a farm-level approach to the methyl ... - getd.libs.uga.edu · a farm-level approach to the methyl...

A FARM-LEVEL APPROACH TO THE METHYL BROMIDE PHASE-OUT: IDENTIFYING

ALTERNATIVES AND MAXIMIZING NET-WORTH USING STOCHASTIC DOMINANCE

AND OPTIMIZATION PROCEDURES

by

MARK MCCULLOH BYRD

(Under the Direction of Cesar L. Escalante)

ABSTRACT

Methyl bromide (MeBr) is an effective yet toxic soil fumigant that is subject to an

accelerated phase-out under the Montreal Protocol on Substances That Deplete the Ozone Layer.

As a result of the phase-out, Georgia farmers now require a substitute fumigant and herbicide

combination that can control for weeds, pests, and diseases while delivering comparable yields to

those of MeBr. This study analyzes the economic viability of alternative production methods on

a representative Georgia pepper farm. The analysis is addressed through stochastic dominance

analysis, an enterprise budget, and the development of a linear programming model. Alternative

fumigants are ranked, analyzed within the budgetary framework, and subjected to production

constraints such as farm acreage, capital structure, and asset allocation within the programming

model. Compared with MeBr, the models prescribe results supporting the economic feasibility

of several alternative fumigants, and thus serve to educate Georgia vegetable farmers on

available input-substitution strategies.

INDEX WORDS: Methyl Bromide, Bell Pepper, Stochastic Dominance, Enterprise Budget, Linear Programming, Optimization-Simulation

iv




by

MARK MCCULLOH BYRD

B.S.A., University of Georgia, 1999

A Thesis Submitted to the Graduate Faculty of The University of Georgia in Partial Fulfillment

of the Requirements for the Degree

MASTER OF SCIENCE

ATHENS, GEORGIA

2005

© 2005

MARK MCCULLOH BYRD

All Rights Reserved




by

MARK MCCULLOH BYRD

Major Professor: Cesar L. Escalante

Committee: Michael E. Wetzstein Esendugue Greg Fonsah

Electronic Version Approved: Maureen Grasso Dean of the Graduate School The University of Georgia August 2005

DEDICATION

This thesis is dedicated to all of my friends within the College of Agriculture: Peter,

Katy, Hongsin, Feng (Frank), Rui (Carolyn), Xiaohui (Sarah), Yingzhuo, Marianna, Laxmi, Joe,

Jordan, Horacio, Katia, Tatiana, Dmitriy, R.J., Vahe, Doris, and Carrie. One of my greatest

future challenges will be to surround myself with such a diverse, kind, and intelligent group of

people. Thank you for all the good times.

iv

ACKNOWLEDGEMENTS

I cannot begin to thank Dr. Michael Wetzstein for his influence on my academic and

personal life. Your enthusiasm and commitment to the teaching profession, along with your

encouragement, respect, intellect, and laughter, is greatly appreciated and will never be

forgotten. To Dr. Cesar Escalante, it is a rarity to work with such a fine person. Your guidance

on this project has been invaluable. Thank you for your patience and support while juggling the

demands of a devoted father and beginning professor. I want to thank Dr. Esendugue Greg

Fonsah, for his valuable insights and contributions to this study.

Next, I want to express my deepest thanks to my parents, Paul McCulloh Byrd and Alice

Kathryn Gatlin, for their support and encouragement during my graduate education. You have

shown me more love than I could ever hope to repay. I love you, and I could not have done this

without you. To my brothers, Paul and Ben, I am proud of your accomplishments, and I love

you both. I wish you the greatest success the world has to offer.

Finally, no words can express my love for you Rhiannon. Your optimism and zest for

life has taught me to see the world in a different way. I will never forget our journey to the Lost

Coast, our plumerias and mangoes in Hawaii, thousands of miles of road, fireworks in Cuba,

backpacking in Canada, jazz clubs in New York City, Sunday reading with endless cups of

coffee, examinations, and the joy and laughter attached to each of them. Thank you for showing

me that unconditional love truly exists.

v

TABLE OF CONTENTS

Page

ACKNOWLEDGEMENTS.............................................................................................................v

LIST OF TABLES....................................................................................................................... viii

LIST OF FIGURES .........................................................................................................................ix

CHAPTER

1 INTRODUCTION .........................................................................................................1

Background Information ...........................................................................................1

Objectives..................................................................................................................7

Organization ..............................................................................................................8

2 LITERATURE REVIEW ..............................................................................................9

Economic Impact Assessments .................................................................................9

Methyl Bromide Alternatives Outreach ..................................................................13

Additional Economic Impact and Viability Studies................................................20

3 COMMERCIAL PEPPER PRODUCTION ................................................................28

United States Pepper Industry .................................................................................28

Georgia Pepper Production .....................................................................................30

Marketing ................................................................................................................31

Farm Management...................................................................................................32

4 EMPIRICAL METHODOLOGY................................................................................41

Expected Utility, Risk Aversion, and Efficiency Criteria .......................................41

vi

Interpretation of Dominance Analysis, Enterprise Budget, and Programming

Model.................................................................................................................53

5 ANALYSIS OF RESULTS .........................................................................................69

Stochastic Efficiency-The Final Four Production Methods ....................................69

Results of the Enterprise Budget Analyses .............................................................71

Programming Solutions...........................................................................................77

6 SUMMARY AND CONCLUSION ............................................................................93

Study Summary .......................................................................................................93

Conclusions .............................................................................................................95

Future Research.......................................................................................................97

BIBLIOGRAPHY..........................................................................................................................98

APPENDICES .............................................................................................................................104

A Bell Pepper Fresh Market Variable Cost Budget for Methyl Bromide .....................105

B Bell Pepper Fresh Market Variable Cost Budget for C35 + KPAM..........................107

C Bell Pepper Fresh Market Variable Cost Budget for C35 + Chloropicrin.................109

D Bell Pepper Fresh Market Variable Cost Budget for Telone II + Chloropicrin.........111

E GAMS Program for the Base-Case Farm Model using Methyl Bromide..................113

vii

LIST OF TABLES

Page

Table 4.1: Pre-Operating Balance Sheet........................................................................................66

Table 4.2: Description of Constraints and Requirements ..............................................................67

Table 4.3: Description of Portfolio Activities................................................................................68

Table 5.1: Efficiency Ranking of Final 8 Production Methods Using Second Degree Stochastic

Dominance.....................................................................................................................83

Table 5.2: Cumulative Distribution Functions for Final 8 Production Methods ...........................84

Table 5.3: Financial Efficiency and Break-even Analyses Using Experimental Yield Results ....85

Table 5.4: Financial Efficiency and Break-even Analyses Using Constant Yield Results ...........86

Table 5.5: Comparison of Costs per Carton for Final Four Production Methods..........................87

Table 5.6: Programming Solutions and Financial Ratios of Five-Year Averages for all

Production Methods ......................................................................................................88

Table 5.7: Yearly Programming Solutions for Decision Variables and Key Financial Measures

Using Methyl Bromide..................................................................................................89


Using C35 + KPAM as Methyl Bromide Substitute .....................................................90


Using C35 + Chloropicrin as Methyl Bromide Substitute ............................................91


Using Telone II + Chloropicrin as Methyl Bromide Substitute ....................................92

viii

LIST OF FIGURES

Page

Figure 3.1: Leading U.S. Bell Pepper Producing States (2002) ....................................................36

Figure 3.2: U.S. Fresh Bell Peppers, Production, Import, Export and Domestic Use (1979-2003)37

Figure 3.3: U.S. Bell Pepper Import Values from Selected Countries of the World (2000-2003)38

Figure 3.4: Georgia Farms, Farm Sizes, and Harvested Acreage (2002) ......................................39

Figure 3.5: U.S. Bell Peppers Season Average Prices and Per Capita Consumption (1979-2003)40

ix

1

CHAPTER 1

INTRODUCTION

1.1 Background Information

1.1.1 Historical Framework

On September 16, 1987, the signatures of 24 countries resulted in the Montreal Protocol

on Substances that Deplete the Ozone Layer. Former President Ronald Regan later committed

the United States to the international agreement on April 5th, 1988. Effective January 1, 1989,

the Protocol’s signatory nations began actively working toward fulfilling the guidelines

established under the Protocol. Between the meetings of the parties, scientists assessed the

toxicity of many chemicals and created an index which identified substances causing the most

damage to the ozone layer. As substances were listed, amendments were added to the original

treaty which recognized the need to accelerate the phase-out of these dangerous substances.

Methyl bromide (MeBr) was identified as one of the most toxic contributors to ozone depletion.

It was recommended during the Ninth Meeting of the Parties (1997) in Montreal, Canada that

MeBr face an accelerated phase-out schedule.

Methyl Bromide is an agricultural fumigant that is used to control weeds, nematodes,

soil-borne pests, and diseases. Methyl Bromide is widely accepted by agricultural producers

because they understand its chemical properties, it is inexpensive, and it is effective in most U.S.

climates. According to the United States Environmental Protection Agency, the U.S has "one of

2

the largest agricultural bases in the world and has historically used more MeBr than any other

country (EPA, 2005)."

A controversy now surrounds the use of MeBr and its phase-out schedule as developed

nations face a complete elimination of the substance by 2005. Within the U.S., California and

Florida consume the largest amounts of MeBr as their combined total domestic usage is greater

than 75% (Carpenter et al., 2000). Thus, Florida and California have conducted most of the

research on the costs and benefits of MeBr because they will absorb most of the total cost

associated with the phase-out. Florida and California use MeBr on fruits and vegetables to fight

microscopic parasitic roundworms known as root-knot nematodes, and major soil-borne diseases

such as bacterial wilt, southern blight, fusarium wilt (fungus), and fusarium crown and root rot.

Weeds are also effectively handled with repeated applications of MeBr (Carpenter et al., 2000).

Many of the problems faced by growers in Florida and California are the same as those faced by

Georgia producers.

This study addresses the potential economic impact to Georgia bell pepper producers at

the farm level under the MeBr phase-out. Georgia ranks third in the U.S. in acreage of fresh

market vegetables planted, and vegetables are the second most valuable crop in Georgia with an

approximate farm-gate value of $901.2 million (Boatright and McKissick, 2003, p.59).

Vegetable growers in Georgia state that eliminating MeBr will reduce yields and increase

production costs as a result of adopting more expensive and less effective alternatives (Seabrook,

2005). They requested that the United Nations Environment Programme (UNEP) consider their

nominations to continue using MeBr as defined under the “critical use exemptions” clause of the

Protocol.

3

1.1.2 Montreal Protocol and “Critical Use”

According to a USDA Economic Research Service (ERS) report (USDA, 2000), over 160

countries have agreed to reduce ozone depleting substances (ODS) by ratifying the Montreal

Protocol. Methyl bromide is individually classified within the Protocol as an Annex E controlled

substance with an ozone depleting potential (ODP) of 0.6 (UNEP, 1995). The ODP indicates the

amount of ozone destroyed by the emission of a particular gas relative to chlorofluorocarbon-11

(CFC-11), a major ozone depletor (Carpenter et al., 2000). Substances with an ODP over 0.2 are

considered Class I ozone depletors and are required to be phased out under the Protocol and the

United States' Clean Air Act. Additionally, Section 602(e) of the Clean Air Act states, “Where

the ozone-depletion potential of a substance is specified in the Montreal Protocol, the ozone-

depletion potential specified for that substance under this section shall be consistent with the

Montreal Protocol (USEPA,1990 )” However, MeBr's contribution to agriculture, and its

integral role in facilitating international trade have lead to a general agreement that there needs to

be allowances for “critical,” “quarantine,” and “pre-shipment” uses.

The term “critical use” was introduced to identify uses pertaining specifically to MeBr.

Decision IX/6 of Appendix 1 to the Protocol lays the foundation of “critical use” as decided upon

at the Ninth Meeting of the Parties (UNEP, 2000). Decision IX/6 states: (a) That a use of methyl

bromide should qualify as “critical” only if the nominating Party determines that:

(i) The specific use is critical because the lack of availability of methyl bromide for

that use would result in a significant market disruption; and

(ii) There are no technically and economically feasible alternatives or substitutes

available to the user that are acceptable from the standpoint of environment and

health and are suitable to the crops and circumstances of the nomination;

4

The Protocol delineates the phase-out schedules of MeBr according to parties considered

“developed” or “developing” nations. Article 2H of the Protocol defines the phase-out schedule

of MeBr to be administered by all Parties identified as “developed” nations, and stipulates that

these Parties “shall ensure…its calculated level of the controlled substance in Annex E does not

exceed, annually, its calculated level of consumption in 1991 (UNEP, 2000).” The 1991 baseline

figure also applies to those parties involved in the production of MeBr. The phase-out schedule

for Article 2H countries is:

25% reduction in 1999




Preshipment and quarantine uses exempt

Critical and emergency uses allocated after 2005.

1.1.3 Methyl Bromide Alternatives

Critical uses not withstanding, Georgia producers must identify alternative production

methods that incorporate chemical substitutes as a replacement to MeBr. As noted, California

and Florida have been the two states to predominantly research alternatives and conduct

economic impact analyses concerning the phase-out. These states rely on MeBr for the benefits

accruing to the production of strawberries and tomatoes. Florida uses 237% more MeBr in the

production of fresh market tomatoes than ten other states combined (Carpenter, 2000).

California uses 277% more MeBr for fresh market strawberry production than nine other leading

states combined (Carpenter, 2000). California alone accounts for half of the 35 million pounds

(active ingredient) of MeBr used annually for preplant fumigation in the U.S. (USDA ERS,

5

2000). Moreover, Florida consumes 30% of the total MeBr used for preplant fumigation in the

U.S. (USDA ERS, 2000). The total values to California and Florida for fresh market

strawberries and tomatoes are estimated at $1.4 billion and $920 million, respectively (USDA

NASS, 2005). These states combined account for 95% and 68%, respectively, of the total U.S.

fresh market value of strawberries and tomatoes (USDA NASS, 2005).

1.1.4 Technically Viable Alternatives

Before conclusions could be made regarding the economic impact of the MeBr phase-out,

substantial resources were invested in identifying technically viable alternatives. The USDA

Economic Research Service (ERS), the University of Florida, and the National Center for Food

and Agricultural Policy (NCFAP) worked together to develop studies that analyzed the

regulatory limitations and economic impacts of technically viable alternatives to MeBr (USDA

ERS, 2000). Scientists and growers are interested in chemicals that can produce yields

exceeding or comparable to MeBr. The majority of experiments that focused on yield were

conducted on California and Florida's strawberry and tomato crops. Thus, yield data concerning

other vegetables, vineyards, orchard, and nursery crops is limited (USDA ERS, 2000).

Early studies focused on the possibility of methyl iodide (MI) (iodomethane) as a

technically viable alternative to MeBr (Hueth et al., 2000). Studies conducted at the University

of California, Riverside indicate that MI is advantageous to MeBr because it is quickly broken

down and lasts in the atmosphere for a period of one to two days as opposed to MeBr's two years

(Zhang et al., 1998). Further, because MI is handled as a liquid and MeBr is handled as a gas,

worker safety is increased throughout the production process. Studies tested the performance of

MI under a broad range of environmental conditions such as soil moisture, temperature, texture,

and fumigation time. They found that MI was," consistently more effective than MeBr, on a

6

molar concentration basis, under a range of soil moistures, temperatures, soil textures and

fumigation times (Zhang et al., 1998, p.78)." An additional study by Becker et al. (1998) tested

the dose responses of MeBr and MI against three parasitic nematodes and one fungal plant

pathogen. California growers are susceptible to nematode infestations and fungi without the use

of an effective fumigant. In every case MI proved more effective at controlling for these pests

than MeBr. Finally, (Ohr et al., 1996) evaluated the effectiveness of MI for the control of fungi,

weeds, and nematodes in a series of fifteen field trials, and again found MI to be as efficacious or

more than MeBr.

Florida growers face different environmental conditions than those in California. In

addition to fungi, nematodes, and bacteria, the most problematic pests for growers in Florida and

the southeast are weeds. Yellow and purple nutsedge are common weeds that thrive in Florida

due to a long growing season and humid climate. Purple nutsedge is considered to be the most

noxious of all weeds found in tropical and sup-tropical regions of the world (Gilreath et al.,

2004), and few herbicides are registered for use throughout the entire growing season. Thus, it is

necessary to apply a combination of plastic mulch, fumigants, and herbicides to crops such as

tomato and pepper (Gilreath et al., 2004). A list of fumigants such as metam sodium (MNa),

chloropicrin (teargas) (Pic), anhydrous ammonia (AHN4), and 1,3-dichloropropene (1,3-D) were

analyzed in comparison to MeBr. These fumigants were tested in combination with the

herbicides napropamide, metolachlor, and pebulate. Gilreath et al. (2003) tested the effects of

these chemicals on bell peppers and cucumbers when applied under plastic polyethylene mulch

with drip irrigation. This work furthered studies performed on tomatoes by Locascio et al.

(1997). Both studies found the combination of 1,3-D + Pic to be effective at controlling for

nematode infestations, but the same combination failed to control for purple and yellow

7

nutsedges. With respect to tomatoes and peppers, combining the fumigants with pebulate

significantly reduced nutsedge levels up to 22% compared to those treatments not receiving

herbicides (Gilreath et al., 1994). The results of the 2003 study suggest that there are, "several

viable alternatives for MeBr in vegetable rotations (Gilreath et al., 1994, p.4)." However,

changes in rotations and planting seasons could affect the observed values. Unfortunately, the

erratic nature of fumigant efficacy on weed control does not allow for a single replacement for

MeBr (Gilreath et al., 2003).

Many field trials and indoor experiments conducted throughout the U.S. explored the

efficacy of alternative fumigants, fungicides, and herbicides with respect to MeBr. It is noted in

the scientific literature that local environmental conditions will dictate which combinations of

these chemicals and their respective application methods are most suitable for growers.

Agricultural economists use these results to conduct economic impact studies that account for the

regional demands and limitations of producers.

1.2 Objectives

The specific objectives of the research are as follows:

1) Identify and classify feasible fumigant alternatives to MeBr for Georgia growers through the

use of stochastic dominance analysis.

2) Combine available alternative fumigants with current production technologies to produce

economically viable production methods for Georgia growers.

3) Provide agricultural extension agents with a tool to educate Georgia vegetable farmers on the

resulting economic impact of varying input substitution strategies.

Stochastic dominance analysis, a tool that considers risk-return tradeoffs in identifying

more efficient methods among alternative production plans, will be used to address the first

8

objective. The second and third objectives are accomplished by employing an enterprise budget

developed for Georgia pepper producers in conjunction with an optimization/simulation

programming model. This combination recognizes the unique factors inherent to pepper

production and offers tailored solutions for Georgia producers. Enterprise budget analysis will

determine the comparative financial feasibility of a pepper production enterprise under

production plans involving MeBr and its substitutes. Optimization-simulation techniques

employed in a multi-period programming framework will determine the relative overall

economic viability and feasibility of optimal production and financial plans prescribed under

production systems that involve MeBr and its substitutes.

1.3 Organization

The remainder of this thesis is divided into five chapters. Chapter 2 examines the

economic literature with respect to stochastic dominance, and the optimization and simulation

procedures related to mathematical programming techniques. Chapter 3 identifies issues related

to commercial pepper production and management in the U.S. and Georgia. The development of

the theoretical model employed in this study is then discussed in Chapter 4. In addition to the

descriptive data and enterprise budgets used as support for the theoretical models, Chapter 5

presents the results of the stochastic dominance analysis and the simulation and optimization

procedures. The final chapter, Chapter 6, summarizes the study, presents the conclusions,

addresses limitations of the study, and offers suggestions for future research.

9

CHAPTER 2

LITERATURE REVIEW

2.1 Economic Impact Assessments

A chemical's technical feasibility is a necessary, but not independently sufficient,

condition for its economic feasibility. Published research on a crop's yield or survival, subject to

chemicals under various environmental or controlled conditions, does not satisfy requirements of

economic feasibility such as regulatory approval, sustainability, or a consistent positive return on

investment (Schaub, 2004). These requirements constitute potential barriers to economic

feasibility and dictate the use of economic assessment studies to evaluate the MeBr phase-out.

As noted in the previous chapter, regional differences in soil type and weather conditions may

produce large discrepancies among the efficacy of alternatives to control for a variety of pests

and diseases. Research conducted in California and Florida examines these differences and

constitutes a majority of the information concentrated on identifying MeBr alternatives

(VanSickle, 2000) (Carpenter, 2000).

In addition to these states, studies cited in this section investigate the economics of

pesticide use in the U.S., the economic values of alternative production methods and chemicals,

and the economic impact of the MeBr phase-out among regions, state, and countries. The

literature provides alternative empirical methods of assessing economic impacts through the use

of partial budgeting analysis, stochastic dominance analysis, and optimization/simulation

procedures. The use of integrated pest management (IPM) is of contemporary importance to

10

producers subject to the phase-out. The literature emphasizes that IPM will likely be a necessary

component of the production process for these growers. Literature surrounding the MeBr phase-

out acknowledges the variability and usefulness of IPM techniques among producers. However,

IPM is usually evaluated as a secondary consideration to the more important task of identifying

viable chemical substitutes (Lynch et al., 1997).

2.1.1 Economic Impact Assessments of Pesticide Use

Fernandez-Cornejo et al. (1998) investigated the economic value of pesticide use in

agriculture. The report summarized the empirical evidence related to the economics of pesticide

use. It emphasized the estimation of the value of pesticides in U.S. agriculture, the economic

effects of reducing or restricting pesticide use, and the promotion of IPM to reduce the potential

health and environmental effects associated with pesticide use. The study identified marginal

productivity calculations, the expected loss to pests, and the economic effect of banning

pesticides as three perspectives used to estimate the economic value of pesticide use.

It is generally recognized that producers' per acre expenditure for pesticides increases as

the value of the crop increases. For example, while wheat and corn producers spent

approximately $6 and $22 per acre in 1995, respectively, on pesticide use, cotton and strawberry

producers spent nearly $48 and $1600 per acre, respectively, during the same time period

(Fernandez-Cornejo et al., 1998). According to Fernandez-Cornejo et al. (1998), estimates of the

value of the marginal product (VMP) of pesticide use provide growers and policymakers with an

indirect measure of the cost "in terms of foregone agricultural output" of reducing pesticide use.

They note that under the usual assumptions, a farmer would maximize profits by increasing

pesticide use up to the point where the expected marginal return (VMP) equals the pesticide

marginal cost (Fernandez-Cornejo et al., 1998). They found that the VMP of pesticides appears

11

to be falling, but the methodology surrounding alternative empirical models is controversial.

Further, when farm risk is assessed in combination with pesticide use, they concede that the

conventional view recognized pesticide use to be risk reducing; however, there was no empirical

consensus that determined the effect of risk on the VMP. Studies found that uncertainty about

output price and yield leads to lower optimal levels of pesticide use by individual farmers,

whereas uncertainty about other variables such as pest density leads to a higher optimal pesticide

use under risk aversion (Fernandez-Cornejo et al., 1998). Thus, the authors concluded that the

VMP of most pesticides in the U.S. is higher than their corresponding price.

Fernandez-Cornejo et al. (1998) examined studies that focused on the expected yield

losses relative to a current or potential yield and the respective impact these losses have on the

value of pesticide use. They recognized that, "estimates of crop yield losses that might result

without the availability of pesticides are difficult to obtain because these losses vary by crop,

soil, and weather condition. In addition, yields may vary by year because of technological

developments (i.e. new plant varieties), changes in cropping practices (destruction of crop

residues), appearance of new pests, and weather (Fernandez-Cornejo et al., 1998, p.470)." The

resulting empirical estimates are highly variable and tend to be based on judgments of experts in

different fields of natural science (Fernandez-Cornejo et al., 1998). The authors summarized

yield losses for several major U.S. crops and concluded that fruits and vegetables have high

losses when subject to the elimination of insecticides and fungicides. They estimated that the

elimination of herbicides results in yield losses ranging from 0% to 53%. Fernandez-Cornejo et

al. (1998) cite a study conducted by the USDA in 1997 that calculated the productivity of

pesticides to be approximately $3 to $4 of pesticide expenditure.

