Citation
Perception of U.S. Peanuts among Quality Control Professionals, Operations Managers, and Purchasing Agents Abroad

Material Information

Title:
Perception of U.S. Peanuts among Quality Control Professionals, Operations Managers, and Purchasing Agents Abroad
Creator:
JOHNSON, JULIE DIANE ( Author, Primary )
Copyright Date:
2008

Subjects

Subjects / Keywords:
Censorship ( jstor )
Farms ( jstor )
Flavors ( jstor )
House brands ( jstor )
Imports ( jstor )
Peanut butters ( jstor )
Peanuts ( jstor )
Prices ( jstor )
Purchasing ( jstor )
Snacking ( jstor )

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright Julie Diane Johnson. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Embargo Date:
12/31/2007
Resource Identifier:
659814091 ( OCLC )

Downloads

This item is only available as the following downloads:


Full Text

PAGE 1

PERCEPTION OF U.S. PEANUTS AMONG QUALITY CONTROL PROFESSIONALS, OPERATIONS MANAGERS AND PURCHASING AGENTS ABROAD By JULIE DIANE JOHNSON A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2006 1

PAGE 2

Copyright 2006 By Julie Johnson 2

PAGE 3

TABLE OF CONTENTS page LIST OF TABLES ...........................................................................................................................5 LIST OF FIGURES .........................................................................................................................7 ABSTRACT .....................................................................................................................................8 CHAPTER 1 INTRODUCTION................................................................................................................. ...9 The Peanut Industry ..................................................................................................................9 The American Peanut Council ..................................................................................................9 Research ..................................................................................................................................10 Objective .................................................................................................................................10 2 LITERATURE REVIEW.......................................................................................................13 Peanut History ........................................................................................................................13 Introduction of the Quota System ....................................................................................13 The 1996 Farm Bill .........................................................................................................14 The 2002 Farm Bill .........................................................................................................15 American Peanut Council .......................................................................................................16 3 DATA DESCRIPTION..........................................................................................................17 Summary of Participants .........................................................................................................17 Demographic Information ...............................................................................................17 Perceptions of Quality, Price and Value ..........................................................................18 Private Label....................................................................................................................20 Important Factors .............................................................................................................20 United States Peanuts ......................................................................................................23 Processing Peanuts ..........................................................................................................26 American Peanut Council ................................................................................................26 4 THEORY AND EMPRICAL MODEL..................................................................................35 Choosing the Model ................................................................................................................35 Theory: The Tobit Model .......................................................................................................35 Background of the Tobit Model ......................................................................................35 The Problems with Trun cating and Censoring ................................................................37 Truncated and Cens ored Distributions ............................................................................40 The Normal Distribution .................................................................................................40 The Truncated Normal Distribution ................................................................................42 The Censored Normal Distribution .................................................................................43 3

PAGE 4

The Tobit Model for Censored Outcomes .......................................................................44 Censoring from Above ....................................................................................................45 Empirical: Building the Model ..............................................................................................45 Explanation of the Independent Variables ......................................................................46 Building the Model ..........................................................................................................47 5 RESULTS AND CONCLUSIONS........................................................................................51 Tobit Model Results ...............................................................................................................51 Peanut Purchasing Simulations ...............................................................................................52 Testing Variable Significance .................................................................................................55 Magnitude of Impact ...............................................................................................................56 Conclusions .............................................................................................................................57 APPENDIX A PEANUT QUESTIONNAIRE...............................................................................................61 B TSP CODE..................................................................................................................... .........82 REFERENCES ..............................................................................................................................88 BIOGRAPHICAL SKETCH .........................................................................................................89 4

PAGE 5

LIST OF TABLES Table page 1-1 Peanut production ..............................................................................................................11 1-2 U.S. peanut exports ............................................................................................................11 1-3 Export quantities by country ..............................................................................................12 3-1 Country of respondent ........................................................................................................27 3-2 Job title of respondent ........................................................................................................27 3-3 Size of respondent by country of respondent .....................................................................28 3-4 Percent of peanut pur chases by production use .................................................................28 3-5 Country of origin for peanuts currently purchased ............................................................28 3-6 Ranking of country of origin for quality. ...........................................................................29 3-7 Ranking of country of origin for price. ..............................................................................29 3-8 Ranking of country of origin for value. .............................................................................29 3-9 County specified for purchases by private label buyers/producers ...................................29 3-10 Quality rank of countries specified by private label buyer ................................................30 3-11 Price rank of countries sp ecified by private label buyer ....................................................30 3-12 Value rank of countries sp ecified by private label buyer ..................................................30 3-13 Rank of country by best tasting .........................................................................................30 3-14 Rank of country by meeting quality specifications ............................................................30 3-15 Rank of country by quality inspection process ..................................................................31 3-16 Rank of country by timely delivery ...................................................................................31 3-17 Rank of country by contribution to profit margin ..............................................................31 3-18 Average ranking .................................................................................................................31 3-19 U.S. support .......................................................................................................................32 3-20 Support from U.S. suppliers ...............................................................................................32 5

PAGE 6

3-21 Total quality inspection savings .........................................................................................32 3-22 Significance of TQI savings ...............................................................................................33 3-23 Overall taste comparison ....................................................................................................33 3-24 Importance of overall taste .................................................................................................33 3-25 Country of origin costs .......................................................................................................33 3-26 Problems for processing p eanuts from which countries ....................................................33 3-27 Aware of American Peanut Council ..................................................................................34 3-28 Effective ways to get information ......................................................................................34 3-29 Most effective ways to get information .............................................................................34 4-1 Variable Names and Descriptions ......................................................................................49 5-1 Tobit model results ............................................................................................................59 5-2 T-values for country ...........................................................................................................59 5-3 T-values for size of the firm ...............................................................................................60 5-4 T-values for what is considered important .........................................................................60 6

PAGE 7

LIST OF FIGURES Figure page 3-1 Number of respondents per size category ..........................................................................18 3-2 User type by country ..........................................................................................................19 4-1 Share of U.S. peanuts .........................................................................................................35 4-2 Latent, censored and truncated variables ...........................................................................37 4-3 Effects of censoring and truncation ...................................................................................39 4-4 Normal distributions with truncation and censoring ..........................................................41 5-1 Change in probability of import preference for U.S. peanuts due to country ....................52 5-2 Change in probability of import preference for U.S. peanuts due to company size ..........53 5-3 Change in probability of import preference for U.S. peanuts due to what is considered important ..........................................................................................................54 5-4 Change in probability of import preference for U.S. peanuts due to type of production ..........................................................................................................................55 5-5 Magnitude of impact of variables ......................................................................................57 7

PAGE 8

Abstract of Thesis Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science PERCEPTION OF U.S. PEANUTS AMONG QUALITY CONTROL PROFESSIONALS, OPERATIONS MANAGERS AND PURCHASING AGENTS ABROAD By Julie Johnson December 2006 Chair: Lisa House Major Department: Food and Resource Economics The 2002 Farm Bill eliminated the “marketing quota” system th at supported domestic peanut prices well above wo rld levels. Since 2002 , prices have dropped and pl anted acreage ha s decreased. The peanut industry must now adjust to a more market oriented and uncert ain environment. In order to aid this process, the American Peanut Council initiated a surv ey to be conduct ed among buyers abroad. The survey, which targeted qua lity control professionals, opera tions managers and purchasing agents at companies that had pr eviously purchased peanuts from th e United States, wa s conducted in nine different countries. Fiftyseven responses were re ceived. These responses were analyzed to determine the perceptions he ld by buyers of U.S. p eanuts. In general, the U.S. scored favorably in areas such as quality and va lue and lower in ar eas such as price and profit margin. An econometric model was also estimated to determine the relationship between factors that may be considered important and actual buy ing behavior. The dependent variable tested was the percent of peanuts purchased from the U. S. Independent variables included factors such as country where respondent works, the size of the firm, factors that the company consider important and type of firm. Using the economet ric model, peanut purchasing simulations were created to show the likelihood of purchasing de cisions relative to the average consumer. 8

PAGE 9

9 CHAPTER 1 INTRODUCTION In 2002, a new farm bill radically changed th e U.S. peanut industry by eliminating the marketing quota system, which had been in place, in some form, since the 1930’s. This change required the U.S. peanut industry to adjust to a more uncertain, market-oriented environment than ever before. In 2003, the value of pea nut production fell by 30% and prices by 25%, compared with 2001. This has caused a rapid exit of producers and planted acreage reached its lowest since 1915 (Dohlman, Hoffman, & Young, 2003). The Peanut Industry The four largest peanut producing countries over the past 10 years have been China, India, Nigeria and the United States, generally in that order (FAOStat, 2006). Table 1-1 shows production quantities for these countries, along with others important to the U.S. peanut industry, for the years 1999 through 2005. The largest peanut exporters are China, Argentina, India and the U.S. Though Nigeria is the third largest prod ucer, the country exports very few peanuts. A detailed description of the history of the U.S. peanut industry will be given in Chapter 2. The history gives an explanation of all of the former peanut programs and the effects of the newest farm bill. The American Peanut Council The American Peanut Council (APC) is the trad e association that represents all segments of the peanut industry. The object ive of the American Peanut Council is to increase exports of all types of U.S. peanuts and pea nut products. In line with this goa l, the APC is taking action to find out what buyers in other coun tries want. They are interested in finding out what factors actually influence buying behavior. The APC is interested mainly in the following countries: Canada, Mexico, France, Germany, Spain, the Unite d Kingdom and Italy. Table 1-2 shows U.S

PAGE 10

peanut exports to the above liste d countries for the years1999 through 2004. The table also gives the total U.S peanut exports for these years. Research The Peanut Council initiated research usi ng a survey of quantitative and qualitative questions to learn the buying beha vior and U.S. perceptions of manufacturers (who use peanuts) overseas. The survey was originally conducted by Rose Research. Additional surveys were conducted in Mexico and Canada by Grupo PM a nd Leger Marketing, respectively. The survey was conducted in seven countries with a total of 57 respondents. A copy of the survey can be found in Appendix A. Objective The following peanut producing countries were considered: Argentina, Brazil, China, India, Mexico, Nicaragua, South Africa and the United States. Table 1-1 shows total peanut production of these countries (i n thousands of metric tons) for the years 1999 through 2005. Total export quantities for these co untries are given in Table 1-3. These tables show that Brazil, Mexico, Nicaragua and South Africa, produce and export significantly less than the other countries considered. While the objective of the Am erican Peanut Council is to increase exports, the objective of this research is to give an unbiased representa tion of the perceptions of buyers and what the buyers consider impor tant. Specific objectives include 1. Learning the perceptions that buyers in other countries have of U.S peanuts and peanuts from the other peanut producing countries. This obj ective will be realized by analyzing the survey responses using univariate and multivariate statistics. 2. A second objective is to determine the re lationship between factors that may be considered important by buyers and actual buying behavior. This will be done by estimating an econometric model with percen tage of imported peanuts purchased from 10

PAGE 11

the U.S. as the dependent variable and factor s such as country where respondent works, what the company considers important, the si ze of the firm and type of the firm as independent variables. The empirical model will be laid out in Chapter 4. Table 1-1. Peanut production Quantity of Groundnuts (1000 tons) 1999 2000 2001 2002 2003 2004 2005 Argentina 486 599 564 517 316 419 593 Brazil 179 184 202 195 188 226 292 China 12,706 14,516 14,472 14,895 13,493 14,410 14,409 India 5,258 6,480 7,200 4,363 8,333 7,000 5,900 Mexico 132 142 120 75 75 75 Nicaragua 68 68 81 60 94 104 South Africa 163 136 222 134 67 128 U.S. 1,737 1,481 1,940 1,506 1,880 1,945 2,113 Nigeria 2,894 2,901 2,683 2,699 2,797 2,937 2,937 Note: Information was not available for Me xico, Nicaragua and South Africa for 2005 Note: Information obtained from FAO websites: http://www.fao.org/es/ess/top/commod ity.html?lang=en&item=242&year=2005 and http://faostat.fao.org/site/340/DesktopDefault.aspx?PageID=340 Table 1-2. U.S. peanut exports Quantity in Tons Partner 1999 2000 2001 2002 2003 2004 World 223311 272815 178869 272558 164320 213632 Canada 70424 76247 63349 75030 61978 74551 Mexico 31980 42011 20733 29215 15832 29274 France 5108 3634 1571 6218 2902 4247 Germany 3970 6083 6002 10309 1602 1854 Spain 9066 9046 7470 8662 6550 7267 U.K. 24355 26434 15037 27462 14575 19957 Italy 2437 3415 1167 9189 959 1576 Note: Information obtained from the American Peanut Council Website: http://admin.peanutsusa.com/documents/Doc ument_Library/Peanut%20Product%20Exports.pdf 11

PAGE 12

Table 1-3. Export quantities by country Groundnut Exports (1,000 tons) Country 1999 2000 2001 2002 2003 2004 Argentina * 730.56 631.94 551.18 647.93 571.73 312.74 Brazil 13.09 1.71 14.49 14.41 32.94 68.66 China 746.86 854.81 1,053.11 1,117.01 1,145.54 1,061.81 India 238.08 236.98 178.26 91.93 305.21 516.77 Mexico 8.9 18.98 21.21 13.52 15.25 7.22 Nicaragua 25.27 65.73 65.14 79.64 72.64 92.35 South Africa 39.45 40.76 54.85 75.75 30.38 30.32 U.S. 314.72 378.37 260.07 375.44 318.52 358.36 Note: Information obtained from FAO website: http://faostat.fao.or g/site/343/DesktopDefault.aspx?PageID=343 * According to the numbers obtained, Argentin a is exporting more than they are producing. FAOstat accounts for this by putting the differe nce into a category cal led “other uses.” 12

