result in a higher prices in the marketplace (Kinnucan, 2003). Communicating cr edence attributes to shrimp consumers via a brightlycolored and easily discernable Fre sh from Florida state-wide ma rketing label could facilitate desirable harvest information indicated to be a credence attribute desired by consumers. Summary A business planning strategy for Florida pr oducers considering shrimp aquaculture production that utilizes local factor inputs and a locally-focused marketing channel may increase the probability of financial success for the firm in the context of new industry creation. By creating a sustainable competitive advantage w ithin a monopolistically competitive, nichemarket environment, potential l ong-term marginal benefits may accrue to the investor. This paper addressed the elements of a differentiati on business strategy for Florida-grown shrimp to include direct-marketing efforts to local restau rants, a price premium reflecting the credence attributes of a locally-grow n, fresh, never-frozen whole-sh rimp product, and a generic advertising campaign to promote quali ty and environmental standards. A direct-marketing effort directed toward lo cal Florida restaurants narrows the market development to a population segment where high consumer demand exists for a fresh (neverfrozen), locally-grown product. However, one obstacle to overcome in developing a directmarket approach is convincing th e first buyer (restaurant manager) that the risk/ reward tradeoff will return positive benefits in the absence of hist orical precedents or tangible evidence (Aldrich & Fiol, 1994). Although market re search identifies a latent de mand for the fresh product with consumers, persuading restaurants to subscribe to more than a one-time purchase of fresh product at premium inventory prices may be a lim iting factor as consumers establish purchasing behaviors. Restaurant managers may not be initially convinced to pay a premium price for what could be considered as a homogenous pr oduct touting perceived-value attributes.
37 The marginal utility gained from consuming a perceived superior product is based on an individuals own tastes and pr eferences. A differentiated-product market structure may communicate an increased marginal willingness to pay (a price premium) for the consumers perceived product value (Lancaster, 1966). In a differentiated-produc t market structure, a large number of producers (both global and local) can co-exist, but each group sells a product that is differentiated from their competition (Shaffner et al., 1998). Creating a niche-market industry with a monopolistically-competitive structure ma y be possible for Florida producers who have access to scarce resources such as land, water permits, and skilled labor. Generic advertising in Florida is promot ed through the FDACS Florida Agriculture Promotion campaign for agriculture products. This type of commodity advertisement of agricultural products has histor ically been shown to have significant positive impacts on producer profits (Wolfe, 1944; Nerlove & Waugh, 1961; Comanor & Wilson, 1967; Thompson & Eiler, 1975; Kinnucan & Fear on, 1986; Forker & Ward, 1993; Kaiser & Lui, 1998; Kinnucan & Miao, 1999; Schmit & Kaiser, 2004). Credence attributes, such as those associated with quality and environmental standards can be tr ansferred to the consumer via a promotional advertising strategy (Ward, 2006) to differentiate fresh (never-frozen), Florida-farmed shrimp from frozen, commodity imports. Using generi c advertising as a com ponent of the overall business strategy may facilitate the dissemination of consumer information and may increase the probability of repetitive purchase behavior for bo th consumers and restaurant managers as a fledgling shrimp industry becomes established in Florida.
38 Figure 2-1. Wholesale and HRI shrimp prices 19932006. Source: (Comtell, 2006) Table 2-1. Factors influencing mark et conduct within an industry. Highly competitive Monopolistic competition Monopoly Low price Low price + value-added premium High price Homogenous product Differentiated product Unique product No barriers to entry Some barriers to entry Total barrier Many sellers Many sellers, each slightly different One seller Figure 2-2. Examples of F DACS Florida Agricultural Pr omotional Campaign material. 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00 Jun-94 Oct-95 Mar-97 Jul-98 Dec-99 Apr-01 Sep-02 Jan-04 May-05 Oct-06 Feb-08 Wholesale HRI DatePrice
39 CHAPTER 3 DEVELOPING A STOCHASTIC SPREAD SHEET MODEL TO ANALYZE THE POTENTIAL OF A SHRIMP AQUACULTURE INVESTMENT Introduction The Pacific marine white shrimp, Litopenaeus vannamei, is indigenous throughout the eastern Pacific from Sonora, Mexico to Tumbes in northern Peru (Perez-Farfante, 1997) and has historically been the primary species of farm ed shrimp cultured in the Western Hemisphere (Lewis et al., 2002). It has rece ntly emerged in Taiwan and Chin a as the leading culture species (Wyban, 2002) due to specific disease resistance lower overall feed protein requirements, as well as increased tolerance to low temperatures, wide-ranging salinities, and very high stocking densities (Briggs et al., 2004). The scarcity of available coastal property within the United States for shrimp culture resulted in non-traditional, low-salinity culture protocols being developed for L. vannamei shrimp (Larramore et al., 2001; Samocha et al., 1998, 2002; Van Wyk et al., 1999). Private enterprises located within the states of Alabama, Arizona, Florida, Illinoi s, Indiana, Michigan, Mississippi, South Carolina, and Texas have su cceeded in producing shrimp in low-salinity conditions (Davis et al., 2004). Recently, grap efruit farmers in the Indian River production region of southeast Florida have expressed interest in the feasib ility of this technology to produce farmed shrimp on part of their fallow acreage. Using permitted well-point access to the Floridan aquifer, farmers have access to highly-mineralized irrigation water which is necessary for shrimp osmoregulation. Despite the continued existence of several private shrimp farms within the United States, the profitability and cash flows associated with a low-salinity shrimp aquaculture investment still remain, in general, undetermined. Although produc tion strategies and tec hnologies continue to be made publicly available, limited analyses have been published specific to the financial
40 feasibility of low-salinity pond shrimp operations in south Florida. Analyti cal tools that exist for the shrimp-farming industry are typically 1) el aborated during disc ussions at conference proceedings (Browdy, 2007; Duncan, 2007; Zazu eta, 2007; Roy et al., 2005; Green, 2004; McMahon & Baca, 2001; Samocha et al., 2001 ; Van Wyk, 2000), 2) developed as commercially-marketable software packages (AQUAFarmerTM, 2004; Hanson & Posadas, 2004; Ernst et al., 2000; Griffin & Treec e, 1996), or 3) site and facil ity-specific (Wirth et al., 2004; Van Wyk et al., 2000). A need exists for financ ial data specific to a shrimp aquaculture investment in south Florida to provide grapefruit farmers with adequate information to consider a shrimp culture diversific ation option within a risk management context. Research Objective The objective of this paper is to present the probabilities of economic success of an investment by a Florida grapefru it grower into shrimp pond aqua culture via a series of eight simulation scenarios. Production is assumed to take place within the Indian River production region of Florida, which includes St. Lucie County. Producer interest in the feasibility of a lowsalinity shrimp aquaculture investment is in response to recent pr oduction risks, including hurricane events and citrus canker disease, which have adversely affected grapefruit producers in this area. This study estimates the probabl e costs and returns fo r a hypothetical shrimp production system operating in St. Lucie County, Florida. Economic success for this investment is defined by the positive net present value of net cash income over a 15-year planning horizon at sp ecified discount rates. In addition to the baseline parameters employed, the effect of varying a complement of management variables will be examined relative to the economic performanc e of the investment. These variables will be examined in Scenarios 1 through 7 and include: Low stocking density
41 High stocking density Average survival rate Increasing a local price premium Reducing capital (construction) costs Varying discount rates Reducing the survival rate Following an analysis of these key management va riables, the financial implications associated with including hurricane and random kill events are examined in Scenario 8. The probabilistic determination of economic success for this inve stment, as measured by the positive net present value of net cash income is generated via a stoc hastic simulation spread sheet accounting model. Simulation Model Methodology Stochastic computer simulation models have be en used to analyze capital expenditures and agricultural management scenar ios under conditions of uncerta inty (Richardson et al., 1976, 2000; Gempesaw II et al., 1992; Crawford & Mil ligan, 1984; Featherstone et al., 1990; Griffin & Thacker, 1994; Fumasi 2005). This type of ag ricultural model can be constructed by first programming a financial accounting spreadsheet to reflect deterministic financial accounting equations that calculate investment financia l performance. Capital and production budgets primarily drive the cash flow process ove r each production period throughout the planning horizon. Also programmed into the accounting ma trix are estimates for market prices and production yields that ultimately output a formulated calculation for annual net cash income and the net present value of net cash income for the investment planning horizon. A sensitivity analysis of financial results can be examined through user-defined adjustments of prices or inputs, or changes that might influence pr oduction yields. The deterministic accounting spreadsheet matrix forms the basic framework for the stochastic simulation model. The construction of a stochastic accounting spreadsheet incorporates user-defined probability distributions for each input variable that may be considered to introduce risk into the
42 investment scenario (Figure 3-1) These variables may include random or uncertain input costs, market prices, shrimp survival probabilities, random kill incidents, or hurricane events. Each of the spreadsheet-programming form ulas, typically estimated via least-squares regression, is modified with an additional es timated probability distribution de signation. The values for these random variables are then computed through a repr esentative accounting spreadsheet to result in the financial calculation of key output variables (KOVs) being c onsidered in the investment decision (Richardson, 1976). The KOVs being cons idered in this study are the discounted net present value of net cash income and annual net cash income levels for a fifteen-year planning horizon. Stochastic simulation software uses an itera tive process to draw a random sample from each stochastic variable distribut ion and output a KOV result from suc cessive iterations (or states of nature) to a separate worksheet. Statistical analysis on the iterate d results discerns the probability of positive NPV of net cash income perf ormance at various rates of return to the investor. The stochastic-spreadsh eet model is constructed to be completely dynamic and allows for user-defined changes to be immediately upda ted throughout the model. The dynamic nature of the spreadsheet model promotes its value as a risk management tool for specific farm-level investment analysis. Production System Methodology The framework for modeling a hypothetical low-salinity shrimp production system is based on the University of Florid a/ Indian River Research and Education Center (UF/ IRREC) Shrimp Demonstration Project. In 2003, the Flor ida Department of Agri culture and Consumer Services (FDACS) funded a comm ercial-scale shrimp demonstra tion project (SDP) constructed at the UF/IRREC site in Ft. Pierce, Florida. With the assistance of a project manager, researchers were commissioned to build and ope rate a commercial-scale shrimp demonstration
43 facility utilizing low-salinity groundwater. The findings of this two-year demonstration project chronicled the farm design, cost consider ations, production processes, and marketing environment for the production and harvest of market-size L. vannamei shrimp in four lowsalinity ponds located in South Florida (Wirth et al., 2004) The hypothetical production system design utilizes four 0.29-acre ponds and a head-start greenhouse stru cture, and calculated shrimp production data assumptions outlined in the following sections: Capital Costs Product Size Growth Function Survival Rates Feed Costs Inflation Market Prices Revenue Calculation Hurricane Probabilities Based on the information provided in the Fina l Project Report for the SDP, the revenues and expenses for a hypothetical 5-acre, low-salini ty shrimp aquaculture production system were assembled into a series of fina ncial accounting spreadsheets for th e purpose of estimating future cash flows for this investment. Hypothetical Production System Design To build the stochastic accounting spreadsheet used in this study, a representative farm was modeled after the commercial-scale SDP. The hypothetical production system was originally based on capital and variable cost budge ts published in the SDP Final Report (Wirth et al., 2004). Collaboration with South Carolinas Department of Natural Resources Waddell Mariculture Center (WMC) provide d component cost data and mana gement insight to determine variable input requirements for weekly feed and pond aeration. Zeigle r feed manufacturers (Zeigler, 2006) provided historical feed costs fo r forecasting stochastic feed prices. A shrimp-
44 production consulting firm provided L. vannamei growth measurements to estimate a growth equation for L. vannamei from 1 gram to 25 grams (Manzo, 2006). Stochastic simulation software was used to estimate the probabilities associated with financial performance of the hypothetical producti on system incorporating several relevant economic and management scenarios. These da ta components, organized into a financial spreadsheet framework, complete the stochastic simulation spreadsheet model for a low-salinity shrimp production system. The combination of cost, production, software, and management input components for this research project re sulted in a whole-farm financial model for estimating annual net cash income and the net pr esent value of net cash income over a 15-year planning horizon. Capital Cost The hypothetical five-acre farm site addresse d in this study consis ts of four 0.29-acre grow-out ponds lined in ethylen e propylene diene monomer (EPDM) rubber. The liners prevent water seepage through the sandy-soil environment of southeast Florida. Prior to pond stocking, a 30-foot by 90-foot aluminum framed Quonset-s tyle greenhouse is used for low-salinity acclimation and head-start growth of juvenile sh rimp from a hatchery-s ize of 0.1 gram cultured to 1.0 gram. Four EDPM-lined rectangular raceways (45x1 4x 3) are contained in the greenhouse and schedule 40 poly vinyl chloride plumbing (PVC) pipe s are used to recirculate water through a 10-cubic foot bead -filter (Wirth et al., 2004). S upplementary structures include a climate-controlled feed storage area and a re tention pond used for water overflow specific to draining the production ponds at ha rvest or to receive overflow water in the case of flooding during storm events (Figure 3-2). Due to a relatively high water table in the In dian River production region, additional fill dirt is needed to increase pond elevation for th e prevention of water intrusion beneath the pond
45 linings (Wirth et al., 2004). The fill requireme nt adds additional construction expenses and results in a need for more acreage than would otherwise be needed at sites with a higher elevation above mean sea level. The minimum unit size for this shrimp pond production model is five acres. The construction costs for the greenhouse (Table 3-1) and ponds (Table 3-2) are estimated to be $90,961, and $403,032 respectively. Product Size Assuming a one-crop production period of 35 week s, the production model corresponds to 4 weeks of greenhouse head start and 31 w eeks of pond production for a 25-gram (wholeweight) shrimp. Pond stocking is assumed to occur in April w ith harvest in November. The non-processed, whole-shrimp product is assumed to be direct marketed to local restaurants in Florida. The product form of harvested shrimp in the production model is assumed to be fresh and non-processed (head-on), however, available time-series price data used in the model reflects a commodity-priced, frozen, headless shrimp form Therefore, it was necessary to convert the 25-gram (head-on) shrimp yield produced in the production system to a pricing standard common in the shrimp industry. A 25-gram (head-on) shrimp corresponds to approximately a 31-35 (headless) tail count (tails per pound) shrimp if a 60% product yield is assumed (Pacific Seafood Group, 2002, Briggs et al., 2004). The size-calcula tion function of shrimp tails pe r pound is given in equation 3-1 60 0 ) ( grams weight shrimp Whole y (3-1) where, y is the tail-meat weight (grams ) of the harvested (head-on) shrimp. grams y 15 (3-2) From this equation, a 25-gram (head-on) shrimp harvested from the production system is estimated to yield 15 grams of tail-meat (equation 3-2).
46 Based on the tail-meat yield, the conversion of whole shrimp necessary to yield one pound (454 grams) of tail meat is given in equation 3-3 grams x 454 15 (3-3) where, x is shrimp tails per pound. pound per tails shrimp x 27 30 (3-4) From equation 3-4, the number of shrimp tails per pound is assumed to approximately correspond to an industry standard, 31-35 tails per pound size-class pricing strategy. No processing or marketing costs are directly included in the spr eadsheet model, although ice is indirectly included as a harvesting procedure cost. Growth Function Feeding regimes and their respective costs are a function of estimated pond biomass and survival, which constitute one of the largest expenses for an aquaculture operation (Rosenberry, 2007). The spreadsheet model calculates marg inal growth and mortality rates over the production period to determine production period feed costs and molasse s application costs based on standard feeding protocols for this sp ecies (Wirth, et.al., 2004). Although published data exists for L. vannamei weighing less than or equal to 20 grams whole weight (Trimble, 1980, Wyban et.al., 1987, Hochman et.al., 1990), lim ited published data exists specific to a growth function up to size 25 grams. Typically, L. vannamei is cultured to 20 grams (wholeweight) due to a decreasing marg inal growth rate for this species (Wyban & Sweeney, 1991). This shrimp species approaches broodstock si ze at approximately 35 grams (Palacios, 2000). Market research indicates, however, that premiu m prices may be received by the producer for a direct-marketed 25 gram, whole-weight (31-35 ta ils per pound) product size class (Wirth, 2003). Premium prices may be obtained from a cultu red shrimp product that is differentiated, by
47 product size, from a commodity ( 20 gram whole-weight) size class. The target size for this analysis is a 25-gram (whole-weight) shrimp. Head-start nursery shrimp growth rates ar e estimated linearly from sampled IRREC nursery weight data (Wirth et al., 2004). This growth function used the beginning data point of 0.11 grams and ending data point (pond stocking weight) of 0.72 grams. A simple linear estimation (for Week 1 through Week 4) for a daily nursed-shrimp growth function used in feed calculations is show n in equation 3-5 x w 021 0 11 0 (3-5) where, w is the nursery shrimp weight (grams), x is days, and the marginal growth coefficient is equal to 0.021 grams of nursery weight gain per day. To estimate the production pond growth curve from 1 gram to 25 grams (Week 5 through Week 35), weekly growth data were obtained from a commercial brood stock-shrimp producer employing recirculating technol ogy during a grow-out period of 36 weeks (Manzo, 2006). Recorded weight data was used to estimate a growth function for L. vannamei from 1 gram to 25 grams using a log-reciprocal regression functi on as recommended by Hochman, et. al., (1990) and shown in equation 3-6. ) / 1 (1 0xe w (3-6) where w is the average weight of the animal in grams, is a growth coefficient, and x is the animal age in weeks. From this function, a simple growth equa tion was estimated using SAS software v.8 (equation 3-7). ,) / 1 ( 36 31 16 4 xe wR2 = 0.992 (3-7)
48 The t-statistic of -59.00 for the independent variable (x, age in weeks) allows us to reject the null hypothesis that the sl ope coefficient equals zero. The correlation coefficient (R2) indicates that 99.2% of the vari ance of the weight variable ca n be explained by the regression equation (Ott & Longnecker, 2004). The estimat ed growth for pond stocked shrimp closely resembles the SDP data up to 15 grams and is used to forecast the marginal weight gain of L. vannamei shrimp from 1 gram to 25 gr ams (whole-weight) (Figure 3-3). The marginal weight gain was used to estimate the marginal feed costs for L. vannamei shrimp, on a weekly basis, for 35 weeks. Th e survival rate for calculating the pond feeding regime is pre-estimated by the farm manager at stocking time. Over time, the managers experience with specific pond c onditions should allow for a more accurate estimate of actual survival rates so as to optimi ze feed costs and avoid overfeeding. Survival Rates Survival rates used to calculate yields in th e model are based on yields expected to result from production skill and experience or with some initial consultant advice. A consultant was budgeted in the spreadsheet model for initial construc tion oversight and ma nagement training. The expected yield for the deterministic scenarios is constant at 80% surv ival. Production yield probability distributions for the stochastic scen arios were based on a GRKS distribution using estimates of minimum, mid-point, and maximum expected survival (Ric hardson et al., 2006). This continuous probability distribution is a s ubstitute for a triangular distribution and can be used to simulate an empirical probability distribution for survival that is based on minimal data inputs. Properties of this dist ribution are that 50% of the obser vations are less than the midpoint and 50% of the observations are above the midpoi nt. Additionally, a se cond property of this distribution is that 95% of the observations are between mini mum and maximum, 2.2% of the observations are below the minimum, and 2.2% of the observations are above the maximum.