12

The article further assessed the economic effects of total or partial bans on pesticide use

in general or in individual case studies. The authors identified partial budgeting and large scale

econometric models as two methods that are generally used to estimate these effects. Partial

budgeting estimated the value of the production lost without pesticides assuming that output

prices remained constant, while the econometric models allowed for input and output

substitution during production (Fernandez-Cornejo et al., 1998). Fernandez-Cornejo et al.

(1998) concluded that a total ban of pesticides would increase annual consumer expenditures by

$228 per household (in 1989 dollars), amounting to approximately $30 billion per year. The

production of fruits and vegetables is examined as well as a corresponding total ban in these

industries. The result of a total ban in these sectors may require producers to increase production

acreage by 2.5 million acres (44%), and increase unit production costs by 75%. In turn,

wholesale prices of fruits and vegetables could increase 45%, returns to producers could decrease

by 30%, retail prices could increase by 27%, and domestic consumption could fall by 11%

(Fernandez-Cornejo et al., 1998). However, the studies received criticism for assuming a total

ban would be realistic, and for failing to acknowledge that future research may minimize

impacts. Fernandez-Cornejo et al. (1998) noted that authors of the original studies agree that

further research needs to identify intermediate points between current practices and a total ban.

The availability of alternative means of pest control influences the value of a pesticide's

use to agricultural producers and consumers (Fernandez-Cornejo et al., 1998). Thus, assessments

of partial bans of pesticides offer more realistic estimates of economic impacts. For example,

Fernandez-Cornejo et al. (1998) reviewed a study conducted in 1991 that assumed a 50%

reduction in pesticide use. This study concluded that the total costs of reduction would be

approximately $1 billion per year although the true impact would be difficult to determine

13

because there are different mechanisms in which to implement a ban (Fernandez-Cornejo et al.,

1998).

Finally, Fernandez-Cornejo et al. (1998) considered the value of integrated pest

management in the context of both governments' attempts to reduce the detrimental effects of

pesticides to human health, and as an option for producers to maintain viable business

operations. IPM uses multiple techniques to," maintain pest infestation at the most economically

sensible level rather than attempting to completely eradicate all pests (Fernandez-Cornejo et al.,

1998, p.478)." Some of the more important techniques adopted by producers are scouting, soil

testing, pheromone traps for pests, and cataloging data on weather patterns that identify the

development and activity of specific pests. Additionally, biological controls of pests such as

natural predators, parasites, and pathogens result in minimal environmental hazards, and may

result in an equivalent amount of pest control in some situations. Fernandez-Cornejo et al.

(1998) stated that the empirical evidence on the effect of IPM on the use of pesticide use is

mixed. The authors cited a study conducted by the National Foundation for IPM Education that

concluded IPM reduced pesticide use by 15% while delivering net returns to producers.

2.2 Methyl Bromide Alternatives Outreach

The Methyl Bromide Alternatives Outreach (MBAO) is an organization focused on

facilitating the exchange of information concerning viable alternatives for MeBr. The

organization began holding the Annual International Research Conference on Methyl Bromide

Alternatives and Emissions Reductions in 1994. Their goal is to establish a forum that promotes

the exchange of research information among agricultural and forestry specialists from

governmental, academic and private institutions. Several years after the beginning of the

14

conference, as scientists had had time to conduct experiments on alternatives and publish their

results, discussions began to incorporate the economic impact surrounding these developments.

The earliest economic impact assessments, analyzed and presented by Lynch et al.

(1997), provided researchers with a base-line from which to evaluate the impact of the MeBr

ban. Lynch et al. (1997) identified changes regarding the costs of technically feasible chemical

and non-chemical alternatives, the initial economic results of studies conducted in California and

Florida, the regulatory environment surrounding alternatives, and comparisons of the efficacy of

alternatives. Results presented that same year by Hueth et al. (1997) focused on the economic

impact of banning MeBr in California agriculture. Hueth's study measured the cost of

prohibiting MeBr use in California by directly measuring the change in consumer and producer

welfare. The methodology was based on data concerning the changes in per acre yields and

production costs resulting from the adoption of available alternatives.

Hueth et al. emphasize that their research was an estimation of the short-term impact of

removing MeBr completely, and that it adds to the literature by providing a more general

analysis of the impacts of banning MeBr. The authors further acknowledged that the impacts

reported in the study were uncertain because of the limited experimental opportunities and the

inability to forecast outcomes related to future innovations. The study concluded that the impact

on growers' profits may experience a wide range of values corresponding to the infiltration of

exotic pests. In good years growers may lose as little as $71 million while in years of high

infestation they may lose as much as $334 million in profits. The net impact, which accounts for

consumer benefits and losses as well as grower losses, exhibited the same dramatic range with

yearly impacts from $60 million to $242 million (Hueth et al., 1997).

15

Presentations made in 1998 and 1999 by Lynch and Carpenter focused on identifying the

most likely alternatives for producers, and investigating the impacts of phasing out MeBr for the

United States, respectively,. The 1998 study sought to identify the impact to producer yields and

costs by gathering information from research documents, expert opinions, and workshops. Their

research concluded that both Florida and California growers would increase usage of the

combination of 1,3-D and Chloropicrin. To maximize profits, the study found that Florida

growers would likely use an accompanying herbicide, and California growers would increase the

percentage of Chloropicrin in the mixture. Lynch and Carpenter (1999) then compare MeBr

with the next best alternative over a range of fruits and vegetables grown not only in California

and Florida, but also in Georgia and North Carolina.

Assuming regulatory restrictions, cost, and environmental factors dictate these

alternatives, the authors, "computed the increase in revenue from using one pound of MeBr

rather than the next best alternative assuming the market prices do not change for the crop

(Lynch and Carpenter, 1998, p.2)." These revenues serve as the value of MeBr to producers on a

per pound basis. After making assumptions for producer usage rates of MeBr, a grower's loss

was then calculated as a function of the amount of MeBr used per acre. The value of MeBr to

fruit and vegetable growers in California ranged from $3.85/lb. for peppers to $41.43/lb. for

southern coast strawberries. The value to Florida fruit and vegetable growers ranged from

$4.25/lb. for tomatoes to $33.48/lb. for strawberries.

Lynch and Carpenter's (1999) study identified these estimates as one approach to

estimating the economic impact of the MeBr phase-out. Their second approach examined the

annual crops that use MeBr extensively, and allowed for adjustments to acreage and prices.

Crops such as tomatoes, strawberries, peppers, watermelon, cucumber, squash, and eggplant are

16

crops primarily grown by horticulturalists in California and Florida with additional production

taking place in South Carolina, North Carolina, and Georgia (Lynch and Carpenter, 1999). The

authors used this presentation to introduce the phase-out as a spatial partial equilibrium problem.

Furthermore, Carpenter et al. (2000) fully developed the model used in the 1999 study

within a comprehensive study published the following year. The report by Carpenter et al.

(2000) was funded by the USDA's Economic Research Service and published through the

National Center for Food and Agricultural Policy. The authors developed a regionally

disaggregated model that emphasized California, Florida, Georgia, and South Carolina as the

primary users of MeBr, and identified Texas and Mexico as their competitors. The model used

calculations of baseline equilibrium production, monthly shipments between production areas

and markets, and monthly consumption given current technologies as a guide to compare with

changes resulting from the phase-out (Carpenter et al., 2000). Crop prices for each market were

introduced, and as MeBr is phased out, production technologies shift; resulting in changes in

production costs and expected monthly yields. The alternative technologies used in the study

were identified through previous research acknowledged during MBAO conferences. The study

defined the best alternative technology as the one with the lowest per-unit cost (Carpenter et al.,

2000). The authors use a simulation approach, rather than econometric, because data for

individual producers was unavailable. Costs were allowed to vary among regions and yields

were assumed to be non-stochastic. The goal of the model was to maximize producers' returns

and consumers' benefits while recognizing constraints on available land in each region, and the

amounts sold to consumers could not be greater than the amount supplied (Carpenter et al.,

2000).

17

California strawberry growers were expected to shift from using MeBr to a combination

of Chloropicrin and Vapam at an additional cost of $97.50 per acre. Florida strawberry and

tomato growers were expected to substitute Telone C-17 and various herbicides at an increased

cost of $227.50 per acre. Georgia and South Carolina growers would switch to Telone C-17 and

pebulate at an excess cost of $13.36 per acre (Carpenter et al., 2000). The study did not present

the additional costs per acre for pepper producers in Georgia, and acknowledged that little

research was conducted into alternatives for peppers in Florida. The authors assumed yield

losses among Florida pepper growers to be approximately 12.5%, and that they would use a

combination of Telone C-17 and herbicides costing $227.50 per acre.

The results of the study were presented in comparison to the pre-ban baseline. Carpenter

et al. (2000) stated that consumers incur the heaviest cost from the ban as consumer surplus is

expected to decrease by $158 million. The decrease in consumer surplus for strawberries and

tomatoes was estimated to be 10.3% and 1.7%, respectively. Consumer surplus for peppers was

expected to decrease $4.5 million, or 1.1% (Carpenter et al., 2000). With respect to producers,

the model predicted that California growers will increase strawberry production with a

corresponding decrease in tomatoes. Florida producers will increase the acreage planted of

strawberries, but will reduce the acreage planted of tomatoes and peppers (Carpenter et al.,

2000). Georgia producers were expected to increase their production of tomatoes, but the study

does not present any figures concerning increases or decreases in peppers. However, total U.S.

pepper production was estimated to decrease 14% (Carpenter et al., 2000). Alternatively, the

authors noted that changes in price may make it profitable for growers to continue operations.

They calculated the price impact to peppers resulting from the ban to be in the range of -$39.82

to $51.96 per ton (Carpenter et al., 2000).

18

Carpenter et al. (2000) analyzed the impact on producer revenues across regions and

compared them with the baseline. Revenues did not include the costs of production, and as a

result it was possible for a producer to experience post-ban revenues that were higher than pre-

ban revenues as a result of price increases (Carpenter et al., 2000). The model estimated that

Florida producers will experience decreases in revenues for tomatoes and peppers of $57.3

million and $12.6 million respectively (Carpenter et al., 2000). In California, strawberry and

tomato producers may experience a respective change of $38.8 million, and -$35 million in

receipts (Carpenter et al., 2000). The model prescribed that shifts would occur in production

among regions to compensate for dramatic losses in specific areas. However, the U.S. grower

that loses is expected to experience a decrease in revenues of $153.9 million (Carpenter et al.,

2000). Finally, the report indicated that if revenues were used to gauge the overall impact to

U.S. consumers and producers, the model estimated a total net decrease in welfare of $76.5

million (Carpenter et al., 2000).

More recent studies presented at the MBAO conferences evaluated alternative production

methods and used current data in an attempt to assess the options and impacts to growers facing

the phase-out. A study conducted by Jovicich et al. (2003) investigated the economic benefit

associated with the greenhouse production of peppers. The study found that growing colored

peppers for the specialty market can result in economically viable operations. The research was

conducted through the Protected Agriculture Project at the University of Florida and emphasized

the use of screened plastic greenhouses with passive venting as an economical alternative to field

production (Jovicich et al., 2003). The cost of these systems was estimated at $2 to $4 per

square foot, and it was demonstrated that fungal diseases decreased, fruit quality improved, and

yields were up to ten times that of field grown peppers (Jovicich et al., 2003). Although the

19

growing season could be extended for producers under this method, start up costs and knowledge

requirements may prohibit greenhouse adoption by established growers.

During the 2004 MBAO proceedings, Sydorovych et al. (2004) presented research

focusing on the economic impact of MeBr alternatives for strawberries. Their objective was to

evaluate the economic viability of chemical alternatives using a partial budget analysis. Several

characteristics of this study are relevant to the goals of our research. For example, researchers

first developed a cost model for a plasticulture production system on a five acre representative

strawberry farm in the Piedmont and Coastal Plain regions of North Carolina (Jovicich et al.,

2003). The production practices used in the study resulted from consultations with research

specialists and agricultural extension agents of North Carolina State University. An important

assumption of the Jovicich et al. (2003) model is that the machinery and equipment used in the

enterprise budget can be applied to other farming enterprises other than strawberries. This is

characteristic of the farming operations of many small to medium size growers found in the

southeastern United States. As a result, machinery expenses, not including fumigation and

irrigation costs, reflect the equipment costs for a total farm business (Jovicich et al., 2003).

Partial budgeting analyzes small production adjustments on farm profitability by

comparing the negative effects of applying a new treatment relative to a base or standard

treatment (Jovicich et al., 2003). The study found that production costs increased if an

alternative treatment resulted in either higher fumigation costs and/or higher yields because

higher yields result in both higher labor and material costs. The chemical alternatives tested in

the study were Telone-C35, Telone II, chloropicrin, Inline, and metam sodium. Methyl bromide

was recognized as the base treatment with respect to yields and revenues. Interestingly, MeBr

resulted in the highest cost to producers per acre at $1,267, while chloropicrin reduced costs

20

relative to the base by $92.00 per acre. Further, chloropicrin and Telone-C35, when used

independently, produced yields greater than MeBr. A comparison of revenues resulted in

chloropicrin, Telone-C35 and metam sodium applied by shank injection each providing

additional returns above MeBr of $1,767.60, $290.79, and $3.18 respectively. The study did not

list returns for Telone II and Inline.

2.3 Additional Economic Impact and Viability Studies

2.3.1 Economic Viability of Methyl Iodide

While most viability studies examining the efficacy and economic feasibility of

alternatives focus on a menu of choices available to producers, researchers may some times

focus on the economic impact of a single chemical alternative. Hueth et al. (2000) compared the

effectiveness of methyl iodide (MI) against MeBr for preplant fumigation, "through an analysis

that combines information on the current world market of iodine, the implications of limited

studies on MI effectiveness, and past analyses of the economic benefits of methyl bromide to

California producers (p.45)." An important assumption of the study was that the researchers

considered 1,3-D and chloropicrin to be the primary alternatives used in place of MeBr for

California growers. Thus, their study of MI was not an endorsement, and MI's demand schedule

was constructed by analyzing its cost and effectiveness against these alternatives (Hueth et al.,

2000). The efficacy of MI compared to MeBr was established through the studies noted in

Chapter one.

Hueth et al. (2000) assessed the potential market for MI by calculating the per-unit value

of MI, and developing demand curves for MI under six different scenarios. The study defined

the value of the marginal product (VMP) of MI as the value resulting from a discrete change

from one control material to another. The VMP represented the benefit of using MI compared to

21

the next best alternative (Hueth et al., 2000). The study restricted the demand of MI to

commodities where the VMP was positive. After calculating the VMP, demand schedules were

mapped out over a range of prices under optimistic, average, and pessimistic scenarios regarding

changes in application rates and non-chemical application costs (Hueth et al., 2000). The

expected price for MI was assumed to be the current price of $11.00/lb., and at least as high as

the $2.25/lb. price paid for MeBr.

The results indicated that as the price per pound of MI increases, the demand for MI

under each scenario decreases rapidly. The study evaluated demand schedules for those

producers using both tarping and non-tarping operations. As expected, the demand for MI

decreased at a greater rate for producers using a low density polyethylene tarp because tarping

added additional costs to operations and provided additional protection against pests. The

demand for MI was less across all prices and alternative scenarios when compared to non-tarping

operations.

The authors acknowledged that their analysis was not representative of the entire

California market, and therefore the potential demand for MI would likely be much higher

(Hueth et al., 2000). Their results indicated that at a price of $11.00/lb. demand could reach

four million pounds in California alone. As demand increased for MI due to the phase-out,

suppliers may be encouraged to innovate and enter this new market (Hueth et al., 2000).

However, limitations surrounding MI are its current costs of extraction and limited supply

compared to MeBr. Current production of MI could supply the demand of niche markets in the

U.S., but additional investments in extraction are needed in order to supply larger regional

markets (Hueth et al., 2000). Overall, the authors identify MI as a viable alternative to MeBr.

22

2.3.2 Economic Impacts within the U.S. Vegetable Industry

Analyses conducted by (Spreen et al., 1996) (VanSickle, 2000) investigate the impact of

the phase-out with respect to the U.S. winter fresh vegetable market and the U.S. vegetable

industry as a whole. Both studies make important contributions to the literature by recognizing

and quantifying the impacts to producers that rely heavily on MeBr.

Spreen et al. (1996) estimated the impact of the phase-out on the winter market for fresh

vegetables such as green peppers, squash, cucumbers, tomatoes, eggplant, and watermelons. The

primary impact of the phase-out to this market will be felt by Florida because they are the

leading domestic supplier of these products (Spreen et al., 1996). Previous research had failed to

assess the impact to Florida independently by region or season. Although California is a major

consumer of MeBr it does not compete with Florida's vegetable market due to the seasonal

differences between the two states. The study’s model focused on months in which Florida is a

supplier of vegetables to the U.S. market (Spreen et al., 1996). Spreen et al.'s model served as

the basis for the model used by (Carpenter et al., 2000). However, in this case researchers

characterize the North American winter fresh vegetable market as a spatial equilibrium problem

rather than the entire U.S. market (Spreen et al., 1996). The authors constructed the winter fresh

vegetable market by disaggregating Florida into four regions of production and acknowledging

Texas and Mexico as dominant suppliers during the winter season occurring from November

through May.

A quadratic programming model was developed to determine factors such as equilibrium

consumption of each commodity in every month and region, optimal levels of shipments

between areas of supply and demand, optimum acreage employed for each cropping system, and

optimum monthly production of commodities by region (Spreen et al., 1996). Solutions for total

23

supplies accounted for the double cropping systems common to Florida growers and the

production of commodities under each system. Total demand was estimated through an inverse

Rotterdam system of demand equations developed in previous studies, but was expanded to

include the additional commodities used in this study (Spreen et al., 1996). For simplicity, all

parameters of the model were assumed to be non-stochastic although the yields of fresh

vegetables were known to be highly variable. The authors noted that constraints imposed on the

model did not include land, labor, and machinery. They justified this approach because, "it could

not be established that effective resource requirements restricted the production of fresh

vegetables in Florida, Mexico, or other U.S. supply areas (Spreen et al., 1996, p.436)."

The model was initially solved for a base-case scenario without accounting for the

elimination or reduction of MeBr. This analysis generated results that conformed to past data

with respect to shipping among supply regions, quantity of the crop produced, and the acreage

employed for production of the six crops. However, the model did fail to produce realistic

results concerning pepper production in Florida. Under the base case, Southwest Florida was

expected to grow no peppers although it has historically been the largest pepper producing region

in the state (Spreen et al., 1996). After modifying the model to account for the loss of MeBr, the

solution indicated that the production of peppers, tomatoes, eggplant, and cucumbers would be

eliminated in one of Florida's four regions. All other regions throughout Florida were expected

to experience adverse effects from the ban due to a severe contraction in production of these

crops (Spreen et al., 1996).

The model predicted that Mexico and Texas would provide for most of the lost

production in Florida as neither region was affected by the ban (Spreen et al., 1996). Of the six

crops analyzed in the study, peppers and tomatoes were subject to the greatest reduction in

24

yields. In terms of revenue, Florida's free on board (FOB) revenues for all six crops combined

were expected to decline by approximately $548 million or 53%. Mexico's total (FOB) revenues

were expected to increase by $368 million while Texas gained revenues of nearly $24 million

solely through its increase in pepper production (Spreen et al., 1996). Spreen et al. (1996)

estimated that the loss of MeBr would have a $1 billion impact on the US winter vegetable

industry, with Florida accounting for nearly all of this impact. However, they emphasized that

viable substitutes may invalidate these conclusions and reduce future impacts to Florida growers.

A study conducted by VanSickle et al. (2000) expanded on the Spreen et al. model

developed in 1996. In the interim between studies, significant research was conducted on

alternatives to MeBr that provided more reliable data in terms of yields, costs, and regulatory

restrictions. The objective of VanSickle et al. (2000) was to evaluate the impacts of MeBr

alternatives on U.S. producers of fresh vegetables. Similar to Spreen et al. (1996), the study

characterized the North American vegetable market as a spatial equilibrium problem. However,

the new model introduced strawberries into the model in addition to the crops from the previous

study. Production areas identified by the model were expanded to include Florida, Mexico,

California, Texas, South Carolina, Virginia and Maryland combined, and Alabama and

Tennessee combined. Florida and California were then separated into regions (VanSickle et al.,

2000).

The model allocated production based on the cost of delivering the crops across regional

markets. The study used a similar system of inverse Rotterdam equations to estimate total U.S.

demand (VanSickle et al., 2000). Production costs were estimated according to the geographic

areas used in the model, and the chemical alternatives were selected by a team of scientists,

industry representatives, and environmental advocates (VanSickle et al., 2000). Strawberry

25

growers in Central Florida were expected to switch from MeBr to a combination of Telone C17

and Devrinol. Growers could expect a decrease in pre-harvest costs of $71 per acre and an

accompanying decline in yield of 15% (VanSickle et al., 2000). Florida tomato growers were

expected to use Telone C17 with the herbicide Tillam. If a double cropping system was

employed, costs ranged from a decrease of $61 per acre for tomatoes and cucumbers to an

increase of $255 per acre tomatoes and squash.

Florida pepper production was assumed to replace MeBr with a combination of Telone

C17 and Devrinol. The impact to costs varied depending on single or double cropping systems.

For single crops, costs ranged from a decrease of $41 per acre for spring peppers produced in

Central Florida to an increase of $397 when peppers were grown close to urban areas. For the

double cropping system, costs were expected to increase by $437 per acre. Additionally, the

associated yields for pepper were expected to decline 15% to 25% for most areas of Florida

(VanSickle et al., 2000).

The VanSickle model initially solved for a base-case that assumed the continued use of

MeBr. The base case allowed researchers to identify the accuracy of the model when compared

to the observed data of prior production seasons. Adjustments made to the parameters of the

base model reflected the changes in production costs and changes in the yield in switching to the

alternatives (VanSickle et al., 2000). Three scenarios were examined with respect to the base

case. Holding the costs constant throughout, Van Sickle et al. (2000) ran the model using the

expected yields identified by scientists, a pessimistic scenario using lower than expected yields,

and an optimistic scenario assuming larger yields than the expected case. The base solutions

approximated the production patterns recorded by the UDSA. VanSickle et al. (2000) found that

26

the total production of crops was not expected to dramatically change, but that changes would

occur in the allocation of production across regions.

The greatest losses occurred to tomato and pepper producers in Florida. Under the

expected scenario, Florida tomato growers lost $68.8 million in shipping point revenues and

pepper growers experienced a 65% decline in acres planted (VanSickle et al., 2000). In addition,

the optimistic scenario for pepper growers in Florida resulted in a $119.9 million decrease in

shipping point revenues (VanSickle et al., 2000). In terms of total losses, California strawberries

were greatest at approximately $192.8 million. In the aggregate, Florida and California could

lose $218.4 million and $218.1 million, respectively, in revenues for all crops under the expected

scenario (VanSickle et al., 2000).

Furthermore, continued research examined how much the impacts of the MeBr

alternatives must be reduced in order to minimize the changes in market share for existing

producers (VanSickle et al., 2000). The authors assumed that a "seamless transition" between

MeBr and its alternatives could occur if the largest impact to any producer's yield were less than

or equal to 10% of the baseline scenario. They maintained the cost impacts assumed in the

expected impacts scenario.

Next, adjustments to yield were made on each crop in the model until the largest impact

felt by any producing area for any single crop was a 10% loss in the baseline market share

(VanSickle et al., 2000). Results indicated that in order for tomato production to experience a

seamless transition, while holding costs equal to those under the expected scenario, Florida

growers would reduce yields by 55% to 60% (VanSickle et al., 2000). The model indicated that

pepper producers will need to reduce yield losses by approximately 93% to make the transition.

27

However, the authors noted that future alternatives may reduce the economic impact to producers

if they cause costs to decrease, yields to increase, or new market windows.

28

CHAPTER 3

COMMERCIAL PEPPER PRODUCTION

3.1 United States Pepper Industry

3.1.1 U.S. Production and Consumption

The bell pepper is a mild sweet pepper that is grown commercially on approximately

57,000 acres throughout the U.S. (USDA ERS, 2004). The primary destination for the wide

variety of bell peppers (yellow, purple, red, green) in the U.S. is the fresh market. The fresh

pepper market is the focus of this study because producers in this market face greater demands

by consumers to provide a consistent quantity of quality peppers for rapid distribution to market.