PAGE 13

CHAPTER 2 LITERATURE REVIEW Peanut History The 2002 Farm Security Act made monumental ch anges to the peanut program. This act eliminated the marketing quota system, which had existed in some form since the 1930s, changing the structure of the peanut industr y. Support of peanut pr oduction began with the Agricultural Adjustment Act of 1933. Designa ted as a basic commodity in 1934, government contracts were made with peanut producers to return their acreag e in return for payment. In 1937, regional growers associations were formed to purchase speci fied quantities of peanuts at government-set support prices. However, this pe anut program failed, most likely because it was voluntary (Chvosta, Thurman, Brown & Rucker, 2002). In 1941 a mandatory program was established and acreage allotments were set. This program penalized producers who produced a dditional acreage, but during World War II, compliance with the program wasn’t enforced. During this time period, plantings soared above the allotment of 1.9 million acres to 3.4 million acres. The next change to policy affecting peanut production came with the Agricultural Act of 1949, which established support prices for certain commodities, including peanuts. Fr om 1949 to 1978, peanuts were guaranteed the support price. During this time period, the deve lopment of new peanut varieties and production techniques caused per-acre yields to grow and the cost of the pr ogram to increase dramatically. Once again, the peanut program failed (Chvosta, Thurman, Brown & Rucker, 2002). Introduction of the Quota System In 1978, a program was established to give producers the support pr ice only on “quota” peanuts. The quota was set annually in poundage terms to meet expected edible demand. From 1978 to 1982 farmers were required to own bot h poundage and acreage allotments, but in 1982 13

PAGE 14

the acreage allotments were abandoned. Although the support price was paid only for quota peanuts, there was no limit to to tal domestic production. Before 1977, all peanuts were sold to the edible market or placed under loan with th e Commodity Credit Corporation (CCC). Either way, farmers would receive the support price. Under the 197 7 program, growers who produced more than their quota had two options. They c ould either make a contract with a handler for export or sale in the crush market or place their peanuts under loan with the area grower association. In either case, the peanuts beyond quota would receive a much lower price (Chvosta, Thurman, Brown & Rucker, 2002). The approval of NAFTA and GATT in the fall of 1993 brought considerable pressure to the peanut program. Many members of congress wa nted substantial change s to the program, and thus, the peanut program was revised as part of the 1996 Farm Bill (Chvosta, Thurman, Brown & Rucker, 2002). The 1996 Farm Bill The goal of the 1996 Farm Bill revision to the peanut program was to “guarantee stable income to peanut producers and to ensure an ample supply for the domestic market” (Chvosta, Thurman, Brown & Rucker, 2002). The bill extended the life of the peanut program for seven years. The goal was to continue to be accomplis hed through a two-tiered price support system and a ban on imports. However, participants su ggested that changes should be made to the program so that there would be no net cost to the government. While peanut proponents argued that the program supported rural communities, and therefore, the structure should be maintained, shellers and manufacturers asked for a significant reduction in support prices, or even an elimination of the program (Chvos ta, Thurman, Brown & Rucker, 2002). The final law agreed on by the house and senate resulted in im portant changes. The peanut support price was lowered from $670 to $610 per ton, marking the first time in history the 14

PAGE 15

support price for peanuts had been lowered. In 1 998, the house rejected an amendment to further lower the quota loan rate to $550 per ton. Th e divide between growers and manufacturers continued to grow. Other changes included the elimination of factor s that had formerly increased the cost of the program, such as removing the allowance to let the price support increase at the same rate as the cost of production. The aggr egate quota was again set to re flect the domestic demand for edible peanuts. In the past, unused quota coul d be transferred to th e next planning season. Under the 1996 Farm Bill, this option was e liminated (Chvosta, Thurman, Brown & Rucker, 2002). The 2002 Farm Bill The 2002 Farm Bill, also known as the 2002 Farm Security Act, again changed the peanut program in an attempt to have the program more closely resemble programs for other crops. The marketing quota system was replaced with a set of supports similar to those of other crops. The changes reduced revenue and price stability and forced the peanut industry to become more market-oriented (Dohlman, Hoffman & Young, 2003) . “The key provisions comprise fixed decoupled payments, counter cyclical paymen ts, and a marketing loan program” (Chvosta, Thurman, Brown & Rucker, 2002). The act terminated the poundage quota. Under the 2002 Farm Bill, all peanut producers, former quota holders and not, are eligible to market their peanuts for domestic edible consumption. All peanut producers are also eligible for marketing assistance loans at a rate of $355 per ton. Producers who were formerly quota producers are also eligible for direct payments of $36 per ton and counter-cyclical payments when prices are below $495 per ton. The producer s who are eligible for direct payments and counter-cyclical payments are not required to produce peanuts to r eceive the payments. They are 15

PAGE 16

only required to keep the land in approved agricultural uses. Quota holders were also eligible for a quota buyout in the amount of $1,100 per short ton (Dohlman, Hoffman, & Young, 2003). Because of the 2002 Farm Bill, the peanut i ndustry has become mo re market-oriented. Since 2002 peanut producers have seen large losses in revenue, and there has been a rapid exit of peanut producers. In 2003, the value of pea nut production fell by 30 percent and prices fell by 25 percent, compared with 2001. Planted acreage in the U.S. reached its lowest level since 1915. As the bulk of peanut production is in just seven states and a small number of counties, the changes in the peanut program impact not only peanut producing families, but also rural communities (Dohlman, Hoffman, & Young, 2003). Futures contracts are not available to peanut producers as they are to producers of many other crops. Thus contract marketing has emer ged as a risk aversion strategy. Revoredo, Giha, Ndolnyak, and Fletcher (2005) studied the clauses of various contracts se t by shellers in an attempt to learn what the motivation is behind the c ontracts. It appears th at these contracts have been set to replace the marketi ng structure that existed prior to the 2002 Farm Bill. “The reduction of transaction costs associated to the need for coordinating a continuous supply of homogeneous quality seems to be the most plau sible explanation.” There are two types of contracts that growers have been observed to enter into. The two are forward contracts for delivery at harvest or a later date and “option to purchase” contracts. American Peanut Council The American Peanut Council (APC) was form ed in 1997 as a merger of the National Peanut Council and the National Peanut Council of America. The APC represents all segments of the peanut industry and is the only U.S. orga nization to do so. The four segments represented include peanut growers, peanut shellers, pean ut product manufacturers and allied, brokers and international ( www.peanutsusa.com ). 16

PAGE 17

CHAPTER 3 DATA DESCRIPTION The following is a summary of the data co llected from telephone interviews of manufacturers who use peanuts. The interviews were conducted by Rose Research from October 2005 to April 2006. Two additional research fi rms were used from April 2006 – May 2006 to reach additional firms in Mexico and Canada. In total, 57 participants from 9 different countries responded. Information on non-response was not pr ovided by the research firms who conducted the surveys. Summary of Participants Demographic Information Participants from nine countries were surveyed, and responses from countries varied, with the majority of respondents from Mexico, Canada, and the UK, in part due to a second effort targeting Mexico and Canada (Table 3-1). The remaining surveys come from countries in the European Union and Scandinavia. At least tw o respondents were generated from each of the nine countries participating in the survey. Respondents were as ked their job title, and answers varied, with the most common answer being ma naging director, followed by operations manager (Table 3-2). Manufacturers surveyed spanned the different size categories, with 13 respondents in the smallest category (under 1,000 tons purchased) a nd 4 purchasing over 25,000 tons of peanuts in a typical year (Figure 3-1). These responses differed by country as shown in Table 3-3. Respondents were asked to identify what the peanuts purchased were used for (Table 3-4). These results also showed that a broad spectru m was sampled, and that responses per country differed (Figure 3-2). Most respondents, 81% (46 respondents), purchased for snack kernel peanuts, including 23 respondents who only pur chased for snack kernel; while only 7 17

PAGE 18

0 2 4 6 8 10 12 14 16 18 20 Under 1,0001,000 less than 5,000 5,000-less than 10,000 10,000-less than 25,000 25,000 or more Tons of peanutsF re quenc y Figure 3-1. Number of re spondents per size category respondents purchased peanuts for confectionary. The other categor ies were in shell and peanut butter, which 21 and 12 respondents, respectivel y, purchased for. Respondents from Canada were more likely to have a firm with no snack kernel use and respondents from the U.K. reported either 100% snack kernel or 100% peanut butter production use. Respondents also identified what countries they currently imported peanuts from. It should be noted that 7 responses from Mexico, 2 from the UK, and 1 from Canada had to be removed from this portion of the survey due to mistakes gathering the data. The remaining 47 responses are reported in Table 3-5. Of those that respond ed, all but two cu rrently purchase peanuts from the United States. The two that did not were both from Canada and purchased 100% of their peanuts from China. Perceptions of Quality, Price and Value Before beginning a sequence of questions abou t quality, respondents were asked to define what quality meant to them. Answers to the question varied, though so me answers appeared consistently. Flavor or taste ap peared in 68% (39) of the res ponses; consistency or uniformity and aflatoxin each appeared in 37% (21) of th e responses; free of foreign material or clean appeared in 35% (20) of the responses; and color and size appeared in 33% and 32% of the 18

PAGE 19

0 10 20 30 40 50 60 70 80 90 100 0123456789Percent of Production Snack In Shell Peanut Butter Confectionary Figure 3-2. User type by country Country Legend: 1 = Canada; 2 = Mexi co; 3 = France, 4 = Scandinavia, 5 = Germany; 6 = Spain; 7 = Holland; 8 = UK; 9 = Italy responses, respectively. Next, resp ondents were asked to rank the countries they imported from (plus the United States if not already included) in terms of quality (Table 3-6). The United States was ranked first 47 out of 56 times . Nicaragua was ranked first 5 times, followed by China twice and Argentina once. One respondent, who purchased 100% of peanuts used from China, did not answer the question, indicating the only factor in purchasing was price, not quality. This respondent was a purchasing agent for a Canadian company that produced 100% snack kernels. Immediately following the quality ranking, respo ndents were asked to rank countries in terms of price (Table 3-7) and value (Table 3-8). China and the United States were ranked first in terms of price, while the United States was ranked first 38 times for value, followed by China with 7 first rankings. Respondents who ranked China first for value pur chased 20 – 100% of their peanuts from China. The respondents that ranked China firs t included Canada (4), the UK 19

PAGE 20

(2), and Spain (1). Nicaragua was ranked first fo r quality, price, and value 5 times by different respondents (only 1 respondent ranked Nicaragua first in all 3 categories). Finally, respondents were asked to identify what made peanuts from the best country best, and the worst country worst. As far as what makes the peanuts from the best country the best, 63% (36 respondents) mentioned different peanut properties such as taste and flavor, 32% mentioned consistency, 19% mentioned safety, 8% price and 7% other responses. When asked what makes the peanuts from the worst country th e worst, 47% gave no answer or did not give a valid answer, 17% mentioned foreign material, 15 % peanut properties, 7% aflatoxin levels, 5% consistency, 7% gave other answers. Private Label Though quality may drive purchasing behavior, most manufacturers were private label buyers (72% of respondents). Of those, 71% indicated the private label buyer/producer specified origins for purchasing peanuts (30 total respondents). In this ca se, the majority indicated the United States, however, 21% of private label buy ers/producers specified China as the source for peanuts (Table 3-9). Respondents were then asked to rank the countri es specified by their private label buyer in terms of quality, price, and value. Answers tended to be very similar to the previous questions about quality, price and value (3-10 to 3-12). Important Factors Respondents were asked which factors were mo st important when selecting a country to import/purchase peanuts from, and then asked to identify which factor was most important. Overall, price and quality were the most fre quently mentioned factors, mentioned by 73% and 64% of the respondents, respectively (note that many respondents mentioned more than one factor). Shipping time or delivery time was me ntioned 29% of the time, followed by availability 20

PAGE 21

14% of the time. Items that scored high in the definition of quality like taste, lack of foreign material, aflatoxin levels, consistency, and size were only mentioned 11%, 5%, 7%, 3%, and 7% of the time, respectively. The most important fa ctor was quality or price, mentioned 75% of the time. Other responses varied greatly. When asked why the respondent’s company imported or purchased peanuts from the United States, the responses were similar, with one added response. Again, quality and price were the most common answers, coming from 43% and 25% of the respondents. An additional reason stated for purchase from the United States was an established relationship. Ten percent of respondents indicated they purch ase from the U.S. because of good past experiences or a relationship. Additionally, 22% i ndicated fast shipping times and increased shelf life as a reason for purchase. Taste or flavor was mentioned by 68% of the re spondents as an important factor in quality, by six respondents when asked for key factors in purchase decisions, and by four respondents when asked for key factors in purchase decisi ons from the United States. Respondents were asked to rank the countries in terms of “best ta sting peanuts” (Table 313). The United States was ranked first 89% of the time (50 responses). Nicaragua was mentioned 3 times, China twice, and Argentina one time. Meeting quality specifications was not directly mentioned in the definition of quality, but uniformity and consistency were mentioned by 37% of the respondents, likely indicating the same intention. The United States was ranked firs t for meeting quality specifications by 95% of the respondents. Nicaragua, China, and Argen tina each received one first place ranking (Table 3-14). Information on the quality inspection process showed similar results (Table 3-15). Timely delivery, though not mentioned by any respondent in terms of quality, was frequently mentioned 21

PAGE 22

as a reason to purchase peanuts (by 16 respondents) and as a reason to purchase peanuts from the United States (by 13 respondents). Ranking of countri es in regard to a country’s performance on timely delivery again resulted in high ranki ngs for the United States (Table 3-16). Then, respondents were asked to identify whic h peanut producing count ries were best in terms of their profit margins (Tab le 3-17). In this case, the Unite d States received 55% (28) of the first place rankings. China followed with 27% (14), Argentina had 3 first place rankings, and Brazil, Nicaragua, and South Africa each had 2 first place rankings. The United States was ranked last 3 times. Respondents were asked to identify what make s the quality inspection process from the country they considered the best better than others. Eighteen per cent did not give an answer and others only said that the inspection process was better instead of gi ving reasons why. Some common answers were better tec hnology, more government involvement, higher standards, more attention to detail and better personnel. Next respondents were asked wh at makes the total quality insp ection from the country that they considered to be worst, worse than others. 50% of resp ondents did not give an answer (some saying “don’t know,” others seemingly not understanding the quest ion). Some common answers were bad personnel, little government involvement, lower quality standards, less attention to detail and little technology. When asked what makes the best country bette r than other countries in terms of offering timely delivery, 14 mentioned logistics/infrastr ucture, 13 location, 12 time, 3 management, 2 technology and 2 communication. When asked what makes the worst country worse in terms of timely delivery, most answered none or not applicable (likely a result of not knowing and not dealing with the country 22