49 Feed Costs Future feed prices for shrimp nursery feed and growout feed were estimated via leastsquares from historical feed pricing data for years 1997-2006. Feeding protocols within the financial spreadsheet model were based on a proj ected 80% survival rate and a feed conversion of 1.8 kilograms of feed per 1 kilogram of shrimp weight gain (1.8:1). The nursery feed prices represent 40% protein content crumbled feed (PL 40-9 ) blend and the pond growout feed prices represent 35% protein cont ent pelleted feed (Zeigler Brothe rs SI-35) at the 400-ton pricing level (Ziegler, 2006). Shipping cost s for feed transport were estim ated from costs reflected in the IRREC data (Wirth et al., 2004). Inflation Inflation rates for base year (2007) costs in cluding fuels (energy) supplies (PLs), and wage rates for years 2006-2015 are from the Un iversity of Missouri s Food and Agriculture Policy Research Institute published inflation data (FAPRI, 2006). Inflation rate estimates for years 2016 through 2021 (Table 3-3) were estimate d via least squares regression from the 20072015 FAPRI data, (SAS v.8). The inflation rates fo r shrimp nursery feed and growout feed costs were extrapolated from historical feed pric ing data for years 1997-2006. The annual inflation rate of agriculture land values in Florida refl ects a 5% annual rate per year based on the 2002 estimated value of land and buildings per farm between 1997 and 2002 (USDA, 2002). Market Prices Twenty years ago Mexico, Central America, and Northern South America were the predominant suppliers of shrimp to the U.S. Currently, Asia and Indonesia are the predominant domestic shrimp suppliers (Adams et al., 2005). A rapid decline in re tail prices during 19992004 reflects rapid expansion in the eastern hemi sphere (Thailand, Vietnam, Indonesia, India, China, Malaysia, Taiwan, Bangladesh, Sri Lanka and the Philippines) (Figure 3-4) after
50 widespread disease outbreaks re duced long-term production yields in the western hemisphere (Mexico, Belize, Ecuador, Honduras, Brazil, Pa nama, Columbia, and Guatemala) (Rosenberry, 2007). Historical market-price data for a 25-gram whole weight shrimp (corresponding to 31-35 tails per pound (headless) product form) were collected from Urne r-Barry seafood price subscription service for the period January 1999August 2006 (Urner-Barry, 2006). These prices represent price data for Ecuadorian L. vannamei shrimp sized 31-35 tails per pound in a raw, headless, block-frozen form (Figure 3-5). The database prices reflect the so-called Hotel Restaurant Institution price (HRI). This HRI pr ice is an index price that characterizes the average price that a hotel, restaurant, or sim ilar institution would pay a seafood distributor for 31-35 tails per pound count shrimp. HRI price data is collected from a cross section sample of nationwide hotels and restaurant s via a weekly nationwide tele phone survey performed by UrnerBarry services. The forecasted prices used in the financial spreadsheet model were estimated via leastsquares regression on monthly historical prices from January 2004 to August 2006 using SAS v.8 and utilizing the Urner-Barry subscripti on database. The correlation coefficient (R2) indicates 25% of the price data can be explained by the regression equation. This period reflects some stabilization in reported retail shrimp prices from January 2004-August 2007, relative to reported prices from January 1999-December 2003. Th is relative price stab ilization may reflect equilibrium supply and demand levels approachin g a market-clearing worl d price in the global shrimp industry (McConnell & Brue, 1993). Revenue Calculation The growth rate and total biomass for each pond is conducted on a whole shrimp weight basis. However, yields for the shrimp harvest for each pond reflect headless (tail-weight) yields
51 for spreadsheet model revenue calculations. The tail-weight harvest amou nt (in pounds) that is used for the investment revenue calculation reflects a 60% meat yield from the total biomass of shrimp produced and harvested annually (P acific Seafood Group, 2002, Br iggs et al., 2004). Although the model assumes a direct-market, w hole-shrimp sales approach with no producer processing, the use of a headless-pricing strategy for revenue calculation is employed due to the availability of historical head less market-price data for market-price forecasts. The forecasted pricing data reflects historical HRI prices from the Urner-Barry subscription market price data for shrimp sold in a headless, frozen block product form. By usi ng a 60% bodyweight tail yield, the spreadsheet model ca lculates investment revenue for a 31-35 shrimp tail per pound product form using prices repres enting a hotel-restaurant pricing strategy, as shown in equation 3-3 R= 0.6Q HRI Price (3-3) where, R is producer revenue, Qis whole-shrimp biomass harvested, and HRI Price equals retail price received for headless, (31-35 tails per pound) product form. The actual price received by the producer in a direct marketing strategy se lling whole shrimp can be calculated by an algebraic manipulation of equation 3-3 as shown in equation 3-4 P = 0.6Q HRI Price (3-4) where, P is producer price received (for a headon, direct-marketed 31-35 tail count shrimp product). The price received by the producer would be lower than the HRI price due to the total biomass calculated for producer revenue. Therefor e, using an approximate meat yield (60% of biomass) for the revenue calculation in the m odel corresponds to the historical Urner-Barry product form pricing for forecasting future sh rimp prices throughout the investment-planning horizon.
52 Hurricane Probabilities In the stochastic spreadsheet model, the proba bility of a future storm event impacting and causing damage in St. Lucie County was estimated from storm event data available from the National Oceanographic and Atmospheric Admini stration Satellite and Information Service database and Unisys Weather System database (NOAA, 2007; Unisys, 2006). The storm event data collected included storm dates, longitude/ latitude values, and re ported wind speeds per county. Windspeed data was recorded for counties directly within the storm path as well as wind speeds from adjacent counties that were impacted (NOAA, 2006). Damage cost estimates to the shrimp production system were derived from actual events surrounding the 2003-04 hurricane season at the IRREC site in St Lucie County. Structural dama ge is correlated to the SaffirSimpson scale (Table 3-4) which uses wind spee d as the determining factor for estimating the severity of potential property damages (NOAA, 2006). For the purpose of forecasting future hurricane probabilities, wind-e ngineering researchers at the University of Florida cons ider historical hurricane event da ta that captures a period of time long enough to include the cyclical pattern of storm events attributed to the Atlantic Multidecadal Oscillation (AMO). The AMO is a se ries of long-duration ch anges in temperature of the North Atlantic Ocean that may last for 20-40 years at a time (NOAA, 2007). These temperature changes may impact the frequency a nd severity of tropical storms during the annual June November hurricane season, which may influence the calculation of hurricane event probabilities. A fifty-year period extendi ng from 1956 to 2005 produced 25 hurricane landfalls that impacted the State of Florida. Historical wind speed data from the storm track and surrounding areas were recorded on a map of Florida demar cated by county. Individu ally categorized storm
53 windspeed data were input into a spreadsheet and resulted in a count of ten tropical storms, three Category 1 hurricanes, and one Cate gory 2 hurricane for St. Lucie County. The probability of each of these storms imp acting St. Lucie County in the future is calculated by dividing the number of class storms by the total number of storms that have hit Florida over the 50-year period. No hurricane force winds of Ca tegory 3, 4, or 5 impacted St. Lucie County during the prescribed period. Altho ugh the probability exists for a higher intensity storm to impact the model area, these probabilities are not reflecte d in the spreadsheet model. Future probabilities and respective damage estimat es to the farm resulting from a tropical storm, Category 1, or Category 2 hurrica ne are given in Table 3-5. Additional consideration was given specifically to the damage inflicted on a standing shrimp crop from the highest intensity wind event that might occur during the one-year production period. If one or more storms impact ed Florida during a single year within the 1956 2005 period, only the highest intensity storm was c onsidered for that year (regardless if more than one storm event impacted the area). From the aquaculturists perspective, this methodology considers the real impact producers face duri ng the production period regarding damages and lost production time over the course of a singl e hurricane season. Property damage estimates reflect actual damages sustained at the IRREC facility during the 2003-2004 hurricane seasons. Crop loss during this same period was minimal, at less than 1.6%, and is not included as a financial damage estimate. Key Output Variables (KOVs) of Financial Performance The economic model operates as a spreadshee t in Microsoft Excel and utilizes Simetar, an Excel-based add-in develope d at Texas A&M University (Ric hardson et al., 2006). This software operates in a dual capacity to calculate both deterministic vari able output as well as stochastic variable output. While the software is operating in the deterministic mode, all
54 stochastic variables default to expected value. The key output variables (KOVs), (e.g. annual net cash income and net present valu e (NPV) of net cash income for the 15-year planning horizon) are calculated as a measurement of the investme nt financial performance. Conversely, when the software operates in the stochastic mode, all stoc hastic variables utilize a probability distribution that incorporates variability, or risk, into the calculation of KOVs. These variables include market prices, feed costs, hurricane events, random kill events, and survival rates (yields). The KOVs measure the financial performance for multip le scenarios in this study. In this study, 8 scenarios (2 deterministic scenarios and 6 stochas tic scenarios) are analyzed. While all price, cost, and yield variables operate in the stochastic mode for Sc enario 3 through Scenario 8 the influence of hurricane and random kill events are only analyzed in Scenario 8. The KOVs for the deterministic scenarios m easure the annual net cash income and net present value (NPV) of net cas h income over a 15-year plan ning horizon for the shrimp aquaculture investment. The calculation of thes e financial performance measurements are given in equations 3-5 and 3-6 (Higgins, 2004). ) ( VC FC TR NI (3-5) where, NIis net cash income, FCare fixed costs, and VCare variable (operating) costs. n t t t n t t tC r C r C NPV1 0 0) 1 ( ) 1 ( (3-6) where, NPVis net present value of cash flow, Cis cash flow, t is the investment horizon of fifteen years, and r is the discount rate. A baseline disc ount rate of 8% is assumed in the calculations corresponding to a 5% in terest rate that is comparable to a 30-day U.S. Treasury Bill (USDT, 2007) plus a 3% in flation premium assumption.
55 Scenario Overview Eight scenarios were examined in the mode l to determine the impact of managerial changes to the NPV calculation of net cash income. A suite of 2 deterministic and 6 stochastic scenarios examined changes regarding increased stocking density, an incr eased price premium, decreased capital construction costs, and decreased survival rate. Hurricane and random kill events are included in the final scenario analysis. Scenarios 1 and 2Deterministic An alysis of Two Stocking Densities Scenarios 1 and 2 incorporate two differe nt stocking densities assumed during the production period; 80 shrimp per meter squared (m2) and 100 shrimp per m2, respectively. These two deterministic analyses include a discussion of the NPV values and annual net cash income and expenses associated with the shrimp enterpri se using baseline assumpti ons (Table 3-6). All stochastic variables default to the expected value for KOV calc ulations. Hurricane and random kill events are not included in Scenarios 1 or 2. Scenario 3 Probability of Positive NPV at a Higher Stocking Density Stochastic Scenario 3 utilizes the 100 shrimp per m2 assumption used in Scenario 2 to incorporate the variability of st ochastic variables in KOV output. Scenario Three incorporates all stochastic variables (except hurricane and rand om kill events) to calculate the probability of economic success given by positive NPV of net cash income values. The results are presented as a cumulative density function (CDF) indicating the NPV range of values calculated via the iterative stochastic process in th e spreadsheet model. The vertical axis of the CDF represents the NPV corresponding probability of occurrence. In S cenario 3, the discount rate is calculated at 8% and hurricane and random kill events are not included.
56 Scenario 4 Higher Price Premium The analyses for Scenarios 1 through Scenar io 3 assume a price premium of $1 per pound of shrimp based on consumer willingness-to-pay data (corresponding to $0.33 received by the producer) for locally-grown shrimp (Wirth & Davis, 2001b). Stoc hastic Scenario 4 considers consumer willingness-to-pay a higher pric e premium of $2 per pound for a direct-market emphasis on the fresh, never-frozen shrimp produc t attribute for Florid a aquacultured shrimp (Davis & Wirth, 2001). The direct-market assu mption utilizes the food cost principle of the General Rule of Three to calcu late premium benefits received by the producer. This principle is simply stated as, the amount charged for a food item must be at least three times the total cost of the ingredient (Herbert, 1985) Based on this information, benefits to the producer are derived by using a 33% food cost percentage of the $2 consumer price premium received by the restaurant operator. The direct benefit receiv ed by the producer is cal culated as a $0.66 price premium for locally-grown, fresh (n ever-frozen) shrimp. The result s of Scenario 4 are presented as a CDF indicating the NPV range of values calculated via the iterative process in the spreadsheet model. The probability results assume the 100 shrimp per m2 stocking density. In Scenario 4, the discount rate is calculated at 8% and hurricane and random kill events are not included. Scenario 5 Reducing Initial Capital Costs Scenario 5 considers the potential of reducing the capital costs associated with constructing the hypothetical shrimp greenhouse and four-pond production system. The original capital costs were derived from the FDACS-sponsored commerc ial-scale shrimp demons tration project (SDP) at UF/IRREC in Ft. Pierce, Florida. The cap ital costs for the gr eenhouse ($90,961) and fourpond ($403,032) production system totaled $493,993. Assu ming that an agriculture investor has access to idle factor inputs (such as labor and equipment), the access to these resources may lead
57 to cost efficiencies during th e construction process. Consid ering that the $493,993 capital outlay for the SDP in Ft. Pierce was the result of a state-sponsored allocat ive process via contract basis, a privately-sponsored construction effort would like ly incur a capital cost less than this amount. Scenario 5 assumes that the cons truction costs to build the shri mp production system are reduced by 50% to $246,997. The results of Scenario 5 ar e presented as a CDF indicating the NPV range of values calculated via the itera tive process in the spreadsheet model. The probability results assume the 100 shrimp per m2 stocking density and the higher price premium of $0.66 per pound of shrimp. In Scenario 5, the discount rate is calculated at 8% and hurricane and random kill events are not included. Scenario 6 Reducing the Discount Rate The NPV calculation has used a baseline di scount rate of 8% for Scenario 1 through Scenario 5. In reality each invest ors required rate of return (as measured by the discount rate) is independently subjective. Scenario 6 considers a range of discount rates us ed as the investors required rate of return on the sh rimp enterprise investment. The results as are presented as a CDF indicating the NPV range of values calculate d via the iterative process in the spreadsheet model. The probability results assume the 100 shrimp per m2 stocking density, the higher price premium of $0.66 per pound of sh rimp, and the reduced capital cost assumption of $246,997 in this scenario. In Scenario 6, hurricane and random kill events are not included. Scenario 7 Below Average Survival Rates Management experience contributes to th e optimization of healthy water quality parameters (e.g. feeding and aeration). Main taining healthy water quality parameters will improve survival rates of shrimp grown in ponds which in turn will lead to higher production yields at harvest (Griffin & Thacker, 1994). Anecdotal eviden ce indicates that the average expected survival rate at harvest time for shri mp producers is 80% with management experience
58 being a major contributor to production yields at harvest time. Scenario 7 examines the impact of reducing average survival rate s over various discount rates to determine the probability of economic success of the shrimp investment. As outlined previously, survival rates at harvest for the stochastic scenarios are based on a GRKS (minimum-median-maximum) probability di stribution. Scenario 1 through Scenario 6 assumes approximately average survival rates (7080-90). Scenario 7 represents a reduction in the survival rate expectations ( 50-70-90). The results are presen ted as a CDF indicating the NPV range of values calculated via the iterative proc ess in the spreadsheet m odel. The probability results assume the 100 shrimp per m2 stocking density, the higher price premium assumption of $0.66, and the reduced capital cost assumption of $246,997 in this scenario. In Scenario 7, hurricane and random kill events are not included. Scenario 8 Random Kill and Hurricane Events Stochastic Scenario 8 is an extension of S cenario 6 with the addition of hurricane and random kill events included in the stochastic si mulation output. The overall survival at the time of shrimp harvest does not contain any continge ncy for random kill events which can strongly affect firm profitability due to random largescale mortality (Griffin & Thacker, 1994). Random kill events differ from expected mortality althoug h both are influenced by severe water quality or disease issues (Thacker & Griffin, 1994). The random kill event probability in the spreadsheet model is assumed to be 6% during the produc tion period of 1 year, based on estimates by Thacker and Griffin (1994). When a random kill ev ent occurs, it is assumed that 50% -75% of the pond population is lost. The probability of each random kill event is calculated independently, via the formulated equations, fo r each of the four ponds within the 5-acre production unit.