Further, the fresh market receives approximately 90% of all peppers produced domestically

(USDA ERS, 2001). Producers for the remaining 10%, the processed market, are mainly

concerned with generating consistent yields, and place less emphasis on quality.

Within the U.S., most domestic production is supplied by only eight states. The leading

five states for pepper production in terms of harvested acreage in 2002 were California, Florida,

North Carolina, New Jersey, and Georgia respectively (Figure 3.1) (USDA ERS, 2004).

California's production increased 9% from 2002 to 2003 and significantly influenced the 5%

growth in total production for the U.S. over the same period (Figure 3.2) (USDA ERS, 2004).

Peppers are grown throughout the entire year in the U.S., and shipments tend to increase during

the months of May and June. Conversely, import shipments usually rise during the winter

months as domestic producers' fall short of domestic demand (USDA ERS, 2001). The trend in

29

production for peppers has been moving upward over the past twenty-five years as total U.S.

production increased from 582 million pounds in 1979 to 1,648 million pounds as of 2003

(Figure 3.2). This represents an approximate total increase in production of 183% (USDA ERS,

2004). U.S. per capita consumption over the same period increased 141.4% and 2003 registered

as the second highest consumption year on record at 7.0 lbs. per person (USDA ERS, 2004).

3.1.2 Imports and Exports

As U.S. consumer demand for peppers trends upward (Figure 3.2), the contribution of

pepper imports to total U.S. supply exhibits the same trajectory. In 2003, the U.S. maintained its

position as a net importer of peppers with imports representing 25% of the total U.S. supply. Of

the total percentage of imports, 20% supply fresh market demand (USDA ERS, 2004). The

majority of peppers being imported for the fresh market are received from Mexico, Canada, and

the Netherlands with their respective contributions to the whole being 67%, 17%, and 9% in

2003 (USDA ERS, 2004).

Canada is the U.S.'s leading trading partner for fruits and vegetables. Pepper exports to

Canada, valued at $69.4 million, represent 94.5% of total U.S. pepper export value (Figure 3.3).

Mexico is our second most important trading partner for pepper exports. In 2003, Mexico

imported $1.3 million worth of peppers from the U.S. However, Mexico is also the primary

supplier of U.S. pepper imports. U.S. pepper imports from Mexico increased 8.75% from the

years 1999 to 2002 (USDA ERS, 2003). Exports as a percentage of total production increased

over the past twenty-five years, but, from 2002 to 2003, fell from 8% to 7% of total supply

(USDA ERS, 2004) (Figure 3.2).

30

3.2 Georgia Pepper Production

All commercial bell peppers grown in Georgia belong to the genus Capsicum annuum,

and approximately half of commercial brand peppers are openly pollinated or a hybrid variety

(UGA CES, 1990). Georgia's growing season usually begins in the spring during late February

and extends into early July. Fall crops are planted in July and may be harvested into November,

or until the first major frost (Kelly, 2003). Peppers grow well in Georgia due to the state's warm

and humid climate. However, producers must institute effective soil management practices and

pest control schemes in order to generate maximum yields.

Eighty-one percent of bell pepper production in Georgia takes place on approximately ten

farms. Each of these farms is greater than one-hundred acres in size. The remaining nineteen

percent of harvested acreage is spread among eighty-one farms, and almost half of these

producers farm approximately 0.1 to 0.9 of an acre (Figure 3.4) (USDA, 2002). Georgia's total

harvested acreage for bell peppers in 2003 was 5,230 with an accompanying farm gate value

estimated at $87.1 million (CAES, 2003). According to the University of Georgia's Cooperative

Extension Service, watermelons, sweet corn, and snap beans, respectively, lead all Georgia

vegetables in the total acreage employed.

In terms of farm gate value, tomatoes and onions lead the State at approximately $122

million and $104 million, respectively (CAES, 2003). These figures reveal that although bell

peppers rank thirteenth in terms of acres harvested among all Georgia vegetables, their ranking

as the third most valuable vegetable in the state reveals their importance to Georgia agriculture

(CAES, 2003).

31

3.3 Marketing

As with other commodities, the general price level for peppers is determined by supply

and demand. Measured in current dollars, average seasonal prices for U.S. bell peppers have

experienced fluctuations in the range of $19.90 to $30.70 per cwt. through years 1979 to 2003

(Figure 3.5) (USDA ERS, 2004).

Wholesale buyers of vegetables rely on industry standards in packaging to ensure

conformity with every purchase. Thus, bell peppers are commonly packaged in either 1 1/9

bushel wire-bound crates (1 1/9 buctns/crts.) or bushel cartons with each representing 25-30

pounds of peppers (Gast, 1991). An average price per carton of twenty-eight pounds is typically

assumed to be the weight of either packaging option. Thus, converting the U.S. prices measured

in ($/cwt.) into ($/crt.) results in a price range of $5.57 to $8.60 per carton for average U.S.

yields from 1979 to 2003. Georgia producers experienced prices above the national average in

2003. Prices for premium Georgia peppers (Jumbos) ranged from a low of $10.09/crt. for the

fall harvest to a high of $16.24/crt. for the spring harvest. Lower quality peppers (US 1&2)

commanded prices in the range of $7.51/crt. for the fall harvest to $9.93/crt. during the spring

harvest (CAES, 2003)

Variations in commercial pepper prices are primarily explained by supply-side factors.

For example, both seasonal changes in weather conditions and yearly changes in acres planted

influence total production. However, the demand for peppers is unlikely to be a motivating

factor for price variation because consumption patterns change little from year to year (CAES,

2003). Consumer preferences for peppers are represented in the quantity and quality of their

purchases. Although the specialty markets demand a variety of different colored peppers,

consumers tend to prefer green, and they are willing to pay a premium for Jumbo peppers.

32

Thus, marketing campaigns should be focused on these preferences (CAES, 2003). At

the same time, Georgia growers should also maintain focus on delivering a reliable, quality

product, to distributors. In a 2002 University of Georgia marketing survey, wholesalers and

distributors ranked the origin of the product last with respect to those factors that were most

important in their purchasing decision (Wolfe and Fonsah, 2002).

3.4 Farm Management

3.4.1 Technical Considerations

Bell peppers can be grown in a variety of soils and climates throughout the U.S.

However, in order to produce high quality peppers growers must consider climate, soil

conditions, and proper soil preparation before beginning operations. Peppers grow best in light,

fertile, well drained soils. Another requirement is that the soils are subject to adequate tillage.

Tillage is a cultural method of pest management that may be used in conjunction with chemical

and biological methods to reduce crop damage and enhance plant growth. Tillage directly

destroys weeds by disrupting or moving the soil within 10 to 12 inches of the soil surface

(USDA ERS, 2003). Because peppers have roots that reach depths of approximately 36 to 48

inches, tillage also supplies the plants with a greater soil volume, nutrients, and water than other

soil preparation methods such as disking (UGA CES, 1990).

In Georgia, growers usually choose to transplant 5 to 6 week old pepper seedlings rather

than directly seeding into the field. Transplanting allows for better weed control, less specialized

planting and bedding, and earlier harvest dates. Transplants in raised beds warm more quickly

and require greater irrigation and soil maintenance during drought conditions (UGA CES, 1990).

However, an advantage of this technique is that the heat accelerates plant growth. Many Georgia

farmers choose a production combination that uses both transplanting and plastic mulch to

33

increase yields. The combination restricts in-row weed growth while conserving water and

fertilizer (UGA CES, 1990). Drawbacks to this method are the additional costs for specialized

equipment and the inability to control for yellow or purple nutsedges, two particularly

problematic weeds for southeast growers (UGA CES, 1990).

Optimum pepper production in Georgia requires substantial irrigation in addition to

chemical methods of pest management. Producers may choose to use either sprinkler or drip

irrigation depending on the size of operations. While most Georgia growers use some type of

sprinkler system, drip irrigation is becoming more popular due to its effectiveness when used

with plastic mulch. Drip irrigation conserves water and can effectively deliver fertilizers directly

into the bed (UGA CES, 1990). In Georgia, fertilizer management programs are site specific,

but usually require liming the soil with calcitic or dolomitic limestone to lower soil acidity, and a

thorough application of nitrogen (UGA CES, 1990).

3.4.2 Diseases and Pest Management

Pepper producers in Georgia face a variety of diseases and pests that cause severe

economic damage to their crops. Fungi such as Cercospora capsici, Vermicularia capsici, and

Sclerotium rolfsii cause damage to the pepper plant through, respectively, defoliation, rotten

fruit, and rotten stems. In addition to fungi, there are viruses such as tobacco mosaic virus

(TMV), tobacco etch virus (TEV), cucumber mosaic virus (CMV), and potato virus Y (PVY)

(UGA CES, 1990). All of these viruses cause what is generally known as pepper mosaic virus.

As a result of this virus, plant growth may become stunted resulting in malformed and

underdeveloped fruits (Brunt et al., 1996). Pseudomonas spp., also known as bacterial leafspot,

is the most serious disease affecting peppers in Georgia because it results in defoliation and

brown spots on the fruit.

34

Moreover, producers must minimize the devastating effects of root-knot nematodes.

These parasites damage the host by disrupting vascular tissues at the root; thus, preventing or

reducing the uptake of water and nutrients necessary for plant growth. Infected plants display

galled roots, poor growth, wilted foliage, and poor yield. The damage can often be aggravated

by the parasite’s interaction with other micro-organisms such as fungi and bacteria (Brunt et al.,

1996). As a result of these problems, integrated pest management strategies should be employed

throughout the growing season. Scouting fields two to three times a week is the most cost

effective strategy for eliminating seedling pests such as cutworms or thrips, as well as foliage

feeders like aphids, flea beetles, potato beetles, and hornworms. On the other hand, weed

control is one of the more problematic issues facing pepper growers, and must be addressed

through seasonal applications of fumigants and herbicides. As noted, when used with black

plastic mulch, weed control can be attained, but nutsedges may persist as a problem.

3.4.3 Production Costs

The costs involved in maintaining or beginning many agricultural enterprises are

extensive. Producers able to effectively generate a consistent, high quality product for the

market must still consider the costs involved in production in order to maintain viable long-term

operations. Many agricultural commodities do not enjoy successful brand recognition, and as a

result, they do not enjoy the premium prices that such schemes deliver. Therefore, minimizing

production costs should be the main avenue for maximizing net income.

A general listing of costs facing pepper producers can be broken down into total variable

costs and total fixed costs. Land can be a variable or fixed cost. Most variable costs will differ

with respect to the acreage employed and the cultural practices of each producer. Some

examples of variable costs are seed, fertilizer, chemicals, fuel, and labor. These can then be

35

divided between pre-harvest and harvest costs. Pre-harvest costs are utilities, vehicles, and

associated user fees. Fixed costs include equipment owned, management, and general overhead

costs (UGA CES, 1990). Many of these costs can dramatically change across producers or

regions if owners enjoy economies of scale due to organization, specialization, or education.

Georgia producers have access to enterprise budgets prepared by agricultural extension

economists and specialists that allow them to estimate production and break-even costs (UGA

CES, 1990).

36

0

10,000

20,000

30,000

40,000

50,000

60,000

U.S. CA FL NC NJ GA

U.S. & STATES

Har

vest

ed A

cres

$0

$100,000

$200,000

$300,000

$400,000

$500,000

$600,000

Valu

e

Acres Harvested Value in 1000 Dollars

Figure 3.1: Leading U.S. Bell Pepper Producing States (2002) Source: USDA ERS 2004

37

0.0

500.0

1,000.0

1,500.0

2,000.0

2,500.0

1979

1980

1981

1982

1983

1984

1985

1986

1987

1988

1989

1990

1991

1992

1993

1994

1995

1996

1997

1998

1999

2000

2001

2002

2003

2004

YEARS

(Mill

ions

lbs.

)

Production Imports Exports Domestic

Figure 3.2: U.S. Fresh Bell Peppers, Production, Import, Export and Domestic Use (1979-2003) Source: USDA ERS 2004

38

0

50,000

100,000

150,000

200,000

250,000

300,000

350,000

1000 Dollars

2000: 2001: 2002: 2003:YEARS

Netherlands Canada Mexico World

Figure 3.3: U.S. Bell Pepper Import Values from Selected Countries of the World (2000-2003) Source: USDA ERS 2004

39

0

5

10

15

20

25

30

35

40

45

50

(0.1 - 0.9) (1.0 - 4.9) (5 - 14.9) (15 - 24.9) (25 - 49.9) (50 - 99.9) (≥ 100)

Farm Size (Acres)

Num

ber o

f Far

ms

0

500

1000

1500

2000

2500

3000

3500

4000

4500

5000

Har

vest

ed A

crea

ge

Number of Farms Total Harvested Acres

Figure 3.4: Georgia Farms, Farm Sizes, and Harvested Acreage (2002) Source: USDA NASS 2002

40

0

5

10

15

20

25

30

35

40

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26

Years

($/C

wt)

0

1

2

3

4

5

6

7

8

(Lbs

)

Season-average price Per Capita Use

Figure 3.5: U.S. Bell Peppers Season Average Prices and Per Capita Consumption (1979-2003)

Source: USDA ERS 2004.

41

CHAPTER 4

EMPIRICAL METHODOLOGY

4.1 Expected Utility, Risk Aversion, and Efficiency Criteria

4.1.1 Expected Utility Theory

The scientific and economic literature cited in the previous chapters provides this study

with a basis from which to extend the economic analysis of the MeBr phase-out. This chapter

describes the analytical methods used to answer the objectives noted in chapter one. This is

accomplished through the use of stochastic dominance analysis and a mathematical

programming model. These two methods are linked using an enterprise budget that represents

the financial operations of a southeastern vegetable grower.

Profit maximization is the goal of many agricultural producers. However, these

producers must successfully manage risk in order to maintain viable operations. Weather, yield,

pests, and price are types of uncertainty characteristic of the growers' decision-making

environment. To help producers minimize risk, agricultural economists integrate theoretical

tools such as expected utility, mean-variance, and stochastic dominance analyses into the actual

decision making process.

The expected utility hypothesis is the basis for much of the theory of decision making

under uncertainty, and provides a general decision rule for identifying agents' preferences (King

and Robinson, 1981). Risk can be defined as the variability in the outcomes associated with a

"state of nature" (Wetzstein, 2005). Each state of nature can be thought of as a lottery (N), and

42

defined as a set of probabilities, summing to 1, associated with all possible outcomes (Wetzstein,

2005). Similar to consumer theory, expected utility theory assumes that agents' preferences are

rational when choosing among states of nature. As such, the theorem relies on several axioms to

define an agent's behavior when faced with a choice that results in a probability distribution of

outcomes. When choices result in single-dimensioned consequences, the axioms of ordering and

transitivity, continuity, and independence are sufficient for deducing a decision maker's utility

function (Anderson et al., 1977).

The axiom of Ordering assumes that people prefer one of two risky choices a1 and a2 or is

indifferent between them (Anderson et al., 1977). Transitivity extends the concept of ordering

when the menu of available choices is greater than two. Transitivity implies "that an agent

preferring choice a1 to a2 (or is indifferent between them) and prefers a2 to a3 (or is indifferent

between them), she/he will prefer a1 to a3 (or be indifferent between them) (Anderson et al.,

1977, p.67)." The Continuity Axiom assumes an agent's ability to effectively determine the

subjective probabilities associated with outcomes involving risk. More formally, "if an agent

prefers a1 to a2 to a3, a subjective probability P(a1) exists other than zero or one such that she is

indifferent between a2 and a lottery yielding a1 with probability P(a1) and a3 with probability 1 –

P(a1) (Anderson et al., 1977, p.67)." This implies that an agent will risk receiving a bad

outcome, when given the choice between a good and a bad outcome, if the probability of

receiving this outcome is sufficiently low (Anderson et al., 1977). The Independence Axiom is a

fundamental assumption of many theories concerning choice under uncertainty (Wetzstein,

2005). The Independence Axiom states:

If N, N', and N'' are states of nature, and ( ρ ) is the probability of an outcome occurring,

43

then N f N' if and only if

ρ N + (1- ρ ) N'' f ρ N' + (1- ρ ) N'' (4.1) The state of nature N'' should not influence an agent's preference when choosing between state of

nature N and state of nature N' (Wetzstein, 2005). Finally, unlike commodities, states of nature

are mutually exclusive and can not be consumed jointly (Wetzstein, 2005).

In the mid-1940s, John von Neumann and Arthur Morgenstern developed the Expected

Utility Theory as a model to measure risk preferences. They expanded the more common utility

function into the expected or von Neumann-Morgenstern utility function. The expected utility

function is the weighted average of utility obtained from alternative states of nature, where the

weights are the probabilities of the states occurring (Wetzstein, 2005). In this model, each state

of nature is assigned its respective probability of occurring ( ρ i), and for all states of nature (Xi ),

the utility received in one state is added to the utility received in another state (Wetzstein, 2005).

For two possible states of nature the expected utility function can be represented as:

U(x1, x2, ρ 1, ρ 2) = ρ 1U(x1) + ρ 2U(x2) (4.2)

Agents make choices concerning the outcomes of each state of nature based on subjective and/or

objective probabilities. Thus, expected utility theory posits that a decision maker should act to

maximize her subjective expected utility if she is to be consistent with her expressed preferences

(Anderson et al., 1977).

4.1.2 Risk Aversion

Agents' preferences can be deduced according to his or her response to risk. Expected

utility employs the statistical concept of variance, a measure of the spread of a probability

distribution, to quantify the variability in the outcomes of alternative states of nature (Wetzstein,

2005). An increase in the variance of an outcome causes an agent's risk, associated with the

44

corresponding state of nature, to increase. The relationship between risk and variance is a

fundamental component of risk theory, and provides a basis for analyzing the concept of risk

aversion with respect to agents' preferences. Agents' preferences can be defined as risk neutral,

risk averse, or risk seeking. It is generally understood that most agents (i.e. firms, households, or

individual consumers) are risk averse. This idea, originally proposed by the mathematician

Daniel Bernoulli in the 1700s, posits that agents with risk-averse preferences will not play

actuarially fair games (Wetzstein, 2005). An actuarially fair game is defined as "a game with an

expected value of zero, or when the cost of playing the game is equal to its expected value

(Wetzstein, 2005, p.590)." Further, an experiment known as St. Petersburg's paradox reveals that

agents may not participate in games yielding a positive expected payoff (Wetzstein, 2005). The

paradox of the game occurs because the theoretical outcome results in both an expected payoff

and variance of infinity. However, no rational agent would be willing to pay an infinite amount

to play the game (Wetzstein, 2005). Moreover, because the variance of the game is infinite, the

paradox emphasizes that uncertain outcomes are worth less in utility terms than certain ones,

even when the expected payoffs may be large or equal.

With respect to an actuarially fair game, agents with risk neutral preferences are

indifferent between either playing or not playing the game, but those with risk seeking

preferences will play the game (Wetzstein, 2005). The measurement of these preferences

requires the derivation of an agent's utility function. These functions relate the outcomes of

choices to "single-valued indices of desirability" and in theory serve as exact representations of

preferences (King and Robinson, 1981, p.510). Although expected utility functions require

explicit information regarding agents' preferences, the inability to accurately measure these

45

preferences does not produce reliable evaluations of alternative choices (King and Robinson,

1981).

4.1.3 Efficiency Criteria

The shortcomings of expected utility theory motivated researchers to develop efficiency

criteria that more accurately determine the optimum choice among a set of risky alternatives.

The basis for this methodology are efficiency criterion, or decision rules, that provide a partial

ordering of choices for agents whose preferences conform to a specified set of conditions (King

and Robinson, 1981). When actual preferences are unknown, researchers determine the risk

efficient set of choices by making assumptions concerning an agent's preferences and/or various

approximations of the decision probability distributions (Wetzstein et al., 1988).

4.1.3.1 Expected Value Analysis

Wetzstein et al. (1988) and King and Robinson (1981) analyze popular efficiency criteria

cited in the literature. Among the most prominent are mean-variance (EV) and multiple versions

of stochastic dominance (SD) analysis. Expected value analysis is a necessary condition, for all

decision criteria under risk aversion, for one distribution to dominate another as it examines the

first moment of the decision density function (Wetzstein et al., 1988). The first through fourth

moments of a density function are defined as the mean, variance, skewness, and kurtosis,

respectively. The first moment measures the central value of the probability distribution. This is

accomplished by taking a weighted average of all the outcomes, where the weights are the

probabilities associated with each random value. Although the expected value, or mean, is the

most commonly used measure of the central value, both the median and mode of a distribution

may be employed when a distribution contains extreme values or researchers want to identify the

random value with the highest probability of occurrence. However, beyond the first moment,

46

necessary and sufficient conditions for identifying dominant alternatives depend on an efficiency

criterion's risk function (Wetzstein et al., 1998).

4.1.3.2 Mean Variance Analysis and Higher Moments

Mean-variance analysis is an efficiency criterion that ranks agents' choices according to

the first two moments of the probability distribution of the random variable (Bodie et al., 2005).

An assumption of mean-variance analysis is that agents are rational and act to maximize his or

her expected utility. In finance and economics, the random variable of interest is commonly the

payoff value of each outcome associated with a menu of risky choices. The probability

distribution is formed by creating a list for all the payoff values and their respective probabilities

of occurrence (Bodie et al., 2005). The expected value of the portfolio of all the uncertain

payoffs can be interpreted as the utility an agent receives as a function of the probability

distribution from his or her wealth:

E(r) =∑ (4.3) =

n

ssrs

1)()Pr(

Where s = 1,…..n represents all the possible scenarios, r(s) is the payoff from scenario s, and

Pr(s) is the probability of each payoff (Bodie et al., 2005). The risk of this portfolio is

determined by analyzing the distribution's higher moments. The variance, as the second central

moment, describes all the individual outcomes' deviations from the mean. The most frequently

used computation of the variance is the expected squared deviation from the mean:

= 22σ ∑=

−n

srEsrs

1)]()()[Pr( (4.4)

Mean-variance employs the first and second moment to determine an agent's utility

maximizing portfolio. More specifically, "for any two distributions G and H, G dominates H if

47

uG ≥ uH and holds with at least one strict inequality, where u and denote a

distribution's mean and variance, respectively (Wetzstein et al., 1988, p.171)." According to

(Samuelson, 1970), the first two moments are adequate to describe an agent's probability

distribution function if the distribution is "compact." Compactness assumes that an agent is able

to control most of the risk inherent to his or her portfolio. Thus, theoretically an agent can

construct a risk efficient set of alternative choices where no other set can provide a higher

expected value given the same level of risk for all sets.

22HG σσ ≤ 2σ

However, mean-variance may not be able to correctly describe utility functions of a

higher power than the quadratic or if the function violates the assumption of normality

(Wetzstein et al., 1988). Because the variance characterizes the risk of a portfolio on the basis of

the random variables' squared deviations from the mean, both good and bad deviations from the

mean are given the same weight (Bodie et al., 2005). If risk preferences are asymmetrical,

agents' do not perceive large rewards and losses to have the same effect on risk and this may lead

to an incorrect ranking of choices under uncertainty (Wetzstein et al., 2005).

Further, because mean-variance is limited to defining a distribution up to its second

moment, it may be unable to completely define distributions which are asymmetric, or skewed,

around the mean. Skewness is non-dimensional measurement describing the asymmetry of a

distribution. It strictly characterizes the shape of the distribution and can be mathematically

defined as:

M3= ∑ 3

=

−n

srEsrs

1)]()()[Pr( (4.5)

It is theoretically possible to have two distributions that have the same mean and variance, but

differing degrees of skewness. In this case, measurement of the third moment emphasizes

48

greater deviations about the mean, and thus gives more weight to the tails of the distributions

(Bodie et al., 2005). Risk-averse agents will act to minimize large negative deviations from the

mean. A positive skewness describes an asymmetric distribution characterized by frequent but

small losses and less frequent but extreme gains. Thus, a negatively skewed distribution can be

described as having a long tail extending outward toward the more negative values of the random

variable. Therefore, a risk-averse agent will choose a positively skewed distribution over a

negatively skewed distribution, assuming the distributions are mirror images around the mean, as

it is less likely to result in large unexpected losses (Bodie et al., 2005).