PAGE 23

they perceive to be the worst). Of those who answered, 13 mentioned time, 6 logistics or infrastructure, 4 location and 4 management. Respondents were next asked what makes th e profit margins from the best country the best. Again many did not answer (20 responde nts). Answers were price (13 respondents), quality (8 respondents), size (5 respondents) and foreign material (3 respondents). Next respondents were asked what makes profit margins from the worst country worse than others. Thirty-eight respondents did not answer. Ten respondents mentioned price and there were a variety of other answers. Respondents were then asked to gi ve the freight cost for 1 metr ic ton of peanuts from each of the countries they import or purchase peanut s from. There were no responses to this question. Respondents were asked to rank a list of items that may affect buying decisions in order of importance, where a 5 means “extremely important” and a 1 means “not important at all.” The highest average ranking was 4.95 fo r aflatoxin levels. The lowest was 3.40 for participation in seed programs. Table 3-18 shows th e average rankings for all items. United States Peanuts At this point in the survey, direction changed to focus on th e U.S. in comparison to other countries. Respondents were first asked about support from the U.S. compared to other countries as much better, somewhat better, about the same , somewhat worse or much worse. Forty nine percent (28 respondents) felt that th e U.S. was much better than ot hers in terms of support, and 30% (17 respondents) felt that the U.S. was somewhat better. Only one respondent indicated that the U.S. was somewhat worse, while no respondents indicated that the U.S. was much worse. Responses to this question are listed in Table 3-19. When asked why they felt this way, there were a variety of answers. Thirty-five percent (20 respondents) did not give an answer (some because they saw no difference, others because 23

PAGE 24

they had no problems, some said no comment and others did not understand the question). Seventeen percent mentioned that the U.S. had professional support, 12% mentioned communication. Other answers included organiza tion (7%), fast support (5%), and availability of APC, USDA and FDA (3.5%). 3.5% said that it depends on the company not the country. There were a variety of other answers (14%). Next respondents were asked about the support from a supplie r perspective. Forty one percent said that the U.S. was much better. Th irty percent said somewh at better, and 18% said about the same. No respondents answered that the U.S. was somewhat worse or much worse, though 6 respondents did answer that they did not know. Table 3-20 shows the distribution of answers to this question. When asked why they felt this way, most res pondents responded with the same answer as for the previous question. Those who did not give the exact same answer tended to give answer very near their answer for the previous question about U.S. support in general. Respondents were asked what advantages and disadvantages are asso ciated with U.S. peanuts. In terms of advantages, 10 % responded with none or don’t know, 26% mentioned peanut properties, 21% good delivery, 19% consiste ncy, 8% low afltatoxin levels, 8% proximity and 7% low amounts, or no, foreign material. Eight an swers were classified as other for a total of 14% of the responses. In terms of disadvantages encountered with peanuts from the U.S., 61% of respondents indicated that there were no disadva ntages with the U.S., 30% mentioned price, 3.5% said that U.S. peanuts were not c onsistent, and 7% gave other answers. When asked if there are any cost savings b ecause of the total quali ty inspection process that U.S. peanuts go through, 81% responded yes and 19% responded no (Table 3-21). Respondents were also asked to qua ntify these savings but were not able to. Those who indicated 24

PAGE 25

that there were savings associated with the U.S. total quality inspection process were then asked if these savings were extremely significant, very significant, somewhat significant, not very significant or not at all significant. Only 14% of respondents said that the savings were extremely significant, 48% said they were very significant, 32% somewhat significant, and 7% not very significant (table 3-22). Those who answered no to the total quality inspection process saving money were asked why it does not save money. It appears that this question was not only asked of those who said that the process did not save them money but also of those who sa id that it did. Most answered that they did not know or gave no answ er. A few said that it didn’t save enough money. Others answered that the total quality inspection proce ss DID save them money by having better quality, which allowed them to sell them faster or that it meant they did not have to complete the inspection process themselves. Next respondents were asked how the overall taste of U.S. peanuts compared to other countries. 30% said that U.S. peanuts were mu ch better, 45% somewhat better, 21% about the same, 0 somewhat worse, 2% much worse, and 2% don’t know (Table 3-23). Adding the categories of much better and somewhat better to gether shows that 75% of respondents consider U.S. peanuts to have better overall taste than other countries. Respondents were then asked why they felt this way. Most repeated that they have better taste or flavor. However, there were a few ot her responses, 10% oil levels, 8% sweeter, 8% fresher, 5% said they learned this from consum er testing, 3.5% said they have a more intensive taste. A few other answers were crunchy, better roasting, and that no one compares. Next respondents were asked to use a 5 point scale to ra nk whether the taste of U.S. peanuts was worth paying for (5 means extrem ely valuable and worth paying more for and 1 25

PAGE 26

means that it is not valuable at all and is not worth paying fo r). Though many respondents said that U.S. peanuts have a bette r taste, few felt it was worth paying for. The majority of respondents (40%) were indifferent and gave answers in the middle. Many more answered this question with either a 1 or 2. See Table 3-24 for the exact distribution. Processing Peanuts The next series of questions focused on p eanut processing. Respondents were asked to describe some of the issues faced when processing peanuts. The most frequent response was presence of foreign materials (17 respondents), fo llowed by aflatoxin (6), sk ins (6), and moisture levels (4). Other problems mentioned were contamination, immature peanuts and damaged nuts. Of these problems, respondents were asked to list the ones that have major cost implications in processing. The biggest were damaged nuts, contaminated nuts and foreign materials at 21%, 18% and 11%. Aflatoxin was the next at 5%. Other answers included immature nuts and loss of output. Respondents were asked if country of origin caused any of the co st increasing factors mentioned. Half said yes. Half said no (Table 3-25). Next respondents were asked which countries they have this problem with. Fortyfive percent mentioned China, 24% U.S. and 17% Argentina. The full distribution of answers to this question is cont ained in Table 3-26. American Peanut Council Respondents were asked whether or not they we re aware of the Ameri can Peanut Council. The majority of respondents (86%) said that they were aware of th e APC (Table 3-27). Respondents were then asked what the most effective ways were to get information. Respondents were asked to “check all that ap ply.” Answers show that all ways listed (newsletters, website/internet, seminars, visits and publications) are effective ways to get information. See Table 3-28 for distribution of answers. Lastly respondents were asked what 26

PAGE 27

the most effective way would be for the American Peanut Council to get information to them. For this question, respondents were asked to “check only one.” Forty-seven percent said websites or internet, 28% said visits, 11% semina rs, 9% publications and 5% newsletters (Table 3-29). This indicates that t hough technology has made the transfer of information easy for most, many still prefer face-to-face meetings. Table 3-1. Country of respondent Country Frequency Percent Canada 10 0.16 Mexico 16 0.14 France 3 0.07 Scandinavia 3 0.07 Germany 3 0.07 Spain 8 0.18 Holland 2 0.05 UK 10 0.23 Italy 2 0.05 Total 57 1 Table 3-2. Job title of respondent Job Title Frequency Percent Quality Control 5 0.09 Operations Manager 12 0.21 Purchasing Agent 4 0.07 Managing Director 14 0.25 Other 22 0.39 Total 57 1 27

PAGE 28

Table 3-3. Size of respondent by country of respondent Under 1,000 1,000 less than 5,000 5,000less than 10,000 10,000less than 25,000 25,000 or more Canada 4 1 0 4 1 Mexico 8 2 5 0 1 France 0 1 0 2 0 Scandinavia 1 1 1 0 0 Germany 0 1 1 1 0 Spain 0 3 4 1 0 Holland 0 0 0 1 1 UK 1 2 6 0 1 Italy 0 0 1 1 0 Table 3-4. Percent of peanut purchases by production use Frequencies Percent Snack Kernel Inshell Peanut Butter Confectionary Other zero 11 36 44 50 54 less than or equal to 25% 2 10 3 4 0 more than 25-50% 10 7 0 1 2 more than 50-75% 7 1 2 1 1 more than 75100% 27 3 7 1 0 Table 3-5. Country of origin for peanuts currently purchased Percent Argentina Brazil China India Mexico Nicaragua South Africa U.S 0% 30 37 21 41 43 35 40 2 25% or less (>0%) 8 6 11 2 0 8 2 5 26 – 50% 5 0 8 0 0 1 1 8 51 – 99% 1 0 1 0 0 0 0 14 100% 0 0 2 0 0 0 0 15 28

PAGE 29

Table 3-6. Ranking of country of origin for quality. Rank Argentina Brazil China India Mexico Nicaragua South Africa U.S. 1 1 0 2 0 0 5 0 47 2 9 1 14 0 0 7 1 5 3 3 3 10 0 0 1 2 2 3 3 3 10 0 0 1 2 2 4 3 2 4 1 0 1 0 0 5 3 1 0 1 0 0 0 1 Table 3-7. Ranking of country of origin for price. Rank Argentina Brazil China India Mexico Nicaragua South Africa U.S. 1 3 1 21 0 0 5 2 21 2 9 4 3 1 0 3 1 14 3 3 0 3 0 0 2 0 11 4 2 0 2 0 0 2 0 3 5 0 0 0 1 0 0 0 3 Table 3-8. Ranking of country of origin for value. Rank Argentina Brazil China India Mexico Nicaragua South Africa U.S. 1 3 0 7 0 0 5 1 38 2 6 3 15 0 0 2 2 7 3 3 1 6 0 0 5 0 5 4 3 1 3 1 0 0 0 2 5 3 1 0 1 0 0 0 0 Table 3-9. County specified for purchas es by private label buyers/producers Countries Frequency Percent Argentina 7 0.15 Brazil 2 0.04 china 10 0.21 india 3 0.06 Mexico 0 0.00 Nicaragua 4 0.08 South Africa 0 0.00 United States 22 0.46 Total 48 1 Note: Some respondents indicated more than one country was specified. 29

PAGE 30

Table 3-10. Quality rank of countries specified by private label buyer Rank Argentina Brazil China India Mexico Nicaragua South Africa U.S. 1 1 0 1 0 0 1 0 21 2 4 1 3 1 0 2 0 2 3 0 1 5 1 0 0 0 0 4 1 0 1 1 0 0 0 0 5 1 0 0 0 0 1 0 0 Table 3-11. Price rank of countries specified by private label buyer Rank Argentina Brazil China India Mexico Nicaragua South Africa U.S 1 4 1 7 1 0 1 0 10 2 1 1 3 2 0 1 0 5 3 1 0 1 0 0 1 0 4 4 1 0 0 0 0 1 0 1 5 0 0 0 0 0 0 0 2 Table 3-12. Value rank of countries specified by private label buyer Rank Argentina Brazil China India Mexico Nicaragua South Africa U.S. 1 3 0 2 0 0 2 0 16 2 1 2 4 2 0 2 0 2 3 1 0 4 0 0 0 0 2 4 0 0 0 1 0 0 0 2 5 2 0 0 0 0 0 0 0 Table 3-13. Rank of country by best tasting Rank Argentina Brazil China India Mexico Nicaragua South Africa U.S. 1 1 0 2 0 0 3 0 50 2 10 1 15 0 0 6 1 5 3 3 4 8 0 0 3 2 0 4 3 0 3 2 0 1 0 1 5 1 1 2 0 0 1 0 0 Table 3-14. Rank of country by m eeting quality specifications Rank Argentina Brazil China India Mexico Nicaragua South Africa U.S. 1 1 0 1 0 0 1 0 52 2 9 2 15 0 0 9 1 1 3 3 2 8 1 0 3 2 1 4 3 1 5 0 0 0 0 1 5 2 1 1 1 0 0 0 0 30

PAGE 31

Table 3-15. Rank of country by quality inspection process Rank Argentina Brazil China India Mexico Nicaragua South Africa U.S. 1 1 0 1 0 0 2 0 52 2 11 1 15 0 0 7 1 2 3 2 3 9 1 0 2 2 1 4 3 1 3 0 0 2 0 0 5 2 1 1 1 0 0 0 0 Table 3-16. Rank of country by timely delivery Rank Argentina Brazil China India Mexico Nicaragua South Africa U.S. 1 1 0 1 0 0 0 0 53 2 9 3 13 0 0 8 2 1 3 5 2 10 0 0 1 1 0 4 3 0 3 1 0 3 0 0 5 0 1 3 1 0 0 0 0 Table 3-17. Rank of country by contribution to profit margin Rank Argentina Brazil China India Mexico Nicaragua South Africa U.S. 1 3 2 14 0 0 2 2 28 2 6 1 8 1 0 3 0 12 3 4 2 4 0 0 1 1 6 4 3 0 0 1 0 3 0 1 5 0 0 1 0 0 1 0 3 Table 3-18. Average ranking Important Factors Avg. Aflatoxin 4.95 Proximity and NAFTA trade rules (respondents in Canada and Mexico only) 4.89 Flavor/better flavor 4.75 Foreign material 4.75 Pesticide residue (food safety) 4.75 Price 4.73 Country import standards/specifications 4.68 Consistent moisture levels 4.53 Overall roasting consistency 4.53 Food safety certifications 4.5 Just in Time shipping 4.5 Shelf life (better olei c/linoleic ratio) 4.45 Consistent oil levels 4.35 Good farm practices (i .e. insect/fungal/weed control, irrigation, etc.) 4.3 31

PAGE 32

Table 3-18 Continued. Important Factors Avg. Origin certification 4.28 Lower % consumer rejects 4.23 Size distribution 4.18 Traceability 4.18 Quality seed varieties 4.05 Cold storage System 3.85 Sugar content 3.83 Nutrients/nutrient content 3.78 Continual peanut variety development 3.75 Standardized Tote bag 3.73 Better farm practices 3.65 Higher vitamin E 3.58 Participation in seed program decisions 3.4 Table 3-19. U.S. support Frequency Percent much better 28 0.49 somewhat better 17 0.30 about the same 8 0.14 somewhat worse 1 0.02 much worse 0 0.00 don't know 3 0.05 Table 3-20. Support from U.S. suppliers Frequency Percent Much better 23 0.41 Somewhat better 17 0.30 about the same 10 0.18 somewhat worse 0 0.00 much worse 0 0.00 don’t know 6 0.11 Table 3-21. Total quality inspection savings Frequency Percent Yes 44 0.81 No 10 0.19 32

PAGE 33

Table 3-22. Significance of TQI savings Frequency Percent Extremely Significant 6 0.14 Very Significant 21 0.48 Somewhat Significant 14 0.32 Not very significant 3 0.07 not at all significant 0 0.00 Table 3-23. Overall taste comparison Frequency Percent Much Better 17 0.30 Somewhat Better 25 0.45 About the same 12 0.21 Somewhat worse 0 0.00 Much worse 1 0.02 don't know 1 0.02 Table 3-24. Importance of overall taste Frequency Percent Extremely Valuable 2 0.04 6 0.11 22 0.40 15 0.27 Not at all valuable 10 0.18 Table 3-25. Country of origin costs COO Cost Frequency Percent Yes 27 0.51 No 26 0.49 Total 53 1.00 Table 3-26. Problems for processi ng peanuts from which countries Frequency Percent Argentina 5 0.17 Brazil 1 0.03 China 13 0.45 India 1 0.03 Mexico 0 0.00 Nicaragua 1 0.03 South Africa 1 0.03 United States 7 0.24 Note: These are the frequencies of total answers. Some responde nts gave more than one answer, while others did not answer. Percenta ges are the percent of answers given. 33