59 Inclement weather events occur in Florida regul arly and with varying levels of severity from tropical storm strength to hurricane status Depending on the severity of the storm event, damages to a shrimp culture facility might range from negligible damage to total structural failure. The probability of tropical storm and hur ricane events are included in the stochastic model, with damages reflecting actual propert y damages sustained by the commercial shrimp demonstration facility in Ft. Pierce, Florid a during the 2004 hurricane season (Table 3-5). Deterministic Analysis Results Baseline Operating Variables The deterministic investment model (5 acres) estimates the expected values of all input variables including market prices for 31-35 count size class shrim p, shrimp growout feed prices, and shrimp yields and baseline operating variables (Table 36). The planning horizon under consideration is fifteen years for a 100% equity-financed farm which assumes that all cash shortfalls are covered through an operating loan at a 9.25% annua l interest rate (Farm Credit, 2007). The two scenarios in the following determ inistic analysis include two different stocking densities while keeping all other baseline parameters constant. D ecreased survival probabilities, random kill events, and the impacts from forecasted hurricane events are not reflected in the deterministic scenarios. The impact of these pa rameters will be discussed in the stochastic portion of the analysis. Scenario 1 Stocking Density 80 Shrimp per m2 In the first scenario, the stocking density is assumed to be 80 shrimp per m2 of pond area. That stocking density is exp ected to yield at total 13,300 pounds (lbs) of shrimp tail meat annually from the four production ponds. Base d on this stocking density the Scenario 1 investment generates a negative net presen t value of -$462,401. The NPV is calculated considering the discounted net cash incomes from Year 1 through Year 15 and the initial capital
60 investment of -$493,993 in constr uction Year 0. The income statements over the 15-year planning horizon (Table 3-7) shows crop va lue of $60,383 and expenses of $68,172, resulting in a farm loss of -$7,789 in Year 1 (construction ta kes place in Year 0). Operating loans are required for all production years. Variability in the annual net cash income values for Year 2 through Year 15 reflects variation in the re placement of required production and harvest equipment and supplies. The estimated useful life and replacement schedule for greenhouse and pond components that are replaced over the invest ment horizon are included in Appendix C. Inflation was built into the model and applie d to feed, labor, fuel, and supplies costs. Scenario 2 Stocking Density 100 Shrimp per m2 Scenario 2 increases the stocking density in the spreadsheet model to 100 shrimp per m2 while holding all of the other baseline variables constant (Table 3-6). This increased stocking density in Scenario 2 results in a negative NP V of net cash income over the 15-year planning horizon of -$343,120. The NPV is calculated consid ering the discounted net cash income flow over Year 1 through Year 15 and the initial capit al investment of -$493,993 in construction Year 0. The annual production from this scenario yi elds 16,667 lbs of shrimp tail meat from four production ponds. The income statement for the pl anning horizon for Year 1 reflects increased revenue (Table 3-8), relative to Sc enario 1 (Table 3-7), as a result of a higher stocking density. Production Year 1 produces a crop valued at $75, 479 and incurs expenses of $76,846 resulting in a negative net cash income in Year 1 of -$1,367. Operating loans are required for Year 1 through Year 8 due to an operating cash shortfa ll at the beginning of the production period in these years. However, expected values as a re sult of the increased stocking density in this scenario return positive values of ne t cash income for Year 2 through Year 15. Interest charges (9.25%) for operating and lo ans in Year 1 through Year 8 increase the total expenses of the enterprise for the first 8 years. In Year 9 through Year 15, the annual
61 beginning cash balance covers all operating costs; therefore, ope rating loans are not required for the remainder of the production horizon. Although the higher stoc king density does not result in a positive NPV, the results of the deterministic an alysis in Scenarios 1 an d 2 provide the investor with preliminary costs and benefit information that may assist in initial consideration of a shrimp aquaculture investment. Scenarios 3 through 8 consider stochastic (ris ky) elements that may impact the shrimp aquaculture investment. In these scenarios, the chance of economic success of the shrimp aquaculture investment is illustrated via proba bilities of positive NPV of net cash income. Scenario results are output as cumulative density functions with the domain of the NPV of net cash income on the horizontal axis and the probability of the NPV is equal to or less than that value on the vertical axis. Stochastic Scenario Framework The stochastic model is developed in a Mi crosoft Excel (Microsoft Corp., Washington, USA) spreadsheet, with the purpose being to forecas t the financial parameters associated with a shrimp aquaculture investment over a 15-year planning horizon. The Excel add-in, Simetar, allows variables within the spreadsheet to be stoc hastic (probabilistic). The stochastic variables include input prices, market prices, survival rate random kill events, and hurricanes events. By creating a historical database of the proposed va riables, least-squares li near regression can be used to forecast the future values. A linear equation using historical variables (independent variables) to forecast future variables (depe ndent variables) may not correspond to an exact linear function of the sample observations. The variability of the predicted estimate of the dependent variable is indicated by a random erro r term in the regression model (Equation 38). x y1 0 (3-8)
62 where, y is the dependent variable, 0 is the intercept, 1 is the slope, x is the independent variable, and is the error term. The error term can be used as the source of future variability for the forecasted values. The error term associ ated with each of the linear estimates has an underlying probability distribution estimated empiri cally from the observed historical values (Richardson, 1976). The combination of the linear estimation equati on and the probability estimation of the error term are what produce the stochastic variables in the spreadsheet model. The Simetar software utilizes a stra tified sampling technique known as a Latinhypercube sampling procedure (Inman, et al., 1981) th at is efficient in se lecting random variables from the probability distribution of a stochas tic variable. During repeated sampling, this procedure avoids clustering and ensures that all areas of the probability distribution are considered (Richardson, 2006). Through an iterati ve procedure, the accounting formulas are calculated by means of random sampling of the probability distributions (Anderson, 1974). Using a random number generator via the Simeta r software, this iterative process (n=500) solves for the key output variable measures of fina ncial performance. A sta tistical analysis of the output results delivers a probability estimate of financial performance of the investment. These results allow an investor to evaluate the risk a ssociated with this par ticular investment. In essence, the risk associated w ith this procedure is reduced to a single variable (Richardson, 1976). In Scenarios 3 through 8, the resulting probabilities repres ent the chance of economic success, as measured by positive values of the discounted net present va lue (NPV) of net cash income. Each of these scenario simulations inco rporates all of the risk parameters for market prices, costs, and yields define d as stochastic in the previous discussion. Only Scenario 8
63 includes the probable impact on economic success of the investment as a result of hurricane and random kill events. The stochastic simulation process assumes the following: 1) Historical price and production variability (ris k) are included in the forecasted values over the planning horizon 2) Impact of random kills, decrease d survival rates, a local pric e premium, and various discount rates can be analyzed with resp ect to their influence over the es timated profitability of this investment. 3) Inclusion of hurricane and random kill event pr obabilities in Scenario 8 adds a realistic dimension to the probabilities of positive NPV estimates at various discount rates. Stochastic Analysis Results Deterministic Scenarios 1 and 2 may be consid ered as an investors first pass at determining the financial viability of a shrimp aquaculture investment under the assumptions defined in the model. However, the determinis tic (expected value) met hod does not take any of the stochastic price risk into consideration relative to the probability of economic success. Random variability of prices and yields based on historic variability are calculated within the spreadsheet equation matrix for each of the 500 st ochastic iterations. In addition to price and yield risk, two other key com ponents of production risk have not been considered in the deterministic scenarios. The addition of the risk associated with reduced shrimp survival rates and the addition of random kill and hurricane event probabilities provide the investor with beneficial information and theref ore contribute to greater insight into the l ong-term profitability, and risk, of a shrimp farm investment in Florida. Scenario 3 Probability of Economic Success at a Higher Stocking Density Scenario 3 results are represented by a Cumu lative Probability Dens ity Function (CDF) of the discounted net present value of cash flows fo r the proposed investment with average (70-8090) expected survival rates and 100 shrimp per m2 stocking density. This scenario illustrates that there is a 0% probability that the investment will yield a positive net present value at these
64 average survival rates at harvest time, holding a ll other management variables constant (Figure 3-6). Additional information presented in the CDF indicates that there is also a 100% probability of a negative NPV of net cash flow s value of -$281,215 as indicated by the vertical height of the CDF function corresponding to this NPV value on the horizontal ax is. The CDF also indicates that there is also a 50% pr obability of a NPV of net cash flows calculation of -$347,000 as indicated by the 50% level at the vertical height of th e CDF function that is corresponding to the -$347,000 NPV value on the horizontal axis Scenario 4 Higher Price Premium In Scenario 4, a $0.66 price prem ium is added to the forecaste d stochastic prices in the spreadsheet model. The stocking density is assumed to be 100 shrimp per m2, with an average survival rate, and assumes original baseline va lues previously discussed (Table 3-6). This scenario indicates that there is a 0% probabi lity that the investment will yield a positive net present value of net cash income (Figure 3-7). Additional information presented in the CDF indicates that there is also a 50% probability of a negative NPV of ne t cash flows value of $292,000 as indicated by the height of the CDF function corresponding to this NPV value on the horizontal axis Scenario 5 Reducing initial capital costs In Scenario 5, it was assumed the possibility that the initial capital costs necessary to construct the shrimp production facility could be reduced by 50% based on the assumption that the investor has access to idle factor inputs and is able to attain co st efficiencies relative to the SDP budgeted costs. In this scen ario, the stocking density is calculated at 100 shrimp per m2 with average survival rates, a $0.66 price premium is included, and baseline assumptions are assumed (Table 3-6). Scenario 5 results indica te that there is a 2% probability that the investment will yield a positive net present value of net cash income (Figure 3-8). Additional
65 information presented in the CDF indicates that there is also a 50% pr obability of a negative NPV of net cash flows value of -$46,000 as i ndicated by the height of the CDF function corresponding to this NPV value on the horizontal axis Scenario 6 Reducing the Discount Rate The NPV calculation has used a baseline disc ount rate of 8% for Scenarios 1 through 5, however, in reality each investo rs required rate of return (as measured by the discount rate) is independently subjective. In this scenario anal ysis, the probability of positive net present value is calculated using lower disc ount rates, ranging from 3% to 7%, required by the investor. Model assumptions include a stocking density of 100 shrimp per m2 with average survival, a $0.66 price premium, baseline variable s (Table 3-6) and 50% reduced capital costs. The specific probability values for each of th e discount rates are indicated by the height of the CDF function corresponding to the NPV value on the horizontal axis. Scenario 6 output reflects a shift toward positiv e values of NPV of net cash income as the discount rate is lowered (Figure 3-9). As expected, with lowe r required rates of return (discount rate) the investors probability of economic succ ess increases. Increasing positive probabilities exist, ranging from a 9% probability at a 7% disc ount rate to a 94% probab ility at a 3% discount rate. Scenario output of NPV of net cash income that is estimated to ha ve a 50% probability of occurring (Table 3-9) ranges from $66,198 to a negative NPV of -$39,502. This quantitative data output indicating a 50% probab ility of occurrence ma y be helpful for risk neutral investors willing to take a 50-50 chance on the shrimp aquaculture investment. Scenario 7 Below Average Survival Rates Scenario 7 represents a reduction in the surv ival rate expectations (50-70-90) that may result from poor water quality conditions due to overfeeding or heterotrophic/ autotrophic
66 imbalances. Assumptions in this scenario includ e: a stocking density calculated at 100 shrimp per m2, capital costs reduced by 50% to $246,997, and a higher price premium of $0.66. Results are calculated using a range of discount rates from 3% to 8%. Reduced survival rates illustrated in the S cenario 7 output reflect a shift away from positive values of NPV of net cash income (Figur e 3-10). Lowered survival rates significantly impact the profitability potential as measured by po sitive values of NPV. There is virtually zero chance of positive NPV of the net cash income if less than average survival rates are encountered over the production planning horizon. A 50% ch ance level for NPV of net cash income ranges from -$375,000 to -$125,000 across all discount rate calculations. Scenario 7 illustrates the considerable impact that survival rates have on the overa ll profitability of the shrimp aquaculture investment. Scenario 8 Random Kill and Hurrica ne Event Probabilities Included Scenario 8 includes risky effects of random kill and hurricane events affecting the financial outcome of the investment. In this scenario, as sumptions include a stocki ng density calculated at 100 shrimp per m2, a $0.66 price premium, average survival rates, original capital costs reduced by 50% to $246,997, and baseline variables (Table 3-6) The NPV of net cash income is given in for a range of discount rates from 3% to 8%. Scenario 8 outputs reflect an improved probability of positive NPV of net cash income (Figure 3-11) relative to Scenario 7 results, which utilized lower average survival (Figure 3-10). However, these results indicate some (small) prob abilities of extremely negative NPV values for the shrimp enterprise at all discount rate levels These losses reflect th e implications of multiple random kill and hurricane events over the 15-year production horizon. The significant financial losses are associated with headstart greenhouse da mages sustained in the hurricane events and the resulting reconstruction cost s for this production component. These extreme negative NPV
67 values range from -$1,136,000 to -$882,000, however the probability of occurrence at these levels is less than 1%. Summary In summary, the various combinations of economic management and production risks including random kill and hurricane events contri bute to the overall probabilities of economic success of the shrimp aquaculture investment. These combinations are summarized within 8 scenario permutations to determ ine the level of impact of exogenous variables on the spreadsheet model results. Deterministic Scenarios 1 and 2 considered stocking densities of 80 shrimp per m2 and 100 shrimp per m2, respectively. In both of these scenar ios the point-estimate NPV of net cash flows were negative at -$462,401 and -$343,120, respectively. The negative NPV indicates that the shrimp aquaculture invest ment would require changes regarding baseline assumptions made within the spreadsheet mode l in order to become profitable. Reducing the capital costs required to build the shrimp production facility could improve the economic success of the enterprise as measured by NPV of net cash income. Stochastic Scenario 3 served as the first stoc hastic analysis of the shrimp investment to determine the probability levels of positive NPV values via a cumulative density function output. The vertical axis of the CDF represents the probabili ty and allows the probability for certain values of NPV to be determine d. 100% of the NPV values fell between $450,000 and -$300,000 at a discount rate of 8%. Stochastic Scenario 4 changed the price pr emium amount received by the shrimp producers from $0.33 to $0.66 based on consumer willi ngness-to-pay a price premium of $2 per pound for fresh, never-frozen shrimp. The pr ice premium translates into a $0.66 revenue margin for the producers based on the Rule of Three that is generally accepted in the restaurant industry regarding food costs. This Rule of Three states that the amount charged for a food item must be at least thr ee times the total cost of the ingredient (Herbert, 1985). Although the pr obability of positive NPV values is 0 for this management strategy, increasing the price premium to $0.66 shif ts the CDF function slightly to the right with NPV values falling between -$350,000 and -$225,000. Stochastic Scenario 5 examined the impact of reducing the capital (construction) costs by 50%. This management strategy assumes that the producer would be able to assemble the factors of production necessary to build the hypothetical pro duction system for less than the cost of construction required for th e FDACS and UF/ IRREC Commercial Scale Shrimp Demonstration Project (Wirth et al., 2004). The CDF results indicate that only a 2% probability of economic success is calcul ated by the spreadsheet model using an 8% discount rate. Additionally, there is a 50% chance that the NPV will be -$46,000.
68 Stochastic Scenario 6 reduces the discount rate calculation for positive NPV probabilities. The baseline discount rate of 8% assumes a 5% interest rate comparable to a 30-day Treasury Bill (USDT, 2007), plus a 3% in flation premium assu mption. Reducing the discount rate calculation effectively correspond s to increased levels of risk acceptance by the producer as the risk premium assumpti on is eroded. Although the probability of negative values of NPV exist for all discount rate calculations, positive probability values exist for discount rates ranging from 3% to 7%. However, a small probability of 5% exists for a negative NPV value of -$65,000 at the 7% discount rate. Stochastic Scenario 7 reduces the average su rvival rate from 80% to 70%. The reduction of survival rates significantly (and negatively) impacts the overall proba bility of economic success, as measured by the NPV of net cash income, for all discount rate calculations ranging from 3% to 8%. Stochastic Scenario 8 utilizes a 100 shrimp per m2 stocking density, a $0.66 price premium, reduced capital costs, and average (8 0%) survival rates. This scenario also includes the impacts of random kill and hurrica ne probabilities. The net change of NPV probability values regarding the average surviv al rates (relative to Scenario Seven) and random kill and hurricane impacts indicates a relative increase in positive NPV probabilities. The initial deterministic scenarios for th e shrimp production system utilizing baseline assumption variables do not indicate positive NPV values (Scenario 1 and 2). However, by changing the baseline assumptions in the model to reflect higher stocking densities, a higher value-added price premium, lower capital costs, a nd average survival rate s, the probability for positive NPV values increases for all discount rates considered from 3% to 8%. The addition of random kill and hurricane events reduces th ese positive NPV probabilities but does not completely eliminate the chance of economic su ccess at every discount rate considered ranging from 3% to 8%.