Similar to the variance, the fourth moment provides a measure of a distribution's extreme

values (Bodie et al., 2005). As with skewness, kurtosis is a pure, non-dimensional number that

describes the peakedness or flatness of a distribution. Measurements of kurtosis indicate the

likelihood of a distribution's extreme values, in addition to its shape, with larger values

indicating greater uncertainty (Bodie et al., 2005). Leptokurtic describes a distribution with

positive kurtosis, or peaked shape, and platykurtic describes a distribution that has a relatively

flat top when compared with the normal or Gaussian distribution (Bodie et al., 2005).

Mathematically the kurtosis is defined:

M4= 4 ∑=

−n

srEsrs

1)]()()[Pr( – 3 (4.6)

where the (− 3) provides for a kurtosis of (0) in the instance of a normal distribution. 4.1.3.3 Stochastic Dominance Analysis

Stochastic dominance analysis (SD) is an efficiency criteria used extensively by

agricultural economists to determine the risk efficient set of alternatives available to producers

when faced with uncertain outcomes. Different producers may experience similar production

49

problems across regions without regard to an operation's size or knowledge. Ranking

alternatives must therefore use a methodology that is not reliant on obtaining producers'

preferences of risk. Stochastic dominance allows for a ranking of alternatives based on a

minimal set of assumptions concerning producers' risk preferences. Researchers have developed

multiple variations of stochastic dominance but its two basic criterions are first-degree and

second-degree stochastic dominance.

4.1.3.3.1 First-Degree Stochastic Dominance

First-degree stochastic dominance (FSD) is the most general of all SD analyses. The

only economic axioms required for FSD are that agents maximize their utility of expected wealth

and are non-satiated in consumption, or agents' prefer more to less. The axiom of non-satiation

requires an agent's utility function to be monotonically increasing where the first derivative is

strictly positive, i.e. U1(y) > 0 (Anderson, 1977). A requirement of FSD is that population

probability distributions of the random variable be stated in terms of cumulative probability

distribution functions (CDFs) (Anderson, 1977). If (y) denotes a random variable with a

distribution function F(y), then the relationship between the distribution function F(y) and the

probability density function for the random variable (y), noted as F'(y) or f(y), can be expressed

as:

F(y) = (4.7) tdtfy∫ ∞−

)(

where t is used as the variable of integration (Wackerly et al., 2002). Comparing two continuous

CDF's F1 and G1 defined over an interval of [0,1], with each representing a set of risky

alternatives, " F is said to dominate G in the sense of first-degree stochastic dominance (FSD) if

F1(y) ≤ G1(y) for all possible (y) in the range [0,1] with at least one strong inequality (i.e., the <

holds for at least one value of (y) (Anderson, 1977, p.282)."

50

For example, if Y is a random variable representing an agent's given level of wealth, and

F1 and G1 represent the cumulative distribution functions of the rates of return on two mutually

exclusive assets, then for any Y, the probability that the rate of return on asset F is less than Y is

less than that for asset G (Huang and Litzenberger, 1988). Graphically, this scenario can be

observed when for a given level of wealth, the cumulative distribution of asset F lies to the right

of not only asset G but all other assets. This does not mean that the realized rate of return for

asset F is always greater than asset G (Huang and Litzenberger, 1988). It only infers that a

necessary condition for FSD is that the expected return of asset F must not be less than the

expected return of all other assets (Huang and Litzenberger, 1988). A necessary and sufficient

condition for FSD to hold is that the entire cumulative distribution of the dominating asset,

alternative, or state lie to the right of the dominated distributions.

The axiom of non-satiation proves insufficiently restrictive to produce a complete

ranking of risky alternatives in many cases of agricultural decision analysis. Although FSD is

transitive, producer preferences must be more strictly defined in order to reduce the number of

undominated alternatives that may result from this efficiency criterion (Huang and Litzenberger,

1988). FSD ranks alternatives by evaluating the first moment of the random variable's

probability distribution, but neglects the risk component when calculating the dominant states.

This may lead to conclusions where producers are not able to unambiguously state that he or she

prefers risky asset F to risky asset G (Huang and Litzenberger, 1988). In such situations the

cumulative distributions of the two assets may cross resulting in one asset having a higher mean

than the other, but also having a higher degree of risk associated with the payoff.

51

4.1.3.3.2 Second-Degree Stochastic Dominance

Second-degree stochastic dominance (SSD) eliminates dominated or inefficient

distributions from the FSD set (Anderson, 1977). This is accomplished by adding the

assumption of risk aversion to the decision making process with respect to agents' preferences.

Huang and Litzenberger note that," risk-averse agents may have utility functions that are not

monotonically increasing (Huang and Litzenberger, 1988, p.45)." However, they prove that

necessary and sufficient conditions required of second-degree stochastic monotonic dominance

are equivalent in theory to those of SSD. Therefore, the presumption that an agent's utility

function over the range [a, b] of possible payoffs is not only monotonically increasing but also

strictly concave can be understood as requirement of the measurements employed in this study

(Anderson, 1977).

Similar to first-degree stochastic dominance, SSD ranks alternative states by first

interpreting their respective probability density functions as CDFs. A distribution may be

considered dominant if it "lies more to the right in terms of differences in area between the CDF

curves cumulated from the lower values of the uncertain quantity (Anderson, 1977, p. 284)."

Huang and Litzenberger mathematically define the necessary and sufficient conditions for SSD

given two mutually exclusive distributions associated with assets A1 and B1. A1 is determined to

be dominant to B1 if and only if:

E [ r~ A] ≥ E [ r~ B] (4.8)

And

S(y) ≡ ∫ ≤−y

BA dzzFzF0

0))()(( ]1,0[∈∀y (4.9)

The inequality (4.8) describes the sufficient condition for SSD where the expectation associated

with the rate of return for asset A must be greater than or equal to the expected rate of return for

52

asset B. S(y), which describes the sum of the areas between the distributions, must be

continuous and negative. A necessary condition for A1 to be SSD over B1 is that S(y) must be

equivalent to the difference between the integrals of the two distributions A1 and B1 where (z)

represents a given state of wealth (Huang and Litzenberger, 1988). SSD works with a more

restrictive set of assumptions about risk preferences to determine the most efficient set of

alternatives, and this implies that any state which is FSD over another state is also guaranteed to

be SSD.

Risk-averse agents seeking to maximize utility will never prefer a dominated distribution

(Anderson, 1977). Therefore, a second-degree stochastically efficient set of alternatives will be

comprised of only non-dominated distributions, and any further reduction of this set will require

additional assumptions concerning risk preferences (Anderson, 1977).

4.1.3.3.3 Ranking Chemical Alternatives

This study used SSD in order to develop an efficient set of alternative chemical choices

for Georgia bell pepper producers faced with the MeBr phase-out. This procedure was chosen

instead of the mean-variance criterion (EV) for the following reasons: As previously noted, EV

may lead to an incorrect ranking if an agent's utility function is greater than a quadratic

functional form. Since SSD does not require a specific functional form of agents' utility

function, its use as a ranking mechanism for a broad section of producers is more justified. Next,

EV's assumption of a normal probability distribution of the random variable may not accurately

characterize agents' risk preferences. SSD overcomes this problem by making valid

generalizations concerning producers' behaviors. The additional assumptions concerning risk

preferences employed by SSD go beyond FSD to further reduce all available chemical choices

into a more efficient set.

53

4.2 Interpretation of Dominance Analysis, Enterprise Budget, and Programming Model

4.2.1 Dominance Analysis Our methodology began by using bell pepper price and yield data obtained from three

separate field trials during the spring and fall harvests of 2002 and 2003. The trials were

conducted at the Ponder Farm, University of Georgia's Experimental Plots Rural Development

Center located in Tifton, Georgia. The Center's agricultural economists, plant pathologists, and

horticulturalists determined the initial set of chemicals to be tested through research of the

scholarly literature: Thomson, 1991; Harris, 1991; Stall, 2000; Csinos et al., 1997 (Culpepper

and Langston, 2000). Each alternative fumigant, or combination of fumigant and herbicides, was

defined as a "production method" available to Georgia producers for the purposes of this study.

Before ranking each method, a database of total revenues corresponding to each method

was constructed using their respective wholesale prices and yield data for each harvest period.

Each trial generated peppers of varying quality ranging from the premium, or "Jumbo", pepper to

the lower grades of peppers classified as US 1&2's. Jumbo peppers are consistently preferred to

US 1&2 peppers and thus command higher prices. Jumbo peppers are categorized according to

size as extra-large, large, or medium with larger sizes commanding premium prices. An

additional factor affecting the market price of peppers and considered in the analysis was the

harvest period. Average prices for all grades of peppers, calculated using 2003 price data for

Georgia peppers under the assumption that this year represented a "normal" or representative

year for producers, were higher during the spring harvest than the fall harvest. The average total

revenues for each method were calculated by multiplying the yield of each grade of pepper by

their respective average prices during each harvest. The average prices served as weights to fully

characterize the distribution of peppers obtained from each method and harvest, and ensured a

54

continuity of prices across trials. The average total revenues of each method were then summed

across trials. The resulting figures served as the ranking criteria within the stochastic dominance

framework.

The initial yield data provided twenty-one different production methods that formed the

basis of this study. However, before ranking these methods assumptions were made in order to

increase the number of observations within each set, or method, and reduce the number sets to be

ranked. Those sets eliminated from the ranking procedure included any set of an individual

fumigant, or combination of fumigant and herbicides which was not tested across all three trial

periods. This resulted because these methods lacked a sufficient amount of observations to be

included in the ranking scheme. Additionally, methods that were composed of identical

fumigants or combinations of fumigant and herbicides across all trials, but differed in their

application methods were consolidated and labeled as an individual production method for the

purpose of the ranking procedure. Chemicals are usually applied in the field via broadcast or an

in-ground drip system. This study assumed that most producers will use the broadcast method

due to their familiarity with the system, its ease of use, and the cost savings accruing from its use

on multi-crop farms. After consolidation and elimination, the initial twenty-one methods were

reduced to eight methods.

The goal was to identify the two or three most efficient methods from this set of eight

methods through stochastic dominance analysis. The study employed the statistical software

package Simetar (Richardson et al., 2004) to perform the SSD analysis. The program

constructed the cumulative distribution functions using the average total revenues of each

method as the random variable. In this case, the average total revenues constituted observations

within a set, and could be interpreted as a measure of a given level of wealth corresponding to

55

each production method. The probabilities assigned to each observation were equally weighted

within each set in proportion to the observation's contribution to the whole set. For example,

given five observations within a set, each observation would be assigned a probability of (ρ=.20)

due to each observations equal likelihood of occurrence. After ranking the eight methods, the

three dominant methods that did not incorporate MeBr as the primary fumigant, as well as the

method incorporating MeBr, were chosen for further analysis in the enterprise budget. The set

employing MeBr was chosen to represent the base case from which to provide comparisons with

the dominant alternative methods.

4.2.2 Budget Analysis

Land grant universities develop enterprise budgets for a variety of agricultural

commodities. The budgets support academic research and provide agricultural producers with a

business decision tool tailored to their financial operations. This study developed four versions

of a pepper enterprise budget that represents production methods involving methyl bromide,

alternative fumigants, and herbicides. These versions use the format and base assumptions found

in the 2003 University of Georgia enterprise budget developed for Georgia's fresh bell pepper

and growers market (Fonsah and Rucker, 2003).

The enterprise budget attempts to control for incomplete knowledge concerning input and

output factors of production by systematically organizing data into a framework conducive to

more accurate decision making (Calkins and DiPietre, 1983; Fonsah and Chidebelu, 1995;

Fonsah and Rucker, 2003). For example, experimental data collected through test plots assign a

value to random variables such as yield per acre, while price forecasts attempt to accurately

predict future returns based on historical data. The budget can not reduce the randomness of

these variables, but it can assist producers by providing a more rigorous analysis concerning the

56

probability of future events. Obtaining the most accurate data available for all input factors is

imperative to implementing a successful budget. The enterprise budget thus provides the optimal

return per unit of enterprise using estimates of the combination of available inputs (Calkins and

DiPietre, 1983; Fonsah and Chidebelu, 1995; Fonsah and Rucker, 2003).

The enterprise budget represents a single point on the production function, and serves as

a piece of a whole farm budget that analyzes a producer's entire operations (Calkins and DiPietre,

1983). Using multiple budgets (i.e. enterprise, partial, and cash flow), producers' goals are to

anticipate all income and expenses of the organization in an effort to maximize net farm income

(Calkins and DiPietre, 1983; Fonsah and Chidebelu, 1995; Fonsah and Rucker, 2003).

Successful budgets allow producers to answer questions concerning farm operations such as the

optimal amounts of machinery and labor to employ in combination, whether to purchase or rent

additional capital inputs, and whether additions to or reductions in capital are needed in alternate

growing seasons (Calkins and DiPietre, 1983; Fonsah and Chidebelu, 1995; Fonsah and Rucker,

2003).

The enterprise budget used in this study is similar to those developed for other

agricultural crops. However, variables included in each component of an enterprise budget may

differ among producers depending upon whether production concerns a new investment, or

whether capital costs are distributed over a multi-crop or mono-crop system. The basic

components of an enterprise budget can be broken down into gross returns, total variable costs,

and total fixed costs per acre.

4.2.2.1 Gross Returns

The expected gross returns were calculated by multiplying estimates of the expected yield

times an expected price per unit (Calkins and DiPietre, 1983; Fonsah and Rucker, 2003). The

57

expected yields used in this study were calculated by summing the average yields for the Jumbo

and US 1&2 peppers under each production method and across trials. Moreover, test plots used

in field trials are not typically developed on an acreage basis. The calculations of a typical

enterprise budget, however, are based on the acre as the standard unit of measurement.

Therefore, original plot yields were extrapolated to conform to the acre unit using measurements

recorded by field scientists.

The expected prices used in this study were calculated by summing the average

wholesale prices obtained for Jumbo and US 1&2 peppers during the 2003 growing season.

Unlike yields, the expected price employed in each of the four final budgets analyzed remained

constant. Average prices and yields for bell peppers were recorded using a bushel-carton as the

standard unit of measurement. This unit is understood by producers to represent approximately

twenty-five to thirty pounds of peppers. This study assumed an average bushel-carton weight of

twenty-eight pounds for all calculations (Gast, 1991).

4.2.2.2 Total Variable Costs

The variable cost section is divided into two main parts: pre-harvest and

harvest/marketing costs. Total variable costs were calculated by summing these two

components. The fumigant and herbicide costs associated with each of the final four production

methods analyzed were addressed within the pre-harvest component of the variable cost section.

More specifically, four separate budgets were created that differed in terms of the yields and

costs associated with the fumigants and herbicides employed for each unique production method.

The marginal cost of fumigants and herbicides may vary according to the size of producers'

operations. However, this study assumes that a producer's cost schedule for fumigants and

herbicides applies to a forty acre representative farm for the purposes of the budget analysis.

58

This is because producers will not invest all of the necessary equipment if only one acre is

farmed. The prices for all chemicals used in this study were obtained from Hendrix and Dale

located in Tifton, Georgia. Chemical prices for 2003 were used in order to ensure continuity

with the wholesale prices received for 2003 peppers. Average chemical prices were calculated

according to a range that is inherent to the pricing structure of many chemical dealers. All prices

were based on typical application rates for Georgia pepper growers. Because growers purchase

chemicals in bulk quantities, those growers with smaller operations will experience a greater loss

of product through waste, thus resulting in a higher total cost, while large growers will benefit

from a input price discount. Additional costs within the pre-harvest section include plants,

fertilizer, insecticide, fungicide, plastic, drip tape, machinery, labor, land rent, irrigation, and

interest on operating capital. The harvesting costs component of the variable cost section

includes picking and hauling, grading and packing, containers, and marketing. The marketing

cost per unit was calculated as 8.50% of the expected price per carton. This value, when

multiplied by the expected yield per acre for each method, resulted in the total marketing cost per

acre.

4.2.2.3 Total Fixed Costs

The fixed cost section in this budget includes machinery, irrigation, land, and overhead

and management costs. Total overhead and management costs per acre were calculated by

multiplying the total pre-harvest variable cost by 15%. After establishing the total variable costs

and the total fixed costs, the total budgeted cost per acre was calculated by summing these

figures. Finally, the gross returns, variable costs, and fixed costs were used to calculate the gross

margins, and the base budgeted net revenue per unit of enterprise. Gross margins were calculated

by subtracting the total variable costs from the gross returns, and the base budgeted net revenue

59

was calculated by subtracting both the total variable and total fixed costs from the gross returns

(Calkins and DiPietre, 1983; Fonsah and Rucker, 2003).

4.2.3 Linking the Budget with the Linear Programming Model

The four risk-efficient production methods obtained from the stochastic dominance

analyses, which included both the MeBr method and the alternative methods, were analyzed

within the framework of their own enterprise budget. Calculations resulting from the budget

analyses provided variables to be transferred and used within a mathematical programming

model. These variables included the gross returns, total variable costs, total fixed costs, labor

costs, equipment costs, and net revenues. These figures and other financial parameters defined by

average farm financial conditions of a sample of crop farms participating under the Georgia

Farm Business Farm Management program served to construct the pre-operating balance sheet

representative of time (T0) within the context of the programming model as seen in (Table 4.1).

Forecasts of these variables, to be used in the programming model, were then developed over a

five year time horizon. Thus, the enterprise budget, representing the static position of producers'

financial operations, provided a platform from which the final analysis of the linear

programming model could be conducted.

4.2.4 The Linear Programming Model

4.2.4.1 Description of Optimization/Simulation Procedure

Mathematical programming models allow both researchers and producers to extend

analyses of agricultural operations beyond the enterprise budget. The goal of these models is to

construct a set of procedures that deals with the analysis of optimization problems (McCarl and

Spreen, 1994). More specifically, in this study the goal of the producer is to optimize his or her

expected utility of accumulated net worth over a specified planning period by solving for the

60

optimal values for a set of decision variables unique to their operations. The programming

model analytically investigates the decision maker's problem by developing a set of algebraic

expressions that attempt to fully characterize the relationships among the decision variables and

constraints inherent to the operation. This study used a linear programming (LP) model where,"

the decision variables were chosen such that a linear function of the decision variables was

optimized and a simultaneous set of linear constraints on the decision variables was satisfied

(McCarl and Spreen, 1994, p.21)."

The necessary components of a LP model are the decision variables, objective function,

and constraints. The resultant levels of a set of decision variables (xj), of which there are n (j=1,

2 …, n), quantify the amount employed of the respective unknowns (McCarl and Spreen, 1994).

The linear programming problem can be stated in matrix format as:

Max Z = CX

s.t. AX ≤ b

X ≥ 0

where Z is a matrix of the total objective value, C is a column vector representing the

contribution of each unit of X to the objective function, A is the use of the items in the ith

constraint by one unit of xj, and b is the upper limit imposed by each constraint (McCarl and

Spreen, 1994). The final inequality imposes a non-negativity constraint on the decision

variables.

The LP model assumes risk-neutrality on the part of the decision maker because the

model strictly accounts for the net changes in final wealth regardless of the agent's risk

considerations (Escalante, 2001). The model operates under a five-year planning horizon in its

determination of the producer's optimal net accumulation of wealth or net income over this

61

period. The final accumulation of net worth is calculated by accruing the values of farmland,

equipment, and cash balance at the end of the planning horizon less all financing charges

contracted over the same period (Escalante, 2001).

The final goal of optimizing producers’ net worth is predicated and works in conjunction

with a simulation procedure that begins with the analysis of a base-case representative farm

model. The base model accounts for the activity measures and financial conditions of a typical

producer’s operations. Adjustments to components of the base model reveal variations in the

profitability and cost structure of the farm’s operations. In this study modifications to the base

model are conducted by adjusting the production method as defined by the fumigants and

herbicides employed as chemical controls on the farm. More specifically, the production method

of MeBr and an accompanying menu of herbicides were chosen to serve as the base-case. The

three additional production methods as ranked by SSD were then tested against this base-case.

4.2.4.2 Characteristics of the Base Farm Model

In lieu of a singular analysis of the objective values resultant of all four production

methods as the sole determinant of efficiency, considerations were also given to measures of

liquidity and leverage when interpreting the viability of operations. The financial ratios used to

identify these measures are the tenure, current, debt-to-asset, and off-farm investment ratios.

The preferred production method will be characterized as the strategy that provides for the

largest accumulated net worth in proportion to the production’s total outstanding obligations.

The Base Farm Model as defined by Escalante (2001) is composed of sets of constraints

and requirements Table (4.2) that are in turn defined by the farm’s financial, marketing, and

production activities Table (4.3). Further, requirements and activities of a representative farm

must account for variability in cash flows over the planning period. Thus, the model assumes

62

two sub-periods for all cash transfers and defines these activities through the two cash transfer

equations (TRANSF1) and (TRANSF2).

4.2.4.3 Activities, Constraints, and Requirements

i) Initial Endowments

This model assumes the production requirements, constraints, and levels characteristic of

a Georgia pepper farm. Producers are subject to a set of resource constraints that define the

endowments of the operation at time T (0), one year prior to the five-year planning period.

These initial constraints are land (LAND0), equipment (EQPT0), existing long-term loans

(FINLAND), existing intermediate-term loans (MEDCRED), ending cash balance (CASH0), and

current asset position (CURAST0) (Escalante, 2001). The cash balances and other current assets

(OTHCUR0) at the end of year T (0) are transferred to the first sub-period of the first year to

insure coverage of operating expenses and investment requirements (Escalante, 2001).

Downpayment rates for both land (LNDDPMT) and equipment (EQDPMT) are assumed to be

20% and serve as scalars throughout the planning period.

ii) Production Constraints and Requirements

Total farm size in terms of production and marketing (ACPROD) throughout the

planning period can be identified as the sum of the number of acres owned at time (t) (ACBUY)

and the amount of acres that are cash rented (ACRENT) during the same period. This model

assumes an initial farm size of 362 acres to be divided between these two activities at time T (0)

as 300 acres purchased and 62 acres rented. Farm real estate values per acre (LANDVAL (t))

were obtained through the USDA, projected over the five-year horizon, and based on an initial

value of $2,150.00 per acre (USDA NASS, 2004). Figures were also obtained for the initial

value of a cash rented acre (RENT (t)), $105.00, and projected over the planning period. Both

63

the farm real estate values and cash rent figures were forecast using a growth rate of 5.6% under

the assumption that values would approximate the previous ten-year trajectory. Adjustments

were then allowed for these activities in each following period.

Other mechanisms to increase farm size other than purchasing or renting were not

included in this model. Further, a property tax (PROPTAX) of approximately $24.00 per acre

per year is assessed against all owned land throughout the planning period, and payments are

made in two installments over each year (t). All expenses accruing to operations other than cash

rents are entered into the model with positive coefficients with the assumption that payments are

made during the first sub-period (Escalante, 2001). Gross-crop revenues are realized during the

second sub-period, and identified as an inflow taking a negative coefficient for the purposes of

the model.

iii) Machinery and Labor Requirements

The model recognizes the machinery requirement imposed upon production activities

both through the owned equipment component (EQBUY) and additional purchases of new

machinery over the planning period. Adjustments to equipment values over the planning period

are made by acknowledging the cumulative effects of both depreciation and inflation (Escalante,

2001). Inflationary effects within the model were imposed at an approximate historical average

of 3%. Whereas machinery requirements are expressed in units of dollars per acre, labor

requirements are expressed in hours per acre. Total labor requirements are a combination of

direct and annual overhead labor (OVRLAB) requirements that can be satisfied through hired

labor (hirelab (t)). Hired labor is assumed to be the only form of labor available to producers in

this study, and the model disburses payments to labor over two sub-periods.

64

iv) Credit, Cash, and Financing Components

Credit is available to the producer for three main types of financing requirements. Short-

term loans (SHTCRED) are available for the immediate financing of operations. The financing

terms for these loans are for six months with the principle and interest payments (1 + i) accruing

over the first period and coming due on the second sub-period of the year. Intermediate-term

loans (MEDCRED) are typically used to finance equipment purchases and are assumed to carry a

fixed interest rate over the five-year planning period. Long-term loans (FINLAND) carrying

fixed interest rates over a 20 year period are used to finance additional purchases of land. Both

the (FINLAND) and (MEDCRED) components are amortized annually (P & I) with payments

due at the second sub-period of the year (Escalante, 2001). Finally, the model assumes a 7%

interest rate for the duration of all outstanding loans.