PAGE 34

Table 3-27. Aware of American Peanut Council Aware of APC Frequency Percent yes 49 0.86 no 8 0.14 Table 3-28. Effective ways to get information Frequency Percent Newsletters 32 0.16 Website/Internet 48 0.24 Seminars 45 0.23 Visit 46 0.23 Public Relations 29 0.15 Total 200 1.00 Note: Respondents were asked to “check all that apply.” Table 3-29. Most effective ways to get information Frequency Percent Newsletters 3 0.05 Website/Internet 27 0.47 Seminars 6 0.11 Visit 16 0.28 Public Relations 5 0.09 Total 57 1.00 Note: Respondents were asked to “check only one.” 34

PAGE 35

CHAPTER 4 THEORY AND EMPRICAL MODEL Choosing the Model Looking at the distribution of the data for th e dependent variable, what percentage of peanuts were purchased from the U. S., it is apparent that there ar e a wide range of answers from 0% to 100%, with a relatively large number of responses at 100% (Fi gure 4-1). Due to the distribution of the dependant variable, a simple linear regression model cannot be used to model the data. These problems will be laid out in this chapter, along with a description of the censored regression, or tobit model, the model chosen to represent this data set. Tobit (U.S.)0 20 40 60 80 100 120 1591317212529333741454953576165 Observation Figure 4-1. Share of U.S. peanuts Theory: The Tobit Model Background of the Tobit Model In the linear regression model, all values are known for the dependent and independent variables. However, in some cases not all inform ation is known. In thes e situations the sample is limited either by censoring or truncation. “Censoring occurs when we observe the independent variable for the en tire sample, but for some obser vations we have only limited 35

PAGE 36

information about the dependent variabl e” (Long, 1997 p. 187). In this case, we know something about the missing dependent variables, su ch as, they are all less than ten, but we do not know exactly what they are. In the case of this research, we have many data points at 100%. Therefore we are able to censor at 100. “Trun cation limits the data more severely by excluding observations based on characteri stics of the dependent variab le” (Long, 1997 p.187). In this case, all cases where the dependent variable is less than 10 would be deleted, or in our model, all observations of 100% would be deleted. It is ea sy to see that deleting all observations at 100%, would change the sample completely. Truncation changes the sample, but censoring does not (Long, 1997). James Tobin, who is the namesake of the tob it model, developed the classic example of censoring. This model is Tobin’s (1958) study of household expenditures. In his study, a consumer maximizes utility by purchasing durable goods, with the constraint that expenditures must not exceed income. Therefore, expenditure s must at least equal the cost of the least expensive good. In the model, the least expensive good is $100. If the household only has $50 after other expenses, they are unable to purchase additional goods. This causes the outcome to be censored because there is no way to know how much the household would have spent if a durable good could have been purchased for $50 or less (Long, 1997). In our model, we are censoring from above instead of from below. There is no way to know how much the respondent’s company would have pu rchased from the U.S. if they were able to purchase more than 100% of their peanuts from the U.S. There are many other examples of censoring such as: hours worked by wives (Quester & Greene, 1982), extramarital affairs (Fair, 1978), and foreign trade and investment (Eaton & Tamura, 1994). 36

PAGE 37

The Problems with Truncating and Censoring To understand the problems of censoring and truncation, consider the following. Let y * be an uncensored dependent variable. Figure 4-2 shows the distribution of y * . Figure 4-2. Latent, censored a nd truncated variables. A) Latent. B) Censored. C) Truncated Reproduced with permission from Long, J.S. (1997). Regression Models for Categorical and Limited Dependent Variables. Thousand Oaks, CA: Sage Publications. In Panel A of figure 4-2, the height of the cu rve gives the relative frequency of a given value of y * . Assume that we do not know the value of y * 1, the shaded region. This means that y * cannot be observed over the entire range. y * is called a latent variable . To define the censored variable we have yi yi* ifyi* 1 0 ify * 1 [4.1] In our model, the shaded region would include anything over 100%. Let’ s call our independent variable P * to represent the percentage purchased from the U.S. Figure 4-2B shows the censored variable, y. Censored cases, the shaded region from Panel A, are st acked at 0. In our model, if P P * is greater than or equal to 100, it will be censored at 100. Figure 4-2C shows the truncated variable, y | y > 1, where the cen sored cases are deleted (Long, 1997) . If we used truncation, all P* P greater than or equal to 100 would be deleted. 37

PAGE 38

Figure 4-3 shows the effects of censoring a nd truncation. Panel A shows the regression without censoring (assuming that all values of th e dependent variable were known). Using an OLS regression of y on x, and including the censored data as 0’s, results in underestimates of the intercept and overestimat es of the slope (long dashed line in Figure 4-3B). This causes inconsistent estimates. For our model, since we will be censoring above at P * 100, the affect of censoring will be different in terms of specifics. However, the outcome is the same in terms of producing inconsistent and biased estimates. Seeing that censoring causes problems, it seems that the next logical choice would be to use the OLS to estimate the regression after truncating the sample. However, truncating will overestimate the intercept and underestimate the slope. This approach causes a correlation between x and , which also produces inconsistent estimates (Long, 1997). In our model there will again be differences in the specifics. However, after making adjustments for censoring from above, the resu lt will be the same. There will be correlation between x and , and inconsistent estimates will be produced. Since both censoring and truncation cause inc onsistent estimates, we must find another approach. The third approach is to estimate the tobit model, also referred to as the censored regression model. The tobit mode l uses all of the information, including information about the censoring. It provides consistent estimates of the parameters. In Panel B of Figure 4-3, maximum likelihood estimates are shown by the solid line. This line is i ndistinguishable from the estimates in Panel A of Figure 4-3 where there was no censoring (Long, 1997). As with censoring and truncation, the resu lts of censoring from above with the Tobit model will be the same as those from censoring from below in terms of consistency. 38

PAGE 39

Figure 4-3. Effects of censoring and trunca tion. A)Regression w ithout Censoring B) Regression with Censoring and Truncation Reproduced with permission from Long, J.S. (1997). Regression Models for Categorical and Limited Dependent Variables. Thousand Oaks, CA: Sage Publications. Since both censoring and truncation cause inc onsistent estimates, we must use another approach. The third approach is to estimate the tobit model, also referred to as the censored 39

PAGE 40

regression model. The tobit mode l uses all of the information, including information about the censoring. It provides consistent estimates of the parameters. In Panel B of Figure 4-3, maximum likelihood estimates are shown by the solid line. This line is i ndistinguishable from the estimates in Figure 4-3A where all dependent variables were known (Long, 1997). Truncated and Censored Distributions In order to consider the tobit model, it is ne cessary to have results about the truncated and normal distributions. These distributions are at the f oundation of most models of truncating and censoring. Results in this secti on will be used to show censoring from below in the tobit model (in terms of censoring and trunca tion, they are results for censo ring and truncation on the left. Formulas are also available for censoring and truncation on the right, and both on the left and right (Long, 1997)). The Normal Distribution To indicate that y * is distributed normally with mean and variance 2 , we write y * ~ N( , 2 ). y * has the pdf: f ( y *|,) 12 exp 1 2 y * 2 [4.2] The pdf is plotted in Panel A of figure 4-4. The cdf is Fy *|, fz |,dz Pr Y * y * y * [4.3] so that Pr Y * y * 1 Fy *| , [4.3] F( | , ) is the shaded region in Panel A and 1F( | , ) is the region to the right of so that the entire area under the curve equals 1. 40

PAGE 41

Figure 4-4. Normal distributions with truncati on and censoring. A) Norm al. B) Truncated. C) Censored. Reproduced with permission from Long, J.S. (1997). Regression Models for Categorical and Limited Dependent Variables. Thousand Oaks, CA: Sage Publications. The standard normal distribution, when = 0 and = 1, is written in the simplified notation: y * fy *| 0, 1 y * Fy *| 0, 1 [4.4] All normal distributions can be writ ten as a function of the standard normal. In this case, the pdf of can be written as fy *|, 12 exp 1 2 y * 2 1 y * [4.5] and the cdf of y * can be written as Pr Y * y * y * [4.6] So that, Pr Y * y * 1 y * [4.7] 41

PAGE 42

The Truncated Normal Distribution When values below are deleted, the variable y | y > has a truncated normal distribution. In this case of our model, trun cation will delete values above and the variable is written as P | P < 100. In terms of Panel A (Figure 4-4), the trun cated distribution considers the distribution of y * in the unshaded region but ignores the cases in the shaded region. Since deleting values below will cause the area of the distribution to no longer be 1, to get the truncated pdf, we must divide the pdf of the original distribution by the region to the right of , forcing the resulting distribution to have an area of 1. The truncated pdf is written as fy | y ,, fy *| , Pr Y * [4.8] The solid line (the curve with the higher peak) in Panel B of figure 4-4 shows the truncated distribution. The shaded region to the right of has been distributed over the area of the left of , making the curve slightly higher in this region. The truncated distribution can be written using equations 4.5 and 4.6 as the following: fy | y ,, 1 y * 1 1 y * [4.9] We are given that the left hand side has been truncated so E(y | y > ) must be larger than E(y * ) = . If y * is normal, the following must be true: Ey | y [4.10] 42

PAGE 43

where ( ) = ( )/ ( ) is the inverse mills ratio, where ( ) is the standard normal pdf and ( ) is the standard normal cdf (Long, 1997). The Censored Normal Distribution When a distribution is censored on the left, observations at or below are set to y : y y * ify * yify * [4.11] Generally = y , but sometimes other values can be usef ul (such as 0). Fi gure 4-4C shows the censored normal variable. The censored observations are indicated by the spike at y = . (In the case of our model, the distribution is censored on the right and th e spike would be at = 100). From equation 4.6, we know that if y * is normal, then the probab ility of an observation being censored is Pr Censored Pr y * [4.12] and the probability of a case not being censored is 1 minus the probability of it being censored, which is written as Pr Uncensored 1 [4.13] Therefore, the expected value of a censored variable equals Ey Pr Uncensored Ey | y Pr Censored Ey | y y y [4.14] 43

PAGE 44

where the last equality uses equation 4.10. The expected value of the censored variable depends on the value of . Notice that as approaches , the probability of being censored approaches 1 and E(y) approaches the censoring value y . As approaches , the probability of being censored approaches 0 and E(y) approaches the uncensored mean (Long, 1997). The Tobit Model for Censored Outcomes The structural equation for the tobit model is yi* x i i [4.15] where ~ N(0, 2 ). The x’s are observed for all cases. y * is a latent variable, observed for values greater than and is censored for valu es less than or equal to . In our case, all x’s are observed, and P * is a latent variable, observed for values less than = 100 and censored for values greater than or equal to 100. In the example for y * as the latent variable, the observed y is defined by the measurement equation: yi yi* ifyi* yify * [4.16] Combining equations 4.15 and 4.16 we get yi yi* xi iifyi* yifyi* [4.17] As has been mentioned but not explained, the to bit model is not only useful in the situation where there is censoring from below. It can also be used in situations where there is censoring from above. In our example all observations grea ter than or equal to 100 are combined into a category for observations that are over 100, the samp le is censored from above. The tobit model, using y as the va riable, is now 44

PAGE 45

yi yi* xi iifyi* yifyi* [4.18] Censoring from Above Censoring from above is the easie st extension of the previous model. Equation 4.18 shows the model for censoring from above. This mode l can be found by simply changing the signs of y. Changing the signs we can censor –y at , which is identical to censoring y at . This simple change produces small differences that will not be explained here. Usi ng the same process laid out previously for censoring for below, we find the tobit model for censoring from above to be what is given in equation 4.18. Empirical: Building the Model The dependent variable in our model is the percent of peanuts purchased from the U.S., which will be called US SHARE in the model (Table 4-1). Independent variables in the model were selected based on prior expectations. Becau se of the limited degrees of freedom, not all data collected can be tested as independent variables. The chosen independent variables are country of the firm, size of the firm, what th e buyer considers importa nt and the type of production of the firm. Dummy va riables are created for each of these independent variable categories by assigning a series of zeros and ones. Again, due to the limited number of observations, as dummy variable s were created, like data wa s grouped into one dummy. For example, we can examine the dummy variable creat ed to represent the count ry of location of the manufacturing firm. If a du mmy were created for each country, we would have 9 dummy categories. However, there were very few obs ervations in many of these categories, most notably from countries in the European Union, not including the United Kingdom. In this case, we created a dummy variable to represent the countries of Canada, Mexico, the United Kingdom, 45

PAGE 46

and the remaining countries, which included mo stly European Union countries. This was done with each of the independent variables named previously (Table 4-1). Explanation of the Independent Variables There were responses from 9 di fferent countries, as shown in table 3-1. However, the majority of the responses came from Canada, Mexico and the United Kingdom (UK). Other surveys came from France, Scandinavia, Germany, Spain, Holland and Italy. These countries were grouped together and called other. Dummy va riables were created fo r the four “countries,” Canada (CAN), Mexico (MEX), the UK and Others (including Fr ance, Scandinavia, Germany, Spain, Holland and Italy). The next variable incl uded was the size of the firm. The size of the firm was broken into four categories on the survey, less than 1,000 tons (TONS1), 1,000-5,000 tons (TONS2), 5,000-10,000 tons (TONS3) and more than 10,000 tons (TONS4). Dummy variables were also created for the firm size categories. The independent variable, what buyers consider important comes from a qualitative survey question where respondents were asked what th e MOST important fact or was when choosing from where to buy peanuts. Information from this question was broken into five categories: price (PRI), safety (SAF), properties (PROP), c onsistency (CONS) and other (OTH). It can be noted that quality is not on the li st. The reason for this is that there was an earlier question where buyers were asked what quality meant to th em. Since respondents had many different perceptions of quality, some saying price, others safety, etc., it would not have been correctly explained to just say that quality was most im portant. Instead answers for what quality meant were inserted into the question of what was most important anytime the re spondent said quality. Safety included responses such as decreased fo reign matter or low aflato xin levels. Properties included responses such as taste, flavor and moisture levels. Consistency included responses about needing peanuts to be similar to make pr oduction easier. Other incl uded responses that did 46