69 Figure 3-1. Flow chart of model component parts. Figure 3-2. Low-salinity shrimp production system diagram. Retention Pond Greenhouse Shrimp Ponds Deep-water well p oint C C A A P P I I T T A A L L B B U U D D G G E E T T S S Deepwater Well Greenhouse Ponds V V A A R R I I A A B B L L E E C C O O S S T T F F L L O O W W S S Weekly Operating Inputs Replacement Schedules P P R R O O D D U U C C T T I I O O N N B B U U D D G G E E T T S S Feeding Regime Weight Data Electricity Post Larvae S S T T O O C C H H A A S S T T I I C C W W O O R R K K S S H H E E E E T T Market Prices Feed Prices Yields Hurricanes Random Kill M M O O D D E E L L O O U U T T P P U U T T Annual Net cash income Annual Cash Flows NPV Cash Flows S S I I M M U U L L A A T T I I O O N N S S O O F F T T W W A A R R E E Random Variables Accounting Spreadsheet 500 Iterations S S t t a a t t i i s s t t i i c c a a l l C C a a l l c c u u l l a a t t i i o o n n o o f f 5 5 0 0 0 0 K K O O V V s s I I n n p p u u t t s s U U s s e e r r d d e e f f i i n n e e d d P P r r o o b b a a b b i i l l i i t t y y D D i i s s t t r r i i b b u u t t i i o o n n s s
70 Table 3-1. Capital costs for greenhouse. Item Amount ($) Total (%) Nursery Greenhouse Construction: 51,585 56.7 Greenhouse Electrical Installation: 10,982 12.1 Greenhouse Equipment General: 3,432 3.8 Water Testing & Shrimp Sampling Equipment 1,751 1.9 Greenhouse Equipment: Pumps & Motors 7,007 7.7 Nursery Transfer Equipment/ Supplies 1,403 1.5 Greenhouse Construction Labor: 14,800 16.3 Total 90,960 100 Table 3-2. Capital cost s for four 0.29-acre ponds. Item Amount ($) Total (%) Engineering and Surveying: 12,329 3.1 Feed Storage Building 24,000 6.0 Well Construction: 62,920 15.6 Well Water Testing: 2,578 0.6 Emergency Generator: 25,531 6.3 Electrical Installation: 20,959 5.2 Earthmoving: Pond and Roadway Construction 96,899 24.0 Pond Liners and Installation: 49,849 12.4 Pond Electrical Installation: 13,904 3.4 Pond Construction Equipment: 7,623 1.9 Pond Construction Labor: 73,546 18.2 Pond Equipment: Paddlewheels 11,143 2.8 Pond Equipment: Water testing & shrimp sampling 1,751 0.4 Total 403,032 100.0
71 Figure 3-3. L. vannamei shrimp growout weight calculation. Table 3-3. Inflation rate sfor inputs used in the model for years 2006-2021. Year Fuels Supplies Wage rates 2006 0.08 0.05 0.03 2007 -0.02 0.02 0.03 2008 -0.02 0.01 0.03 2009 -0.03 0.01 0.03 2010 -0.04 0.01 0.03 2011 -0.03 0.01 0.03 2012 -0.02 0.01 0.03 2013 0.02 0.02 0.02 2014 0.02 0.01 0.02 2015 0.01 0.02 0.02 2016 0.02 0.005 0.02 2017 0.03 0.011 0.02 2018 0.03 0.011 0.02 2019 0.04 0.011 0.02 2020 0.05 0.011 0.02 2021 0.05 0.011 0.02 R-Square 0.58 0.99 0.74 p-value 0.014 2.57E-09 0.001 Source: (FAPRI, 2006, for years 2007-2015) 0 5 10 15 20 25 30 35 0 10 203040 50 WeeksWeight (Grams)w = e4.16 -31.36(1/x)
72 Figure 3-4. Shrimp imports (all product form s) by hemisphere for years 1989-2006. (Source: NMFS, 2007). Figure 3-5. Retail (HRI) historical shrimp prices January 1999August 2006 (31-35 tails/ lb). (Source: Comtell, 2006). 0200 400 600 800 1,000 1989 1992 1995199820012004 Year West East Million Pounds 0.00 2.00 4.00 6.00 8.00 10.00 Jul-98 Dec-99Apr-01Sep-02Jan-04May-05Oct-06Feb-08 DatePrice ($)
73 Table 3-4. Saffir-Simpson wind dama ge scale. Source: (NOAA, 2006). Category Damage Windspeed (MPH)Description Tropical Storm None or Minimal 39-73No real damage to building structures. Damage primarily to unanchored mobile homes, shrubbery, and trees. Category 1 Minimal 74-95No real dama ge to building structures. Damage primarily to unanchored mobile homes, shrubbery, and trees. Category 2 Moderate 96-110Some roof ing material, door, and window damage of buildings. Considerable damage to shrubbery and trees with some trees blown down. Considerable damage to mobile homes, poorly constructed signs, and piers. Category 3 Extensive 111-130Some structural damage to small residences and utility buildings. Mobile homes and poorly constructed signs are destroyed. Category 4 Extreme 131-155Some complete roof structure failures on small residences. Shrubs, trees, and all signs are blown down. Complete destruction of mobile homes. Extensive damage to doors and windows. Category 5 Catastrophic >155Complete ro of failure on many residences and industrial buildings. Some complete building failures with small utility buildings blown over or away. All shrubs, tr ees, and signs blown down. Complete destruction of mobile homes. Severe and extensive window and door damage. Table 3-5. Storm event probability and da mage description for St. Lucie County. Storm type Probability (%) Damage description Tropical Storm 20 Plastic Cover of Greenhouse Category 1 Hurricane 6 Plastic Cover of Greenhouse Category 2 Hurricane 2 Greenhouse Structure + 2% Mortality
74 Table 3-6. Baseline variables for a five-acre shrimp farm investment. Item Value Unit Stocking Density Scenario One 80 m2 Stocking Density Scenario Two 100 m2 Survival Rate 80 % Feed Conversion Ratio (FCR) 1.8 : 1.0 Pounds of feed : kilograms shrimp harvested Discount Rate for NPV 8 % Retail Price 4.20 $ per lb (tail-weight) Local Markup Price Premium 0.33 $ per lb PL Cost 11 $ per 1,000 Growout Feed Price 0.40 $ per lb of feed Labor Cost 15 $ per hour
75Table 3-7. Scenario 1 Net cash income for a 5-acre shrimp farm investment at 80 shrimp per m2 stocking density. Income Statement Year 0Year 1Year 2Year 3Year 4 Year 5Year 6Year 7 Total Market Receipts 60,38363,40666,42969,453 72,47675,49978,523 Greenhouse (GH) Expenses GH Construction Cost 90,960 Post larvae 5,9215,9956,0326,104 6,1406,2116,317 GH Feed 1,1351,1371,1401,143 1,1461,1481,151 GH General Labor 3,5523,6513,7483,841 3,9334,0224,091 GH Electrical 2,1832,1432,0821,997 1,9321,8981,944 GH Harvest Labor 1,9772,0322,0862,138 2,1892,2382,277 GH Supplies 1,3539401,2531,084 1,2869851,455 Sum GH Construction Costs 90,960 Sum GH Variable 16,12115,89816,34116,307 16,62616,50217,235 (4) 0.29-acre Pond Expenses Pond Construction Cost 403,032 Pond # 1 Feed 4,2164,2734,3314,389 4,4474,5064,564 Pond # 2 Feed 4,2164,2734,3314,389 4,4474,5064,564 Pond # 3 Feed 4,2164,2734,3314,389 4,4474,5064,564 Pond # 4 Feed 4,2164,2734,3314,389 4,4474,5064,564 Aerator Electrical 7,0746,9446,7456,472 6,2596,1506,298 General Electrical 587576559537 519510522 Ponds General Labor 19,35019,35019,35019,350 19,35019,35019,350 Ponds Harvest Labor 2,1602,1602,1602,160 2,1602,1602,160 Pond Supplies 1,6711,6921,7028,445 1,8521,7538,739 Harvest Supplies 70964117468 119281539 Recurring Annual 51525253 535354 Sum Pond Construction Costs 403,032 Sum Pond Variable 48,46647,93048,00955,041 48,10048,28155,918 Carryover Loan Interest 08761,4721,896 2,8972,9822,693 Operating Loan 3,5853,9944,3795,108 5,6035,1705,389 Total Expenses 68,17268,69870,20178,352 73,22672,93581,235 Net Cash Income -7,789-5,292-3,772-8,899 -7502,564-2,712
76Table 3-7. Continued. Income Statement Year 8Year 9Year 10Year 11Year 12 Year 13Year 14Year 15 Total Market Receipts 81,54684,56987,59290,61693,639 96,66299,685102,709 Greenhouse (GH) Expenses GH Construction Cost Post larvae 6,3866,4896,5186,5836,648 6,7126,7756,838 GH Feed 1,1541,1571,1591,1621,165 1,1681,1701,173 GH General Labor 4,1764,2594,3344,4064,475 4,5414,6044,664 GH Electrical 1,9892,0222,0692,1302,206 2,2962,4002,518 GH Harvest Labor 2,3242,3702,4122,4522,490 2,5272,5622,595 GH Supplies 1,0121,3591,1691,3791,054 1,5461,0741,432 Sum GH Construction Costs Sum GH Variable 17,04117,65617,66118,11218,038 18,79018,58519,220 (4) 0.29-acre Pond Expenses Pond Construction Cost Pond # 1 Feed 4,6234,6824,7424,8014,861 4,9214,9815,042 Pond # 2 Feed 4,6234,6824,7424,8014,861 4,9214,9815,042 Pond # 3 Feed 4,6234,6824,7424,8014,861 4,9214,9815,042 Pond # 4 Feed 4,6234,6824,7424,8014,861 4,9214,9815,042 Aerator Electrical 6,4436,5506,7036,9037,148 7,4387,7758,157 General Electrical 534543556573593 617645677 Ponds General Labor 19,35019,35019,35019,35019,350 19,35019,35019,350 Ponds Harvest Labor 2,1602,1602,1602,1602,160 2,1602,1602,160 Pond Supplies 1,8021,8319,0171,8581,876 9,2851,9121,930 Harvest Supplies 6912550035471 397250132 Recurring Annual 5556565757 585859 Sum Pond Construction Costs Sum Pond Variable 48,90549,34357,31050,45950,699 58,98952,07452,633 Carryover Loan Interest 2,9982,1841,0032980 000 Operating Loan 5,3624,8845,3563,9533,287 2,277982973 Total Expenses 74,30674,06781,33072,82272,024 80,05671,64172,826 Net Cash Income 7,24010,5026,26217,79421,615 16,60628,04429,883
77Table 3-8. Scenario 2 Net cash income for a 5-acr e shrimp farm investment at 100 shrimp per m2 stocking density. Income Statement Year 0Year 1Year 2Year 3Year 4 Year 5Year 6Year 7 Total Market Receipts 75,47979,25883,03786,816 90,59594,37498,153 Greenhouse (GH) Expenses Greenhouse Construction Cost 90,960 Postlarvae 8,5128,6188,6718,775 8,8268,9299,081 GH Feed 1,4181,4221,4251,429 1,4321,4361,439 GH General Labor 3,5523,6513,7483,841 3,9334,0224,091 GH Electrical 2,1832,1432,0821,997 1,9321,8981,944 GH Harvest Labor 1,9772,0322,0862,138 2,1892,2382,277 GH Supplies 1,3539401,2531,084 1,2869851,455 Sum GH Construction Costs 90,960 Sum GH Variable 18,99518,80619,26519,264 19,59819,50820,287 (4) 0.29-acre Pond Expenses Pond Construction Cost 403,032 Pond # 1 Feed 5,2705,3425,4155,487 5,5605,6335,706 Pond # 2 Feed 5,2705,3425,4155,487 5,5605,6335,706 Pond # 3 Feed 5,2705,3425,4155,487 5,5605,6335,706 Pond # 4 Feed 5,2705,3425,4155,487 5,5605,6335,706 Aerator Electrical 7,8257,6807,4607,158 6,9236,8026,966 General Electrical 587576559537 519510522 Ponds General Labor 19,35019,35019,35019,350 19,35019,35019,350 Ponds Harvest Labor 2,1602,1602,1602,160 2,1602,1602,160 Pond Supplies 2,0462,0722,0858,832 2,2412,1479,140 Harvest Supplies 70964117468 119281539 Recurring Annual 51525253 535354 Sum Pond Construction Costs 403,032 Sum Pond Variable 53,80853,32253,44360,506 53,60553,83561,555 Carryover Loan Interest 015400 000 Operating Loan 4,0414,0994,0334,207 4,0542,8432,202 Total Expenses 76,84476,38176,74183,977 77,25776,18684,044 Net Cash Income -1,3652,8776,2962,839 13,33818,18814,109
78Table 3-8. Continued. Income Statement Year 8Year 9Year 10Year 11Year 12 Year 13Year 14Year 15 Total Market Receipts 101,932105, 711109,490113,269117,049 120,828124,607128,386 Greenhouse (GH) Expenses Greenhouse Construction Cost Postlarvae 9,1809,3279,3699,4639,556 9,6489,7399,829 GH Feed 1,4421,4461,4491,4531,456 1,4601,4631,466 GH General Labor 4,1764,2594,3344,4064,475 4,5414,6044,664 GH Electrical 1,9892,0222,0692,1302,206 2,2962,4002,518 GH Harvest Labor 2,3242,3702,4122,4522,490 2,5272,5622,595 GH Supplies 1,0121,3591,1691,3791,054 1,5461,0741,432 Sum GH Construction Costs Sum GH Variable 20,12320,78320,80221,28321,237 22,01821,84222,504 (4) 0.29-acre Pond Expenses Pond Construction Cost Pond # 1 Feed 5,7805,8545,9286,0026,077 6,1526,2276,303 Pond # 2 Feed 5,7805,8545,9286,0026,077 6,1526,2276,303 Pond # 3 Feed 5,7805,8545,9286,0026,077 6,1526,2276,303 Pond # 4 Feed 5,7805,8545,9286,0026,077 6,1526,2276,303 Aerator Electrical 7,1277,2457,4147,6357,906 8,2278,6009,023 General Electrical 534543556573593 617645677 Ponds General Labor 19,35019,35019,35019,35019,350 19,35019,35019,350 Ponds Harvest Labor 2,1602,1602,1602,1602,160 2,1602,1602,160 Pond Supplies 2,2072,2429,4302,2752,297 9,7112,3412,363 Harvest Supplies 6912550035471 397250132 Recurring Annual 5556565757 585859 Sum Pond Construction Costs Sum Pond Variable 54,62255,13763,17856,41256,742 65,12858,31258,976 Carryover Loan Interest 00000 000 Operating Loan 1,2480000 000 Total Expenses 75,99375,92083,98077,69577,979 87,14680,15481,480 Net Cash Income 25,93929,79125,51035,57439,070 33,68244,45346,906
79 Figure 3-6. Scenario 3 CDF of net present va lue of net cash income with 100 shrimp per m2 stocking density. 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 (500,000) (400,000) (300,000) (200,000) (100,000) 0 NPV of Net Cash Income $-281,215 $-347,000
80 Figure 3-7. Scenario 4 CDF of net present va lue of net cash income with $0.66 price premium. 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 (400,000) (350,000) (300,000) (250,000) (200,000) (150,000) (100,000) (50,000) 0 N PV of Net Cash Income
81 Figure 3-8. Scenario 5 CDF of net present valu ed of net cash income with 50% reduced capital costs. 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 (120,000) (100,000) (80,000) (60,000) (40,000) (20,000) 0 20,000 40,000 N PV of Net Cash Income
82 Figure 3-9. Scenario 6 CDF of net present value of net cash income at varying discount rates. Table 3-9. Scenario 6 Probabil ity values for calculated discount rates ranging from 3% to 7% and corresponding 50% NPV values. Discount rate Positive NPV probability NPV of net cash income 50% probability (%) (%) ($) 3 9966,198 4 94 52,785 5 7252,785 6 39 2,300 7 9 (39,502) 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 (150,000) (100,000) (50,000) 0 50,000 100,000 150,000 200,000 3% 4% 5% 6% 7%
83 Figure 3-10. Scenario 7 CDF of net present value of net cas h income with reduced shrimp survival at varying discount rates. 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 (450,000) (400,000)(350,000) (300,000) (250,000) (200,000) (150,000) (100,000) (50,000) 0 50,000 3% 4% 5% 6% 7% 8%
84 Figure 3-11. Scenario 8 CDF of net present value of net cash income with random kill and hurricane events at varying discount rates. 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 (1,200,000) (1,000,000) (800,000) (600,000) (400,000) (200,000) 0 200,000 3% 4% 5% 6% 7% 8%
85 CHAPTER 4 PROBABILISTIC FINANCIAL ANALYSIS FOR TWO AGRICULTURAL ENTERPRISES: APPLICATION FOR A GRAPEFRUIT & SHRIMP DIVERSIFICATION STRATEGY Introduction Grapefruit producers in Florida have been expos ed to multiple production risks recently. These producers have been particularly affected by the landfall of f our hurricanes during the 2004 hurricane season that drastically reduced fruit yi elds and facilitated th e spread of citrus canker disease. In addition to the hurrican es immediate impact on 2004-2005 grapefruit yields, the protocols for a state-mandated Citrus Canke r Eradication program (CCEP) resulted in thousands of citrus acres being eradicated in an attempt to c ontrol the disease. Although the CCEP was halted in January 2006 (W hite & van Blokland, 2006), a re duction in tree health and vigor are expected to be long-term impacts resulting from the spre ad of disease. The resulting short and long-term financial uncertainty em anating from the 2004-2005 hurricane season has begged the question of risk-management alternatives for grapefruit producers. A crop diversification strategy has important financial impli cations for producers seeking to mitigate production and price risks associated with a grapefruit-only production strategy. This risk management strategy decreases the producers e xposure to risk if activ ities are selected that tend to have good and poor performance in differe nt years (Olsen, 2004). Diversification seeks to reduce the dispersion of overall returns by combining activ ities that are low or negatively correlated. These correlations may include financia l implications associated with disease or pest incidents, weather events, or input and market prices (Hardaker et al ., 2004). Low-salinity shrimp aquaculture producti on has been discussed by many grapefruit producers as a diversification strategy. However, the lack of financial data re quired for an economic analysis has impeded further consideration of this diversification strategy.
86 The mitigation of production risks gained th rough diversification is dependent upon the degree of production and price co rrelation between the crop ente rprises under consideration; negative correlations can mitigate variability of potential returns to the producer over time (Olsen, 2004). Returns from traditional agricult ural crops typically demonstrate a strong positive correlation due to similar production and marketi ng parameters including weather and rainfall events, as well as input prices, and market pr ices. Shrimp production, as a diversified cropping strategy for grapefruit producers, suggests low or negatively correlated returns due to dissimilarities in production practices, input a nd output price trends, and harvest schedules. Market prices for shrimp and grapefruit exhibit properties of a moderate negative correlation, as indicated by a Pearson-productmoment correlation of -0.47. Problem Setting The producers diversification problem regard ing a specific multi-cr opping strategy takes on many of the analytical aspects traditionally used by financial investors facing a portfolio selection problem (Johnson, 1967). Portfolio selecti on seeks to determine if the variability in expected returns from a grapefru it-only producing enterp rise can be abated via a diversification investment in low-salinity shrimp aquaculture The expected value and variability of the discounted net present value of cash flows a nd annualized net cash in come are determined through a farm-level aquaculture simulation mo del developed in Microsoft Excel. The simulation software, Simetar, can run stochastical ly to pattern the annual income statements, cash flows, and other financial information includ ing discounting the net present value (NPV) of cash flows (Richardson et al., 2004) for the grapef ruit enterprise, the shrimp enterprise, and a combination grapefruit/ shrimp enterprise. The enterprise scenarios under consideration in the model include: a five-acre shri mp-only investment, a fifty-acre grapefruit-only investment, and a fifty-acre combined grapefruit (45-acre) and shrimp (5-acre) investment.
87 Due to the inclusion of stochastic variables defined for this model, additional information exists for the decision maker when risk is consider ed to be part of the decision-making process. This study discusses results from the spreadsheet model as both a portfolio selection problem and as a stochastic simulation problem. The addition al risk information provided within the findings of a stochastic probability solution includes fi nancial impacts associated with forecasted financial consequences from hurricane events. The analysis contains results with respect to three cropping scenarios: a grapefruit-only investme nt, a shrimp-only investment, and a combined grapefruit/ shrimp investment. The portfolio sele ction approach is presented first, followed by the stochastic probability solution. Portfolio Selection Problem Microeconomic theory suggests that a profit-maximizing firm chooses inputs and outputs with the sole consideration of maximizing profit (Nicholson, 2002) In a portfolio analysis context, however, this hypothesis fails to consider diversification benefits that may reduce risk exposure to the investor regard ing income variability over ti me (Markowitz, 1952). Portfolio selection for the agricultural producer makes th e assumption that minimizing the variability (variance) of overall returns of an enterprise mi x, at an acceptable expected return, would be preferred to potentially higher prof its at higher probabilities of fina ncial loss. Ideally, it is purely theoretical that all variance can be eliminated and realistically the investor can gain expected return by taking on variance, or reduce variance by giving up expected retu rn (Markowitz, 1952). The mean-variance (EV), or efficiency rule, is a method of decision analysis that can be used in third-party model development when the produce rs risk preferences are unknown (Hardaker et al., 2004). This rule allows the producer to examine the expected returns and variance associated with an enterprise mix to determine acceptable profits that may accrue with a diversification strategy.
88 Agricultural investors using EV for portfolio selection can utilize the suggested steps (Markowitz, 1952; Sharpe, 1963) to utilize EV for portfolio selection purposes: Step 1) Make a probability estimate of future performances of each investment enterprise, Step 2) Analyze the estimates to determine an efficient set of selected enterprises, and Step 3) Select from that set the choice that best suits each investo rs risk preferences Farm decision makers are generally assumed to be risk averse in this analysis; however, risk preferences and utility theory (Step 3) are not addr essed in this analysis. The portfolio analysis of the investment enterprises unde r consideration in this study illu strates Step 1 and Step 2 in considering the expected performance of a shrimp -only investment, a grap efruit-only investment, and a combined shrimp and grapefruit investme nt. The potential EV results for the cropping combinations may be illustrated geometrically on the expected value and variance axis relative to the efficient frontier (Figure 4-1). The effici ent frontier would be located at a point where expected values are relatively high and the variability of returns are relatively low. Stochastic Simulation Problem The steps used to create the stochastic model used for this analysis are based on the modeling technique developed for using probabilistic cash flows to incorporate risk into financial decisions (Richardson & Mapp, 1976). The first step in quantifying the risky elements associated with an enterprise investment is to identify the critical, stochastic variables that are expected to influence the success or failure of the investment over the planning horiz on. These stochastic variables are defined as uncertain quantities within a definite rang e of values that can be attained within a defined probability distributi on (Uspensky, 1937). A mathemati cal (often, a least-squares estimate) equation is used to generate a de terministic forecast equa tion of the critical variables. In the second step, a probability distribution fo r these stochastic variab les is defined by the sample set of historical trend data. The third step is to link the pr obability distribution to a foreca st of the stochastic variable mathematical (often econometric) equation. Fo r example, the yield for grapefruit can be forecast via a least squares estimate using historical yield data and a probability distribution that is define d by the historical data.