The model further allows for purchases of equipment and land through cash components

(casheq (t)) and (cashland (t)), respectively. However, financing involving cash is subject to a

minimum down payment requirement and upper limits on distributions according to the

producer’s cash position (Escalante, 2001). Thus, solutions of the model may allow for

financing of up to 90% of purchases, or self-financing according to the producer’s current cash

position.

v) Consumption and Non-farm Investments

The expenditures (famcon (t)) and off-farm investments (offinv (t)) of the representative

farm family modeled in this study are based on the historical levels of family living expenses and

the historical yields of United States Treasury bills. Family consumption is accounted for in

each sub-period of each year, constrained by a fixed upper bound in each year, and assumed to

increase over each period (t) (Escalante, 2001).

65

The total off-farm income generated from off-farm investments (i.e. government

securities) is not a significant activity measure in this model. Greater emphasis is given to on-

farm revenue generating activities (Escalante, 2001). However, capital gains (1 + i) are

recognized by the cash transfer equations as inflows in the second sub-period resulting from the

outflow of investment made during the first sub-period.

vi) Taxable Net Income

The producer’s gross taxable income is calculated by summing net operating margins and

off-farm income generated through investments. The taxable net income (txincm (t)) considers

allowable deductions for such items as the interest payments on intermediate and long-term

loans, property taxes, cash rents, depreciation of business equipment, and wages (Escalante,

2001). A tax rate of 40% accounts for all applicable local, state, and federal taxes imposed on

the farm enterprise. The value of the taxable net income for time T (0) was $20,171.

66

TABLE 4.1

Pre-Operating Balance Sheet

Current Assets Current Liabilities $241,688 Cash $124,255 Non-Farm Investments $100,000 Other Current Assets $379,861 Intermediate Liabilities $422,010 Total Current Assets $604,116

Long-Term Liabilities

$350,099Intermediate Assets Machinery & Equip. $468,900 Total Liabilities $1,013,797

Fixed Assets Land $645,000

Buildings $125,218$770,218 Net Worth $829,437

Total Assets $1,843,234 Total $1,843,234

67

TABLE 4.2

Description of Constraints and Requirements

Sign of

Constraint Unit Right

Hand Side Value

INITIAL RESOURCE AND DEBT LEVELS

=

Acres

B

Initial Land Owned = Dollars B Initial Equipment Owned = Dollars B Initial Long-Term (Real-Estate) Loan = Dollars B Initial Intermediate-Term (Equipment) Loan = Dollars B Initial Cash Available = Dollars B Initial Amount of Current Assets = Dollars B LAND REQUIREMENT Number of acres produced and marketed ≤ Acres 0 Long-term financing for real estate purchases = Dollars 0 Land downpayment requirement in year t ≥ Dollars 0 EQUIPMENT REQUIREMENT Machinery requirement to prod. & mkt. crops in year t

≤ Dollars 0

Intermediate financing for equipment purchases in year t

= Dollars 0

Machinery downpayment requirement in year t ≥ Dollars 0 LABOR REQUIREMENT ≥ Hours 0 FAMILY LIVING EXPENSES = Dollars B OFF-FARM INCOME = Dollars B CASH FLOW PER YEAR (TRANSFERS) = Dollars 0 TAXABLE INCOME = Dollars 0

68

TABLE 4.3

Description of Portfolio Activities

Code Unit Description

ACPROD (t) Acres Number of acres of peppers produced and marketed in year t ACBUY (t) Acres Number of acres purchased in year t ACRENT (t) Acres Number of acres rented in year t FINLAND (t) Dollars Long-term loan at fixed rate over 20 years used in year t EQBUY (t) Dollars Amount of equipment purchased in year t MEDCRED (t) Dollars Intermediate-term loan at fixed rate used in year t SHTCRED (t) Dollars Short-term (operating) loan used in year t cashland (t) Dollars Cash used to purchase land in year t casheq (t) Dollars Cash used to purchase equipment endcash (t) Dollars Ending cash balance at period p of year t famcon (t) Dollars Family consumption in year t offarm (t) Dollars Off-farm income in year t txincm (t) Dollars Taxable net income at year t offinv (t) Dollars Investments in treasury bills (off-farm investments) at year t hirelab (t) Hours Amount of hired labor in year t

69

CHAPTER 5

ANALYSIS OF RESULTS

5.1 Stochastic Efficiency-The Final Four Production Methods

As discussed in Chapter four, the initial set of twenty-one production methods was

consolidated into a more efficient set of eight production methods Table (5.1). Two of these

eight methods consisted of 1) MeBr as a stand alone fumigant and 2) MeBr used in combination

with a menu of herbicides. The remaining six methods were composed of three classes of

fumigants without the use of herbicides, corresponding to three mutually exclusive production

methods, and the same three classes in combination with an identical menu of herbicides as

found in the base-case using MeBr. Therefore, the four classes of fumigants identified in the set

of eight production methods are as follows:

i) Methyl Bromide

ii) KPAM in combination with C35

iii) Telone II in combination with Chloropicrin

iv) C35 in combination with Chloropicrin

Employing SSD as an efficiency criteria resulted in the ranking scheme identified in Table

(5.1). The resultant ranking scheme is consistent with the expectation that those fumigants

employed in combination with a menu of herbicides will lead to greater yields as a result of their

ability to more efficiently control for a host of weeds and pests. None of the four production

methods not employing herbicides ranked as one of the final four efficient methods.

70

An expected result of the analysis was that the method employing MeBr with herbicides

would rank as the most efficient set out of the final four sets. However, the final four efficient

methods resulted in the ranking scheme listed below as identified by their coefficients of

variation (C.V.). The following listed numbers in bold italics correspond to the methods’

identification numbers in Tables (5.1) and (5.2), and hereafter will uniquely identify the methods

within the text.

3. C35 + KPAM (.058)

1. Methyl Bromide (.268)

5. Telone II + Chloropicrin (.308)

7. C35 + Chloropicrin (.323)

The coefficient of variation is a traditional economic measurement that allows for comparing an

investments standard deviation to its expected return. In this case, the expected return or mean is

the average total revenue per test plot per production method. The (C35+KPAM) method ranked

as the most dominant method with its average revenues exceeding those of MeBr by $1.77,

(Telone II and PIC) by $3.17, and (C35 + PIC) by $3.25.

The cumulative distribution chart shown in Table (5.2) gives a graphical representation of

SSD. As can be seen, methods (3, 1, 5, & 7) lie significantly to the right of methods (2, 4, 6, &

8) in terms of the total areas under the cumulative functions. In this graph, the (X) axis

represents the average total revenues per plot and the (Y) axis represents the probability of

receiving this amount of wealth or less. It is obvious that the conditions for FSD do not hold due

to the crossing of the cumulative functions. However, for method (3) one can observe that for

most levels of wealth (x), the probability of obtaining that amount or less is greater than almost

71

all other methods. After performing the SSD calculations, methods (3, 1, 5, & 7) were analyzed

within an enterprise budget.

5.2 Results of Enterprise Budget Analyses

The results of the enterprise budget allow for a comparison of the costs, revenues, and

financial efficiency measures among the final four production methods. Similar to the

programming model, analyses of the budgets considered the production operation employing

MeBr as the primary fumigant as the base-case model from which the other three methods were

compared. Tables (5.3) and (5.4), respectively, identify the activity measures, efficiency ratios,

and breakeven statistics when the budgets were analyzed using experimental yield results and

when expected yields were held constant. The purpose of separate analyses was to isolate key

activity measures within the production process and assess their individual contribution to the

entire budget.

Activity measures analyzed for each budget were the representative farm’s yield, gross

returns, total variable costs, total fixed costs, and net returns. Total variable costs were broken

down into two components, fumigant costs and other variable costs, in order to examine each

fumigant’s contribution to the farm’s economic viability. These contributions were appraised

through three farm efficiency ratios; 1) Net returns / Gross Returns, 2) Variable Cost / Gross

Returns, and 3) Fumigant Cost / Gross Returns.

Finally, break-even statistics were calculated for each method. Break-even analysis is

traditionally a type of simulation or “what if” analysis with the objective of determining what

volume of production, value of sales, or value of production inputs will result in the firm

experiencing zero net revenue or “breaking-even.” This study calculated the farm’s break-even

revenues, yields, and prices according to each production method under both the experimental

72

and constant yield scenarios. All activity measures, efficiency ratios, and break-even statistics

were calculated on a per acre basis.

5.2.1 Examination of Costs

5.2.1.1 Other Variable Costs

The final four production methods analyzed in the budget were subject to identical

variable cost measures. Most pre-harvest variable cost values other than fumigants remained

through each budget. Variations in the pre-harvest variable costs were primarily attributed to the

differences in the interest charged on operating capital. Appendices (A-D) give a visual picture

of the basic components required for the representative pepper farm. Each acre required 174

pepper plants, costing $5.50 a plant, or $957 an acre. Lime was applied at a rate of one ton per

acre with an associated cost of $26/acre. The combination of base and sidedress fertilizers cost a

total of approximately $214/acre. This study assumed that the prices attributed to insecticides,

fungicides, nematicides, and herbicides remained constant for all four methods and cost the

grower an approximate total value of $996/acre. However, producers may in actuality

experience discounts in these areas due to business negotiations, the ability to store extra

product, or bulk price reductions.

Low-density polyethylene plastic (LDPE), which is used to control for weeds and pests,

is charged to growers by the roll, and when combined with the cost of its removal cost nearly

$267/acre. Operations required 8700 feet of drip tape per acre at a cost of $.02 / foot or

$174/acre. Machinery costs are calculated at five hours per acre, $21/acre, for a total of

$105/acre. The labor component of the budget is broken into two parts; transplant labor and

non-transplant labor. For this study the cost of both were calculated at $8.00/hour in accordance

with prevailing regional prices. A sum total of 53 man-hours were required per acre resulting in

73

a total cost of $424/acre. Land-rents were expected to cost the grower $105 per acre while

irrigation, consisting of machinery and labor costs, held a total cost per acre of $65. Finally,

differences in the interest owed on producers’ operating capital were significant with MeBr

paying ($168), C35 + KPAM ($185), C35 + PIC ($213.48), and Telone II + PIC paying ($174).

Thus, pre-harvest variable costs for all methods ranged from a low of $3,905 for MeBr to a high

of $4,534 for the C35+PIC method.

Next, the total harvest and marketing costs vary according to yield as greater yields

require more picking and hauling (P & H), grading and packing, containers, and marketing. The

cost per unit for all components remained the same for all methods. Measured in cartons, the

price of (P &H), $1.25, accounts for approximately 20% of the $6.29 total harvesting and

marketing costs. Grading and packing, at a price of $2.75/ctn. represents 44% of the total,

containers ($1.30/ctn.) equal 21%, and marketing at $0.99/ctn makes up 16%.

5.2.1.2 Fixed Costs

The fixed costs for the producer consist of machinery, irrigation, land, and overhead and

management (O & M). Overhead and management, costing $645/acre, is the largest fixed cost

for producers and represents approximately 70% of all total fixed costs ($919/acre). The second

largest cost component is irrigation at $221 per acre, or 24% of the total. Fixed machinery costs

of $54/acre represent 6% of total fixed costs. The total budgeted cost per acre is calculated by

summing both the total variable costs and the total fixed costs per acre. The range of this cost

over all methods was a low of $12,107 for MeBr to a high of $13,388 for C35+KPAM.

5.2.1.3 Costs per Carton

The enterprise budget then calculates for several of the previously listed costs on a per

carton basis (c.p.c.). Table (5.5) illustrates the differences in costs per carton among the four

74

methods. Differences in the pre-harvest variable costs, due to the discrepancies in the prices of

the fumigants, and the fixed costs, caused by the fluctuating charges associated with overhead

and management, produce a range for the total budgeted cost per carton of $10.23 for C35 +

KPAM to $11.05 for C35 + PIC.

5.2.2 Net Revenues

In this study, calculations of gross and net revenues were dependent on a constant

expected price of $11.68 received by producers. Gross revenues were calculated by multiplying

expected yield times expected price. Base budgeted net revenues for each method were

calculated after deducting all variable and fixed costs from the gross figures. Gross and net

revenues differed for both the experimental and constant yield scenarios due variations in the

overall cost structures of each method.

5.2.3 Experimental Yield Results

Experimental yields were obtained through field trials conducted in Tifton, Georgia.

Yields indicated in Table (5.3) indicate the number of cartons per acre. C35 + KPAM produced

the largest yields with 1,298 cartons per acre. Methyl bromide and C35 + PIC varied slightly in

yield at 1,167 ctn./acre and 1,115 ctn./acre respectively. Telone II + PIC produced the smallest

yield of 1,119 ctn./acre. Although the C35 + KPAM method incurred total variable costs that

were approximately $1,200 more than MeBr, its greater yield resulted in net returns that were

$248 more than MeBr and $1,036 greater than the C35 + PIC combination, which experienced

the lowest net returns at $732 per acre.

5.2.3.1 Efficiency Ratios-Experimental Yields

The efficiency ratios indicated in Table (5.3) measure a method's ability to generate net

revenue, or profits, from its gross returns, as well as inefficient cost allocations among the four

75

methods. For example, MeBr was able to generate a net return of $0.13 for every $1.00 of gross

returns while C35 + PIC only generated $0.02 per $1.00 from approximately the same base of

gross returns. This results predominantly from the discrepancies in fumigant costs between the

two methods. MeBr incurs a fumigant cost of $405/acre while C35 + PIC cost 2.5 times more

per acre at $1,007. Although C35 +KPAM only earned $0.12 profit for every $1.00 of gross

revenues, and experienced total variable costs nearly 2 times that of MeBr at $786/acre, its

ability to generate 131 more cartons of peppers compensated this method for its higher cost

structure. Telone II + PIC generated $0.09 per $1.00 of gross revenues but its variable costs as a

percentage of gross revenues exceeded MeBr and C35 + KPAM by 2% and 3% respectively.

5.2.3.2 Break-even Statistics-Experimental Yields

Break-even statistics for these four methods give an indication of the units of production

that are needed at a given price in order to maintain economic viability. Consequently, the

reverse question also allows producers to calculate the price that is necessary, for a given level of

production, in order to ensure the farm's profitability. The goal of break-even analyses is to

determine what value of a given variable (x), holding other variables constant, assures a net

return to the firm of zero. From this value, producers can determine what values of (x) must be

attained in order to maintain profitability. Break-even revenues can be found by calculating the

total budgeted cost per acre. Revenues in excess of this figure will lead the firm into profitability

while revenues below this figure will eventually cause the firm to fold. In this study, break-even

yields were calculated holding the expected price of $11.68 per carton constant. In order for

producers to maintain viable operations (MeBr) must yield a minimum of 1,037 ctns. /acre, (C35

+ KPAM) 1,146 ctns. /acre, (C35 + PIC) 1,092 ctns. /acre, and (TII + PIC) must yield 1,024 ctns.

/acre. Alternatively, holding each method's unique experimental yield constant, (MeBr) must

76

receive a minimum of $10.38 per carton, (C35 + KPAM) $10.32, (C35 + PIC) $11.05, and (TII +

PIC) must receive $10.68 per carton. Due to its comparatively high cost of production, (C35 +

KPAM) must generate revenues of $13,388 to maintain profitability. This figure is $1,281

greater than (MeBr), $633 greater than C35 + PIC, and $1,433 more than TII + PIC.

5.2.4 Constant Yield Results

The study assumed constant yields of 1200 cartons per acre as a result of consultations

with field scientists and producers in Georgia concerning realistic assumptions involving the

expected yields generated by MeBr. Because MeBr was taken as the base-case for this study,

holding these expected yields constant allows producers to completely isolate costs within the

budget and determine the effects that these costs have on net revenues (Table 5.4). As with the

analysis on experimental yields, pepper prices were again held constant at $11.68 / ctn. for these

calculations. A result of holding both the price and yield constant was that the gross revenues for

each method were identical in this analysis. However, in this experiment MeBr replaced C35 +

PIC as the production method generating the largest net returns. More specifically, MeBr

generated net returns of $1,700/acre, (C35 + KPAM) produced $1,242/acre, (C35 + PIC)

produced $976/acre and (TII + PIC) generated $1,552 per acre. The total fixed costs remained

the same for each method due to the constant yields.

5.2.4.1 Efficiency Ratios-Constant Yields

Under a constant yield scheme, net returns as a percent of gross returns increased 5% for

(C35 + PIC), 2% for (TII + PIC), and 3% for (C35 + KPAM), but decreased 1% for (MeBr).

Fumigant costs as a percent of gross returns increased 1% for (C35 + KPAM), decreased 1% for

(C35 + PIC) and remained unchanged for (TII + PIC) against the figures from the experimental

analysis.

77

5.2.4.2 Break-even Statistics-Constant Yields

The break-even revenues, as expected, did not change dramatically from the experimental

case. All the variation among these figures was due to the interest charged on operating

expenses contained in the variable cost component of the budget. As expected, break-even

yields for all methods except (C35 + KPAM) increased compared to the experimental yields.

Under the new scenario, (MeBr) needed to yield 1,054 ctns. /acre, (C35 + PIC) 1,116 ctns. /acre,

and (TII + PIC) needed 1,067 ctns. /acre. Assuming yields were held at 1200 ctns./acre, (MeBr)

had to receive $10.26 per carton, (C35 +KPAM) needed $10.64 per carton, (C35 + PIC) required

$10.87 per carton, and (TII + PIC) needed to receive $10.39 per carton to maintain profitability

(Table 5.4). The break-even figures for price reveal that if production were able to generate at

least 1,200 cartons of peppers per acre, growers should have little problem maintaining viability

as the highest price required for profitability out of all four methods was $10.87, which was

$0.81 below the average price received by growers across spring and fall harvests of year 2003.

5.3 Programming Solutions

The LP model delivers solutions to the optimization problem for each period throughout

the five year time horizon. Additions and/or reductions to both the farm’s assets and liabilities

dictate adjustments to the final value of accumulated net worth. These adjustments can be made

through increases or decreases in owned land, cash-rented acreage, equipment, investments, debt,

or any combination of these various decision variables bound within the constraints of the model.

Each simulation of the four production methods began with an identical set of

assumptions concerning certain attributes of the representative farm. For example, beginning

land values and cash rent levels, equipment costs, family consumption, off-farm income and

yields on off-farm investments, labor costs, depreciation schedules, and interest on credit

78

facilities were constant values at time T(0) for all methods. Variables such as the projected

variable costs per acre, gross returns per acre, overhead costs per acre, and net margins per acre

were forecast over the planning period based on calculations made in each method’s enterprise

budget. The enterprise budget analyses employed for each method indicated the economic

viability of operations under all four scenarios. Therefore, the expectation for the solutions to

the programming model was that the farms would increase production over the planning period

until maximum production could be achieved according to the constraints.

5.3.1 Solutions-Activity Measures

5.3.1.1 Production Solutions

Conveying the LP solutions for the production variables as five-year averages, Table

(5.6) indicates that the two dominant methods (MeBr) and (C35 + KPAM) yield similar solutions

to the production problem over the planning period. The LP set an upper limit constraint on total

production of 1,000 acres per method. Although both (MeBr) and (C35 + KPAM) produced the

limit of 1,000 acres each year, (MeBr) purchased seven acres more and rented seven acres less

than (C35 + KPAM) in year T(1).

Note that the solutions for activity measure (ACBUY) indicated in Tables (5.7, 5.8, 5.9,

5.10) are incremental measures that express the additional acres purchased in excess of the initial

endowment of 300 acres. The values for (ACRENT) do not accumulate during the planning

period. Producers rent land for a period of one year, and at the beginning of the next period must

again decide how much acreage to devote to renting. This figure is then carried over and added

to purchases made during the next period, with this process being repeated over the life of the

planning period. Thus, the total owned acres for a given period is the prescribed value for

(ACPROD) less (ACRENT).

79

During year T (2) the model predicted that (MeBr) purchased an additional 93 acres, nine

less than that purchased by (C35 + KPAM). However, (MeBr) rented two more acres than the

latter over the same period. For the remaining three years T (3) through T (5), the amount of

land purchased and rented by each method remained nearly identical. Finally, it is expected that

relaxing the constraint on farm size would result in increased production under (MeBr) and (C35

+ KPAM). However, the results of the model suggest growers would plant more under the (C35

+ KPAM) method when compared to (MeBr) due to a more favorable financial position over the

planning period.

The model solutions for (C35 + PIC) and (T2 + PIC) suggested increases in production

through cash-renting although neither reached the 1,000 acre limit. Over the planning period

(C35 + PIC) cash-rented an average of 586 acres per year and (T2 + PIC) rented an average of

587 acres per year. The farm size solutions for these methods were nearly identical. Thus, the

model did not suggest any additional purchases of land for either method over the planning

period.

Of the four financial ratios provided in Table (5.6), the tenure ratio illustrates the

relationship between the total land owned and total land produced. Tenure ratios of (0.4530) and

(0.4350) for (MeBr) and (C35 + KPAM) indicate that the returns and cash flow available to

producers using these methods encourage land ownership. (C35 + PIC) and (T2 + PIC) realize

tenure ratios of (0.3386) and (0.3382) suggesting that the higher costs associated with ownership,

acquisition costs and taxes, force these methods to rely more heavily on cash-renting.

Further, the model did not prescribe any additional allocations to equipment for methods

(C35 + KPAM) and (MeBr) over five years. For (C35 + PIC), solutions indicated that the

80

producer averaged additional purchases of equipment of $22,416 over five years, and (T2 + PIC)

averaged $32,276 over the same period.

5.3.1.2 Financial Solutions

Discrepancies among solutions for the financial decision variables were pronounced

between the methods (MeBr)/(C35 + KPAM) and (C35 + PIC)/(T2 + PIC). Each of the four

farms began with an initial allocation of off-farm investments totaling $100,000. The former two

methods realized off-farm investments of approximately $1.5 million per year over five years

Table (5.6). Methods (C35 + PIC) and (T2 + PIC) were prescribed equal solutions of $393,840

invested in off-farm activities per year when averaged over the planning period.

All farms were prescribed total current assets, assets that can be easily converted into

cash in less than one year, at time T (0) of $604,116 from a total asset pre-operating base of

$1.85 million. The model revealed that growers using methods (MeBr) and (C35 + KPAM) will

realize cash, non-farm investments, and other current assets positions over the five-year period

up to an approximate average of $45.5 million for (MeBr) and $48.5 million for (C35 + KPAM).

The solutions for the farms’ total assets, or amount of current and fixed assets, for the same two

methods increased to approximately $47 million and $50 million, respectively, as averaged over

the planning period.

However, the current assets position increased less on average for (C35 + PIC) and (T2 +

PIC) to approximately $9.9 million and $16.3 million over five years. Increases in total assets to

$18 million and $17.6 million, respectively, were generated by the model for these two methods.

During the planning horizon decisions are made by producers to either increase investment in

machinery and land, or to maintain his or her current capital position. For (C35 + PIC), heavy

investments were made during years one and five in machinery and equipment. Alternatively,

81

although the five-year averages for the total assets are similar, producers using method (T2 +

PIC) distributed investments in fixed assets more evenly over the planning horizon.

The current liabilities facing producers represent all short-term loans, interest, unpaid

expenses, accounts payable, and other debts coming due within one year. Current liabilities

under each method at time T (0) were $241,688. Note that the figures for (CULIAB) indicated

in the tables are marginal values representing the short-term liabilities of the operation. These

liabilities are not carried over from one period to the next, but are incurred at the beginning of

the period and repaid at the end of each year. However, intermediate-term (MEDCRED) and

long-term (FINLAND) liabilities accounts follow an amortization schedule, and their respective

figures at time T (5) indicate the final balance on these loan accounts. The average five-year

balance on short-term loans decreased to $146,386 for (MeBr) and (C35 + KPAM), while

decreasing to $156,714 for (C35 + PIC) and $162,110 for (T2 + PIC). Total liabilities also

decreased for all methods from a pre-operating base of approximately $1 million. (MeBr) and

(C35 + KPAM) averaged total liabilities of $420,840 over five years. (C35 + PIC) averaged

$453,946, and (T2 + PIC) averaged $469,268 during each period.