PAGE 47

not fit into the previous categories, such as cu stomer demand, reliable shipping and longer shelf life. Dummy variables were also created for these variables. On the survey, type of production was broken in to four categories: snack kernel, peanut butter, in-shell and confectionary . Because the question actually asked respondents to give the percentage purchased for each type, it was not enough to create a dummy variable for each, for example to say that the respondent s’ firm either purchased peanut s for peanut butter or did not. It could be questioned whether there was a significant difference between someone who purchased 50% for peanut butter and someone who purchased 100% for peanut butter. Histograms were used to make decisions on how to break up these variables. Snack was broken into three categories, those who purchased zero peanuts for snack (SNACKNO), those who purchased somewhere between 0 and 100% for snack (SNACKSOME) and those who purchased peanuts only for snack (SNACK100). Peanut butte r was divided in the same way for those who purchased none (PBNO), some (PBSOME) or all for peanut butter (PB100). In-shell was divided into to only two categorie s, those who purchased for in-s hell (IN) and those who did not (INNO). Confectionary was also broken into only two categories for those who purchased for confectionary use (CONF) and those who did not (CONFNO). Building the Model Using dummy variables can cause what is known as the dummy variable trap. If this is not accounted for, dummy variables are likely to ca use collinearity, meani ng linear relationships among the variables (Gujarati, 2003). We acc ount for this by putting in one fewer dummy variable than there are in its category. For example, we have five categories for what buyers consider important. In this cas e, we must only put four dummy variables in the model for what buyers consider important. There are a few diffe rent ways to do this. We will explore two options to account for the dummy variable trap : dropping one variable and the averaging method. 47

PAGE 48

The method chosen will dictate how the results are interpreted. If we drop one variable, significance in the model is against the dropped variable. For ex ample, for the question on firm size, if we drop the last variab le, firms using more than 10,000 tons of peanuts, significance of all other variables will be against those firms. Th erefore, if we find that the coefficient for the smallest firm size is significant, that means that it is significantly different from largest firms. It is not always convenient to compare all variables to a dropped variable. The averaging method tends to be considered mo re convenient. In this case, we take the average of the firm. The average now becomes th e benchmark so if we say that small firm size is significant, it is significantly different from the average. Because of the convenience in interpreting the re sults, this method was chosen. The method for taking the average of the firm is to subtract the last variable in the group from each of the previous variables. This will ca use the last variable to drop off so that we now have one fewer variable. For example, for country of the firm we have: AVGCAN = CAN – OTHER AVGMEX = MEX – OTHER AVGUK = UK – OTHER DROPPED = OTHER – OTHER = 0 We see that the last dummy category always equa ls zero, so it drops out of the model and we are left with only 3 variables for country. The same method is used for averaging each of the other dummy variable categories. See table 51 for variable names and descriptions. Once the averages have been taken, it is time to estimate the tobit model using statistical software. We run the tobit model as: US SHARE = f [CONSTANT, AVGCAN, AVGMEX, A VGUK, AVGTONS1, AVGTONS2, AVGTONS3, AVGPRI, AVGSAF, AVG PROP, AVGCONS, AVGSNACKNO, AVGSNACKSOME, AVGPBNO, AVG PBSOME, AVGIN, AVGCONF] Results for the model will be given in Chapter 5. 48

PAGE 49

Table 4-1. Variable Names and Descriptions Variable Name Variable Definition CAN 1 if respondent is locate d in Canada; 0 otherwise CANAVG defn = CAN-OTHER MEX 1 if respondent is locate d in Mexico; 0 otherwise MEXAVG defn = MEX-OTHER UK 1 if respondent is locate d in the UK; 0 otherwise UKAVG defn = UK-OTHER OTHER 1 if respondent is located in Fr ance, Germany, Holland, Italy, Scandinavia, or Spain; 0 otherwise DROPPEDCOUN defn = OTHER-OTHER = 0 TONS1 1 if respondent's firm purchases less than 1,000 tons; 0 otherwise AVGTONS1 defn = TONS1-TONS4 TONS2 1 if respondent's firm purchases between 1,000 and 5,000 tons; 0 otherwise AVGTONS2 defn = TONS2-TONS4 TONS3 1 if repondent's firm purchases between 5,000 and 10,000 tons; 0 otherwise AVGTONS3 defn = TONS3-TONS4 TONS4 1 if respondent's firm purchases over 10,000 tons ; 0 otherwise DROPPEDTONS defn = TONS4-TONS4 = 0 PRI 1 if respondent consider s price the most importa nt factor; 0 otherwise AVGPRI defn = PRI-OTH SAF 1 if respondent considers safety the most important factor; 0 otherwise AVGSAF defn = SAF-OTH PROP 1 if respondent considers peanut pr operties most important factor; 0 otherwise AVGPROP defn = PROP-OTH CONS 1 if respondent considers consistency most import ant factor; 0 otherwise AVGCONS defn = CONS-OTH OTH 1 if respodnet considers any other factor most important; 0 otherwise DROPPEDFACT defn = OTH-OTH = 0 SNACKNO 1 if repondent's firm purchases no peanuts for snack; 0 otherwise AVGSNACKNO defn = SNACKNO-SNACK100 SNACKSOME 1 if respondent's firm purchases so me peanuts for snack; 0 otherwise AVGSNACKSOME defn = SNACKSOME-SNACK100 SNACK100 1 if respondent's firm purchases 100% for snack; 0 otherwise DROPPEDSNACK defn = SNACK100-SNACK100 = 0 PBNO 1 if respondent's firm purchases no p eanuts for peanut butter; 0 otherwise AVGPBNO defn = PBNO-PB100 PBSOME 1 if respondent's firm purchases some peanuts for peanut butter; 0 otherwise AVGPBSOME defn = PBSOME-PB100 Table 4-1 continued. 49

PAGE 50

Variable Variable Description PB100 1 if respondent's firm purchases 100 % for peanut butter; 0 otherwise DROPPEDPB defn = PB100-PB100 = 0 IN 1 if respondent's firm purchases peanuts for in-she ll; 0 otherwise AVGIN defn = IN-INNO INNO 1 if respondent's firm purchases no peanuts for in-shell; 0 otherwise DROPPEDINSHEL defn = INNO-INNO = 0 CONF 1 if respondent's firm purchases peanuts for confectionary use; 0 otherwise AVGCONF defn = CONF-CONFNO CONFNO 1 if respondent's firm purchases peanuts for in-she ll; 0 otherwise DROPPEDCONF defn = CONFNO-CONFNO = 0 50

PAGE 51

CHAPTER 5 RESULTS AND CONCLUSIONS Using data collected from a survey of ma nufacturers abroad, the model specification explained in the previous ch apter and maximum likelihood pr ocedures, a tobit model was estimated. The independent variable shows the percentage of peanuts used that each respondent’s company imports from the United States. A tobit model was used because the dependent variable was restricted to 100% or less (Figure 4-1). As is explained in Chapter 4, this leads us to using the censored tobit model. Tobit Model Results Table 5-1 shows a description of the estimated coefficient, st andard error, and significance levels for each variable from the tobit model. Significance is shown in the p-value. A p-value less than .05 shows that the variable is signi ficantly different from the average at a 95% confidence level. A significant p-value, along with a positive coefficient, suggests the U.S. share will increase with the level of that explanatory variable if everything else is held constant. However, very few of the variables, only pea nut properties and confec tionary production, are shown to be significantly different from the averag e at a 95% confidence level or higher. Only a few more variables, firms purchasing 1,000 to 5,000 tons per year, consistency as the most important factor and in-shell as the type of production, are signif icant at a 90% confidence level. Though this seems to show that the majority of this research is insignificant, it is possible to show that some of the other variables are signifi cantly different from each other. This will be explained later in the chapter. Notice also that the p-value for sigma is shown to be significant. This shows that the toibt model was needed to ex plain the dependent variab le. If sigma was not significant, the least squares method could have been used. 51

PAGE 52

Peanut Purchasing Simulations Simulations were created to give predicted probabilities. The predicted probabilities, found using the simulation data, show the like lihood of purchasing deci sions relative to the average consumer, who purchases 90.38% of their pe anuts from the U.S. While it is possible to show marginal effects for any combination of th e variables, it is more convenient to use the average as is explained in Chapter 4. Figures 5-1 through 5-4 show the deviation from the mean (90.38%) for the change in proba bility of U.S. import preference. Those above the mean are more likely to import a larger percent of pea nuts from the U.S. than those below the mean Figure 5-1 shows buyer preference according to the country wh ere the respondent worked. Respondents in Mexico are likely to purchase the most from the U.S. at 6.65 percentage points above the average. Respondents in the United Ki ngdom are likely to purchase the lowest share from the U.S. at 22.42 percentage points below th e average. Canada and other countries (France, Germany Holland, Italy, Scandinavia and Spain) also show predicted probabilities below the average, 11.7 and 7.18 percentage points below the average, respectively. Change in probability of import preference for U.S. peanuts due to country -25 -20 -15 -10 -5 0 5 10 Canada Mexico Other UKCountry Figure 5-1. Change in probability of import preference for U.S. peanuts due to country Figure 5-2 shows the change in probability of U.S. import pref erence due to company size. As the graph shows, the largest operations, th ose who purchase over 10,00 0 tons of peanuts per 52

PAGE 53

year, are likely to purchase a lower share than av erage (1.7 percentage points) from the U.S. It should be noted that the scale for fi gure 5-2 is much different than that of 5-1. In this case, no size category has a preference far below the average of 90.38. Those who purchase between 1,000 and 5,000 tons are most likely to buy from the U.S. at 1.69 percentage points above the average. The other two categor ies, those who purchase below 1, 000 tons per year and those who purchase between 5,000 and 10,000 tons, have pr eference at 0.24 and 0.29 percentage points below the average, respectively. Change in probability of import preference for U.S.peanuts due to company size -2 -1.5 -1 -0.5 0 0.5 1 1.5 2 Under 1000 tons 1000 to 5000 tons 5000 to 10000 tons over 10000 tonsCompany Size Figure 5-2. Change in probability of import preference for U.S. peanuts due to company size The third group of variables tested was what f actors the buyers felt we re most important in their decision from where to purchase peanuts (Fi gure 5-3). There are more positives in this case where those who value peanut pr operties, such as taste, flavor and moisture levels, (1.94 percentage points above average) are the most lik ely to purchase from the U.S. Those who value safety, factors such as decreased foreign materi al and low aflatoxin le vels, (1.12 percentage) were also above average, while those who valu e price and consistency were below average at 0.87 and 4.96 percentage points below average, re spectively. Those who value other factors, 53

PAGE 54

customer demand, reliable shipping and longer shel f life, were also more likely to purchase peanuts from the U.S. at 0.67 per centage points above the average. Change in probability of import preference for U.S. peanuts due to what is considered important-6 -5 -4 -3 -2 -1 0 1 2 3Price Safety Properties Consistency OtherCategories of Important Factors Figure 5-3. Change in probability of import preference for U.S. peanuts due to what is considered important The fourth, and last, group of variables tested was type of production (F igure 5-4). In this case those who produced 100% snack peanuts and 100% peanut butter were the most likely to purchase peanuts from the U.S. at 2.93 and 4.04 percentage points above the average. This likely coincides with the previous test where those who believed peanut properties, such as flavor and taste, were the most important factors (assuming that those who produce snack peanuts and peanut butter would believe that ta ste, flavor and moisture levels were more important than those who produ ce confectionary products). Those likely to purchase a lower share of peanuts from the U.S. were those w ho produce some snack peanuts or some peanut butter and those who produce no pea nut butter at 4.31, 4.91 and 4.11 percentage points below the average, respectively. Those who use peanuts fo r in-shell and confectionary production are also likely to purchase a lower share of U.S. peanut s than the average at 3.45 and 2.91 percentage points below the average. 54

PAGE 55

Change in probability of import preference for U.S. peanuts due to production type -6 -5 -4 -3 -2 -1 0 1 2 3 4 50% Snack 1-99% Snack 100% Snack 0% PB 1-99% PB 100% PB Inshell ConfectionaryProduction Type Figure 5-4. Change in probability of import preference for U.S. peanuts due to type of production Testing Variable Significance Tables 5-2 through 5-4 show the significance of the variables in each group relative to the other variables in that group. These matrices give the t-stat istics for the significance between the variables. The t-statistics are calculated by t estimator parameter erro r 22 xi 2 where the error stands for th e estimated standard error, 2 is the estimator, 2 is the parameter and is the estimated standard deviation. Using a t-table, we find that if the absolute value of the t-statistic is greater than 2.0345, the two vari ables are significantly different from each other at a 95% confidence level. 2.0345 is called the critical t-value. Si nce few variables are significant at a 95% confidence level, significan ce will also be given at 90% and 85%. The critical t-value for a 90% confidence level at 33 degrees of freedom is 1.692 and for an 85% confidence level is 1.474. Table 5-2 shows that Canada and Mexico are not significantly different from any of the other countries tested. However, the U.K. and other (France, Germany, Holland, Italy, Scandinavia and Spain) are signif icantly different from each other at a 95% confidence level. 55

PAGE 56

This may be explained by the fact that these are all European countries. Table 5-3 shows that the different firm sizes are not significantly different from each othe r at a 95% confidence level. At a 90% confidence level, firms that purchase 1,000 to 5,000 tons of peanuts per year are significantly different than firms that purchas e more than 10,000 tons of peanuts per year. The next group of variables is what factors buyers consider important . As explained in Chapter 4, these variables come from a qualitati ve question where respondents were asked to give the most important factor when choosi ng from where to buy peanuts. The covariance matrix for what is considered important (Table 5-4) shows that at a 95% confidence level peanut properties, such as taste flavor and moisture levels are significan tly different from both price and consistency. Consistency is also significantly different from safe ty, comprised of factors such as dcreased foreign material and lo w aflatoxin levels. Other respons es (responses that did not fit into the other categories such as customer demand, reliable shi pping and longer shelf life) are not found to be significantly different from any of the other variables. Magnitude of Impact Figure 5-5 shows the magnitude of impact of each variable on share of U.S. peanuts or the range of the share purchased from the U.S. depe nding on each variable group. For example, the share purchased from the U.S. depending on country ranges over 29.08 per centage points, while changes in firm size only cause the share to range over 3.4 percentage poin ts. This shows that country has the highest magnitude of impact, while the size of the firm has the lowest magnitude of impact. Change in type of production and what is considered important causes share to range over 8.96 and 6.9 percentage points, respectively. 56