89 The fourth step is to link the defined st ochastic variables to a financial accounting framework for the proposed investment. The Simetar software utilizes a stra tified sampling technique known as a Latinhypercube sampling procedure (Inman et al., 1981) th at is efficient in se lecting random variables from the cumulative density function of the stoc hastic variable distribution. The Latin hypercube sampling procedure segments the probability distribution into n components, with n being the number of iterations used, to ensure that all portions of the distribu tion are sampled. This procedure has advantages over th e Monte Carlo sampling technique that randomly selects values from the probability distribution, thereby ove r-sampling the means, while under-sampling the tails (Richardson, 2006). During repeated sampling in an iterative process, this procedure avoids clustering and ensures that all ar eas of the probability distribu tion are considered (Richardson, 2006). Through this iterative procedure (n = 500), the accounting formulas are calculated by means of random sampling of the probability di stributions (Anderson, 1974) to produce financial outcomes such as net cash income or net present value. The financial outcomes are called Key Output Variables (KOVs) and are the variables be ing considered as a measure of risk in the analysis. The iterations represen t 500 individual states of natu re: each being one solution that generates a value for the key out put variables in the model (Ric hardson, 2006). From the iterated output, simple statistics descri be the probabilities of economic success in terms of NPV of net cash income and annual net cash incomes for each i nvestment scenario under consideration. In essence, the risk associated w ith this procedure is reduced in to a single variable (Richardson, 1976).
90 Model Development Shrimp Production System The framework for modeling a hypothetical low-salinity shrimp production system is based on the University of Florid a/ Indian River Research and Education Center (UF/ IRREC) Shrimp Demonstration Project. In 2003, Florida Department of Agriculture and Consumer Services (FDACS) funded a comm ercial-scale shrimp demonstrati on project constructed at the UF/IRREC site in Ft. Pierce, Florida. With the assistance of a project manager, researchers were commissioned to build and operate a commercialscale shrimp demonstration facility utilizing low-salinity groundwater. The findings of this two-year demonstration project chronicled the farm design, cost considerations, production pr ocesses, and marketing environment for the production and harvest of market-size L. vannamei shrimp in four low-salinity ponds located in South Florida (Wirth et al., 2004). Based on the in formation provided in th e Final Project Report for the SDP, the revenues and expenses fo r a hypothetical 5-acre, low-salinity shrimp aquaculture production system were assembled into a series of financial accounting spreadsheets for the purpose of estimating future cash flows for this investment. For the purposes of this study, the minimum size sh rimp culture facility investment is five acres, which includes four plastic -lined production ponds, a nursery headstart greenhouse, and a retention pond area. A portion of the surrounding ar ea is used as a source of fill material to increase the elevation of the ponds above the local water table. The area used for fill is assumed to be rendered as a wetlands area due to water in trusion from the high water table in the St. Lucie County area. Although one five-acre parcel is denoted as a minimum unit of production, results will be given fo r a per acre basis. The original capita l costs were derived from the FDACS-sponsored commercial-scale shrimp demonstration project (SDP) at UF/IRREC in Ft. Pierce, Florida. The capital costs for
91 the SDP production system totaled $493,993 with specific costs of $90,961 for the headstart greenhouse and $403,032 for the four-pond production system. Assuming that a grapefruit producer considering a shrimp aquaculture investme nt has access to idle factor inputs (such as labor and equipment), the access to capital inputs may lead to cost efficiencies during the construction process. Considering that the $493 ,993 capital outlay for the SDP in Ft. Pierce was the result of a state-sponsored allocative process via contract basis, a privately-sponsored construction effort would likely in cur a capital cost less than this amount. The portfolio analyses in this study assume a 50% reduc tion of the SDP capital costs. The capital costs assumed for the portfolio analyses in this study total $246,997 to build a head start greenhouse and four-pond production system. The specific UF/ IRREC costs ar e detailed in Appendix A and B. No initial cash reserves are assumed for the shrimp produc tion enterprise as operating loans cover cash shortfalls at a 9.25% interest rate. Production acreage is valued at $17,000 (Connelly, 2007) in the spreadsheet model and assumed to be owned by the investor; no land rental rates apply. The expected values of calculated variable costs for the shrimp production enterprise (per pound of tails) over the 15-year planning horizon are given in Table 4-1. Shrimp production is assumed to take place wi thin the Indian River production region of Florida, which is a major grapefruit-producing ar ea. The one-crop produc tion period of thirtyfive weeks in the model corresponds to four week s of greenhouse head start and thirty-one weeks of pond production for a 25-gram whole shrimp. Pond stocking is assumed to occur in April with harvest in November. The product form of harvested shrimp is fresh (head-on), while the size class approximately corresponds to 31-35 ta il count shrimp. The non-processed, wholeshrimp product is assumed to be direct ma rketed to local restaurants in Florida.
92 Grapefruit Production System Commercial grapefruit producti on in Florida ranges from a few small parcels to groves consisting of thousands of acres. The representati ve size of the grapefruit grove used in this analysis is assumed to be 50 acres and is represen tative of a small land-holder proprietorship in Florida (Skvarch, 2007). The production acreage is valued at $17,000 (Connelly, 2007) in the spreadsheet model and assumed to be owned by the investor; no land rental rates apply. The land value of the 50 acre grove is $850,000. No initial cash reserves are assumed for the grapefruit enterprise as operati ng loans cover cash s hortfalls at a 9.25% a nnual interest rate. The elements used to construct the grapefru it portion of the spreadsheet accounting model incorporate budget data published by the University of Floridas Citrus Research and Education Center (UF/ CREC) in Lake Alfr ed, Florida. The grapefruit budge ting cost and return data for the Indian River production region of Florida is developed from survey data collected from custom operators, input suppliers, growers, and sc ientists at the UF/ CRE C and UF Indian River Research and Education Center in Ft. Pierce, FL (UF/ IRREC) (Muraro & Hebb, 2005). The variable cost data included in the spreadsheet model for a 50-acre, custom-managed grapefruit grove includes fertilizer, herb icide, pruning, irrigation and drainage, management, and harvest costs (Table 4-2). Forecast of the Critical Variables In the financial spreadsheet m odel, risk can be quantified by examining critical variables that significantly influence the variability of cash flows for each enterpri se under consideration. Exogenous variables that are considered to be stochastic in the model include shrimp and grapefruit yield, shrimp and grapefruit market prices, shrimp feed prices, shrimp random kill events, and hurricane events. Grapefruit price an d yield estimates were forecasted by economic researchers with the Florida Department of Citrus (FDOC, 2007). A $0.66 price premium is
93 added to the forecasted shrimp prices calculated in the spreadsheet model. This price premium is based on consumer willingness-to-pay up to an additional $2 per pound for fresh, never-frozen shrimp (Davis & Wirth, 2001). A direct-marketing strategy to Florida restaurants utilizes the food cost principle of the General Rule of Thre e to calculate premium benefits received by the producer. This principle is simply stated as, t he amount charged for a food item must be at least three times the total cost of th e ingredient (Herbert, 1985). Base d on this information, benefits to the producer are derived by using a 33% food co st percentage (of the $2 additional consumer price charged by the restaurant operator) that computes a $0. 66 price premium for the fresh, never-frozen shrimp aquacultured shrimp product. Average shrimp yields were estimated by a shrimp consulting service (Manzo, 2006). Random kill events were forecasted from estimations within aquaculture stochastic modeling literature (Griffin & Thacker, 1994) Hurricane event estimates were forecasted via a historical archive of storm wind data (NOAA, 2006). The futu re shrimp market prices and shrimp feed prices were extrapolated via least-squares estim ates using historical tim e series data, assuming that future variability is similar to the hist orical variability as re vealed by Equation 4-1 1 0 x y (4-1) where, y is an estimate of the dependent variable, 0 is the intercept estimate, 1is an estimate of the slope coefficient, x is the independent variable, and is the prediction error estimate of unexplained variability of the critical variable. This classic least-s quares method chooses 0 and 1 to minimize the prediction error given in Equation 4-2 (Ott & Longnecker, 2004): 2 1 0 2)] ( [ ) ( i i i ix y y y (4-2)
94 where, iy is the observed variable and iy is the predicted vari able (Figure 4-2). Forecast of Stochastic Distributions When operating in the deterministic mode, the software calculates the least-squares estimates that minimize the error terms of all fo rmulated equations in the spreadsheet accounting model. The deterministic mode generates the best linear prediction (e xpected value) of all formulated variables for an NPV output calculati on. However, when the software mode is changed from deterministic to stochastic mode, the error term and its assigned probability distribution are then calculated as random variable s (Table 4-3). The variability (error term) of the historical data is then used as an additiona l component to the least-squares equation formulas. The stochastic forecasting procedure using the Sime tar software is two-fold. First, future grapefruit and shrimp prices and grapefruit yiel ds are estimated via least-squares regression. Second, the probability distributions are then de termined via an empiri cal distribution function from the historical data. For example, the GR KS probability distribution used to model shrimp yields, utilizes three user-defined data points (minimum: 0.70, mode: 0.80, and maximum: 0.90) to independently calculate the harvest yield for each of the four production ponds (Richardson et al., 2006). The three data points us ed as yield estimates reflect anecdotal yield results provided by a shrimp consulting service (Manzo, 2006). Th e probability of a random kill event occurring in a single pond is assumed to be 6% of the standing crop per pond (Griffin & Thacker, 1994). When a pond experiences a random kill, the GRKS distribution for the mortality ranges from 50% -75% as defined by the data points (min imum: 0.50, mode: 0.625, and maximum: 0.75) (Manzo, 2006). The hurricane probabilities we re calculated from an archived hurricane windspeed database with results indicating a 6% probability of tropical storm or category 1 hurricane impacting the St. Luci e County area of Florida and a 2% probability of a category 2
95 hurricane impact based on the Saffir-Simpson Hurricane Scale (NOAA, 2006). The probability of hurricane occurrence is calculated in the stocha stic mode via a discrete empirical probability. Expected Value Calculation The expected value and variance of the Key Output Variables (KOVs) (e.g. net present value of net cash income and annual net cash in come) can be plotted on the EV axis as an illustration of the investment risk associated with an investment in each enterprise strategy under consideration. The calculation of each KOV is illustrated in equations 4-3 and 4-4 (Higgins, 2004) n t n t t tC r C r C NPV01 0 2 2) 1 ( ) 1 ( (4-3) where, NPVis net present value of cash flows C is cash flows, t is time, n is the investment horizon of 15 years, and r is the ba seline discount rate of 5%. Depreciation is not included in the net cash income calculation. Net cash income re presents cash available to farm operators to meet expenses and debt payments (USDA, 2003). An initial capital (construction) investment is assumed to occur in t = 0 for the shrimp investment. This capital investment is calculated as a negative value at time zero. 15 1) (tVC FC TR NI (4-4) where, NIis net cash income, FCare fixed cash costs, and VCare variable costs (cash operating costs). When the software operates in the determin istic mode, the calcul ated value of each stochastic variable is the central expected valu e that is defined by the probability distribution assigned to that variable (Richard son, 2006). These expected values are calculated to be either 1) the mean for empirical distri butions (grapefruit pr ices, grapefruit yields, shrimp prices, and
96 shrimp feed prices), 2) the mode for GRKS distributions (shrimp yields and random kill mortality), or 3) the weighted probability of occurrence for the discrete empirical estimation (hurricane events). The expected value for the stochastic variables, as explained by Uspensky (1937) in Equation 4-5, assumes that a st ochastic variable possesses n values (nx x x2 1,) and (np p p2 1,) denotes the respective probabilities (ip) of x By definition, the mathematical expectation of ) (x E is n nx p x p x p x E 2 2 1 1) ( (4-5) The variability of the NPV of net cash in come and annual net cash incomes for each enterprise is measured via the coefficient of va riation (C.V.) to account for the scale discrepancy between the representative grapefruit and shrimp investments. Using the C.V. as a standardized measure of variability gives results as a per centage of the mean of NPV values for each enterprise (Finney, 1980) as illustrated in Equation 4-6 100 .y V C (4-6) where, is the standard deviation and y is the mean of the NPV calculation. The C.V. is calculated via the scenario simulation of each ente rprise; however, in the EV portfolio analysis the random kill and hurricane probabilities are not included in the calculations. The results of the NPV calculations from the spreadsheet model are illustrated graphically in a portfolio analysis context. The returns for each enterprise investment are presented as an expected value calculation and the variability of each enterprise is measured by the coefficient of variation. The results of the NPV calculations in the portfolio analysis are presented on a per acre basis for the three investment scenarios. Shrimp, grapefruit, and combined shrimp and
97 grapefruit enterprises are illust rated graphically according to the traditional portfolio analysis concept. Stochastic Probability Calculations In the stochastic setting, the simulation re sults of the KOV calculations are presented as probabilities of occurrence. The introduction of uncertainty transforms net returns from a single value to a probability distribu tion related to stochastic va riables affecting the proposed investment (Richardson & Mapp, 1976). Simulate d results are presented as a cumulative density function of the net present value of net cash income for each investment strategy. Results Per acre NPV of net cash income for the three investment scenarios indicate that the mean calculation for the grapefruit investment ($17, 864) is greater than the mean for shrimp investment ($3,505) (Figure 4-3). Additionally, the coefficient of variation for the shrimp investment (151) is greater than the coefficient of variation for the grapefruit investment (20). The diversified (combined) enterprise investme nt return NPV calculati on ($16,491) is less than the grapefruit-only investment. Portfolio selection for the agricultural producer makes the assumption that minimizing overall return variability via a diversified crop mix, at an acceptable return, would be preferred to poten tially higher profits at higher probabilities of financial loss. However, no benefits of diversification appear to accrue to the investor from a combined shrimpgrapefruit enterprise due to the lower returns an d higher variability associated with the shrimp investment. Model output is presented as a cumulative pr obability density function (CDF) for the NPV net cash income for each of the investment scen arios (Figure 4-4). The CDF for the shrimp investment reveals a 72% probabil ity that the NPV will be positive. The CDF for the grapefruit and combined grapefruit/ shrimp enterprise investments both produce 100% probabilities of
98 positive NPV values for the 15-year planning horizon. Additionally, the relatively large slope of the shrimp enterprise CDF in comparison to the smaller slope for the grapefruit and combined enterprise CDF is indicative of a wider range of potential NPV values on the horizontal axis at every probability on the vertical axis. The larger ratio corresponds to a higher variability (C.V.) in NPV calculations that was illustrated in the portfolio analysis (Figure 4-3). Summary This paper discusses two different approaches to analyzing the dive rsification decision facing Florida grapefruit producer s regarding the investment deci sion of shrimp aquaculture. This study utilizes the capabilities of stochastic simulation softwa re in an accounting spreadsheet framework. The stochastic mode considers the probability of financial performance, as measured by a CDF of the NPV of net cash in come, for a shrimp enterprise, a grapefruit enterprise, and a combined shrimp and grapefruit enterprise. The portfolio approach strategy examines th e relative trade-off between the expected returns per acre, as measured by the NPV net cash income, and variabil ity of returns, as measured by the coefficient of variation, for each en terprise mix. The calculated discount rate is 5% which is a rate that is comparable to the risk-free return on U.S. Treasury Bills without risk or inflation premium. At hi gher discount rates the shrimp a quaculture investment is not considered to be economically successful as m easured by probability of positive NPV of net cash income. The per acre expected value of NPV of net cash income for the shrimp, grapefruit, and combined shrimp and grapefruit enterprises are $3,505, $17, 864, and $16,491, respectively. The expected return on the grapefruit investment reflects FDOC grapefruit price forecasts which considered yields reductions with in this industry as a result of the 2004 hurricane season. The forecasts indicate higher prices fo r grapefruit (relative to previous years) due to supply shocks
99 within the marketplace. Based on the expected va lue of NPV for each investment, the grapefruit scenario returns the highest per acre disc ounted dollar value for the investor. Additionally, benefits may accrue to the inve stor in a diversifi cation strategy if the expected return variability from a crop invest ment can be minimized with the addition of a second crop. However, the variability of e xpected returns (measure by the coefficient of variation) of the shrimp investment at 151 is high er than the grapefruit variability of 20 and a combined shrimp/ grapefruit cropping strategy in creases the overall vari ability to 22 for the diversified crop mix. The results of the side-b y-side returns for each enterprise investment indicate that the investment in a shrimp investment would be a counterproductive investment strategy for Florida grapefruit farmers, base d on model assumptions. The results of the diversification strategy using a por tfolio approach indicates that shrimp aquaculture returns a lower expected value of net cash returns and a higher variability of those returns compared to the grapefruit investment. Therefore, the portfolio analyses indicate no positive benefits associated with a shrimp aquaculture investment by Flor ida grapefruit producers using the assumptions outlined in this study.