The current ratio is a measure of a firm’s liquidity, or an indication of a firm’s ability to

cover its short-term obligations. It is calculated by dividing the firm’s current assets by its

current liabilities. The measure provides a method for comparison among the alternative

production methods. (C35 + KPAM) with a current ratio of (331.59) had the highest liquidity.

(MeBr) was also highly liquid with a ratio of (310.72). (C35 + PIC) measured (221.17), and (T2

+ PIC) registered the lowest liquidity of all methods at (100.60). The ratios for each method

were well above the critical value of two times liabilities and indicated that each had the ability

to quickly pay off short-term debts.

82

Measurements of each method’s total debt to total assets (debt-to-asset ratio) provide a

value from which to compare firms’ financial risk by indicating how much of a firm’s assets

have been financed by debt (Bodie et al., 2005). Debt-to-asset ratios exceeding 65% indicate

excessive debt. Due to the current assets position as prescribed by the model, no method

incurred a significant amount of debt in proportion to its overall financial portfolio. Debt-to-

asset ratios ranged from less than one percent (.008) for (C35 + KPAM) to a high of

approximately (.027) for (T2 + PIC). Finally, the off-farm investment ratio reveals that more

money is able to be allocated outside of the primary revenue generating activities of the farm for

(MeBr) and (C35 + KPAM).

Finally, the accumulated net worth of each simulated farm increased over the planning

horizon. The pre-operating net worth attributed to all methods was $829,437. The net worth

solution prescribed by the model for (C35 + KPAM) ranked first among all methods and

indicated that a producer employing this method would realize an average of $46.6 million over

the planning period. The net worth solution for (MeBr) determined in the first year of simulation

was approximately $11,688,000, and after five years increased to a maximum value of $80.8

million in T (5), indicating an average five-year value of $46.7 million. The third highest five-

year average net worth was prescribed to (C35 + PIC) at $17.5 million. The model designated

the net worth of (T2 + PIC) to be $17.1 million over the planning period. The dominant net

worth positions prescribed by the model for (MeBr) and (C35 + KPAM) can be attributed to the

gains resulting from greater production. A final comparison of the solutions specified by the

model establishes (C35 + KPAM) as not only a viable alternative to MeBr in terms of liquidity,

leverage, and tenure, but also the preferred method in terms of the accumulated value of net

worth.

83

TABLE 5.1

Efficiency Ranking of Final 8 Production Methods Using Second Degree Stochastic Dominance

Production Methods*

(1) (2) (3) (4) (5) (6) (7) (8)

Methyl

Bromide with

Herbicides

Methyl

Bromide without

Herbicides

C35 + KPAM

with Herbicides

C35 + KPAM without

Herbicides

Telone II +

Chloropicrin with

Herbicide

Telone II +

Chloropicrin without

Herbicide

C35 +

Chloropicrin with

Herbicide

C35 +

Chloropicrin without

Herbicide Mean (u)

14.84

12.96

16.61

10.65

13.44

9.98

13.36

9.30

Standard Deviation (S.D.)

3.97

4.70

0.96

5.53

4.14

5.81

4.31

4.79

Coefficient of Variation (C.V.)

0.268

0.363

0.058

0.519

0.308

0.432

0.323

0.515

Ranking**

( 2 )

( 5 )

( 1 )

( 8 )

( 3 )

( 6 )

( 4 )

( 7 )

* Production methods ranked according to the total revenues generated per plot across all trials ** Increasing rank number corresponds to decreasing level of efficiency

84

TABLE 5.2Cumulative Distribution Functions for Final 8 Production Methods

0.00

0.10

0.20

0.30

0.40

0.50

0.60

0.70

0.80

0.90

1.00

0 2 4 6 8 10 12 14 16 18 20

Cum

ulat

ive

Prob

abili

ty

1 2 3 4 5 6 7 8

PRODUCTION METHODS

85

TABLE 5.3 Financial Efficiency and Break-even Analyses Using Experimental Yield Results

Production Methods Methyl

BromideC35 &KPAM

C35 &

Chloropicrin Telone II &

ChloropicrinActivity Measures Yield (cartons) 1,167

1,298

1,155 1,119

Gross Return $13,627 $15,156

$13,487 $13,070

Fumigant Cost $405 $786

$1,007 $528

Other Variable Cost $10,869 $11,682

$10,794 $10,548

Total Variable Cost $11,247 $12,468

$11,801 $11,076

Total Fixed Cost $860 $920

$954 $879

Net Returns $1,520 $1,768

$732 $1,115

Efficiency Ratios (Rounded to the nearest % )

Net Returns / Gross Returns 13.0% 12.0%

2.00% 9.00% Variable Cost / Gross Returns 83.0% 82.0%

88.0% 85.0% Fumigant Cost / Gross Returns 3.0% 5.0%

8.0% 4.0% Break-even Statistics

Break-even Revenues $12,107 $13,388

$12,755 $11,955

Break-even Yield 1,037 1,146

1,092 1,024

Break-even Price $10.38 $10.32

$11.05 $10.68

86

TABLE 5.4 Financial Efficiency and Break-even Analyses Using Constant Yield Results

Production Methods

Methyl Bromide

C35 &KPAM

C35 &

Chloropicrin Telone II &

Chloropicrin

Activity Measures

Yield (cartons)* 1,200

1,200

1,200 1,200

Gross Return $14,016 $14,016

$14,016 $14,016

Fumigant Cost $405 $786 $1,007 $528

Other Variable Cost $11,051 $11,068 $11,078 $11,057

Total Variable Cost $11,456 $11,854

$12,085 $11,585

Total Fixed Cost $860 $920

$954 $879

Net Returns $1,700 $1,242

$976 $1,552

Efficiency Ratios (Rounded to the nearest % )

Net Returns / Gross Returns 12.0% 9.0%

7.0% 11.0% Variable Cost / Gross Returns 82.0% 85.0%

86.0% 83.0% Fumigant Cost / Gross Returns 3.0% 6.0%

7.0% 4.0% Break-even Statistics

Break-even Revenues $12,316 $12,774

$13,040 $12,464

Break-even Yield 1,054 1,094

1,116 1,067

Break-even Price $10.26 $10.64

$10.87 $10.39

*Yields of 1200 ctns./acre have an associated probability of occurrence of (p=.50)

87

TABLE 5.5

Comparison of Costs per Carton for Final Four Production Methods

Production Methods

Cost per Carton MeBr C35 + KPAM C35 + PIC Telone II + PIC

Pre-harvest Variable Cost

$3.35

$3.32

$3.93

$3.60

Harvest and Marketing

$6.29

$6.29

$6.29

$6.29

Fixed Cost

$0.74

$0.71

$0.83

$0.79

Total Budgeted Cost

$10.38

$10.23

$11.05

$10.68

88

TABLE 5.6

Programming Solutions and Financial Ratios of Five-Year Averages for All Production Methods

Five Year Averages Production Methods

Activity Measures MEBR C35 + KPAM C35 + PIC T2 + PIC ACPROD 1,000 1,000 886 887ACBUY (cumulative) 435 435 300 300ACRENT 565 565 586 587EQBUY $0 $0 $22,416 $32,276OFFINV $1,528,000 $1,520,800 $393,840 $393,840CASSETS $45,485,200 $48,541,400 $9,857,504 $16,309,720TOTASSTS $47,081,400 $50,137,600 $17,961,600 $17,622,800CULIAB $146,386 $146,386 $156,714 $162,110TOTLIAB $420,840 $420,840 $453,946 $469,268NETWRTH $46,660,800 $49,717,000 $17,507,660 $17,153,500 Financial Ratios

Tenure 0.4530 0.4350 0.3386 0.3382

Current 310.7210 331.5987 62.9012 100.6090

Debt-to-Asset 0.0089 0.0084 0.0253 0.0266

Off-farm Investment 0.0325 0.0303 0.0219 0.0223

89

TABLE 5.7

Yearly Programming Solutions for Decision Variables and Key Financial Measures Using Methyl Bromide

Time Periods

T(1)

T(2)

T(3)

T(4)

T(5) Activity Measures

ACPROD 1,000 1,000 1,000 1,000 1,000 ACBUY* 55 93 0 11 7 ACRENT 645 552 552 541 534 EQBUY $0 $0 $0 $0 $0 OFFINV $0 $1,873,500 $1,876,100 $1,921,200 $1,969,200 CASSETS $10,910,000 $29,618,000 $45,459,000 $62,057,000 $79,382,000 TOTASSTS $12,310,000 $31,240,000 $47,065,000 $63,707,000 $81,085,000 CULIAB* $144,210 $155,090 $144,210 $144,210 $144,210 TOTLIAB $622,200 $530,100 $406,890 $284,530 $260,480 NETWRTH $11,688,000 $30,710,000 $46,658,000 $63,423,000 $80,825,000 * Marginal values

90

TABLE 5.8

Yearly Programming Solutions for Decision Variables and Key Financial Measures Using C35 + KPAM as Methyl Bromide Substitute

Time Periods

T(1)

T(2)

T(3)

T(4)

T(5)

Activity Measures

ACPROD 1,000 1,000 1,000 1,000 1,000 ACBUY* 48 102 0 10 7 ACRENT 652 550 550 540 533 EQBUY $0 $0 $0 $0 $0 OFFINV $0 $1,837,500 $1,876,100 $1,921,200 $1,969,200 CASSETS $11,597,000 $31,436,000 $48,465,000 $66,300,000 $84,909,000 TOTASSTS $12,981,000 $33,062,000 $50,076,000 $67,954,000 $86,615,000 CULIAB* $144,210 $155,090 $144,210 $144,210 $144,210 TOTLIAB $622,200 $530,100 $406,890 $284,530 $260,480 NETWRTH $12,359,000 $32,532,000 $49,669,000 $67,670,000 $86,355,000

* Marginal values

91

TABLE 5.9

Yearly Programming Solutions for Decision Variables and Key Financial Measures Using C35 + Chloropicrin as Methyl Bromide Substitute

Time Periods

T(1)

T(2)

T(3)

T(4)

T(5)

Activity Measures ACPROD 949 905 892 863 822 ACBUY* 0 0 0 0 0 ACRENT 649 605 592 563 522 EQBUY $0 $0 $77,461 $34,620 $0 OFFINV $0 $0 $0 $0 $1,969,200 CASSETS $604,120 $9,134,400 $13,038,000 $13,283,000 $13,228,000 TOTASSTS $10,410,000 $14,304,000 $14,583,000 $14,541,000 $35,970,000 CULIAB* $144,210 $144,210 $160,270 $167,440 $167,440 TOTLIAB $622,200 $519,220 $463,370 $353,970 $310,970 NETWRTH $9,787,300 $13,785,000 $14,120,000 $14,187,000 $35,659,000 * Marginal Values

92

TABLE 5.10

Yearly Programming Solutions for Decision Variables and Key Financial Measures Using Telone II + Chloropicrin as Methyl Bromide Substitute

Time Periods

T(1)

T(2)

T(3)

T(4)

T(5)

Activity Measures

ACPROD 949 838 912 890 844 ACBUY* 0 0 0 0 0 ACRENT 649 538 612 590 544 EQBUY $0 $0 $109,040 $52,341 $0 OFFINV $0 $0 $0 $0 $1,969,200 CASSETS $8,972,600 $11,752,000 $13,152,000 $13,205,000 $34,467,000 TOTASSTS $10,248,000 $13,019,000 $14,481,000 $14,558,000 $35,808,000 CULIAB* $144,210 $144,210 $166,810 $177,660 $177,660 TOTLIAB $622,200 $519,220 $486,400 $384,900 $333,620 NETWRTH $9,625,500 $12,500,000 $13,994,000 $14,173,000 $35,475,000 * Marginal Values

93

CHAPTER 6

SUMMARY AND CONCLUSION

6.1 Study Summary

This study began by addressing the phase-out of methyl bromide within the historical

framework of the Montreal Protocol on Substances that Deplete the Ozone Layer. It was

recognized that United States’ agricultural production has come to rely on this fumigant for the

benefits accruing to both pre and post-harvest activities. Many states have now proposed the

continued use of this chemical through the applications of “critical use exemptions.” These

exemptions have provided a short-term solution for a problem whose conclusion will likely be a

complete phase-out by member countries within the immediate future. As developed countries

begin to eliminate MeBr completely, U.S. producers will be forced to identify alternative

methods of production that result in economically viable operations.

Further analysis of the problem reveals that Georgia pepper producers face significant

losses from the phase-out as growers struggle to identify alternative combinations of agricultural

chemicals that provide a comparative yield structure to that of MeBr. Therefore, the objective of

this study was three-fold. First, yield and price data obtained by plant scientists and agricultural

economists was analyzed using stochastic dominance analysis in order to classify an efficient set

of feasible alternative fumigants for Georgia growers. Next, the results of this analysis were

combined with current production technologies within the context of enterprise pepper budgets

with the expectation of identifying a set of economically viable production methods. Finally,

94

results obtained from the enterprise budget were subjected to the simulation/optimization

procedures of a linear programming model. The results provided for by the General Algebraic

Modeling System (GAMS) software will be used to provide Georgia growers with a comparative

analysis of additional input-substitution strategies.

The related literature surrounding the problem was then reviewed in an attempt to both

identify and eliminate sources of error within the study. It was recognized that studies conducted

in Florida and California composed the bulk of the research with regard to both technically and

economically feasible alternatives to MeBr (Deepak et al., 1996; Carpenter et al., 2000;

VanSickle et al., 2000). However, these analyses predominantly focused on optimizing a

“spatial equilibrium” problem that dealt with cross-boundary impacts on a national or

international scale. Additional studies provided a biological assessment of the technical

feasibility of MeBr substitutes (Gilreath et al., 1994; Becker et al., 1998).

Although these studies identified substitutes for MeBr, regional growing conditions and

environmental policies dictate that plausible solutions can not always be implemented that satisfy

the production requirements of every U.S. producer. As a result, a general assessment of U.S.

pepper production was coupled with a more thorough investigation of Georgia pepper

production, marketing, and farm management.

The study then considered the analytical tools needed to address the problem, and

described the efficiency criteria employed by agricultural economists to determine the optimal

choice among a set of risky alternatives. Expected value, mean-variance, and stochastic

dominance analyses were compared and contrasted according to their various assumptive

restrictions. Second-degree stochastic dominance was chosen for use in this study due to the

imposition of more restrictive assumptions that served to reduce an initially broad set of

95

alternative chemicals into a more efficient set. Each method composing this final set was

analyzed within an enterprise pepper budget for variations in gross returns, variable costs, total

costs, and base budgeted net revenues. A determination of the economic viability of each

production method allowed for an initial comparison that would be more fully scrutinized within

the linear programming model.

The programming model was developed according to the characteristics of a

representative Georgia pepper farm. The simulation/optimization procedures of the model

prescribed solutions that were conditional on an initial set of production and financial

constraints. These constraints, although tailored for pepper production, addressed a full menu of

components inherent to many farm operations such as acres produced, equipment and machinery

purchases, cash and credit allowances, and labor. Expectations were that the model would

prescribe a feasible set of solutions for each alternative method from which to make a

comparative analysis based on the maximization of producers’ accumulated net worth.

6.2 Conclusions

The comparative analysis of alternative production technologies was based on methyl

bromide representing the base-case due to its current position as the dominant fumigant choice

for pepper growers throughout Georgia. Our initial expectations were that production methods

employing MeBr either as a stand-alone fumigant or in combination with a set of herbicides

would dominate within the final efficient set.

After consolidating an initial set of twenty-one production methods into a set of eight, the

latter set was ranked using Simetar, a software package that provides for second-degree

stochastic dominance analysis, into a final set of four methods. The ranking of these eight

methods according to their average total revenues per experimental plot did not conform to one

96

of our initial expectations. Of the average total revenues for the final four fumigant

combinations, C35 + KPAM dominated all other methods by a minimum of $1.77 per test plot.

Further, comparisons according to the methods’ coefficients of variation revealed that C35 +

KPAM remained the dominant method followed by MeBr, T2 + PIC, and finally C35 +PIC.

However, the ranking did substantiate our expectation that those methods incorporating a menu

of herbicides into their production technologies would dominate non-herbicide technologies.

Each of the final four dominant methods employed a corresponding menu of herbicides.

These rankings followed through to the enterprise budget analysis where C35 + KPAM

resulted in the largest base budgeted net revenue ($1,768/acre) of all methods. MeBr resulted in

net revenues per acre of ($1,520) while the final two methods T2 + PIC and C35 + PIC resulted

in net revenues of ($1,115) and ($732) per acre, respectively. After constructing a pre-operating

balance sheet from the budget data, projections were made of specific components of the budget

to be incorporated as variables within the five-year planning horizon of the programming model.

Finally, the linear programming model prescribed solutions against initial expectations.

Net worth solutions averaged over the five-year planning horizon resulted again in C35 + KPAM

being prescribed as the dominant production strategy with a value of approximately $49.7

million. MeBr was prescribed a solution of nearly $46.6 million while T2 + PIC and C35 + PIC

resulted in average net worth’s of approximately $17.1 million and $17.5 million respectively.

Although both SSD and the enterprise budget accorded T2 + PIC the dominant position when

compared with C35 + PIC, the linear model reversed this trend and produced solutions that

recognized the latter as the preferred method for producers.

Financial ratios were created from these solutions that measure the firm’s liquidity and

solvency, as well as its asset allocations. The current ratios suggested that each firm was highly

97

liquid as assets were well in excess of liabilities for each method. The two leading methods,

MeBr and C35 + KPAM, had ratios above the critical value of (2X) liabilities at approximately

311 and 332 respectively. Debt-to-asset ratios for all four methods ranged from a low of (0.008)

for C35 + KPAM to a high of (0.025) for C35 + PIC. These figures indicate that none of the

firms relied heavily on the use of debt to finance operations as the critical value for this ratio is

typically 65%. As a result of the highly liquid positions associated with methods C35 + KPAM

and MeBr, the tenure ratio suggests that growers using these options will choose to add

additional acres to production through purchases of land rather than cash-renting. On the other

hand, growers employing C35 + PIC and T2 + PIC will choose to cash-rent as the overall

financial characteristics of these farms do not suggest either additional land purchases or

maximizing the bound production constraints of prescribed by the model. Finally, the dominant

current assets position of C35 + KPAM and MeBr as compared with C35 + PIC and T2 + PIC

over the five-year planning period allows growers to dedicate a larger proportion of their overall

financial portfolio to off-farm investments. The model prescribes solutions for off-farm

investments that constitute approximately 3.0% of the growers total net-worth for C35 + KPAM

and MeBr, and approximately 2.0% for C35 + PIC and T2 + PIC.

Our results suggest that economically viable alternatives exist to replace MeBr.

However, growers have not moved to adopt solutions recognized in the literature due to several

reasons. First, growers may not have access to information concerning the alternatives cited in

scholarly journals. Next, irrigation levels, soil conditions, diseases, or pests may negate the

efficacy of the alternatives cited both in the literature and this study. This forces producers to

uniquely identify the fumigant or combination of fumigants and herbicides that results in the

same yield and quality levels as MeBr. An advantage of MeBr is its ability to eradicate diseases,

98

weeds, and pests over a wide range of environmental conditions and growing conditions.

Finally, MeBr has maintained consistent yields for producers. Although alternatives have been

found to be equally effective in experimental trials, there is not enough information that they can

deliver consistent yields over the long-term.

6.3 Future Research

Additional research concerning methyl bromide's role in Georgia pepper production can

expand on the findings of this study through several outlets. First, continued research into

economically viable fumigant alternatives must begin with the generation of additional field trial

data from which to conduct economic impact assessments. A limiting factor in this study was

the availability of historical yields and prices for Georgia bell peppers. In addition to yield and

price data, future studies must focus on the residual effects of alternatives over several harvests

and the effects crop rotations have on these substitutes.

A benefit of MeBr is that growers may apply the fumigant once during the growing

season. Subsequent crops then benefit from its ability to remain active in the soil. Tests on

alternatives have not confirmed the same results and may indicate higher costs to producers over

the long-term. Moreover, comprehensive data would allow researchers to more accurately

characterize the operations of a representative Georgia pepper farm. For example, the linear

programming model used in this study was limited by its inability to account for producers'

varying degrees of risk-aversion.

Next, the development of a quadratic programming model as employed by (Escalante,

2001) would provide for the effects of risk on such components of the farm model as debt-levels,

share-leasing, and credit reserves. Moreover, initial rankings can be subjected to a more rigorous

analysis within the framework of stochastic dominance. Stochastic dominance with respect to a

99

function accounts for an Arrow-Pratt risk aversion coefficient when evaluating the ranking

scheme. Finally, as advancements are made into the research and development of alternatives,

enterprise budgets and programming models can be modified to account for the changing cost

structure of these new substitutes. Georgia growers will then be able to adjust to these changes

in an attempt to remain profitable and competitive on a regional and international scale.

100

BIBLIOGRAPHY

Anderson, J.R., J.L. Dillon, and B. Hardaker. Agricultural Decision Analysis, 1st ed. Ames, Iowa: Iowa State University Press, 1977. pp. 65-108. Bertelson, D. “The U.S. Strawberry Industry” USDA, ERS. Statistical Bulletin Number 914, January 1995. Bodie, Z., A., Kane, and A. Marcus. Investments, 6th ed. New York: McGraw Hill Book Co., 2005. Brooke, A., D. Kendrick, and A. Meeraus. GAMS: A User's Guide, Redwood City, California: The Scientific Press, 1988. Brunt, A.A., K. Crabtree, M. Dallwitz, A.J. Gibbs, and L., Watson. Viruses of Plants: Descriptions and Lists from the VIDE Database. United Kingdom, C.A.B. International, 1996. p.1484. Calkins, P. H., and D.D. DiPietre. Farm Business Management-Successful Decisions in a Changing Environment, New York: Macmillan Co., 1983. Carpenter J., L. Gianessi, and L. Lynch. "The Economic Impact of the Scheduled U.S. Phaseout of Methyl Bromide." National Center for Food and Agricultural Policy, 2000. Chyen, D., M.E. Wetzstein, R.M. McPherson, and W.D. Givan. "An Economic Evaluation of Soybean Stink Bug Control Alternatives for the Southeastern United States." Southern Journal of Agricultural Economics 24, no. 2 (1992): 83-94. Csinos, A.S., W.C. Johnson, A.W. Johnson, D.R. Sumner, R.M. McPherson, and R.D. Gitaitis. "Alternative fumigants for methyl bromide in tobacco and pepper transplant production." Crop Protection 16 (1997): 585-594. Culpepper, S. and D.B. Langston. "Replacement of Methyl Bromide by Integrating the Use of Alternative Soil Fumigants, Cultural Practices, and Herbicides for Tomato, Pepper, and Watermelon." Proposal for CSREES/IFAFS for Program Area Entitled Critical and Emerging Pest Management Challenges. Pro. No.14. (2000). Deepak M.S., T.H. Spreen, and J.J. Van Sickle. "An Analysis of the Impact of a Ban of Methyl Bromide on the U.S. Winter Fresh Vegetable Market." Journal of Agricultural and Applied Economics 28, no. 2 (1996): 433-443.