PAGE 57

Magnitude of impact of each variable 05101520253035 Size Importance Type CountryVariableMagnitude of Impact Figure 5-5. Magnitude of impact of variables Conclusions Now that the results have been laid out, the question is, what does this mean to the peanut industry? The objectives, given in Chapter 1, we re to learn the percep tions of buyers and to determine which factors actually influence buying behavior. The perceptions of buyers are given in detail in Chapter 3 and the re sults of the test to determin e which factors influence buying behavior are given earlier in this chapter. The data description in Chapter 3 shows that the U.S. is considered favorably by buyers in many ways. In general, buyers pe rceive the quality and value of U.S. peanuts to be high, while price is considered somewhat lower. China tended to be ranked high in te rms of price but low in terms of quality. The United States also ranked hi gh in terms of best tasting peanuts, meeting quality specifications, quality inspection pro cess and timely delivery but somewhat lower in terms of profit margins. Buyers also feel that U.S. support is be tter than that in other countries. The results given earlier in this chapter can be used by the U.S. peanut industry to see where to focus. In the model we learned that the United Kingdom was to purchase a lower share of peanuts from the U.S. It can also be noted that the U.K. was found to be significantly different from other, which includ ed mostly European countries. This may give the U.S. peanut 57

PAGE 58

industry insight as to how much focus should be placed on the U.K. In order to market peanuts to the U.K., it would be useful to have more information on what buyers in the U.K. consider important. These results also show that the largest firms are likely to purchase a lower share of peanuts from the United States. This may be due to the fact that many larger operations are influenced much more by price than quality. The size of firms likely to purchase the largest share from the U.S. are those who purchase betw een 1,000 and 5,000 tons of peanuts per year. This may tell us that those firms are influenced more by things that the U.S. does well. The question of import preference due to what the buyer considers im portant shows that firms who value consistency are le ss likely to purchase peanuts from the U.S. than firms that value other factors. This is in line with survey results showing that the U.S. is not perceived to have the most consistent peanuts. Results al so show that firms who produce 100% snack peanuts or 100% peanut butter are likely to purchase a la rger share from the U.S. than other types of producers. This could be based on the fact that firms who focus more on peanut properties such as taste and flavor are most likel y to purchase from the U.S. It is likely that firms who produce peanut butter or snack peanuts need peanuts that have a better ta ste or flavor than some other types of producers. This research can be very important to the U.S. peanut industry in deciding how U.S. peanuts should be marketed. The industry must choose a marketing plan that accounts for its strengths and weaknesses, many of which have been given in this paper. 58

PAGE 59

Table 5-1. Tobit model results PARAMETER ESTIMATE STANDARD ERROR P-VALUE CONSTANT -55.84 40644.70 0.999 CANADA 56.18 40644.70 0.999 MEXICO -176.70 121934.00 0.999 OTHER 41.67 40644.70 0.999 UNDER 1,000 TONS 2.20 12.53 0.860 1,000 5,000 TONS -18.51 10.60 0.081 5,000 10,000 TONS 2.62 9.68 0.787 PRICE 7.49 8.87 0.399 SAFETY -11.45 12.44 0.357 PROPERTIES -21.78 10.80 0.044 CONSISTENCY 32.30 18.88 0.087 NO SNACK 7.94 198984.00 1.000 SOME SNACK 29.13 99492.10 1.000 NO PEANUT BUTTER 28.14 99492.10 1.000 SOME PEANUT BUTTER 32.08 99492.10 1.000 IN-SHELL 24.63 13.10 0.060 CONFECITIONARY 21.49 9.86 0.029 SIGMA 26.54 3.61 0.000 Table 5-2. T-values for country T-Values for Country Canada Mexico Other UK Canada 0.0000 0.0014 0.7588 -0.8640 Mexico 0.0014 0.0000 -0.0013 -0.0016 Other 0.7588 -0.0013 0.0000 -2.1366 UK -0.8640 -0.0016 -2.1366 0.0000 59

PAGE 60

Table 5-3. T-values for size of the firm T-values for Size of the Firm (tons of peanuts) Under 1,000 1,000 to 5,000 5,000 to 10,000 More than 10,000 Under 1,000 0.0000 1.0246 -0.0213 -0.6781 1,000 to 5,000 1.0246 0.0000 -1.4217 -1.9801 5,000 to 10,000 -0.0213 -1.4217 0.0000 -0.7437 More than 10,000 -0.6781 -1.9801 -0.7437 0.0000 Table 5-4. T-values for what is considered important T-values for What is Considered Important Price Safety Properties Consistency Other Price 0.0000 1.1906 2.3762 -1.0276 0.8095 Safety 1.1906 0.0000 0.5933 -1.6623 -0.2389 Properties 2.3762 0.5933 0.0000 -2.1143 -0.7916 Consistency -1.0276 -1.6623 -2.1143 0.0000 1.2750 Other 0.8095 -0.2389 -0.7916 1.2750 0.0000 60

PAGE 61

APPENDIX A PEANUT QUESTIONNAIRE Rose Research I.D.#______________ Boca Raton, Florida Study #2535 September, 2005 APC Trade Study Universal Screener Hello, I’m ______________ from Rose Research, a global market research company, and we’re conducting a survey about peanuts among quality control professionals, operation managers and purchasing agents. Your name was given to us by the American Peanut Council (who represent the U.S. peanut industry) and you were chosen b ecause of your expertise in the category. The information that you provide will be invaluable to the American Peanut Council and help them better serve their customers. Please be assured that your answers will be kept strictly confidential. The answers that you provide will be combined with others who are participating in the exact same survey. Th ere are no right or wrong answers. All we want is your honest opinions about the topics we will be discussing. The survey will take about 30 minutes and we’d like to know if you would be will ing to participate. Is now a good time to talk? If not, then please let me know when you could be available for about 30 minutes and I’ll call you back. It could be anytime during the day – befo re, during or after working hours. (IF RESPONDENTS CANNOT TALK NOW, RESCHEDULE INTERVIEW) ________________ ________________ (write in day) (write in time) INTERVIEWER ASK: What is your job function/title? Quality Control ( )-1 Operations Manager ( )-2 Purchasing Agent ( )-3 Managing Director ( )-4 Other ____________________________________ ( )-5 (write in) INTERVIEWER: CHECK OFF COUNTRY IN WHICH RESPONDENT WORKS: Canada ( )-1 Mexico ( )-2 France ( )-3 Scandinavia ( )-4 Germany ( )-5 Spain ( )-6 Holland ( )-7 UK ( )-8 Italy ( )-9 61

PAGE 62

NAME__________________________________________________ ADDRESS_______________________________________________ COUNTRY_______________________________________________ TITLE___________________________________________________ COMPANY_______________________________________________ TELEPHONE _____________________FAX____________________ E MAIL ADDRESS________________________________________ GO TO QUESTIONNAIRE 62

PAGE 63

Rose Research I.D.#_____________ Boca Raton, Florida Study #2535 September, 2005 APC Trade Study Questionnaire Thank you for agreeing to participate in our study. Let’s begin. The first couple of questions are for classification purposes only. 1a. Approximately how many tons of peanuts doe s your company purchase in a typical year? Please remember, we only want to know about the amount of peanuts your company purchases. (READ CHOI CES, CHECK ONE ONLY) Under 1,000 tons ( )-1 1,000 – less than 5,000 tons ( )-2 5,000 – less than 10,000 tons ( )-3 10,000 – less than 25,000 tons ( )-4 25,000 tons or more ( )-5 1b. If your total peanut/peanut product purchasing equals 100%, what percent do each of the following account for? (READ CHOI CES. ANSWERS MUST EQUAL 100%) Snack kernel peanuts ________ % In-shell peanuts __ ______ % Peanut butter __ ______% Confectionery products (w/peanuts) __ ______ % Other _____________________ ________ % (specify) Other _____________________ ________ % (specify) _______________ TOTAL 2a. Thinking about peanuts in general, how would you define qua lity? (PROBE:) Specifically, what words or attributes would yo u use to describe high quality peanuts? (PROBE:) Is there anything else you can think of? (PROBE FOR SPECIFICS) 63

PAGE 64

2b. INTERVIEWER: HAVE THE RESPONDENT RANK THEI R ANSWERS IN Q.2a FROM “MOST IMPORTANT” TO “LEAST IMPORTANT” AND WRITE IN ORDER BELOW: MOST IMPORTANT 1. _________________________________ 2. _________________________________ 3. _________________________________ 4. _________________________________ 5. _________________________________ 6. ________________________________ 7. _________________________________ 8. _________ ________________________ 9. _________ ________________________ LEAST IMPORTANT 10. _________ ________________________ 64

PAGE 65

3a. And to the best of your ability, please break down the percentage of peanuts that your company imported/purchased/used from the following countries in the past year? (INTERVIEWER: READ LIST OF COUNTRIES. MAKE SURE %’S ADD TO 100%) (INTERVIEWER: IF RESPO NDENT IS UNABLE, OR UNWILLING, TO ESTIMATE BY PERCENTAGES IN Q.3a, ASK THEM TO RANK THE FOLLOWING COUNTRIES. THE COUNTRY THAT SUPPLIES THEIR COMPANY WITH THE MOST PEANUTS SHOULD BE RANKED A ”, THE SECOND MOST A ”, ETC. USE APPROPRIATE COLUMN TO FILL IN ANSWERS) (INTERVIEWER: ASK Q’S 3b – 3d AMONG COUNTRIES THE RESPONDENT MENTIONED IN Q.3a. IF U.S. IS NOT MENTIONED, PLEASE ADD IT TO THE LIST OF COUNTRIES IN Q’S 3b – 3d). 3b. Thinking about the countrie s you import/purchase/use peanut s from, I’d like you to rank them in terms of quality. For example, the country th at produces the highest quality peanuts should be ranked #1, the second hi ghest quality #2, etc. (READ COUNTRIES MENTIONED IN Q.3a. AND THE U.S. IF IT’S NOT MENTIONED. REPEAT IF NECESSARY) 3c. And, how about price? Please rank the following countri es from lowest to highest in terms of overall peanut prices. The country th at offers the lowest price should be ranked #1, the second lowest price #2, etc. (READ COUNTRIES MENTIONED IN Q.3a. AND THE U.S. IF IT’S NOT MENTIONED. REPEAT IF NECESSARY) 3d. Can you also rank the following peanut producing countries in terms of best overall value? Again, the country that offers the best overall va lue should be ranked #1, the second best overall value #2, etc. (R EAD COUNTRIES MENTIONED IN Q.3a. AND THE U.S. IF IT’S NOT MENTIONED. REPEAT IF NECESSARY) CHART FOR Q.3a – 3d ANSWERS IS ON THE NEXT PAGE 65

PAGE 66

Q.3a % Breakdown Q.3a Ranking (OPTIONAL) Q.3b Quality Q.3c Price Q.3d Value Argentina Brazil China India Mexico Nicaragua South Africa United States (ALWAYS READ) Other______________ (specify) Other______________ (specify) Other______________ (specify) Don’t know (DO NOT READ) 66

PAGE 67

ASK Q.4a ABOUT COUNTRY THAT PRODUCES HIGHEST QUALITY PEANUTS IN Q.3b 4a. What makes the peanuts from (INSERT ANSWER FROM Q.3b) better than other countries? (PROBE:) Are there any other reasons you ca n think of? (PROBE FOR SPECIFICS) ASK Q.4b ABOUT COUNTRY THAT PRO DUCES POOREST QUALITY PEANUTS IN Q.3b 4b. And, what makes the peanuts from (IN SERT ANSWER FROM Q.3b) worse than other countries? (PROBE:) Are there any other reasons you ca n think of? (PROBE FOR SPECIFICS) Still thinking about peanuts 67

PAGE 68

4c. Is your company a private label buyer/producer? Yes ( )-1 (CONTINUE) No ( )-2 (SKIP TO Q.5a) 4d. Does your private label buyer/producer specify origins when buying peanuts? Yes ( )-1 (CONTINUE) No ( )-2 (SKIP TO Q.5a) Don’t know ( )-3 (SKIP TO Q.5a) 4e. Which countries does your private label/p roducer buyer specify peanuts from? (READ LIST, CHECK ALL THAT APPLY) (ASK Q.4f, 4g AND 4h ABOUT COUNTRIES MENTIONED IN Q.4e) 4f. Can you rank the countries specified by your private label buyer in terms of quality – with a ” meaning the best quality, a ” second best, etc.? 4g. How about price? Can you rank the countri es specified by your private label buyer in terms of price – with a ” meaning the lowest price, a ” second lowest, etc.? 4h. And, can you rank the countries specifie d by your private labe l buyer in terms of overall value – with a ” meaning the best overall value, a ” second best, etc.? Q.4e Countries Specified From Q.4f Quality Q.4g Price Q.4h Value Argentina Brazil China India Mexico Nicaragua South Africa United States (ALWAYS READ) Other______________ (specify) Other______________ (specify) Other______________ (specify) Don’t know (DO NOT READ) 68

PAGE 69

5a. When deciding from what country you will import/purchase/use peanuts, what key factors influence your decision? (PROBE:) Are there any other reasons you can think of? (PROBE FOR SPECIFICS) 5b. And, which one factor is most important to you when deciding from what country to import/purchase/use peanuts? (WRITE IN ONE COMMENT ONLY) 69

PAGE 70

ASK Q.6a AMONG THOSE WHO ME NTIONED THE U.S. IN Q 3a 6a. Why does your company import/purchase/use peanuts from the U.S.? (PROBE:) Are there any other reasons you can thin k of? (PROBE FOR SPECIFICS) ASK Q.6b AMONG THOSE WHO DID NOT MENTION THE U.S. IN Q 3a 6b. Why doesn’t your company import/purchase/us e peanuts from the U.S.? (PROBE:) Are there any other reasons you can thin k of? (PROBE FOR SPECIFICS) 70