100 Figure 4-1. Geometric interpretation of cr opping combinations for portfolio analysis. Table 4-1. Calculated shrimp variable cost s (V.C.) per pound (tail-we ight) at two stocking densities, 80 shrimp/ m2 and 100 shrimp/ m2. YEAR V.C. at 80 shrimp/ m2 V.C. at 100 shrimp/ m2 Year 1 4.73 4.28 Year 2 4.81 4.35 Year 3 4.83 4.37 Year 4 5.64 5.02 Year 5 5.64 5.03 Year 6 5.00 4.52 Year 7 5.49 4.92 Year 8 5.25 4.73 Year 9 5.14 4.66 Year 10 6.57 5.80 Year 11 5.14 4.67 Year 12 5.58 5.02 Year 13 5.83 5.24 Year 14 5.33 4.84 Year 15 7.42 6.52 Efficient Combinations E= (Expected value) V= Variance Attainable Combinations
101 Table 4-2. Grapefruit variable costs (VC) per acre, assuming 95 trees planted per acre. Year V.C. per acre Year 1 1,386 Year 2 1,385 Year 3 1,376 Year 4 1,367 Year 5 1,358 Year 6 1,363 Year 7 1,389 Year 8 1,410 Year 9 1,434 Year 10 1,452 Year 11 1,481 Year 12 1,513 Year 13 1,551 Year 14 1,595 Year 15 1,645 Figure 4-2. Prediction error illustration. Table 4-3. Type of probability distributi ons defined for each stochastic variable. Stochastic variable Type of probability distribution Grapefruit Prices Empirical Grapefruit Yields Empirical Shrimp Prices Empirical Shrimp Yields GRKS Feed Prices Empirical Hurricane Events Discrete Empirical Random Kill GRKS 1 0 y i iy y X Y
102 Figure 4-3. Portfolio frontier for NPV of net cash income for three investment scenarios. Figure 4-4. CDF of net present value of cash flows (NPV) for each enterprise. 0 50 100 150 200 0 5,000 10,00015,00020,000 N PV per AcreC.V. Shrimp Grapefrui t Combined (3,505, 151) 16,491, 22 17,864, 20 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 (200,000) 0 200,000 400,000 600,000 800,000 1,000,000 1,200,000 1,400,000 1,600,000 Shrimp Grapefruit Combined 28% (-) Vertical Axis at NPV=0 72% (+) (16491, 22) (17864, 20)
103 CHAPTER 5 SUMMARY, CONCLUSIONS, AND IMPLICATIONS Disease outbreak and hurricane events intr oduced unprecedented production risks recently for grapefruit producers in the Indian River produ ction region of Florida. Low-salinity shrimp aquaculture, as a crop diversification and risk management strategy, has been considered as a potential investment for grapefru it producers in the Indian Rive r production region of Florida. Although generalized production tec hnologies have been well-documented for the culture of Litopenaeus vannamei shrimp in low-salinity conditions, the financial and risk implications of this investment can be difficult to generalize and therefore have not been widely disseminated. This lack of conclusive financial analyses a ssociated with low-salinity shrimp aquaculture production are due to the many perm utations regarding site-specifi c facility construction costs (shrimp greenhouse and production ponds) and assu med management decisions (shrimp stocking density and assumed survival for feed calculation s) that can be made during the production and marketing process. Additionally, uncertain (stochastic) variables exist that, when considered, contribute to the probability of fi nancial success resulting from an investment in a shrimp culture enterprise. Consideration of these variables ma y include grapefruit yiel ds and prices, shrimp yields and prices, shrimp feeds, survival rates, random kill and hurricane events. In three distinct studies, the objective of this thesis focuses on: 1) Defining the business strategy environment that may lead to competitive advantages via a direct-marketing effort to local restaurants 2) Defining parameter size and scope for a repr esentative low-salinity shrimp production enterprise and the creation of a computer-based stochastic spreadsheet production model to analyze the crop diversification decision 3) Analyzing the risk management decision re garding crop diversific ation from grapefruit monocropping strategy to a combined gr apefruit and shrimp enterprise
104 Together, these three obj ectives are used to address the hypotheses initially posed in this thesis regarding an investment by grapefruit producers into low-salinity shrimp aquaculture in the Indian River production region of Fl orida. Restated, the hypotheses are: An investment into shrimp aquaculture pr oduction in low-salinity ponds is relatively capital intensive and may not be a profitable investment Probability of acceptable financ ial performance (as measured by the net present value of net cash income) is dependent upon premium market prices, lowered capital construction costs, higher survival rates, the discount rate calculation used, and the impact of uncertain variables including weather and random kill events A shrimp aquaculture investment may not mini mize the variance of returns associated with a diversification strate gy for grapefruit produ cers in the Indian Ri ver production region By using historical data to estimate futu re probabilities of fi nancial performance, investment risk can be quantified and analyzed via stochastic simulation to reflect their implications on investment returns. Conclusions Economic theory suggests that within a monopolistically competitive environment, producers who sell a slightly diffe rentiated product can receive ma rginal benefits if consumers perceive value in the differentiated product. In a global shrimp-c ommodity market, the producers are price takers and suppl y and demand equilibrium dictates market prices. Therefore, agricultural producers who are able to communicate additional value attributes to their customers may be able to receive a market price higher than world price. Consumer research data indica tes that consumers in Florid a are willing to pay a price premium up to $1.00 per pound for locally-grown shrimp (Wirth and Davis, 2001b) and up to $2.00 more per pound for fresh, never-frozen shri mp (Davis and Wirth, 2001). A directmarketing approach to local Florida restaurants may be a feasible strategy for Florida shrimp producers who are able to communicate the qua lity benefits perceived by consumers to restaurant managers who are their primary buyers in the direct-marketing strategy. Use of a
105 collaborative state and producer-sponsored marketi ng campaign, such as the Florida Agricultural Marketing Campaign may be cost effective in delivering consumer education information regarding the availability of fres h, locally-grown shrimp at local Florida restaurants. Based on typical restaurant food-cost calculations, the pr oducer may expect to negotiate an approximate price premium of $0.33 to $0.66 per pound (tail-meat ) of shrimp sold. This price premium is added to the commodity shrimp price in the calc ulation of shrimp invest ment revenues generated by the hypothetical shrimp production system. A stochastic spreadsheet accounting model wa s developed based on the production results of the hypothetical shrimp producti on system. Estimated capital outlay for construction costs to build a low-salinity shri mp aquaculture production facility was $493,993. Cash shortfalls were assumed to be covered through operating loans. Processing costs were not directly included in the cost estimates although the cost of ice was included in the model as a harvesting cost to account for the whole-shrimp, direct-mar keting strategy assumed in the model. The NPV of net cash income for the fifteen -year planning horiz on, considering two stocking densities in the deterministic analys is, were -$462,401 and -$343,120 for the stocking densities of 80 shrimp per m2 and 100 shrimp per m2, respectively. The discount rate used in the deterministic NPV calculation was 8%. Changi ng the management variables to positively influence the profitability of th e investment was the focus of the various management scenarios within this study. These variables included a higher price premium received by the producer, lower capital costs, and varying le vels of the discount rate calc ulation. All scenarios were analyzed at the higher stocki ng density of 100 shrimp per m2. Random kill and hurricane impacts were examined after the various management scenarios were analyzed.
106 The analysis of capital costs and returns indicates that the shrimp aquaculture investment is relatively capital intensive. Additionally, this in vestment was only profitable with a combination of a local price premium, lower capital costs, hi gher survival rates, a nd at discount rates lower than 8%. This may not be an attractive inve stment for agricultural pr oducers as the risk-free investment alternative return for U.S. Treasury Bills is 5.25% (USDT, 2007). However, non-risk adverse investors with available acreage may prefer to invest in low-salinity shrimp aquaculture if a solid marketing strategy and long-term be nefits are perceived from this investment. The diversification decision can be approach ed via traditional portfolio analysis and stochastic analysis. The portfolio analysis incl udes the all of the risk elements provided by the stochastic variables and provides a measure of the expected valu e-variance tradeoff that may be realized with a diversification investment from grapefruit to shrimp production. However, the expected value of returns for the shrimp investme nt is lower than the ex pected value of returns for the grapefruit investment. Additionally, the variance associated with the shrimp returns is higher than the variance associated with the grap efruit returns. Based on the portfolio theory, no beneficial tradeoff effects between expected va lue and variability (ove r the long-term, fifteenyear planning horizon) appear to reduce the risk to the investor via crop diversification. The stochastic analysis indicates that the gr apefruit investment returns a 100% probability of positive NPV considering all of the stochastic variables. The shrimp investment enterprise returns a 72% probability of a pos itive NPV of net cash income. A risk averse investor may not wish to consider an investment in shrimp aqu aculture with a 28% probability of negative NPV. Crop diversification into shrimp aquaculture does not minimize the va riance of returns associated with a diversification strategy for grapefruit producers in the Indian River production region of Florida.
107 Implications Production technologies for cultu ring the marine shrimp, Litopenaeus vannamei, in lowsalinity water continue to evolve in Florida. The development of this financial accounting spreadsheet utilized producti on parameters and operating variables assumed to affect the investment decision regarding a lo w-salinity shrimp aquaculture investment in Florida. It appears that the shrimp aquaculture investment decision is capital intensive and less profitable than an investment in grapefru it production in the Indian Rive r production region of Florida. However, technological changes may increase the probability of economic success of this investment. A risk-preferring invest or may perceive potential benefits from investing in this type of production enterprise, especially if factors of production (land, skilled labor, equipment) are available to reduce capital costs. Recent hurricane events in Florid a serve as a reminder that th e production risks associated with hurricane events are valid a nd that crop and facility destruct ion is of prudent consideration with regard to Florida investors. The financial implications that include potential losses due to hurricane or random kill events indicate greater ov erall loss potential for the shrimp aquaculture investment (relative to the grap efruit investment) as measured by NPV of net cash income. The shrimp investment scenario output indicates a lo w percentage of high-value loss associated with rebuilding the greenhouse structure due to hurricane impacts. The capital re-inves tment required for continued operations may not be a feasible strategy for some investors. Technological advancements in structural design may reduce th e risk of greenhouse failure and increase the overall probability of economic success of this type of shri mp production system investment. This study outlined a direct-marketing business strategy for Florida shrimp farmers to consider for the purpose of establishing a premiu m pricing strategy for a fresh, never-frozen and locally-grown product. Establis hing an industry based on a pr oducts credence attributes may
108 prove challenging. However, factor inputs are av ailable in the St. Luci e County area specific to highly-mineralized fresh water and skilled labor av ailability. The Fresh from Florida state marketing campaign may also lend positive economic benefits for the establishment of this industry. Anecdotal evidence within the aqu aculture industry indicate s that establishing a marketing channel for aquacultured product, prio r to pond stocking, is essential to minimizing potential investment loss at harv est time due heavy bio-loads asso ciated with shrimp stocked at 100 per m2 and the impact on water chemistry (e.g., oxyge n, ammonia, and nitrite). The delay in harvest scheduling while investig ating potential markets increases the investors risk of a catastrophic pond die-off and subseque nt loss of profit. Due care s hould be taken to investigate the long-term marketing potential for aquacultured shrimp prior to the initial capital outlay for pond system construction. Finally, the scenario output indicates that low survival rates significantly impact the shrimp investment returns. The learning curve for aquacultured products is less forgiving than traditional agrarian crops due to the complexity of changing water chemistry at high stocking densities required to return accepta ble profits. Consultant advice, especially at the onset of this type of investment, may significan tly reduce the investors potentia l for crop loss associated with poor water quality conditions and non-optimized va riable costs associated with feed and labor inputs. Experienced consultant advice may also provide the investor with timely insights regarding alternative marketing and distribution channels that may become necessary at some point in the future for expans ion or to efficiently move pr oduct out of the production system should primary market channels become unstable. Remember that the adag e is that, you arent a shrimp aquaculturist until youve killed a million shrimp. An experienced consultant has already paid many dues rega rding this perspective of the learning curve.
109 Future Research Needs According to the model, the probabilistic forecasts for a low-salinity shrimp aquaculture investment in Florida require a considerable degree of risk ac ceptance by investors. However, the development of a stochastic spreadsheet mode l has the potential to provide interdisciplinary researchers with a tool aimed at optimizi ng biological and economi c performance based on quantifiable input parameters. Th e spreadsheet model may be used to analyze cost efficiencies associated with construction costs as these cost s seem to greatly influence the probability of economic success of the investment. For example, economies of size and scale may be investigated via computer simulation prior to a capital intensive investment. Increasing the pond size incrementally from the 0.29 acre size up to 5 acres or increasing the number of ponds per production unit may reduce the capital costs per unit of output. Additionally, this research model may provide a template for other species currently being considered as candidates for aquaculture investment. Potentia l dual-cropping investments (i.e., fish and shrimp) may make the capital investment more profitabl e. Utilizing the ponds in the c ool weather with a cool-water species may improve NPV values for the aquacult ure investment. Finally, the development and dissemination of the model components and findi ngs may contribute to the development of different design and input parameters to measur e the financial performance for risk-accepting investors who continue to search for informa tion regarding the devel opment of a low-salinity shrimp production industry in Florida. These desi gn parameters could include structural changes to the greenhouse design to withstand hurricane impacts and minimize rebuilding costs. Using computer modeling and stochastic analysis as a farm or extensi on tool to quantify investment risks associated with aquaculture investments ma y promote cost-efficient i nvestment alternatives for Florida agricultural producers.
110 APPENDIX A HYPOTHETICAL SHRIMP PRODUCTION SYSTEM CONSTRUCTION COSTS Table A-1. Greenhouse Capital Cost Budget. Greenhouse budget item Item cost Total Construction and equipment costs $90,961 Earthmoving 2,000 PVC materials 664 Greenhouse permit fees 93 Quonset greenhouse package and assembly 17,786 3"PVC, screws, fittings 719 50 treated 2'x12'x12' 749 PVC pipe and fittings 1,966 Fittings, valves, bits, PVC glue 553 TEK screws (400) #14x3" 26 TEK screws, bolts, nuts, washers 70 Floor, set sump, drain pipe/fixtures 22,624 Lumber, treated 1,440 TEK screw (600) #14x2.5" 48 Sacrete, bolts, turnbuckles, washers 45 Lumber treated & Portland cement 102 Lumber, treated 37 Reimburse Grainger strapping kit 63 Fittings, screws, screw eyes 189 Nylon screening standpipes 108 Banding & banding buckles 139 Reimburse PVC materials 12 Omni threaded ball valve 50 Pulleys, ratchets for transfer system 22 White rock gravel 15 PVC and tap screws 34 PVC and mesh for media boxes 202 EPDM liners 1,829 Nursery Greenhouse Construction: $51,585 Greenhouse Electrical Installation: $10,982 Proline bacteria fresh and salt 74 Rustol, saw blades, chalk, knife 38 Channel lock plier and 4" vise 61 Band tool 84 Washers, tap screws, nylon ties 17 Kaldness bio-filter media 999
111 Table A-1. Continued. Greenhouse budget item Item cost Total O2 regulator 96 Rubber Maid garden cabinet 189 Reel mount, rubber plug & cord 77 Hose & reel, electric cord, nozzles 112 PVC elbow and pipe 35 Sacrete (168) and lumber 630 Sacrete, screws, PVC slips/adapters 725 4' fiberglass stepladder 118 Big wheel 50-gal trash can 152 11' black nylon ties 25 Greenhouse Equipment General: $3,432 DOmeter, pH, YSI Spectrophotometer and reagents 1,320 Ohaus scale, refractometer 261 Ohaus Scout scale 170 Water Testing and Shrimp Sampling Equipment $1,751 Sweetwater and centrifugal pumps 2,082 Blowers (2) and float switches (2) 1,290 Bead filter, controllers, relays 3,426 Flotec 1.5hp pump 209 Greenhouse Equipment: Pumps and Motors $7,007 Juvenile transport tank and cage fittings 322 O-ring for transfer tank 3 PVC for transfer tank 15 PVC cement for transfer assembly 16 PVC Suction hose 6" x 100' 786 PVC, couplings, ties for transfer 261 Nursery Transfer Equipment/ Supplies $1,403 Project Mgt (2 months for G.H. construction) 10,000 Skilled Labor 4,800 Greenhouse Construction Labor: $14,800
112 Table A-2. Four 0.29-Acre Ponds Capital Cost Budget Pond budget item Item cost Total Total site and pond construction costs $403,032 Surveying services 1,809 Easement engineering 1,500 Well, electric, generator engineering 5,490 Miscellaneous fees 180 Engineering & drafting as-built plans 3,350 Engineering and Surveying: 12,329 Feed Storage Building 24,000 Deep well 57,943 Deep well pump 4,977 Well Construction: 62,920 Deep well 800' 617 Deep well 800' 51 Deep well 940' 765 Deep well 1070 TDS 130 Bioassays on deep well 250 Deep well 1070 765 Well Water Testing: 2,578 Emergency generator, delivered 20,576 Generator slab and extension, wall 4,955 Emergency Generator 25,531 Electrical Installation 20,959 Total Site Prep Fixed Costs $148,317
113 Table B-1. Continued. Pond budget item Item cost Total Earthmoving contract 48,032 Site clearing & road const 4,423 15'x30' alum culvert 249 Coquina FL rock-563.71 yds@18 10,147 Coquina rock delivery; 52.5hrs@$50 2,625 Seeding pond levees 1,549 Borrow pit expand 3,999 Contract change order 1,961 Sod pond levees 4,550 Sod retention pond levee 1,664 12'x40' culvert/flap gate 1,600 Culverts 4,879 Donated heavy equipment services 10,000 Gate materials 686 Tools and hardware 535 Earthmoving: Pond and Roadway Construction 96,899 Transportation of HDPE 2,500 30-mil HDPE 15,000 Install 88,626 ft2 liner 27,244 Portable toilets 105 Anchor trenching-ponds 1&2 5,000 Pond Liners and Installation 49,849 Pond Electrical Installation 13,904 Rope and pipe for dividers 2,824 Lumber for dividers 286 PVC couplings, valves, netting 2,202 Dock materials-ponds and retention 800 Lumber, hanger nails, bolts, washers 126 Gate valves 113 Rope, screws, washers, nylon ties 159 Gate and ball valves 800 8" PVC and clamps for standpipes 136 Liquid nails for standpipe 4 PVC, nuts, washers for standpipes 12 PVC and cleaner-pond water valves 35 Rope 0.5" white nylon (600') 126 Pond Construction Equipment: 7,623
114 Table B-1. Continued. Pond budget item Item cost Total Project Mgmt 10 months (c onstruction and training) 50,000 Install concrete pads in ponds 726 water lines-well to 1&2 987 Dock labor at WCC 1,375 Install docks-ponds and retention 2,200 Install posts and dividers, ponds 1&2 2,658 Construction manager prof it-well, generator, elec. 6,000 Skilled labor 9,600 Pond Construction Labor: 73,546 12 paddlewheel aerators 10,632 Paddlewheel spare motor 276 Paddlewheel spare gearbox 235 Pond Equipment: Paddlewheels 11,143 DO and pH meter, YSI spect rophotometer and reagents 1,320 Ohaus scale, refractometer 261 Ohaus Scout scale 170 Pond Equipment: Water testing & shrimp sampling 1,751 Total Pond Construction and Equipment Costs $254,715
115 APPENDIX B HYPOTHETICAL SHRIMP PRODUCTION SYSTEM REPLACEMENT SCHEDULE Table B-1. Greenhouse Compone nt Estimated Useful Life Greenhouse component Useful life Nursery Greenhouse Plastic Covering 4 years Kaldness Bio-filter media 20 years O2 regulator 15 years Rubber Maid Garden Cabinet 10 years Reel mount, rubber plug & cord 5 years Hose & reel, electric cord, nozzles 5 years 4' fiberglass stepladder 5 years Big wheel 50-gal trash can, garbage can, and straps 3 years DOmeter, pH, YSI Spec & reagents 4 years Ohaus scale, refractometer 4 years Ohaus Scout scale 2 years Sweetwater and cent. pumps 5 years Blowers (2) & float switches (2) 5 years Bead filter, controllers, relays 20 years Flotec 1.5hp pump 3 years Juv. transport tank and cage fittings 5 years O-ring for transfer tank Annually PVC cement for transfer tank and assembly 5 years PVC Suction hose 6" x 100' 15 years PVC, couplings, ties for transfer 5 years Trash can, WD-40, caulk, glue gun 2 years Rope for shadecloth 2 years garden hoses (2) & nozzle 2 years Push brooms (2) 2 years Hose adaptors 2 years Handles (2) push brooms 2 years 10' 2x4s, coater broom 2 years Instant Ocean Salt Annually Chlorine Annually Proline Bacteria fresh H2O (4 x gal) Each stocking Aqua ammonia and acid Each stocking Fittings and batteries 6 months Roof coater brooms (2) 2 years Paper towels 2 months Dustpan, brush, screwdriver set 3 years
116 Table B-2. Pond Component Estimated Useful Life Pond component Useful life Rope & pipe for dividers 5 years Lumber for dividers 5 years PVC couplings, valves, netting 5 years Dock materials ponds and retention 3 years Lumber, hanger nails, bolts, washers 3 years gate valves 5 years Rope, screws, washers, nylon ties 5 years gate and ball valves 10 years 8" pvc & clamps for standpipes 4 years Liquid Nails for standpipe 4 years PVC, nuts, washers for standpipes 4 years PVC & Cleaner pond water valves 5 years Rope 0.5" white nylon (600') 3 years 12 Paddlewheel Aerators 15 Years DOmeter, pH, YSI Spec & reagents 4 years Ohaus scale, refractometer 4 years Ohaus Scout scale 2 years Nitrite, pH, Potassium starter kit Annually Nitrite reagent replacement kit Annually Molasses Annually Paddlewheel Spare Motor 3 years Paddlewheel Spare Gearbox 3 years Pillow block 3 years Paddlewheel oil (1/2 Gallon or 1.8 Liters) Aerator on Pails and handles 2 years Two 2-bu. Baskets 2 years Tap screws, washers, nuts, hose clamps 3 years Big wheel 50-gal trash cans 3 years Door pulls, wooden stakes Annually hose clamps & PVC male adapters 3 years PVC Ls, hose clamps, tap screws 3 years adapter/shrimp harvest 5 years well screen 3 years tracking trailers (2) 3 years Aquaculture Certification: Annually
117 LIST OF REFERENCES Aaker, D.A. (2001) Developing business strate gies (6th Edition). John Wiley & Sons, New York, New York. Adams, C., Kiethly, W.J. & Versaggi, S. (2005) The shrimp import controversy. In: International Agricultural Trade Disputes, Case Studies in North America, (eds. A. Schmitz, C.B. Moss, T.G. Schmitz & W.W. Koo), PP 225-243. University of Calgary Press, Alberta, Canada. Aldrich, H.E. & Fiol, C.M. (1994) Fools rush in? The institutional context of industry creation. Academy of Management Review, 19(4), 645-670. Allshouse, J., Buzby, J.C., Harvey, D. & Zorn, D. (2003) International trade and seafood safety. In: International trade and food safety: Economic theory and case studies (ed J. Buzby), Electronic report from the Economic Research Service, United States Department of Agriculture, Agricultural Econom ic Report No. 828, November 2003. AQUAFarmerTM V2 (2004) Fisheries Victoria, Departme nt of Primary Industries, Victoria, Australia. Anderson, J.R. (1974) Simulati on: Methodology and applicati on in agricultural economics. Review of Marketing and Agricultural Economics, 42(1), 3-55. Bailey, D.V. & Richardson, J.W. (1985) Analysis of selected marketing strategies: A whole farm simulation approach. American Journal of Agriculture Economics, 67, 813-820. Briggs,M., Funge-Smith, S., Subasinghe, R.& Phill ips, M.. (2004). Introductions andmovement of Penaeus vannamei and Penaeus stylirostr is in Asia and the Pacific. Foodand Agricultural Organization of the United Nati ons Regional Office for Asia and thePacific. RAP Publication 2004/10, Bangkok, Thailand. Browdy, C.L., Atwood, H., McAbee, B., Leffler, J. Stokes, A.D., Samocha, T., Losordo, T., Hanson, T.R., & Posadas, B.C. (2007) Resear ch and development of sustainable superintensive shrimp production systems: perspec tives on commercialization potential for US growers, Proceedings of the World Aquacultu re Society, Aquaculture 2007: Science for Sustainable Aquaculture, p. 117, February 26March 2, 2007, San Antonio, TX. Caswell J.A. & Mojduszka, E.M. (1996) Using info rmational labeling to in fluence the market for quality in food products. American Journal of Agricultural Economics, 78(5), 1248-1253. Carlton, D. & Perloff, J. (2005) Modern industrial organization (4th Edition). Pearson Prentice Hall, Boston, Massachusetts. Clemson University (2005) S outh Carolina Shrimp Fishery Assistance, Fourth Semiannual Progress Report, Retrieved April 9, 2007 (http://www.clemson.edu/scshrimp/pdf/progrep4all.pdf ).