101

Escalante, C.L. "Farm-Level Evidence on Risk Balancing Strategies of Illinois Grain Farmers Against Increasing Income Risk." PhD Dissertation, University of Illinois at Urbana-Champaign, 2001. Fernandez-Cornejo, J., S. Jans, and M. Smith. "Issues in the Economics of Pesticide Use in Agriculture: A Review of the Empirical Evidence" Review of Agricultural Economics 20, no. 2 (1998): 462-488. Fonsah, E.G., and Angus SND Chidebelu. “Economics of Banana Production and Marketing in the Tropics,” Minerva Press Publishers, London, pp.1-230. Fonsah, E.G., and K. Rucker. “Bell Pepper on Plastic Enterprise Budget,” Department of Agricultural and Applied Economics, University of Georgia, 2003. Gast, K.L.B. "Containers and Packaging: Fruits & Vegetables," Kansas State University, Agricultural Experiment Station, MF-979, March 1991. Gilreath, J.P., J.P. Jones, and A. J. Overman. "Soilborne pest control in mulched tomato with alternatives to methyl bromide." Proc. Fla. State Hort. Soc. 107: pp.156-159. 1994. Goodhue, R.E., S.A. Fennimore, and H.A. Ajwa, University of California, Davis. "Economics of VIF Use with Inline and PIC EC in California Strawberries." 2004. http://mbao.org/2004/Proceedings04/099%20GoodhueR%20MBAOVIF.pdf(Last Accessed July 2005) Harris, D.C.. "A Comparison of Dazomet, Chloropicrin and Methyl Bromide as Soil Disinfectants for Strawberries." Journal of Horticultural Science 66, no.1 (1991): 51-58. Huang, C., and R. Litzenberger. Foundations for Financial Economics. Upper Saddle River, New Jersey: Prentice-Hall, 1988. Hueth, B. M., L. Lynch, B. McWilliams, M. Roberts, J. Siebert, D. Sunding, and D. Zilberman. "Economic Impact of Banning Methyl Bromide Use in California Agriculture." Annual International Research Conference on Methyl Bromide Alternatives and Emissions Reductions. 1997 http://www.epa.gov/ozone/mbr/airc/1997/030hueth.pdf (Last Accessed July 2005) Hueth, B., B. McWilliams, D. Sunding, and D. Zilberman. "Analysis of an Emerging Market: Can Methyl Iodide Substitute for Methyl Bromide?" Review of Agricultural Economics 22, no. 1 (2000): 43-54. Jorn Ole Becker, Howard D. Ohr, Nigel M. Grech, Milton E. McGriffen, Jr & James J. Sims. "Evaluation of Methyl Iodide as a Soil Fumigant in Container and Small Field Plot Studies." Pesticide Science. 53 (1998): 71-79.

http://mbao.org/2004/Proceedings04/099 GoodhueR MBAOVIF.pdf

http://www.epa.gov/ozone/mbr/airc/1997/030hueth.pdf

102

Jovicich, E., D.J. Cantliffe, N. Shaw, M. Smither-Kopperl, and P. A. Stansley. University of Florida, IFAS, Protected Agriculture Project, Gainesville, FL 32611. SWFREC, Immokalee, FL 34142. Economic Benefit of Greenhouse Production of Peppers in Soilless Culture. 2003. http://mbao.org/2003/027%20Smither-KopplerM2MBA%20gh%20pepper.pdf(Last Accessed July 2005) Kelly, W. "Land and Resources-Vegetable Production." New Georgia Encyclopedia. University of Georgia Press, 2003. http://www.georgiaencyclopedia.org/nge/Article.jsp?id=h-1005 (Last Accessed July 2005) Kelly, W. "Quiet, deliberate growth: Georgia vegetable production advances." Southeast Farm Press, December 2002, p.1. King, R.P., and L.J. Robinson. "An Interval Approach to Measuring Decision Maker Preferences." American Journal of Agricultural Economics 63, no. 3 (1981): 510-520. Locascio, S.J., D.W. Dickson, and T.A. Kucharek. "Evaluation of multi-purpose soil fumigants for root-knot nematode and soilborne disease management on tomato in Florida." Proc. of the International Resource Conference on Methyl Bromide Alternatives Emissions Reduction, Orlando, FL, pp.3-4. 1994. Lucier G., C. Plummer, and A. Jerardo. USDA, ERS. Vegetables and Melons Situation and Outlook Yearbook. July 2004. http://ers.usda.gov/publications/vgs/Jul04/VGS2004.pdf(Last Accessed July 2005) Lynch, L., B. McWilliams, and D. Zilberman. "Economic Implications of Banning Methyl Bromide: How Have They Changed With Recent Developments?" 1997. http://www.epa.gov/ozone/mbr/airc/1997/029lynch.pdf (Last Accessed July 2005) Lynch, L., C. Greene, and C. Kramer-LeBlanc. USDA, ERS. Proceedings of the Third National IPM Symposium/Workshop. May 1997. http://www.ers.usda.gov/Publications/mp1542/(Last Accessed July 2005) Lynch, L. and J., Carpenter. The Economic Impacts of Banning Methyl Bromide: Where Do We Need More Research? 1999 http://www.epa.gov/ozone/mbr/airc/1999/55lynchl.pdf(Last Accessed July 2005) Lynch, L. and J., Carpenter. Economic Impacts of Switching From Pre-Plant Fumigation with Methyl Bromide. 1998 http://www.epa.gov/ozone/mbr/airc/1998/039lynch.pdf (Last Accessed July 2005) McCarl, B.A., and T.H. Spreen. Applied Mathematical Programming Using Algebraic Systems. College Station, Texas: Texas A&M University, 1994.

http://mbao.org/2003/027 Smither-KopplerM2MBA gh pepper.pdf

http://www.georgiaencyclopedia.org/nge/Article.jsp?id=h-1005

http://www.georgiaencyclopedia.org/nge/Article.jsp?id=h-1005

http://ers.usda.gov/publications/vgs/Jul04/VGS2004.pdf

http://www.epa.gov/ozone/mbr/airc/1997/029lynch.pdf

http://www.ers.usda.gov/Publications/mp1542/

http://www.epa.gov/ozone/mbr/airc/1999/55lynchl.pdf

http://www.epa.gov/ozone/mbr/airc/1998/039lynch.pdf

103

Ohr, H.D., J. J. Sims, N. M. Grech, J. O. Becker, and M. E. McGiffen, Jr. "Methyl Iodide, an Ozone-Safe Alternative for Methyl Bromide as a Soil Fumigant." Plant Disease 80(1996):731-35. Osteen, C. "Agricultural Resources and Environmental Indicators: Pest Management Practices." USDA, ERS. No. AH722. September 2000. http://ers.usda.gov/publications/arei/ah722/ arei4_3/DBGen.htm (Last Accessed July 2005) Richardson, J.W., K. Schumann, and P. Feldman. Simetar “Simulation for Excel to Analyze Risk.” Version 2004.01.20 Beta. Agricultural and Food Policy Center at Texas A&M University. January 2004. Samuelson, P.A. "The Fundamental Approximation Theorem of Portfolio Analysis in Terms of Means, Variances, and Higher Moments." Review of Economic Studies 37, no. 4 (1970): 537-542. Schaub, J. D. "Economic Forces and Progress with Post-harvest Methyl Bromide Alternatives" Paper presented at the International Research Conference on Methyl Bromide Alternatives and Emissions Reductions, Orlando FL, 1-18 November 2004. Seabrook, C. "Vegetable growers rely on ozone-depleting gas." The Atlanta Journal and Constitution. February, 2005. Stall, W. Methyl Bromide Alternatives Field Day. Live Oak, FL: North Florida Research and Education Center-Suwannee Valley, 2000. Sydorovych, O., C. D. Safley, E. B. Poling, L. M. Ferguson, G. E. Fernandez, P. M. Brannen, and F. J. Louws. Departments of Agricultural and Resource Economics, Horticultural Science, and Plant Pathology, North Carolina State University, Raleigh NC 27695; University of Georgia, Plant Pathology, Athens GA 30602. Economic Evaluation of Methyl Bromide Alternatives for Strawberry Production. 2004. http://mbao.org/2004/Proceedings04/028%20SafleyC%20Louws %20economicsl%202004.pdf (Last Accessed July 2005) Thompson, W.T. Agricultural Chemicals: Book IV-Fungicides. Fresno, CA: Thompson Publications, 1991. United Nations Environmental Program, Report of the Seventh Meeting of the Parties to the Montreal Protocol on Substances that Deplete the Ozone Layer. New York, 1995. United Nations, Environment Programme. The Montreal Protocol on Substances that Deplete the Ozone Layer. New York, 2000. University of Georgia, College of Agricultural and Environmental Sciences. 2003 Georgia Farm Gate Vegetable Survey Report. Compiled by Susan R. Boatright and John C. McKissick. SR-04-01 (pp.1-10)

http://ers.usda.gov/publications/arei/ah722/arei4_3/DBGen.htm

http://ers.usda.gov/publications/arei/ah722/arei4_3/DBGen.htm

http://mbao.org/2004/Proceedings04/028 SafleyC Louws economicsl 2004.pdf

http://mbao.org/2004/Proceedings04/028 SafleyC Louws economicsl 2004.pdf

104

University of Georgia, Cooperative Extension Service. Commercial Pepper Production. Bulletin No. 1027, February 1990. http://pubs.caes.uga.edu/caespubs/pubcd/b1027-w.html(Last Accessed July 2005) USDA. Vegetables Final Estimates 1998-2003. Statistical Bulletin # 987. May 2004. http://usda.mannlib.cornell.edu/reports/general/sb/sb987.txt (Last Accessed July 2005) USDA, ERS Publications and Outlook Reports 2004. http://www.ers.usda.gov/publications/ so/view.asp?f=/specialty/vgs-bb/ (Last Accessed July 2005) USDA, ERS. Vegetables and Melons Situation and Outlook Yearbook. July 2004. http://www.ers.usda.gov/publications/so/view.asp?f=/specialty/vgs-bb/ (Last Accessed July 2005) USDA, ERS. Vegetables and Melons Outlook. December 2001. http://www.ers.usda.gov/publications/so/view.asp?f=/specialty/vgs-bb/(Last Accessed July 2005) USDA, ERS. Vegetables and Melon. June 2003. http://www.ers.usda.gov/publications/so/view.asp?f=/specialty/vgs-bb/(Last Accessed July 2005) USDA, NASS. Land Values and Cash Rents, 2004 Summary. August 2004. http://usda.mannlib.cornell.edu/reports/nassr/other/plr-bb/land0804.pdf(Last Accessed July 2005) USDA, NASS. Vegetables Annual Summary. January 2005. http://usda.mannlib.cornell.edu/reports/nassr/fruit/pvg-bb/2005/(Last Accessed July 2005) USDA, NASS. 2002 Census of Agriculture, State Data. (http://www.nass.usda.gov/census/census02/volume1/ga/st13_1_035_035.pdf)] (Last Accessed July 2005) USDA, NASS. Statistics of Vegetables and Melons. 2004. http://www.usda.gov/nass/pubs/agr04/04_ch4.pdf (Last Accessed July 2005) USDA, NASS. Vegetables, Final Estimates 1998-2003. Statistical Bulletin Number 987. May 2004. http://usda.mannlib.cornell.edu/usda/reports/general/sb/sb987.pdf (Last Accessed July 2005) U.S. Environmental Protection Agency. Methyl Bromide Alternative Case Studies. Volumes 1, 2, and 3. http://www.epa.gov/spdpublc/mbr/casestudies/ (Last Accessed July 2005)

http://pubs.caes.uga.edu/caespubs/pubcd/b1027-w.html

http://usda.mannlib.cornell.edu/reports/general/sb/sb987.txt

http://www.ers.usda.gov/publications/ so/view.asp?f=/specialty/vgs-bb/

http://www.ers.usda.gov/publications/ so/view.asp?f=/specialty/vgs-bb/

http://www.ers.usda.gov/publications/so/view.asp?f=/specialty/vgs-bb/



http://usda.mannlib.cornell.edu/reports/nassr/other/plr-bb/land0804.pdf

http://usda.mannlib.cornell.edu/reports/nassr/fruit/pvg-bb/2005/

http://www.nass.usda.gov/census/census02/volume1/ga/st13_1_035_035.pdf

http://www.usda.gov/nass/pubs/agr04/04_ch4.pdf

http://usda.mannlib.cornell.edu/usda/reports/general/sb/sb987.pdf

http://www.epa.gov/spdpublc/mbr/casestudies/

105

U.S. Environmental Protection Agency. Methyl Bromide Fact Sheet. May 2005. http://www.epa.gov/ozone/mbr/CueFRfactsheet.html (Last Accessed July 2005) U.S. Environmental Protection Agency, “Clean Air Act, Section 602. Listing of Class I and Class II Substances,” 1990. http://www.epa.gov/ozone/ods.html(Last Accessed July 2005) VanSickle, J.J., C. Brewster, and T.H. Spreen. "Impact of a Methyl Bromide Ban on the U.S. Vegetable Industry." Bulletin No. 333, University of Florida, 2000. Wackerly, D., and W. Mendenhall III, and R.L. Scheaffer. Mathematical Statistics with Applications, 6th ed. Pacific Grove, California: Duxbury Resource Center, 2002. Wetzstein, M.E. Microeconomic Theory-Concepts and Connections, 1st ed. Mason, Ohio: Thomson Southwestern, 2005. pp.584-618. Wetzstein, M.E., P.I. Szmedra, R.W. McClendon, and D.M. Edwards. "Efficiency Criteria and Risk Aversion: An Empirical Evaluation." Southern Journal of Agricultural Economics 20, no. 1 (1988): 171-178. Wolfe, K. and E.G. Fonsah. "Wholesale and Distributors Outlook for Fruit and Vegetables Produced in Georgia" GFVGA News Vo. 7, No. 4, Fall 2002. Zhang, W., M.E. McGriffen, J.O. Becker, H.D. Ohr, J.J. Sims, and S. Campbell. "Effects of Soil Physical Factors on Methyl Iodide and Methyl Bromide." Pesticide Science. 53 (1998): 71-79

http://www.epa.gov/ozone/mbr/CueFRfactsheet.html

http://www.epa.gov/ozone/ods.html

106

APPENDICES

107

APPENDIX A

Bell Pepper Fresh Market

Variable Cost Budget for Methyl Bromide Number of acres (1)

Expected Yield (cartons) 1167 Expected Price per carton $11.68 Item Unit Quantity Price Amount per acreVariable Costs Plants Hundred 174.00 $5.50 $957.00Lime, applied Ton 1.00 $26.00 $26.00Base Fertilizer Lbs. 12.00 $9.50 $114.00Sidedress Fertilizer Gal. 0.54 $184.00 $99.36Insecticide Acre 1.00 $186.22 $186.22Fungicide Acre 1.00 $231.36 $231.36Nematicide Acre 1.00 $510.00 $510.00Herbicide Acre 1.00 $68.08 $68.08Plastic Roll 2.80 $68.50 $191.80Plastic removal Acre 1.00 $75.00 $75.00Drip Tape Ft. 8700.00 $0.02 $174.00Fumigation Acre 1.00 $405.00 $405.00Machinery Hour 5.00 $21.00 $105.00Transplant Labor Hour 20.00 $8.00 $160.00Labor Hour 33.00 $8.00 $264.00Land Rent Acre 1.00 $105.00 $105.00Irrigation (Machinery and Labor)

Acre

1.00 $65.08 $65.08Interest on Operating Capital $ 3716.89 $0.09 $168.16Pre-Harvest Variable Costs $3,905.05 Harvest and Marketing Costs Unit Quantity Price Amount per acre Picking and Hauling Ctn. 1167 $1.25 $1,458.33Grading and Packing Ctn. 1167 $2.75 $3,208.33Container Ctn. 1167 $1.30 $1,516.67Marketing Ctn. 1167 $0.99 $1,158.27Total Harvest and Marketing $6.29 $7,341.60Total Variable Costs $11,246.65

108

APPENDIX A (cont'd)


Variable Cost Budget for Methyl Bromide Number of acres (1)

Item Unit Quantity Price Amount per acre Fixed Cost

Machinery

Acre

1.00 $53.59 $53.59

Irrigation

Acre

1.00 $220.65 $220.65

Land

Acre

1.00 $0.00 $0.00

Overhead and Management

$ 3,905.05 $0.15 $585.76

Total Fixed Costs

$860.00

Total Budgeted Cost / Acre

$12,106.64

Costs per Carton Pre-Harvest Variable Cost per Carton $ 3.35 Harvesting and Marketing Cost per Carton $ 6.29 Fixed Cost per Carton $ 0.74 Total Budgeted Cost per Carton $ 10.38 Base Budgeted Net Revenue $1,520

109

APPENDIX B


Variable Cost Budget for C35 + KPAM Number of acres (1)


Acre

1.00 $65.08 $65.08Interest on Operating Capital $ 3716.89 $0.09 $185.29Pre-Harvest Variable Costs $4,302.93 Harvest and Marketing Costs Unit Quantity Price Amount per acrePicking and Hauling Ctn. 1298 $1.25 $1,622.02Grading and Packing Ctn. 1298 $2.75 $3,568.45Container Ctn. 1298 $1.30 $1,686.90Marketing Ctn. 1298 $0.99 $1,288.25Total Harvest and Marketing $6.29 $8,165.63Total Variable Costs $12,468.57

110

APPENDIX B (cont'd)


Variable Cost Budget for C35 + KPAM Number of acres (1)


Machinery

Acre 1.00 $53.59 $53.59

Irrigation

Acre 1.00 $220.65 $220.65

Land

Acre 1.00 $0.00 $0.00


$ 4302.93 $0.15 $645.44

Total Fixed Costs

$919.68


$13,388.25

Costs per Carton Pre-Harvest Variable Cost per Carton

$3.32

Harvesting and Marketing Cost per Carton

$6.29

Fixed Cost per Carton

$0.71

Total Budgeted Cost per Carton

$10.23

Base Budgeted Net Revenue $1,768

111

APPENDIX C


Variable Cost Budget for C35 + Chloropicrin Number of acres (1)

Expected Yield (cartons) 1155 Expected Price per carton $11.68 Item Unit Quantity Price Amount per acreVariable Costs Plants Hundred 174.00 $5.50 $957.00Lime, applied Ton 1.00 $26.00 $26.00Base Fertilizer Lbs. 12.00 $9.50 $114.00Sidedress Fertilizer Gal. 0.54 $184.00 $99.36Insecticide Acre 1.00 $186.22 $186.22Fungicide Acre 1.00 $231.36 $231.36Nematicide Acre 1.00 $510.00 $510.00Herbicide Acre 1.00 $68.08 $68.08Plastic Roll 2.80 $68.50 $191.80Plastic removal Acre 1.00 $75.00 $75.00Drip Tape Ft. 8700.00 $0.02 $174.00Fumigation Acre 1.00 $1,007.05 $1,007.05Machinery Hour 5.00 $21.00 $105.00Transplant Labor Hour 20.00 $8.00 $160.00Labor Hour 33.00 $8.00 $264.00Land Rent Acre 1.00 $105.00 $105.00Irrigation (Machinery and Labor)

Acre 1.00 $65.08 $65.08

Interest on Operating Capital $ 4743.94 $0.09 $213.48Pre-Harvest Variable Costs $4,534.19 Harvest and Marketing Costs Unit Quantity Price Amount per acrePicking and Hauling Ctn. 1155 $1.25 $1,443.45Grading and Packing Ctn. 1155 $2.75 $3,175.60Container Ctn. 1155 $1.30 $1,501.19Marketing Ctn. 1155 $0.99 $1,146.42Total Harvest and Marketing $6.29 $7,266.66Total Variable Costs $11,800.86

112

APPENDIX C (cont'd)


Variable Cost Budget for C35 + Chloropicrin Number of acres (1)


Machinery

Acre

1.00 $53.59 $53.59

Irrigation

Acre

1.00 $220.65 $220.65

Land

Acre

1.00 $0.00 $0.00


$ 4534.19 $0.15 $680.13

Total Fixed Costs

$954.37


$12,755.23

Costs per Carton Pre-Harvest Variable Cost per Carton $ 3.93 Harvesting and Marketing Cost per Carton $ 6.29 Fixed Cost per Carton $ 0.83 Total Budgeted Cost per Carton $ 11.05 Base Budgeted Net Revenue $732

113

APPENDIX D


Variable Cost Budget for Telone II + Chloropicrin Number of acres (1)


Acre 1.00

$65.08 $65.08

Interest on Operating Capital $ 3859.89 $0.09 $173.70Pre-Harvest Variable Costs $4,033.59 Harvest and Marketing Costs Unit Quantity Price Amount per acrePicking and Hauling Ctn. 1119 $1.25 $1,398.81Grading and Packing Ctn. 1119 $2.75 $3,077.38Container Ctn. 1119 $1.30 $1,454.76Marketing Ctn. 1119 $0.99 $1,110.97Total Harvest and Marketing $6.29 $7,041.92Total Variable Costs $11,075.51

114

APPENDIX D (cont'd)


Variable Cost Budget for Telone II + Chloropicrin Number of acres (1)


Machinery

Acre

1.00 $53.59 $53.59

Irrigation

Acre

1.00 $220.65 $220.65

Land

Acre

1.00 $0.00 $0.00


$

4033.59 $0.15 $605.04

Total Fixed Costs

$879.28


$11954.79

Costs per Carton Pre-Harvest Variable Cost per Carton

$3.60

Harvesting and Marketing Cost per Carton

$6.29

Fixed Cost per Carton

$0.79

Total Budgeted Cost per Carton

$10.68

Base Budgeted Net Revenue $1,115

115

APPENDIX E

GAMS Program for the Base-Case Farm Model using Methyl Bromide

*GAMS PROGRAM FOR PEPPER OPTIMIZATION $OFFSYMXREF OFFSYMLIST OPTION LIMCOL=0; OPTION LIMROW=0; *SIMULATION BASE MODEL FOR RISK BALANCING STUDY SETS T YEARS IN PLANNING HORIZON /T0*T5/ X DECISION VARIABLES /ACPROD, ACBUY, EQBUY, FINLAND, ACRENT, MEDCRED, SHTCRED/ ALIAS (T,J) ALIAS (X,Y); SCALARS LAND0 INITIAL LAND OWNED IN ACRES /300/ EQPT0 INITIAL EQUIPMENT OWNED /468900/ CASH0 INITIAL ENDING CASH BALANCE /124255/ CURAST0 INITIAL TOTAL CURRENT ASSETS /604116/ INV0 INITIAL AMOUNT OF NON-FARM INVESTMENTS /100000/ OTHCUR0 OTHER CURRENT ASSETS /379861/ BUILD0 INITIAL VALUE OF BUILDINGS /125218/ OVRLAB ANNUAL OVERHEAD LABOR /900/ LABREQ LABOR REQUIREMENT MANHOURS PER ACRE /53.0/ LNDDPMT LAND DOWNPAYMENT RATE /0.20/ EQDPMT EQUIPMENT DOWNPAYMENT RATE /0.20/ STRATE STOCK PURCHASE RATE /0.02/ PROPTAX PROPERTY TAXES PER ACRE /24.19/ XPRATE PERCENT USE OF EXPENSES PER SUBPERIOD /0.5/ INCTAX INCOME TAX RATE /0.4/ LIMSTOCK LIMIT ON FCS STOCK /1000/ LIMSP LIMIT ON OFF FARM INVESTMENTS /200000/ UPSIZE UPPER LIMIT ON FARM SIZE /1000/; PARAMETER LANDVAL(T) PROJECTED LAND VALUE PER ACRE /T0 2150 T1 2270 T2 2397.54 T3 2531.80 T4 2673.59 T5 2823.31/;

116

PARAMETER RENT(T) PROJECTED LAND CASH RENT LEVELS /T0 105 T1 110.88 T2 117.09 T3 123.65 T4 130.57 T5 137.88/; PARAMETER EQCOST(T) PROJECTED EQUIPMENT COST PER ACRE /T0 1304.25 T1 1343.38 T2 1383.68 T3 1425.19 T4 1467.94 T5 1511.98/; PARAMETER FACONS(T) FAMILY CONSUMPTION /T1 114417 T2 116477 T3 118923 T4 121777 T5 124822/; PARAMETER MOFFINC(T) MAXIMUM OFF-FARM INCOME /T1 63177 T2 64314 T3 65665 T4 67241 T5 68922/; PARAMETER VARCOST(T) VARIABLE COSTS PER ACRE OF OWNED LAND /T0 10674.17 T1 10994.40 T2 11324.23 T3 11663.95 T4 12013.87 T5 12374.29/; PARAMETER LABCOST(T) PROJECTED LABOR COST PER MANHOUR /T0 8.00 T1 8.24 T2 8.49 T3 8.74 T4 9.00 T5 9.27/; PARAMETER RETURN(T) PROJECTED GROSS RETURNS PER ACRE /T0 14016.00 T1 14436.48 T2 14869.57 T3 15315.66