PAGE 71

INTERVIEWER: ASK Q’S 7a – 7d ABOUT COUNTRIES MENTIONED IN Q.3a. AGAIN, IF THE U.S. IS NOT MENTIONED IN Q.3a, PLEASE INCLUDE IT IN THE LIST OF COUNTRIES IN Q’S 7a – 7d 7a. Can you please rank the following countries in terms of offering the best tasting peanuts. The country that offers the best tas ting peanuts should be ranked #1, the second best #2, etc. (READ COUNTRIES MENTIONED IN Q.3a. REPEAT IF NECESSARY) 7b. How about in terms of meeting quality specifications? The country that offers the best quality specifications should be ranked #1, the second best #2, etc. (READ COUNTRIES MENTIONED IN Q.3a. REPEAT IF NECESSARY) 7c. And, how would you rank the follo wing countries in terms of their total quality inspection process? The country that offers the be st total quality inspection process should be ranked #1, the second best #2, etc. (READ COUNTRIES MENTIONED IN Q.3a. REPEAT IF NECESSARY) (INTERVIEWER: IF RESPONDENT IS UNSURE OF THE MEANING OF “TOTAL QUALITY INSPECTION PROCESS”, PROBE: “I MEAN THE TOTAL QUALITY INSPECTION PROCESS FROM THE FARM TO EXPORT”) 7d. And, please rank the following countries in terms of the timely delivery they offer for peanut orders. The country that offers th e best timely delivery should be ranked #1, the second best #2, etc. (R EAD COUNTRIES MENTIONED IN Q.3a. REPEAT IF NECESSARY) 7e. And, can you please rank the following p eanut producing countries in terms of your profit margins? The country that offers your compa ny the best profit margins should be ranked #1, the second best #2, etc. (READ COUNTRIES MENTIONED IN Q.3a. REPEAT IF NECESSARY) INTERVIEWER: USE CHART ON THE NEXT PAGE TO FILL IN RESPONDENTS’ ANSWERS TO Q’s 7a – 7e 71

PAGE 72

Q.7a Best Tasting Peanuts Q.7b Meeting Quality Specifications Q.7c Total Quality Inspection Process Q.7d Timely Delivery Q.7e Profit Margins Argentina Brazil China India Mexico Nicaragua South Africa United States (ALWAYS READ) Other______________ (specify) Other______________ (specify) Other______________ (specify) Don’t know (DO NOT READ) ASK Q.8a ABOUT COUNTRY THAT IS BEST IN TERMS OF THEIR TOTAL QUALITY INSPECTION PROCESS IN Q.7c 8a. What makes the total qual ity inspection process from (INSERT ANSWER FROM Q.7c) better than other countries? (PROBE:) Are there any othe r reasons you can think of? (PROBE FOR SPECIFICS) ASK Q.8b ABOUT COUNTRY THAT IS WORS T IN TERMS OF THEIR TOTAL QUALITY INSPECTION PROCESS IN Q.7c 8b. And, what makes the total quality insp ection process from (INSERT ANSWER FROM Q.7c) worse than other countries? (PROBE:) Are there any other reasons you can think of? (PROBE FOR SPECIFICS) 72

PAGE 73

ASK Q.8c ABOUT COUNTRY THAT IS BEST IN TERMS OF PROVIDING TIMELY DELIVERY IN Q.7d 8c. What makes (INSERT ANSWER FROM Q.7d) better in terms of offering timely delivery than other countries? (PROBE:) Ar e there any other reasons you can think of? (PROBE FOR SPECIFICS) ASK Q.8d ABOUT COUNTRY THAT IS WO RST IN TERMS OF PROVIDING TIMELY DELIVERY IN Q.7d 8d. And, what makes (INSERT ANSWER FROM Q.7d) worse in terms of offering timely delivery than other countries? (PROBE:) Ar e there any other reasons you can think of? (PROBE FOR SPECIFICS) ASK Q.8e ABOUT COUNTRY THAT OFFERS THE BEST PROFIT MARGINS IN Q.7e 8e. What makes your profit margins from (INSERT ANSWER FROM Q.7e) better than other countries? (PROBE:) Are there any other reasons you can think of? (PROBE FOR SPECIFICS) ASK Q.8f ABOUT COUNTRY THAT OFFERS THE WORST PROFIT MARGINS IN Q.7e 73

PAGE 74

8f. And, what makes your profit margins from (INSERT ANSWER FROM Q.7e) worse than other countries? (PROBE:) Are there any other reasons you can think of? (PROBE FOR SPECIFICS) (INTERVIEWER: ONLY ASK Q.8g AMONG PURCHASING AGENTS/BUYERS) (ASK Q.8g AMONG COUNTRIES MENTIONED IN Q.3a. IF THE U.S. IS NOT MENTIONED IN Q.3a, PLEASE INCLUDE IN THE LIST OF COUNTRIES READ IN Q.8g) 8g. What is the freight cost for 1 metric ton of peanuts from each of the countries you import/purchase/use peanuts from? (READ EACH COUNTRY AND RECORD FREIGHT COST BELOW) Freight Cost Don’t know (DO NOT READ) Argentina Brazil China India Mexico Nicaragua South Africa United States (ALWAYS READ) Other______________ (specify) Other______________ (specify) Other______________ (specify) 74

PAGE 75

ASK ALL RESPONDENTS 9. Using a 5 point scale, where a 5 means “extremely important” and a 1 means “not important at all,” how impor tant are the following when purchasing/importing/ using peanuts? (READ CHOICES AND REPEAT SCALE AS NECESSARY) Extremely important Not important at all 5 4 3 2 1 Origin certification Proximity and NAFTA trad e rules (respondents in Canada and Mexico only) Just in Time shipping Traceability Country import standard s/specifications Standardized Tote bag Cold storage System Pesticide residue (food safety) Aflatoxin Flavor/better flavor Price Nutrients/nutrient content Shelf life (better olei c/linoleic ratio) Size distribution Sugar content Consistent oil levels Foreign material Consistent moisture levels Higher vitamin E Overall roasting consistency Lower % consumer rejects Quality seed varieties Participation in seed program decisions Better farm practices o Food safety certifications o Good farm practices (i.e. insect/fungal/weed cont rol, irrigation, etc.) o Continual peanut variety development Other _________________________________ (specify) Other _________________________________ (specify) 75

PAGE 76

Now, let’s talk about U.S. peanuts for a few minutes 10a. Compared to other countries/peanut producers, do you think that the support when importing/purchasing/using U.S. peanut s is? (READ CHOICES, CHECK ONE ONLY) Much better than othe r countries ( )-1 Somewhat better than other countries ( )-2 About the same as other countries ( )-3 Somewhat worse than other countries ( )-4 Much worse than other countries ( )-5 Don’t know (DO NOT READ) ( )-6 10b Why do you feel that way? (PROBE:) Ar e there any other reasons you can think of? (PROBE FOR SPECIFICS) 11a. And, how about from a supplier perspective? Is the support from U.S. suppliers when importing/purchasing/using peanuts? (READ CHOICES, CHECK ONE ONLY) Much better than suppliers from other countries ( )-1 Somewhat better than suppliers fr om other countries ( )-2 About the same as suppliers from other countries ( )-3 Somewhat worse than suppliers from other countries ( )-4 Much worse than suppliers from other countries ( )-5 Don’t know (DO NOT READ) ( )-6 11b. Why do you feel that way? (PROBE:) Are there any other reasons you can think of? (PROBE FOR SPECIFICS) 76

PAGE 77

12a. What advantages do U.S. pea nuts provide to you that other countries don’t? (PROBE:) Are there any other reasons you can think of? (PROBE FOR SPECIFICS) 12b. And, what disadvantages, if any, might be encountered when importing/purchasing/using U.S. peanuts? (PROBE:) Are there any other reasons you can think of? (PROBE FOR SPECIFICS) (ASK Q.13a AMONG THOSE WHO USE U.S. PEANUTS IN Q.3a) 13a. Are there any cost savings to the total qua lity inspection process that U.S. peanuts go through – specifically, does the inspection pr ocess that U.S. p eanuts go through save companies money in the long run? Yes ( )-1 (ASK Q.13b & 13c) No ( )-2 (SKIP TO Q.13d) 77

PAGE 78

13b. You said that the total qua lity inspection process for U.S. peanuts saves you money. How significant would you say that the cost savings due to this inspection process is? Would you say that it is? (R EAD CHOICES, CH ECK ONE ONLY) Extremely significant ( )-1 Very significant ( )-2 Somewhat significant ( )-3 Not very significant ( )-4 Not at all significant ( )-5 13c. And, can you quantify the amount of money your company has saved due to the total quality inspection process that U.S. peanut s go through – can you give me a specific dollar amount? ________________________ write in dollar amount Don’t know ( )-2 (DO NOT READ) 13d. In your opinion, why doesn’t the total quali ty inspection process that U.S. peanuts go through save companies money? (PROBE:) Are there any other reasons you can think of? (PROBE FOR SPECIFICS) 14a. How does the overall taste of U.S. peanuts compare to other countri es/areas? Would you say that U.S. peanuts taste? (READ CHOICES, CHECK ONE ONLY) Much better than other c ountries/areas ( )-1 Somewhat better than other countries/areas ( )-2 About the same as other countries/areas ( )-3 Somewhat worse than other countries/areas ( )-4 Much worse than other c ountries/areas ( )-5 Don’t know (DO NOT READ) ( )-6 14b. Why do you feel this way? (PROBE:) Are there any other reasons you can think of? (PROBE FOR SPECIFICS) 78

PAGE 79

14c. Using a 5 point scale, where a 5 means that it is “extremely valuab le and worth paying more for” and a 1 means that it is “not valuable at all and not worth paying more for”, in general, how valuable is the overall taste of U.S. peanuts? You may use any number in between 1 and 5? (REPEAT SCALE IF NECESSARY, CIRCLE ONE ANSWER ONLY) 5 4 3 2 1 Extremely valuable Not at all valuable and worth paying for and not worth paying for Now, I’d like to ask you some questions about processing peanuts. 15. What are some of the issues or problem s you face when you process peanuts? (PROBE:) Is there anything else you can think of? (PROBE FOR SPECIFICS) 16. Of the issues or problems you just men tioned, which have major cost implications when processing peanuts? (PROBE:) Is there a nything else you can think of? (PROBE FOR SPECIFICS) 17. Are any of the cost factors you just mentioned caused by country of origin? Yes ( )-1 (ASK Q.18) No ( )-2 (SKIP TO Q.19) INTERVIEWER: FOR EACH COST FACTOR MENTIONED IN Q.17, ASK: 79

PAGE 80

18. Which country do you have this problem with? (CHECK AS MANY COUNTRIES THAT APPLY) ( INSERT COST FACTOR FROM Q.17) Argentina ( )-1 Brazil ( )-2 China ( )-3 India ( )-4 Mexico ( )-5 Nicaragua ( )-6 South Africa ( )-7 United States ( )-8 Other ________________ ( )-9 (specify) Don’t know (DO NOT READ) ( )-10 None (DO NOT READ) ( )-11 ___________________________ (INSERT COST FACTOR FROM Q.17) Argentina ( )-1 Brazil ( )-2 China ( )-3 India ( )-4 Mexico ( )-5 Nicaragua ( )-6 South Africa ( )-7 United States ( )-8 Other ________________ ( )-9 (specify) Don’t know (DO NOT READ) ( )-10 None (DO NOT READ) ( )-11 ___________________________ (INSERT COST FACTOR FROM Q.17) Argentina ( )-1 Brazil ( )-2 China ( )-3 India ( )-4 Mexico ( )-5 Nicaragua ( )-6 South Africa ( )-7 United States ( )-8 Other ________________ ( )-9 (specify) Don’t know (DO NOT READ) ( )-10 None (DO NOT READ) ( )-11 19. Are you aware of the Amer ican Peanut Council (APC)? 80

PAGE 81

Yes ( )-1 No ( )-2 20a. Which of the following ways is effectiv e for the American Peanut Council to provide information to you? (READ LIST, CHECK ALL THAT APPLY) 20b. And, finally, what is the most effective way for the American Peanut Council to provide information to you? (READ CHOICES MENTIONED IN Q.20a AND CHECK ONE ONLY) Q.20a Aided Q.20b Most effective 22082 Newsletters Website/the Internet Seminars Visits by representatives Public relations activities/ information Other _________________ (specify) Other _________________ (specify) Other _________________ (specify) THAT WAS OUR LAST QUESTION. I HOPE THAT YOU FOUND OUR DISCUSSION INTERESTING. THANK YOU FOR YOUR PARTICIPATION. 81

PAGE 82

APPENDIX B TSP CODE Options memory=50; Title 'Peanut Preferences by Country-of-Origin'; ? OUT 'D:\ZSTUDENTS\JULIEDAVENPORT\PEANUT_TSP'; ? READ(FORMAT=EXCEL,FILE='D:\ZSTUDENTS\JULIEDAVENPORT\PEANUTDATA.XLS'); ? OUT; IN 'c:\tsp44\files\davenport\PEANUT_TSP'; ? IN 'k:\ZSTUDENTS\JULIEDAVENPORT\PEANUT_TSP'; ? DBLIST 'D:\ZSTUDENTS\JULIEDAVENPORT\PEANUT_TSP'; ? OUT 'D:\ZSTUDENTS\JULIEDAVENPORT\PEANUT_TSP'; ? Q5bProperties = (ID=24) + (ID^=24)*Q5bProperties; ? OUT; select id=45; peanutbut=0; select 1; list zvarz ID Country Tons Snack InShell PeanutBut Confection Othertype SHARE1 SHARE2 SHARE3 SHARE4 SHARE5 SHARE6 Q1 Q2 Q3 Q4 Q5 Q6 P1 P2 P3 P4 P5 P6 V1 V2 V3 V4 V5 V6 Q5bPrice Q5bSafety Q5bProperties Q5bconsistent Q5bOther Origin JIT Traceability Count Tote ColdStore Pesticide Aflatoxin Flavor Price Nutrients ShelfLife Size Sugar Oil ForeignMat Moisture VitE Roasting Rejectperc seedQ seedprog betterfarm foodsafe farmprac varietydev; ? OUT 'D:\ZSTUDENT\JULIADAVENPORT\PEANUT_TSP'; DOC ID 'Firm number'; DOC Country 'Country code'; DOC Tons '1 under 1000tons, 2=1000/5000, 3=5000/10000, 4=10000/25000,5=25000+ tons'; DOC Snack 'percent of product in snack'; DOC InShell 'percent of product in inshell'; DOC PeanutBut 'percent of product in peanut butter'; DOC Confection 'percent of product in confection'; DOC Othertype 'percent of product in other types'; DOC SHARE1 'Argentine share of peanuts'; DOC SHARE2 'Brazil share of peanuts'; DOC SHARE3 'China share of peanuts'; DOC SHARE4 'Nicaragua share of peanuts'; DOC SHARE5 'U.S.share of peanuts'; DOC SHARE6 'Other share of peanuts'; DOC Q1 'Argentine quality ranking'; DOC Q2 'Brazil quality ranking'; DOC Q3 'China quality ranking'; DOC Q4 'Nicaragua quality ranking'; DOC Q5 'U.S. quality ranking'; DOC Q6 'Other quality ranking'; DOC P1 'Argentine price ranking'; DOC P2 'Brazil price ranking'; DOC P3 'China price ranking'; DOC P4 'Nicaragua price ranking'; DOC P5 'U.S. price ranking'; DOC P6 'Other price ranking'; 82