118 Collette, R.L. (2006) Country Reports, United States, Washington, D.C. FDA Antibiotics, In: The Shrimp News (ed B.Rosenberry), Retrieved, February 27, 2007, (http://www.shrimpnews.com/FreeNewsB ackIssues/FreeNews%20Feb%203,%202006.ht ml). Comanor, W.S. & Wilson, T.A. (1967) Advertis ing market structure and performance. The Review of Economic s and Statistics, 49(4), 423-440. Comtell (2006) Frozen seafood prices per pound. Ur ner-Barry Publications, Inc. Internet price subscription service, Toms River, New Jersey. Conner, C. (2006). Letter to Charles H. Br onson, Commissioner of Agriculture, Florida. Department of Agriculture and Consumer Services, Tallahas see, FL, from the United States Department of Ag riculture, Office of the Secretary, Washington, D.C. (January 10) Retrieved, February 7, 2007 (http://www.doacs.state.fl.us/pdf/Canker%20%20Bronson.pdf ). Cramer, G.L., Jensen, C.W., & Southgate, Jr., D.D. (1997) Agricultural Economics and Agribusiness (7th Edition). John Wiley & Sons, Inc., New York, New York. Crawford, E.W. & Milligan, R.A. (1982) A multiyear, stochastic farm simulation model for northern Nigeria: An expe rimental design approach. American Journal of Agricultural Economics, 64(4),728-737. Davis, D., Samocha, T., & Boyd, C. (2004). A cclimating Pacific White Shrimp, Litopenaeus Vannamei, to inland, low-salinity waters, So uthern Regional Agriculture Center, SRAC Publication No. 2601, Stoneville, Mississippi. Davis, K.J. & Wirth, F.F. (2001) Market analysis. In: Identifying and assessing potential direct markets for farm-raised shrimp grown on small farms. pp. 47 98 Florida Department of Agriculture and Consumer Serv ices, Division of Marketing and Development, Tallahassee, Florida. Drummond, H.E. & Goodwin, J.W. (2004) Agricultural Economics (2nd Edition). Prentice Hall, Upper Saddle River, New Jersey. Duncan, P.L. Off season inland culture of Paci fic white shrimp Litopenaeus Vannamei in low salinity recirculating aquaculture systems, Pro ceedings of the World Aquaculture Society, Aquaculture 2007: Science of Sustainable A quaculture, p. 263, February 26-March 2, 2007, San Antonio, TX. Erdem, T. & Swait, J. (1998) Br and equity as a signaling phenomenon, Journal of Consumer Psychology, 7(2), 131-157. Ernst, D.H., Bolte, J.P. & Nath, S.S. (2000) AquaFarm: simulation and decision support for aquaculture facility design and management planning, Aquacultural Engineering, 23, 121179.
119 Farm Credit North Florida, Alachua Florida, February 15, 2007. Featherstone, A.M., Preckel, P.V. & Baker, T.G. (1990) Modeling farm financial decisions in a dynamic and stochastic environment. Agricultural Financial Review, 50, 80-99. Finney, D.J. (1980) Statistics for biologists. Chapman and Hall Ltd., New York, New York.Florida Department of Agriculture and Consumer Serv ices (FDACS) (2007) Division of Marketing and Development, Fresh From Florida Agricultu ral Marketing Campaign.Retrieved April 16, 2007 (http://www.florida-agricult ure.com/marketing/fapc.htm ). FDACS (2005) Aquaculture Best Management Practices Rule Prepared for the Florida Department of Agriculture and Consumer Se rvices, Division of Aquaculture, Charles H. Bronson, Commissioner. Retrieved February 7, 2007 (http://www.floridaaquaculture.com ). FDACS (1992) Aquaculture species research series shrimp. Flor ida Department of Agriculture & Consumer Services, Bob Cr awford, Commissioner, Division of Marketing, Tallahassee, Florida. Florida Department of Citrus (FDOC) (2007). Mark Brown, Senior research economist. Food and Agriculture Organization of the Unite d Nations (FAO) (1997) Yearbook of fishery statistics. Volume 81, Rome, Italy. Food and Agricultural Policy Research Institu te (FAPRI) (2006) FAPRI 2006 U.S. baseline briefing book. Food and Agricultur al Policy Research Institute,. FAPRI-UMC Technical Data Report No. 01-06, University of Missouri at Columbia, Columbia, Missouri. Forker, O.D. & Ward, R.W. (1993) Commodity advertising: The economics and measurement of generic programs. Lexington Books, New York, New York. Fumasi, R. J. (2005) Estimating the impacts of differing pric e-risk management strategies on the net income of Salinas Valley lettuce produ cers: A stochastic simulation approach. Masters Thesis, California Polytechnic State University, San Luis Obispo, California. Gaskill, L.R., Van Auken, H.E. & Manning, R.A. (1993) A factor analytic study of the perceived causes of small business failure. Journal of Small Business Management. 31(4), 18-31. Gempesaw II, C.M., Bacon, J.R. & Wirth, F.F. (1992) Economies of pond size for hybrid stripe bass growout. Journal of the World Aquaculture Society 23(1), 38-48. Global Aquaculture Alliance (GAA) (2007) Functions. Retrieved, March 7, 2007 (http://www.gaalliance.org/miss.html ). Graham, J.H., Gottwald, T.R., Cubero, J. & Achor, D.S. (2004) Pat hogen profileXanthomonas axonopodis pv. citri : Factors affecting succe ssful eradication of citrus canker. Molecular Plant Pathology. 5(1), 1-15.
120 Green,B.W.(2004) Shrimp culture in low-salinity water in Arkansas, Proceedings of the World Aquaculture Society, Aquaculture 2004: An Ecologically Sustainable and Profitable Venture, p. 187, March 15, 2004, Honolulu, Hawaii. Griffin, W., Richardson, J., Ampudia-Peters, L. & Lawrence, A.L. (1987) Economic evaluation of stocking strategies using large nursery ponds in shrimp culture.Presented at the World Aquaculture Society 18th Meeting, January, 19-23, 1987, Guayaquil, Ecuador. Griffin, W.L. & Thacker, S.G. (1994) Combined indoor/ outdoor red drum aquaculture: A stochastic sensitivity analysis. Journal of Applied Aquaculture 4(3), 1-22. Griffin, W.L. & Treece, G. (1996) A Guide to th e Financial Analysis of Shrimp Farming, Texas A&M University, Sea Grant College Program, College Station, Texas. Grinyer, J.R. (1985) Inflation and capital budgeting decisions: A comment. Journal of Business Finance & Accounting, 1(1), 149-155. Hanson, J.S. & Posadas, B.C. (2004) Bio-Econom ic Recirculating Shrimp Analysis Software. Mississippi State University, De partment of Agricultural Economics & University of Southern MississippiGulf Coast Research Laboratory, Ocean Springs, Mississippi. Hanson, J.S., Griffin, W.L., Richardson, J.W. & Nixon, C.J. (1985) Economic feasibility of shrimp farming in Texas: An investment analysis for semi-intensive pond grow-out, Journal of World Mariculture Society, 16, 129-150. Hardaker, J.B., Huirne, R.B.M., Anderson, J.R. & Lien, G. (2004). Coping with Risk in Agriculture. CABI Publ sihing, Cambridge, MA. Harvey, D. (1998). Aquaculture outlook. United States Department of Agriculture (USDA) Economic Research Service Electronic Re port LDP-AQS-8, Retrieved March 6, 2007, (http://usda.mannlib.cornell.edu/usda/c urrent/LDP-AQS/LDP-AQS-10-052006.pdf ). Harvey, D. (2006). Aquaculture outlook. United States Department of Agriculture (USDA) Economic Research Service Electronic Report LDP-AQS-24, Retrieved March 6, 2007, (http://usda.mannlib.cornell.edu/usda/c urrent/LDP-AQS/LDP-AQS-10-05-2006.pdf ) Haswell S. & Holmes, S. (1989), Estimating the small business failure rate: A reappraisal, Journal of Small Business Management, 27(July), 68-74. Herbert, J. (1985) Creating a Successful Restaurant. St. Martins Press, New York, New York. Higgins, R.C. (2004) Analysis for Financial Management. McGraw-Hill/ Irwin. New York, New York. Hochman, E., Leung, P., Rowland, L., & Wyba n, J. (1990). Optimal scheduling in shrimp mariculture: A stochastic growing inventory problem. American Journal of Agriculture Economics, 72(2), 382-393.
121 Inman, R.L., Helton, J.C., & Campbell, J.E. (198 1) An approach to se nsitivity analysis of computer models Introduction, input vari able selection, and preliminary variable assessment. Journal of Quality Technology. 13(3), 174-183. Johnson, S.R. (1967) A Re-examination of the farm diversification problem. Journal of Farm Economics, 49(3), 610-621. Johnston, R., Wessells, C.R., Donath, H., & Asche, F. (2001) Measuring consumer preferences for ecolabeled seafood: An international comparison. Journal of Agricultural and Resource Economics, 26(1), 20-39. Jolly, R.W. (1983) Risk management in agricultural production. American Journal of Agricultural Economics, 65(5), 107-113. Kaiser, H.M. & Liu, D.J. (1998) The effectiv eness of generic versus brand advertising. Review of Agricultural Economics, 20(1), 69-79. Karp, L., Sadeh, A. & Griffin, W.L. (1986) Cycles in agricu ltural production: The case of aquaculture, American Journal of Agricultural Economics, 68(3), 553-561. Kinnucan, H.W., Asche, F., Oystein, M. & Roheim C.A. (2003) Advances in economics of marketing and implications for aquaculture development. Aquaculture Economics & Management 7(1/2), 35-53. Kinnucan, H.W. & Fearon, D. (1986) Effects of ge neric and brand advertising of cheese in New York City with implications for allocations of funds. North Central Journal of Agricultural Economics, 8(1), 93-107. Kinnucan, H.W. & Miao, Y. (1999) Media specific returns to generic advertising: The case of catfish. Agribusiness, 15(1), 81-99. Lancaster, K.J. (1966) A new a pproach to consumer theory. Journal of Political Economy, 74(2), 132-157. Laramore, S., Laramore, C.R. & Scarpa, J. (2001) Effect of low salinity on growth and survival of postlarvae and juvenile Litopenaeus vannamei. Journal of the World Aquaculture Society 32, 385-392. Larson, CM. & Clute, R.C. (1979) The failure syndrome. American Journal of Small Business, 4 (October), 35-43.
122 Lewis, R.R., Phillips, M.J., Clough, B. & Macint osh, D.J. (2002) Thematic review on coastal wetland habitats and shrimp aquaculture. A Report Prepared for the World Bank, Network of Aquaculture Centres in Asia-Pacific, Wo rld Wildlife Fund and Food and Agriculture Organization of the United Nations Consortium Program on Shrimp Farming and the Environment, Retrieved February 22, 2007, (http://184.108.40.206/search ?q=cache:cY7uKOHy9HMJ:www.enaca.org/Shrimp/Case/Th ematic/FinalMangrove.pdf+thematic+review+of+coastal+wetlands+habitats&hl=en&ct=cl nk&cd=1&gl=us). Mabiso, A., Sterns, J., House, L. & Wysocki, A. (2005) Estimating cons umers willingness-topay for country-of-origin labels in fresh apples and tomatoes: A double-hurdle probit analysis of American data using factor scor es. Selected paper, American Agricultural Economics Association Annual Me eting, Providence, Rhode Is land. Retrieved February 15, 2007, (http://agecon.lib.umn.edu/cg i-bin/detainview.pl?paperid=16215). Manzo, Jose (2006) Jose Manzo Shrimp C onsulting Service, Sebastian, FL, Personal communication, March 2006. Markowitz, H.M. (1952) Po rtfolio selection. Journal of Finance 7, 77-91. Mason, E.S. (1939) Price and producti on policies of largescale enterprise, American Economic Review, supp.29, 61-74. Mason, E.S. (1949) The current state of th e monopoly problem in the United States. Harvard Law Review, 62, 1265-1285. McConnell, C.R. & Brue, S.L. (1993) Economics. McGraw-Hill Inc. New York, New York. McMahon, D.Z. & Baca, B. (2001) The effects of various salinities on mort ality and growth of Florida pink shrimp Farfantepenaeus duorar um and Pacific white shrimp Litopenaeus vannamei. Proceedings of the World Aquaculture Society, Aquaculture 2001: A Fantasy Come True, p. 427, January 21-25, 2 001, Lake Buena Vista, Florida. Medley, P.B., Nelson, R.G., Hatch, L.U., Rouse, D.B. & Pinto, G.F. (1994) Economic feasibility and risk analysis of Australian Red Claw Cray fish Cherax quadricarinatus aquaculture in the southeastern United States. Journal of the World Aquaculture Society, 25(1), 135-146. Miller, A., Dobbins, C., Pritchett, J., Boehlje, M. & Ehmke, C. (2004) Risk management for farmers. Staff Paper 04-11, Department of Agricultural Economics, Purdue University, West Lafayette, Indiana. Muraro, R.P. & Hebb, J.W. (2005) Budgeting co sts and returns for Indian River citrus production 2004-05 EDIS FE-630, University of Florida, IFAS, Food and Resource Economics Department, Gainesville, FL. March 2005. Retrieved October 12, 2006 (http://edis.ifas.ufl.e du/pdffiles/FE/FE63000.pdf ).