117

T4 15775.13 T5 16248.39/; PARAMETER INVGAIN(T) PROCEEDS FROM NONFARM INVESTMENTS /T1 1.035 T2 1.035 T3 1.035 T4 1.035 T5 1.035/; PARAMETER YIELD(T) YIELD ON NONFARM INVESTMENTS /T1 0.035 T2 0.035 T3 0.035 T4 0.035 T5 0.035/; PARAMETER OVRHD(T) PROJECTED OVERHEAD COSTS /T0 856.86 T1 882.57 T2 909.04 T3 936.31 T4 964.40 T5 993.34/; PARAMETER MARGIN(T) PROJECTED NET MARGINS PER ACRE /T0 1723.63 T1 1775.34 T2 1828.60 T3 1883.46 T4 1939.96 T5 1998.16/; PARAMETER PRINTS(T) PRINCIPAL AND INTEREST PER DOLLAR SHORT-TERM CREDIT /T1 1.070 T2 1.070 T3 1.070 T4 1.070 T5 1.070/; PARAMETER INPMTS(T) INTEREST PAYMENT PER DOLLAR SHORT TERM CREDIT /T1 0.0725 T2 0.0750 T3 0.0775 T4 0.080 T5 0.080/;

118

TABLE EQDEP(T,J) EQUIPMENT DEPRECIATION RATE (ACRS10) T0 T1 T2 T3 T4 T5 T1 0.00 0.10 0.00 0.00 0.00 0.00 T2 0.10 0.18 0.10 0.00 0.00 0.00 T3 0.18 0.14 0.18 0.10 0.00 0.00 T4 0.14 0.12 0.14 0.18 0.10 0.00 T5 0.12 0.10 0.12 0.14 0.18 0.10; TABLE DEPVALEQ(T,J) DEPRECIATED VALUE OF EQUIPMENT T0 T1 T2 T3 T4 T5 T1 1.0000 0.9000 0.0000 0.0000 0.0000 0.0000 T2 0.9000 0.7380 0.9000 0.0000 0.0000 0.0000 T3 0.7380 0.6350 0.7380 0.9000 0.0000 0.0000 T4 0.6350 0.5580 0.6350 0.7380 0.9000 0.0000 T5 0.5580 0.5030 0.5580 0.6350 0.7380 0.9000; TABLE FPRIN20(T,J) PRNCPL AND INTRST 20 YR FIXED LOAN T0 T1 T2 T3 T4 T5 T1 0.0996 0.0989 0.0000 0.0000 0.0000 0.0000 T2 0.0996 0.0989 0.1019 0.0000 0.0000 0.0000 T3 0.0996 0.0989 0.1019 0.1027 0.0000 0.0000 T4 0.0996 0.0989 0.1019 0.1027 0.1027 0.0000 T5 0.0996 0.0989 0.1019 0.1027 0.1027 0.1027; TABLE INTF20(T,J) INTEREST PAYMENTS 20 YR FIXED TERM LOAN T0 T1 T2 T3 T4 T5 T1 0.0522 0.0761 0.0000 0.0000 0.0000 0.0000 T2 0.0485 0.0744 0.0801 0.0000 0.0000 0.0000 T3 0.0446 0.0725 0.0784 0.0811 0.0000 0.0000 T4 0.0403 0.0705 0.0765 0.0793 0.0811 0.0000 T5 0.0358 0.0683 0.0744 0.0775 0.0793 0.0811; TABLE PRF20(T,J) PRINCIPAL PAYMENTS 20 YR FIXED TERM LOAN T0 T1 T2 T3 T4 T5 T1 0.0474 0.0228 0.0000 0.0000 0.0000 0.0000 T2 0.0511 0.0246 0.0218 0.0000 0.0000 0.0000 T3 0.0550 0.0264 0.0236 0.0216 0.0000 0.0000 T4 0.0593 0.0284 0.0255 0.0233 0.0216 0.0000 T5 0.0638 0.0306 0.0275 0.0252 0.0233 0.0216;

119

TABLE PROUTF20(T,J) PRINCIPAL OUTSTDG 20 YR FIXED TERM LOAN T0 T1 T2 T3 T4 T5 T1 0.6300 0.9772 0.0000 0.0000 0.0000 0.0000 T2 0.5789 0.9526 0.9782 0.0000 0.0000 0.0000 T3 0.5239 0.9262 0.9546 0.9784 0.0000 0.0000 T4 0.4646 0.8978 0.9291 0.9551 0.9784 0.0000 T5 0.4008 0.8672 0.9016 0.9298 0.9551 0.9784; TABLE PRINTMED(T,J) PRNCPL AND INTRST PMTS MEDIUM CREDIT T0 T1 T2 T3 T4 T5 T1 0.2591 0.2558 0.0000 0.0000 0.0000 0.0000 T2 0.2591 0.2558 0.2584 0.0000 0.0000 0.0000 T3 0.2591 0.2558 0.2584 0.2591 0.0000 0.0000 T4 0.2591 0.2558 0.2584 0.2591 0.2591 0.0000 T5 0.2591 0.2558 0.2584 0.2591 0.2591 0.2591; TABLE INTMED(T,J) INTRST PMTS ON MEDIUM TERM CREDIT T0 T1 T2 T3 T4 T5 T1 0.0930 0.0880 0.0000 0.0000 0.0000 0.0000 T2 0.0776 0.0732 0.0920 0.0000 0.0000 0.0000 T3 0.0607 0.0572 0.0767 0.0930 0.0000 0.0000 T4 0.0422 0.0397 0.0600 0.0776 0.0930 0.0000 T5 0.0220 0.0207 0.0417 0.0607 0.0776 0.0930; TABLE PRMED(T,J) PRNCPL PMTS MEDIUM TERM CREDIT T0 T1 T2 T3 T4 T5 T1 0.1661 0.1678 0.0000 0.0000 0.0000 0.0000 T2 0.1815 0.1825 0.1664 0.0000 0.0000 0.0000 T3 0.1984 0.1986 0.1817 0.1661 0.0000 0.0000 T4 0.2169 0.2161 0.1985 0.1815 0.1661 0.0000 T5 0.2371 0.2351 0.2167 0.1984 0.1815 0.1661; TABLE PROUTMED(T,J) PRNCPL OUTSTNDG MEDIUM CREDIT T0 T1 T2 T3 T4 T5 T1 0.8339 0.8322 0.0000 0.0000 0.0000 0.0000 T2 0.6524 0.6497 0.8336 0.0000 0.0000 0.0000 T3 0.4539 0.4511 0.6518 0.8339 0.0000 0.0000 T4 0.2371 0.2351 0.4534 0.6524 0.8339 0.0000 T5 0.0000 0.0000 0.2367 0.4539 0.6524 0.8339;

120

TABLE PRINTM3D(T,J) PRNCPL AND INTRST PMTS MEDIUM CREDIT T0 T1 T2 T3 T4 T5 T1 0.1627 0.1627 0.0000 0.0000 0.0000 0.0000 T2 0.1627 0.1627 0.1627 0.0000 0.0000 0.0000 T3 0.1627 0.1627 0.1627 0.1627 0.0000 0.0000 T4 0.1627 0.1627 0.1627 0.1627 0.1627 0.0000 T5 0.1627 0.1627 0.1627 0.1627 0.1627 0.1627; TABLE INTMED3(T,J) INTRST PMTS ON MEDIUM TERM CREDIT T0 T1 T2 T3 T4 T5 T1 0.0617 0.1000 0.0000 0.0000 0.0000 0.0000 T2 0.0516 0.0937 0.1000 0.0000 0.0000 0.0000 T3 0.0404 0.0868 0.0937 0.1000 0.0000 0.0000 T4 0.0282 0.0792 0.0868 0.0937 0.1000 0.0000 T5 0.0148 0.0708 0.0792 0.0868 0.0937 0.0100; TABLE PRMED3(T,J) PRNCPL PMTS MEDIUM TERM CREDIT T0 T1 T2 T3 T4 T5 T1 0.0101 0.0627 0.0000 0.0000 0.0000 0.0000 T2 0.0111 0.0690 0.0627 0.0000 0.0000 0.0000 T3 0.0122 0.0759 0.0690 0.0627 0.0000 0.0000 T4 0.0134 0.0835 0.0759 0.0690 0.0627 0.0000 T5 0.0148 0.0918 0.0835 0.0759 0.0690 0.0627; TABLE PROUTM3D(T,J) PRNCPL OUTSTNDG MEDIUM CREDIT T0 T1 T2 T3 T4 T5 T1 0.5159 0.9372 0.0000 0.0000 0.0000 0.0000 T2 0.4047 0.8682 0.9372 0.0000 0.0000 0.0000 T3 0.2825 0.7923 0.8682 0.9372 0.0000 0.0000 T4 0.1480 0.7088 0.7923 0.8682 0.9372 0.0000 T5 0.0000 0.6169 0.7088 0.7923 0.8682 0.9372; TABLE PRINTM2D(T,J) PRNCPL AND INTRST PMTS MEDIUM CREDIT T0 T1 T2 T3 T4 T5 T1 0.2538 0.2538 0.0000 0.0000 0.0000 0.0000 T2 0.2538 0.2538 0.2538 0.0000 0.0000 0.0000 T3 0.2538 0.2538 0.2538 0.2538 0.0000 0.0000 T4 0.2538 0.2538 0.2538 0.2538 0.2538 0.0000 T5 0.2538 0.2538 0.2538 0.2538 0.2538 0.2538;

121

TABLE INTMED2(T,J) INTRST PMTS ON MEDIUM TERM CREDIT T0 T1 T2 T3 T4 T5 T1 0.0850 0.0850 0.0000 0.0000 0.0000 0.0000 T2 0.0707 0.0707 0.0850 0.0000 0.0000 0.0000 T3 0.0551 0.0551 0.0707 0.0850 0.0000 0.0000 T4 0.0382 0.0382 0.0551 0.0707 0.0850 0.0000 T5 0.0119 0.0119 0.0382 0.0551 0.0707 0.0850; TABLE PRMED2(T,J) PRNCPL PMTS MEDIUM TERM CREDIT T0 T1 T2 T3 T4 T5 T1 0.1688 0.1688 0.0000 0.0000 0.0000 0.0000 T2 0.1831 0.1831 0.1688 0.0000 0.0000 0.0000 T3 0.1987 0.1987 0.1831 0.1688 0.0000 0.0000 T4 0.2156 0.2156 0.1987 0.1831 0.1688 0.0000 T5 0.2339 0.2339 0.2156 0.1987 0.1831 0.1688; TABLE PROUTM2D(T,J) PRNCPL OUTSTNDG MEDIUM CREDIT T0 T1 T2 T3 T4 T5 T1 0.8312 0.8312 0.0000 0.0000 0.0000 0.0000 T2 0.6481 0.6481 0.8312 0.0000 0.0000 0.0000 T3 0.4494 0.4494 0.6481 0.8312 0.0000 0.0000 T4 0.2339 0.2339 0.4494 0.6481 0.8312 0.0000 T5 0.0000 0.0000 0.2339 0.4494 0.6481 0.8312; TABLE PRMEDD(T,J) ADVANCED PRNCPL PMTS MEDIUM TERM CREDIT T0 T1 T2 T3 T4 T5 T0 0.1661 0.1678 0.0000 0.0000 0.0000 0.0000 T1 0.1815 0.1825 0.1664 0.0000 0.0000 0.0000 T2 0.1984 0.1986 0.1817 0.1661 0.0000 0.0000 T3 0.2169 0.2161 0.1985 0.1815 0.1661 0.0000 T4 0.2371 0.2351 0.2167 0.1984 0.1815 0.1661 T5 0.0000 0.0000 0.2367 0.2169 0.1984 0.1815; TABLE PRF20D(T,J) ADVANCED PRINCIP PYMNTS 20 YR FIXED TERM LOAN T0 T1 T2 T3 T4 T5 T0 0.0474 0.0228 0.0000 0.0000 0.0000 0.0000 T1 0.0511 0.0246 0.0218 0.0000 0.0000 0.0000 T2 0.0550 0.0264 0.0236 0.0216 0.0000 0.0000 T3 0.0593 0.0284 0.0255 0.0233 0.0216 0.0000 T4 0.0638 0.0306 0.0275 0.0252 0.0233 0.0216 T5 0.0687 0.0329 0.0297 0.0273 0.0252 0.0233

122

VARIABLES WLTH CHANGE IN NET WORTH OVER PLANNING PD; POSITIVE VARIABLES decvar(X,T) Decision Variables begcash(T) Amount of cash existing in first subperiod endcash(T) Amount of cash existing in second subperiod cashland(T) Cash used to purchase land casheq(T) Cash used to purchase equipment famcon(T) Family consumption txincm(T) Amount of taxable income invest(T) Amount invested in Treasury bills offarm(T) Amount of off farm income offinv(T) Amount of off farm investments hirelab(T) Amount of hired labor taxes(T) Amount paid in taxes in year t casset(T) Amount of Current Assets iasset(T) Amount of Intermediate Assets *ownland(T) Acreage owned lasset(T) Amount of Long Term Assets tasset(T) Amount of Total Assets culiab(T) Amount of Current Liabilities intliab(T) Amount of Intermediate Liabilities lgliab(T) Amount of Long Term Debt totliab(T) Total Debt nworth(T) Net Worth cureq(T) Current Equity inteq(T) Intermediate Equity lgteq(T) Long Term Equity mscredit(T) Maximum Short Term Credit Reserves micredit(T) Maximum Intermediate Term Credit Reserves mlcredit(T) Maximum Long Term Credit Reserves interest(T) Total interest charges paid depre(T) Depreciation; EQUATIONS OBJ Objective Function PRODMARK Acreage Produced and Marketed FINALAND Long-Term Financing for Land LABORCON Labor Constraint LANDDWP Land Downpayment Requirement MACHREQT Machinery Use Constraint FINAMACH Machinery Financing MACHDWP Machinery Downpayment Requirement FAMICONS Family Consumption OFFARINV Off-Farm Investment OFFARINC Proceeds from Off-farm Investments OFFIDENT Off-Farm Income Identity TRANSF1 Cash Transfer in First Subperiod of the year TRANSF2 Cash Transfer in Second Subperiod of the year

123

TAXINCOM Taxable Income Constraints CURRAS `Current Assets INTAS Intermediate Assets LNGTAS Long Term Assets TOTAS Total Assets CUDEBT Short Term Debt INTDEB Intermediate Term Debt LNGDEBT Long Term Debt TOTDEB Total Debt EQUITY Owner's Equity TAXPAYMT Total Taxes Paid in a year DEPN Depreciation Charges FSIZE Upper Limit on Farm Size CEQUITY Current Equity IEQUITY Intermediate Term Equity LEQUITY Long Term Equity MAXSDEBT Maximum Short-Term Debt MAXIDEBT Maximum Intermediate-Term Debt MAXLDEBT Maximum Long-Term Debt INTREST Interest Charges paid; OBJ.. WLTH =E= SUM(T,(nworth(T)-nworth(T-1))); PRODMARK(T)$(ORD(T) GT 1).. decvar("ACPROD",T)-decvar("ACRENT",T) -SUM(J$(ORD(J) LE ORD(T)),decvar("ACBUY",J))=E=0; FINALAND(T)$(ORD(T) GT 1).. LANDVAL(T)*decvar("ACBUY",T) -cashland(T)-decvar("FINLAND",T)=E=0; LABORCON(T)$(ORD(T) GT 1).. hirelab(T)- (LABREQ*(decvar("ACRENT",T)-SUM(J$(ORD(J) LE ORD(T)), decvar("ACBUY",J))))=G=OVRLAB; LANDDWP(T)$(ORD(T) GT 1).. cashland(T)- (LNDDPMT*LANDVAL(T)*decvar("ACBUY",T))=G=0; MACHREQT(T)$(ORD(T) GT 1).. EQCOST(T)*(decvar("ACRENT",T) -SUM(J$(ORD(J) LE ORD(T)),decvar("ACBUY",J)))- SUM(J$(ORD(J) LE ORD(T)),DEPVALEQ(T,J)*decvar("EQBUY",J))=L=0; FINAMACH(T)$(ORD(T) GT 1).. (1-EQDPMT)*decvar("EQBUY",T) -decvar("MEDCRED",T)=E=0; MACHDWP(T)$(ORD(T) GT 1).. casheq(T)- EQDPMT*decvar("EQBUY",T)=G=0; FAMICONS(T)$(ORD(T) GT 1).. famcon(T)=E= FACONS(T); OFFARINV(T)$(ORD(T) GT 1).. offinv(T)-invest(T)=E=0; OFFARINC(T)$(ORD(T) GT 1).. offinv(T)*YIELD(T)-MOFFINC(T)=L=0;

124

OFFIDENT(T)$(ORD(T) GT 1).. offinv(T)*YIELD(T)- offarm(T)=E=0; TRANSF1(T)$(ORD(T) GT 1).. VARCOST(T)*(decvar("ACRENT",T) -SUM(J$(ORD(J) LE ORD(T)),decvar("ACBUY",J))) +0.5*RENT(T)*decvar("ACRENT",T)+OVRHD(T) +0.5*PROPTAX*SUM(J$(ORD(J) LE ORD(T)), decvar("ACBUY",J)) +cashland(T) +invest(T) -endcash(T-1) +begcash(T)-decvar("SHTCRED",T) +XPRATE*famcon(T)+casheq(T) +(LABCOST(T)*(hirelab(T)/2))=E=0; TRANSF2(T)$(ORD(T) GT 1).. -RETURN(T)*(SUM(J$(ORD(J)LE ORD(T)), decvar("ACBUY",J))+decvar("ACRENT",T)) +0.5*RENT(T)*decvar("ACRENT",T) +0.5*PROPTAX*SUM(J$(ORD(J) LE ORD(T)),decvar("ACBUY",J)) +SUM(J$(ORD(J) LE ORD(T)),FPRIN20(T,J)*decvar("FINLAND",J) +PRINTMED(T,J)*decvar("MEDCRED",J)) +PRINTS(T)*decvar("SHTCRED",T) -invest(T)*INVGAIN(T) -SUM(J$(ORD(J)LE ORD(T)), EQDEP(T,J)*decvar("EQBUY",J)) +endcash(T) -begcash(T)+XPRATE*famcon(T) +INCTAX*txincm(T-1) +(LABCOST(T)*(hirelab(T)/2))=E=0; TAXINCOM(T)$(ORD(T) GT 1).. -MARGIN(T)* (decvar("ACRENT",T) -SUM(J$(ORD(J) LE ORD(T)),decvar("ACBUY",J))) +PROPTAX*SUM(J$(ORD(J) LE ORD(T)), decvar("ACBUY",J)) +RENT(T)*decvar("ACRENT",T) +SUM(J$(ORD(J) LE ORD(T)),INTF20(T,J)*decvar("FINLAND",J) +INTMED(T,J)*decvar("MEDCRED",J))+INPMTS(T)*decvar("SHTCRED",T) +SUM(J$(ORD(J)LE ORD(T)), EQDEP(T,J)*decvar("EQBUY",J)) +LABCOST(T)*hirelab(T) -offarm(T) +txincm(T)=E=0; CURRAS(T)$(ORD(T) GT 1).. casset(T)-endcash(T)-OTHCUR0 -SUM(J$(ORD(J) LE ORD(T)), invest(T)) =E=0; INTAS(T)$(ORD(T) GT 1).. iasset(T)- SUM(J$(ORD(J)LE ORD(T)), DEPVALEQ(T,J)*decvar("EQBUY",J))=E=0; LNGTAS(T)$(ORD(T) GT 1).. lasset(T)- LANDVAL(T)*SUM(J$(ORD(J) LE ORD(T)), decvar("ACBUY",J)) -BUILD0 =E=0;

125

TOTAS(T).. tasset(T)-casset(T)-iasset(T)-lasset(T)=E=0; CUDEBT(T)$(ORD(T) GT 1).. culiab(T) -SUM(J$(ORD(J) LE ORD(T)), PRINTMED(T,J)*decvar("MEDCRED",J) +FPRIN20(T,J)*decvar("FINLAND",J)) -INCTAX*txincm(T)=E=0; INTDEB(T)$(ORD(T) GT 1).. intliab(T)- SUM(J$(ORD(J) LE ORD(T)),PROUTMED(T,J)*decvar("MEDCRED",J)) +SUM(J$(ORD(J) LE ORD(T)), PRMEDD(T,J)*decvar("MEDCRED",J)) =E=0; LNGDEBT(T)$(ORD(T) GT 1).. lgliab(T)- SUM(J$(ORD(J)LE ORD(T)), PROUTF20(T,J)*decvar("FINLAND",J)) +SUM(J$(ORD(J) LE ORD(T)), PRF20D(T,J)*decvar("FINLAND",J)) =E=0; TOTDEB(T)$(ORD(T) GT 1).. totliab(T)-culiab(T)-intliab(T)-lgliab(T)=E=0; CEQUITY(T)$(ORD(T) GT 1).. cureq(T)-casset(T)+culiab(T)=E=0; IEQUITY(T)$(ORD(T) GT 1).. inteq(T)-iasset(T)+intliab(T)=E=0; LEQUITY(T)$(ORD(T) GT 1).. lgteq(T)-lasset(T)+lgliab(T)=E=0; EQUITY(T)$(ORD(T) GT 1).. nworth(T)-tasset(T)+totliab(T)=E=0; MAXSDEBT(T)$(ORD(T) GT 1).. mscredit(T)-5*cureq(T) -MARGIN(T)*(decvar("ACRENT",T) -SUM(J$(ORD(J) LE ORD(T)),decvar("ACBUY",J))) +RENT(T)*decvar("ACRENT",T)+LABCOST(T)*hirelab(T) =E=0; MAXIDEBT(T)$(ORD(T) GT 1).. micredit(T)-2*inteq(T) -2*SUM(J$(ORD(J) LE ORD(T)), PRMED(T,J)*decvar("MEDCRED",J)) -0.30*MARGIN(T)* (decvar("ACRENT",T) -SUM(J$(ORD(J) LE ORD(T)),decvar("ACBUY",J))) +0.30*RENT(T)*decvar("ACRENT",T)+0.30*LABCOST(T)*hirelab(T) -2*SUM(J$(ORD(J) LE ORD(T)), DEPVALEQ(T,J)*decvar("EQBUY",J)) =E=0; MAXLDEBT(T)$(ORD(T) GT 1).. mlcredit(T)-2*lgteq(T) -0.12*MARGIN(T)* (decvar("ACRENT",T) -SUM(J$(ORD(J) LE ORD(T)),decvar("ACBUY",J))) +0.12*RENT(T)*decvar("ACRENT",T) +0.12*LABCOST(T)*hirelab(T) -2*SUM(J$(ORD(J) LE ORD(T)), PRF20(T,J)*decvar("FINLAND",J)) -.05*LANDVAL(T)*SUM(J$(ORD(J) LE ORD(T)), decvar("ACBUY",J)) =E=0;

126

TAXPAYMT(T)$(ORD(T) GT 1).. taxes(T)- PROPTAX*SUM(J$(ORD(J) LE ORD(T)), decvar("ACBUY",J)) -INCTAX*txincm(T)=E=0; INTREST(T)$(ORD(T) GT 1).. interest(T)-INPMTS(T)*decvar("SHTCRED",T) -SUM(J$(ORD(J) LE ORD(T)), INTF20(T,J)*decvar("FINLAND",J) +INTMED(T,J)*decvar("MEDCRED",J))=E=0; DEPN(T)$(ORD(T) GT 1).. SUM(J$(ORD(J) LE ORD(T)), EQDEP(T,J)*decvar("EQBUY",J)) -depre(T)=E=0; FSIZE(T)$(ORD(T) GT 1).. decvar("ACPROD",T)-UPSIZE=L=0; decvar.FX("ACBUY","T0")=LAND0; decvar.FX("ACPROD","T0")=362; decvar.FX("EQBUY","T0")=EQPT0; decvar.FX("FINLAND","T0")=350099; decvar.FX("MEDCRED","T0")=422010; decvar.FX("SHTCRED","T0")=241688; decvar.FX("ACRENT","T0")=362-LAND0; endcash.FX("T0")=cash0; casset.FX("T0")=CURAST0; iasset.FX("T0")=EQPT0; lasset.FX("T0")=BUILD0+(LAND0*2150); nworth.FX("T0")=829437; txincm.FX("T0")=20171; invest.FX("T0")=100000; cureq.FX("T0")=362428; inteq.FX("T0")=46890; lgteq.FX("T0")=420119; MODEL MEBR METHYL BROMIDE STUDY /ALL/; SOLVE MEBR MAXIMIZING WLTH USING LP;

a farm-level approach to the methyl ... - getd.libs.uga.edu · a farm-level approach to the methyl...

Documents