123 Muraro, R.P., Hebb, J.W. & Stover, E.W. (2003) Budgeting costs and returns for Indian River citrus production, 2002-2003. Electronic Da ta Information Source (EDIS) FE 433, University of Florida, Institute of Food and Agricultural Sciences (IFAS), Food and Resource Economics Department, Gainesvill e, Florida, September 2003. Retrieved October 15, 2006 (http://edis.ifas.ufl.edu/pdffiles/FE/FE43300.pdf ). National Oceanic and Atmospheric Admini stration (NOAA) (2006) The Saffir-Simpson hurricane scale. National Hurricane Cent er (NHC), Retrieved September 5, 2007 (http://www.nhc.noaa. gov/aboutsshs.shtml ). NOAA (2007) National Environmenta l Satellite, Data, and Info rmation Services (NESDIS), National Climatic Data Center, Image and P ublication System, Local climatological data, Retrieved October 17, 2006 (http://www7.ncdc.noaa.gov/IPS/index.jsp ). NOAA National Marine Fisher ies Service (NMFS) (2006) NOAA aquaculture program. Retrieved February 7, 2007 (http://www.nmfs.noaa.gov/ mediacenter/aquaculture/ ). NOAA NMFS (2007) Fisheries of the United States 2005. NOAA Fisheries: Office of Science & Technology, Foreign Trade, U.S. Imports http://www.st.nmfs.gov/st1/ fus/fus05/06_trade2005.pdf NOAA NMFS (2006) The Saltonsta ll-Kennedy grant program: Fisheries research and development report. Retrieved March 7, 2007 (http://220.127.116.11/search?q =cache:R4iRPhHAgFAJ:www.nmfs.noaa.gov/mb/financial_ services/skpdfs/SK%2520Annual%2520Report%25202006.pdf+wild+ame rican+shrimp+incorporated&hl=e n&ct=clnk&cd=16&gl=us ). NOAA NMFS (2004) Seafood consumption statistics. Fisheries of the United States, Office of Science and Technology, Fisheries Statistic s and Economics Division, Silver Spring, Maryland. Retrieved March 8, 2007 (http://www.noaanews.noaa.gov/stories 2005/pdf/seafood-consumption11-09-2005.pdf). Nerlove, M. & Waugh, F.V. (1961) Advertising wit hout supply control: Some implications of a study of the advertising of oranges. Journal of Farm Economics, 43(4/1):813-837. Nicholson, W. (2002) Microeconomic theory, (8th Edition). South-Western Thompson Learning, Stamford, Connecticut. Obreza, T.A. & Collins, M.E. (2002) Common soils us ed for citrus production in Florida. EDIS SL-193, University of Florida, Institute of Food and Agricult ural Sciences, Soil and Water Science Department, Gainesville, Florida, Retrieved April 18, 2007 (http://edis.ifas.ufl.edu/pdffiles/SS/SS40300.pdf ). Olsen, K.D. (2004) Farm Management: Principles and Strategies. Iowa State Press, Ames, Iowa.
124 O'Neil, H.M., & Duker, J. (1986), Survival and failure in small business, Journal of Small Business Management, 24(January), 30-37. Ott, R.L. & Longnecker, M.T. (2004) A First Co urse in Statistical Methods. Brooks/ ColeThompson Learning, Belmont, California. Pacific Seafood Group (2002) Warm wa ter shrimp. Retrieved May 3, 2007 (http://www.pacseafood.com/products/ww_shrimp.html). Palacios, E., Ibarra, A.M. & Racotta, I.S. (2000) Tissue biochemical composition in relation to multiple spawning in wild and pond-reared Penaeus vannamei broodstock. Aquaculture, 185(3/4), 353-371. Parker, R. (1995) Aquaculture Science, Delmar Publishers/ Thompson Learning, Florence, Kentucky. Perez-Farfante, I. & Kensley, B. (1997) Penaeo id and Serestioid shrimps and prawns of the World. Keys and diagnoses for the Families and Genera, Memories DuMuseum National DHistorie Naturelle, Paris, France. Porter, M.E. (1996) What is strategy? Harvard Business Review, 74(6), 61-78. Powers, T. (1988) Introduction to Management in the Hospitality Industry (3rd Edition). John Wiley & Sons, New York, New York. Rhodes, .R.J., Manzi, J.J. & Hopkins, J.S. (1987) Potential molluscan culture by South Carolina marine shrimp farms: A preliminary financial analysis. Presented at 18th Annual Meeting World Aquaculture Society, 19-23 Jan. 1987, Guayaquil, Ecuador. Richardson J.W., Schumann, K.D. & Feldman, P.A. (2006) Simetar Simulation for Applied Risk Management, Version 2006 Texas A&M University, College Station, Texas. Richardson, J.W. & Mapp, Jr, H.P. (1976) Us e of probabilistic cash flows in analyzing investment under conditions of risk and uncertainty. Southern Journal of Agricultural Economics 8(2), 19-24. Richardson, J.W., Klose, S.L., & Gray, A.W. (2000) An applied procedure for estimating and simulating multivariate empirical (MVE) probab ility distributions in farm-level risk assessment and policy analysis. Journal of Agriculture and Applied Economics, 32, 299315. Richardson, J.W. & Mapp, Jr., H.P. (1976) Us e of probabilistic cash flows in analyzing investment under conditions of risk and uncertainty. Southern Journal of Agricultural Economics. 8(2), 19-24.
125 Roberts, M.J., Osteen, C. & Soule, M. (2004) Risk, government programs, and agriculture. Economic Research Service, Technical Bulletin 1 908, U.S. Department of Agriculture, Washington, D.C., Retrieved January 16, 2007 (http://www.ers.usda.gov/publi cations/tb1908/tb1908. pdf March 2004 ). Robinson, L. J. & Brake, J. R. (1979) Applicat ion of portfolio theory to farmer and lender behavior. American Journal of Agricultural Economics, 61(1), 158-169. Rosenberry, B. (2005). An oral hi story of shrimp farming in th e western hemisphere. Shrimp Farming News. Retrieved February 6, 2007 (http://www.shrimpnews.com/History.html ). Rosenberry B. (2007) About shrimp farming. Shri mp News International, Retrieved April 9, 2007 (http://www.shrimpnews.com/About.html ). Roy, L.A., Davis, D.A., Saoud, I.P. & Henr y, R.P. (2005) Supplementation of potassium, magnesium, and sodium chloride in pr actical diets for Litopenaeus vannamei reared in low salinity waters. Proceedings of the World Aquaculture Society, Aquaculture America 2005: Aquaculture~Image, Sustai nability &All That Jazz, p. 384, January 17-20, 2005, New Orleans, Louisiana. Sadeh, A., Pardy, C.R., Griffin, W. & Lawrence, A.L. (1986) Uncer tainty consideration resulting from temperature variation on growth of Penaeus stylirostris in ponds. The Texas Journal of Science, 38(2), 159-173. Samocha, T.M., Hamper, L., Emberson, C.R., Davi s, A.D., McIntosh, D., Lawrence, A.L.& Van Wyk, P.M. (2002) Review of some recent deve lopments in sustainable shrimp farming practices in Texas, Arizona, and Florida. Journal of Applied Aquaculture 12, 1-42. Samocha, T.M., Lawrence, A.L., Collins, C.R., Em berson, C.R., Harvin, J.L. & Van Wyk, P.M. (2001) Development of integrated, envir onmentally sound, inland shrimp production technologies. In: The New Wave, Proceedings of the Special Session on Sustainable Shrimp Farming, Aquaculture 2001 (eds C.L. Browdy and D.E. Jory), pp 64-75 The World Aquaculture Society, Baton Rouge, Louisiana. Samocha, T. M., Lawrence, A. L. & Pooser, D. (1998) Growth and survival of juvenile Penaeus vannamei in low salinity water in a semi-closed recirculating system. Israeli Journal of AquacultureBamidgeh 50, 55-59. Sandifer, P.A., Hopkins, J. S. & Stokes, A.D. (1988). Intensification of shrimp culture in earthern ponds in South Carolina: Progress and prospects. Journal of World Aquaculture Society 19(4), 218-226. Say, J.B. (1803) Trait dconomie Politique (2nd Edition). Paris:Deterville, France. Schmit, T.M. & Kaiser, H.M. (2004) Decomposing th e variation of generic advertising response over time. American Journal of Agricultural Economics, 86(1), 139-153.
126 Schumpeter, J.A. (1934) The Th eory of Economic Development. Harvard University Press, Cambridge, Massachusetts. Schubert, T. S. & Sun, X. (2003) Bacterial citr us canker. Plant Pathol ogy Circular No. 377, 5th revision, FDACS, Division of Plant Industry, Gainesville, Florida. Shaffner, D.J., Schroder, W.R., & Earle, M.D. (1998) Food Marketing: An International Perspective. WCB/ McGraw Hill, Boston, Massachusetts. Sharpe, W.F. (1963) A simplified model for portfolio analysis. Management Science 9(2), 277293. Spreen, T.H., Barber, Jr., R.E., Brown, M.G., Hodges, A.W., Malugen, J.C., Mulkey, W.D.,Muraro, R.P., Norberg, R.P., Rahmani, M., Roka, F.M. & Rouse, R.E. (2006) An economic assessment of the future prospects for the Florida citrus indu stry. Special report, University of Florida IFAS, Food and Re source Economics Department, Gainesville, Florida, Retrieved February 18, 2007 (http://www.fred.ifas.ufl.edu/file s/economic_assess_flcitrus_indus.pdf ). Spreen, T.M., Zansler, M.L. & Muraro, R.P. ( 2005) Florida citrus canke r eradication program (CCEP): Annual economic impact on Florida s fresh and processed grapefruit industry. EDIS FE-534, University of Florida, IF AS, Food and Resource Economics Department, Gainesville, Florida, Re trieved January 17, 2007 (http://edis.ifas.ufl.edu/FE534 ). Stickney, R. (1996) Aquaculture in the United States: A Historical Survey. John Wiley & Sons, Hoboken, New Jersey. Thacker, S. & Griffin, W. ( 1994). Indoor intensive red drum aquaculture: A stochastic sensitivity analysis. Journal of the World Aquaculture Society 25, 86-100. Thompson, S.R. & Eiler, D.A. (1975) Producer returns from increased milk advertising. American Journal of Agricultural Economics, 57(3), 505-508. Trimble, W.C. (1980) Production trials for m onoculture and polyculture of white shrimp (Penaeus vannamei) or blue shrimp (P. Sty lirostris) with Florid a pompano(Trachinotus carolinus) in Alabama, 1978-1979. Procedings of the World Mariculture Society, 11, 4459. Unisys Corporation (2006) Hurricane tropical data archive. Retrieved September 19, 2006 (http://weather.unisys.com/hurricane/archive/ ). United States Department of Agriculture (USDA) 2002. National Agriculture Statistics Service (NASS), 2002 Census of agriculture, qui ck facts, Retrieved November 18, 2006 (http://www.nass.usda.gov/Census_of_Agri culture/2002/Quick_Fact s/valueof land.asp ).
127 USDA, Florida Agricultural Statistics Servi ce (FASS) (2007) Florid a commercial citrus production areas. Retrieved, March 7, 2007 (http://www.nass.usda.gov/Stat istics_by_State/Florida/Public ations/Citrus/cci/2006/Citrus %20Areas.jpg). USDA, (FASS) (2004). Comm ercial citrus inventory, prelimin ary report. Retrieved, March 7, 2007,(http://www.nass.usda.gov/Statistics_by_State /Florida/Publications /Citrus/ccipre/cci pr04.pdfOrlando, FL.). USDA (2003) Farm financial picture looks pr omising for 2003. Amber Waves, The Economics of Food, Farming, Natural Resources, a nd Rural America, Retrieved May 17, 2007, (http://www.ers.usda.gov/AmberWaves /Feb03/Findings/financialpicture.htm ). United States Department of Commerce (U SDOC), National Oceanic and Atmospheric Administration. (2005) Atlantic multidecadal oscillation. Physical Oceanography Division, Retrieved September 27, 2006, (http://www.aoml.noaa.gov/phod/amo_faq.php ). United States Department of the Treasury (USDT) (2007) Interest rate st atistics. Daily Treasury bill rates, Retrieved May 17, 2007, (http://www.treas.gov/offices /domestic-finance/debtmanagement/interest-rate/daily_treas_bill_rates.shtml ). University of Florida, Interdisciplinary Center for Biotechnology Resear ch, (2007) Biotech Plant Containment Facility, Citrus canker, March 13, 2007 (http://www.biotech.ufl.edu/PlantContainment/ ). Urner Barry Subscription Serv ices (2006) Urner Barry Comte ll with seafood price current. Historical price databa se, Retrieved August 2006. Uspensky, J.V. (1937) Introduction to Mathematical Probability, McGrawHill Book Company, Inc., New York New York. Van Anrooy, R., Secretan P., Lou, Y. Roberts, R. & Upare M. (2006) Review of the current state of the world aquaculture insurance, F AO Fisheries Technical Paper 493, Food and Agriculture Organization of the United Nations, Rome, Italy. Van Wyk, P.M. (2000) Economics of shrimp culture in a freshwater recirculating aquaculture system. Proceedings of the Third Internati onal Conference on Recirculating Aquaculture. Virginia Polytechnic University and Stat e University, Roanoke, Virginia, p.318-323. Van Wyk, P.M., Davis-Hodgkins M., Laramore, R ., Main, K.L., Mountain, J. & Scarpa, J. (2000) Production of marine shrimp in fres hwater recirculating aquaculture systems, Florida Department of Agriculture and Cons umer Services. Bob Crawford, Commissioner. Tallahassee, Florida. Van Wyk, P.M., Davis-Hodgkins, M., Laramore, C.R ., Main, K.L., Mountain, J. & Scarpa, J. (1999) Farming marine shrimp in recirculat ing freshwater system s, FDACS contract #4520, Florida Department of Agriculture and Consumer Services, Tallahassee, Florida.
128 Walker, J.R. (2002) Introduction to Hospitality. Prentice Hall, Upper Saddle River, New Jersey. Ward, R. (2006) Commodity checkoff programs and generic advertising, Choices Magazine, The Magazine of Food, Farm, and Resource Issues 21(2), Retrieved March 7, 2007, (http://www.choicesmagazine.org /2006-2/checkoff/2006-2-02.htm ). Wessels, C.R., Cochrane, K., Deere, K., Willis, P. & Willmann, R. (2001) Product certification and ecolabeling for fisheries sustainabili ty, United Nations Food and Agriculture Organization Fisheries Techni cal Paper 422, Rome, Italy. Wessells, C.R., Johnston, R.J. & Donath, H. (1999) Assessing cons umer preferences for ecolabeled seafood: The influence of speci es, certifier, and ho usehold attributes. American Journal of Agricultural Economics, 81(5):1084-1089. White, J. & vanBlokland, P. (2006) Trends in Fl orida citrus with emphasis on Indian River, St. Lucie, Martin, and Palm Beach Counties. Electronic Data Information Source (EDIS) FE636, Institute of Food and Agricultural Sc iences, Department of Food and Resource Economics University of Florida, Gaines ville, FL. Retrieved February 7, 2007 (http://edis.ifas.ufl.edu/Fe636 .). Wichmann, H. (1983), Accounting and ma rketingKey small business problems. American Journal of Small Business 7(Spring), 19-26. Wirth, F.F., Love, L.A. &Palma, M.A. (2007) Pu rchasing shrimp for at-home consumption: The relative importance of credence versus physical product features. Aquaculture Economics and Management. 11(1), 17-37. Wirth, F.F. (2003) The United States mark et for shrimp. Unpublished manuscript. Wirth, F.F. & Davis, K.J. (2001a) Seafood restau rant shrimp purchasing behavior and attitudes toward farm-raised shrimp. Staff Paper Series SP01-19 University of Florida, Institute of Food and Agricultural Sciences (IFAS), Food and Resource Economics Department, Gainesville, Florida. Wirth, F.F. & Davis, K.J. (2001b) Assessing pote ntial direct consumer markets for farm-raised shrimp. Staff Paper Series SP01-13 University of Florida, Institute of Food and Agricultural Sciences (IFAS), Food and Re source Economics Department, Gainesville, Florida. Wirth, F.F., Dugger, D. & Creswell, L. (2004) Commercial scale Penaei d shrimp demonstration in inland freshwater systems, Final Projec t Report for Cost Reimbursable Contract 007188 between FL DACS, Division of Aquaculture and University of Florida, IFAS, Gainesville, Florida. Wolfe, A.F. (1944) Measuring the e ffect of agricultu ral advertising. Journal of Farm Economics, 26(2), 327-347.
129 World Aquaculture Society (WAS), World Aqu aculture Society, Aquaculture America 2007, Science for Sustainable Aquaculture, San An tonio, Texas, February 26 March 2, Conference Attendee. Wyban, J.A., Lee, C.S., Sato, V.T., Sweeny, J.N. & Richards, Jr, W.K. (1987) Effect of stocking density on shrimp growth rate s in manure-fertilized ponds. Aquaculture, 61, 23-32. Wyban, J.A., Ogle, J., Pruder, G.D., Rowland, L.W. & Leung, P.S. (1987a) Design, operation, and comparative financial analysis of shrimp farms in Hawaii and Texas. Oceanic Institute Technical Report #87-6, Waimanalo, Hawaii. Wyban, J.A. & Sweeney, J. N. (1989) Intens ive shrimp growout trials in a round pond, Aquaculture, 76, 215-225. Wyban, J.A. & Sweeney, J. N. (1991) The Oceanic Institute shrimp manual: Intensive shrimp production technology. The Oceanic Institute, Makapu u Point, Waimanalo, Hawaii, USA. Wyban, J.A. (2002) White shrimp boom continues. Global Aquaculture Advocate. December, p.18-19, Retrieved March 8, 2007 (http://www.hihealthshrimp.com/pdf/gaa-wybandec02.pdf). Zansler, M.L., Spreen, T.H. & Muraro, R.P. ( 2005) Floridas citrus can ker eradication program (CCEP): Summary of annual costs and benefits, Electronic Data Information Source (EDIS) FE532. Institute of Food and Agricultural Sciences, Department of Food and Resource Economics University of Florida, Ga inesville, Florida, Re trieved February 7, 2007, (http://edis.ifas.ufl.edu/FE532 ). Zazueta, E.S. (2007) Economic risk assessment of production strategies in aquaculture: Case study of a semi-intensive white shrimp Litopenaeus vannamei farm in Mexico, Proceedings of the World Aquac ulture Society, Aquaculture 2007: Science of Sustainable Aquaculture, February 26-March 2, 2007, San Antonio, TX. Ziegler Brothers Feed Manufacturer s (2006) Gardners, Pennsylvania. Zucker, D.A. & Anderson, J.L. (1999) A dynami c stochastic model of a land-based summer flounder Paralichthys dentatus aquaculture firm. Journal of the World Aquaculture Society, 30(2), 219-235.
130 BIOGRAPHICAL SKETCH Jennifer Clark is a native Floridian, born in Wi nter Haven and raised in Vero Beach. After completing an Associate of Science degree in aquaculture production technology from Indian River Community College and Harbor Branch Oceanographic Institution in 1999, she accepted a position as department leader at a research and development aquaculture facility in Florida. During a brief training period in Mexico, she developed the t echnical skills necessary to implement a high-health shrimp broodstock progra m for this firm in a biosecure, recirculating environment. During her employment with this firm, Jenni fer continued her education and completed prerequisite courses to transfer into the Food and Resource Economics Agribusiness Management program at the University of Florid a. She graduated with honors in 2004 with a Bachelor of Science degree in Agribusiness Mana gement degree and entere d the University of Floridas Food and Resource Economics graduate pr ogram that year to pursue her Master of Science degree. After graduation, she plans to remain in Gainesville, Florida where she has accepted a faculty lecturing posit ion within the Food & Resource Economics Department at the University of Florida.