PAGE 83

DOC V1 'Argentine value ranking'; DOC V2 'Brazil value ranking'; DOC V3 'China value ranking'; DOC V4 'Nicaragua value ranking'; DOC V5 'U.S. value ranking'; DOC V6 'Other value ranking'; DOC Q5bPrice 'Most important factor price'; DOC Q5bSafety 'Most important factor safety'; DOC Q5bProperties 'Most important factor peanut properties'; DOC Q5bconsistent 'Most important factor -size and consistency'; DOC Q5bOther 'Most important factor other'; DOC Origin 'Importance of (1=not to 5=extremely important) Origin '; DOC JIT 'Importance of (1=not to 5=extremely important) JIT '; DOC Traceability 'Importance of (1=not to 5=extremely important) Traceability '; DOC Count 'Importance of (1=not to 5=extremely important) Count '; DOC Tote 'Importance of (1=not to 5=extremely important) Tote '; DOC ColdStore 'Importance of (1=not to 5=extremely important) ColdStore '; DOC Pesticide 'Importance of (1=not to 5=extremely important) Pesticide '; DOC Aflatoxin 'Importance of (1=not to 5=extremely important) Aflatoxin '; DOC Flavor 'Importance of (1=not to 5=extremely important) Flavor '; DOC Price 'Importance of (1=not to 5=extremely important) Price '; DOC Nutrients 'Importance of (1=not to 5=extremely important) Nutrients '; DOC ShelfLife 'Importance of (1=not to 5=extremely important) ShelfLife '; DOC Size 'Importance of (1=not to 5=extremely important) Size '; DOC Sugar 'Importance of (1=not to 5=extremely important) Sugar '; DOC Oil 'Importance of (1=not to 5=extremely important) Oil '; DOC ForeignMat 'Importance of (1=not to 5=extremely important) ForeignMat '; DOC Moisture 'Importance of (1=not to 5=extremely important) Moisture '; DOC VitE 'Importance of (1=not to 5=extremely important) VitE '; DOC Roasting 'Importance of (1=not to 5=extremely important) Roasting '; DOC Rejectperc 'Importance of (1=not to 5=extremely important) Rejectperc '; DOC seedQ 'Importance of (1=not to 5=extremely important) seedQ '; DOC seedprog 'Importance of (1=not to 5=extremely important) seedprog '; DOC betterfarm 'Importance of (1=not to 5=extremely important) betterfarm '; DOC foodsafe 'Importance of (1=not to 5=extremely important) foodsafe '; DOC farmprac 'Importance of (1=not to 5=extremely important) farmprac '; DOC varietydev 'Importance of (1=not to 5=extremely important) varietydev '; ? out; 83

PAGE 84

list zvar1 JIT tote ColdStore; list zvar2 Origin Traceability Count; list zvar4 seedQ seedprog betterfarm farmprac varietydev; list zvar5 Flavor Nutrients ShelfLife Size Sugar Oil ForeignMat Moisture VitE Roasting Rejectperc; list zvar6 Pesticide Aflatoxin foodsafe; dot 1 2 4 5 6; prin zvar.; enddot; hist(discrete) share5; ? U.S. market truncation at 100 percent; TSHARE5 = 100 SHARE5; COF= (COUNTRY=1) + (COUNTRY=2)*2 + (COUNTRY=3 | COUNTRY=4 |COUNTRY=5 |COUNTRY=6 |COUNTRY=7 | COUNTRY=9)*3 + (COUNTRY=8)*4; DUMMY COF; ? COUNTRY OF THE FIRM; DOT 1 2 3; DCOF. = (COF. COF4); ENDDOT; ? TONS 1 -5; ZTONS = (TONS<=1) + (TONS=2 | TONS=3)*TONS + (TONS>=4)*4; DUMMY ZTONS; ? TONS IN RANGES 1 under 1000tons, 2=1000/5000, 3=5000/10000, 4=10000 PLUS; DOT 1 2 3; DTONS. = ZTONS. ZTONS4; ENDDOT; ? MOST IMPORTANT THING WHEN CHOSING TO IMPORT FROM; PURIMP= (Q5bPrice=1)*1 + (Q5bSafety=1)*2 + (Q5bProperties=1)*3 + (Q5bconsistent=1)*4 + (Q5bOther=1)*5; HIST(DISCRETE) PURIMP; DUMMY PURIMP; DOT(INDEX=K) 1 2 3 4 ; DIMP.=(PURIMP. PURIMP5); ENDDOT; DOT 1-5; HIST(DISCRETE) SHARE.; ENDDOT; HIST(DISCRETE) SNACK INSHELL CONFECTION; ZSNACK = (SNACK=0) + (SNACK>0 & SNACK<100)*2 + (SNACK=100)*3; DUMMY ZSNACK; DOT 1 2; DSNACK. = (ZSNACK. ZSNACK3); ENDDOT; Zpeanutbut = (peanutbut=0) + (peanutbut>0 & peanutbut<100)*2 + (peanutbut=100)*3; DUMMY Zpeanutbut; DOT 1 2; Dpeanutbut. = (Zpeanutbut. Zpeanutbut3); ENDDOT; ZINSHELL = (INSHELL=0); DUMMY ZINSHELL; DINSHELL = (INSHELL=0) + (INSHELL>0)*(-1); ZCONFECTION=(CONFECTION=0); DUMMY ZCONFECTION; DCONFECTION = (CONFECTION=0) + (CONFECTION>0)*(-1); ? TOBIT MODEL USING THE DROPING ONE DUMMY METHOD; 84

PAGE 85

? TOBIT TSHARE5 C COF1-COF3 ZTONS1-ZTONS3 PURIMP1-PURIMP4 ZSNACK1 ZSNACK2 ZPEANUTBUT1 ZPEANUTBUT2 ZINSHELL1 ZCONFECTION1; ? TOBIT MODEL USING THE AVERAGE FIRM METHOD; TOBIT TSHARE5 C DCOF1-DCOF3 DTONS1-DTONS3 DIMP1-DIMP4 DSNACK1 DSNACK2 DPEANUTBUT1 DPEANUTBUT2 DINSHELL DCONFECTION; MAT TOBCOEF=@COEF; PRINT TOBCOEF; DOT DCOF1 DCOF2 DCOF3 DTONS1 DTONS2 DTONS3 DIMP1 DIMP2 DIMP3 DIMP4 DSNACK1 DSNACK2 DPEANUTBUT1 DPEANUTBUT2 DINSHELL DCONFECTION; SET SIM.=0; ENDDOT; ?=====================; ?CREATING MATRICES; ?=====================; ?====================; ?COUNTRY MATRIX; ?====================; SET BC1=@COEF(1) + @COEF(2); SET BC2=@COEF(1) + @COEF(3); SET BC3=@COEF(1) + @COEF(4); SET BC4=@COEF(1) -[ @COEF(2) + @COEF(3) + @COEF(4) ]; SET VCOV14= 4*@VCOV(2,2) + @VCOV(3,3) + @VCOV(4,4) + 4*@VCOV(2,3) + 4*@VCOV(2,4) + 2*@VCOV(3,4); SET VCOV24= 4*@VCOV(3,3) + @VCOV(2,2) + @VCOV(4,4) + 4*@VCOV(2,3) + 2*@VCOV(2,4) + 4*@VCOV(3,4); SET VCOV34= 4*@VCOV(4,4) + @VCOV(2,2) + @VCOV(3,3) + 4*@VCOV(2,4) + 4*@VCOV(3,4) + 2*@VCOV(2,3); SET VCOV44= @VCOV(2,2) + @VCOV(3,3) + @VCOV(4,4) + 2*@VCOV(2,3) + 2*@VCOV(2,4) + 2*@VCOV(3,4); MFORM(TYPE=GENERAL,NROW=4,NCOL=4) MCNTY=0; SET MCNTY(1,2)= (@COEF(2) @COEF(3))/SQRT[@VCOV(2,2) + @VCOV(3,3) 2*@VCOV(2,3)]; SET MCNTY(1,3)= (@COEF(2) @COEF(4))/SQRT[@VCOV(2,2) + @VCOV(4,4) 2*@VCOV(2,4)]; SET MCNTY(1,4)= ( 2*@COEF(2) + @COEF(3) + @COEF(4) )/SQRT(VCOV14); SET MCNTY(2,3)= ( @COEF(3) @COEF(4) )/ SQRT[@VCOV(3,3) + @VCOV(4,4) 2*@VCOV(3,4)]; SET MCNTY(2,4)= ( 2*@COEF(3) + @COEF(2) + @COEF(4) )/SQRT(VCOV24); SET MCNTY(3,4)= ( 2*@COEF(4) + @COEF(2) + @COEF(3) )/SQRT(VCOV34); SET MCNTY(1,1) = @COEF(2)/SQRT(@VCOV(2,2)); SET MCNTY(2,2) = @COEF(3)/SQRT(@VCOV(3,3)); SET MCNTY(3,3) = @COEF(4)/SQRT(@VCOV(4,4)); SET MCNTY(4,4) = -( @COEF(2) + @COEF(3) + @COEF(4) ) /SQRT(VCOV44); SET MCNTY(2,1)= MCNTY(1,2); SET MCNTY(3,1)= MCNTY(1,3); SET MCNTY(4,1)= MCNTY(1,4); SET MCNTY(3,2)= MCNTY(2,3); SET MCNTY(4,2)= MCNTY(2,4); SET MCNTY(4,3)= MCNTY(3,4); Print mcnty; ?====================; ?TONS MATRIX; ?====================; SET BT1=@COEF(1) + @COEF(5); SET BT2=@COEF(1) + @COEF(6); SET BT3=@COEF(1) + @COEF(7); 85

PAGE 86

SET BT4=@COEF(1) -[ @COEF(5) + @COEF(6) + @COEF(7) ]; SET VCOV14= 4*@VCOV(5,5) + @VCOV(6,6) + @VCOV(7,7) + 4*@VCOV(5,6) + 4*@VCOV(5,7) + 2*@VCOV(6,7); SET VCOV24= 4*@VCOV(6,6) + @VCOV(5,5) + @VCOV(7,7) + 4*@VCOV(5,6) + 2*@VCOV(5,7) + 4*@VCOV(6,7); SET VCOV34= 4*@VCOV(7,7) + @VCOV(5,5) + @VCOV(6,6) + 4*@VCOV(5,7) + 4*@VCOV(6,7) + 2*@VCOV(5,6); SET VCOV44= @VCOV(5,5) + @VCOV(6,6) +@VCOV(7,7) + 2*@VCOV(5,6) + 2*@VCOV(5,7) + 2*@VCOV(6,7); MFORM(TYPE=GENERAL,NROW=4,NCOL=4) MTONS=0; SET MTONS(1,2)= (@COEF(5) @COEF(6))/SQRT[@VCOV(5,5) + @VCOV(6,6) 2*@VCOV(5,6)]; SET MTONS(1,3)= (@COEF(5) @COEF(7))/SQRT[@VCOV(5,5) + @VCOV(7,7) 2*@VCOV(5,7)]; SET MTONS(1,4)= ( 2*@COEF(5) + @COEF(6) + @COEF(7) )/SQRT(VCOV14); SET MTONS(2,3)= ( @COEF(6) @COEF(7) )/ SQRT[@VCOV(6,6) + @VCOV(7,7) 2*@VCOV(6,7)]; SET MTONS(2,4)= ( 2*@COEF(6) + @COEF(5) + @COEF(7) )/SQRT(VCOV24); SET MTONS(3,4)= ( 2*@COEF(7) + @COEF(5) + @COEF(6) )/SQRT(VCOV34); SET MTONS(1,1) = @COEF(5)/SQRT(@VCOV(5,5)); SET MTONS(2,2) = @COEF(6)/SQRT(@VCOV(6,6)); SET MTONS(3,3) = @COEF(7)/SQRT(@VCOV(7,7)); SET MTONS(4,4) = -( @COEF(5) + @COEF(6) + @COEF(7) ) /SQRT(VCOV44); SET MTONS(2,1)= MTONS(1,2); SET MTONS(3,1)= MTONS(1,3); SET MTONS(4,1)= MTONS(1,4); SET MTONS(3,2)= MTONS(2,3); SET MTONS(4,2)= MTONS(2,4); SET MTONS(4,3)= MTONS(3,4); Print MTONS; ?====================; ?IMPORTANCE MATRIX; ?====================; SET BT1=@COEF(1) + @COEF(8); SET BT2=@COEF(1) + @COEF(9); SET BT3=@COEF(1) + @COEF(10); SET BT4=@COEF(1) + @COEF(11); SET BT5=@COEF(1) -[ @COEF(8) + @COEF(9) + @COEF(10) + @COEF(11)]; SET VCOV15= 4*@VCOV(8,8) + @VCOV(9,9) + @VCOV(10,10) + @VCOV(11,11) + 4*@VCOV(8,9) + 4*@VCOV(8,10) + 4*@VCOV(8,11) + 2*@VCOV(9,10) + 2*@VCOV(9,11) + 2*@VCOV(10,11); SET VCOV25= 4*@VCOV(9,9) + @VCOV(8,8) + @VCOV(10,10) + @VCOV(11,11) + 4*@VCOV(8,9) + 4*@VCOV(9,10) + 4*@VCOV(9,11) + 2*@VCOV(8,10) + 2*@VCOV(8,11) + 2*@VCOV(10,11); SET VCOV35= 4*@VCOV(10,10) + @VCOV(8,8) + @VCOV(9,9) + @VCOV(11,11) + 4*@VCOV(8,10) + 4*@VCOV(9,10) + 4*@VCOV(10,11) + 2*@VCOV(8,9) + 2*@VCOV(8,11) + 2*@VCOV(9,11); SET VCOV45= 4*@VCOV(11,11) + @VCOV(8,8) + @VCOV(9,9) + @VCOV(10,10) + 4*@VCOV(8,11) + 4*@VCOV(9,11) + 2*@VCOV(10,11) + 2*@VCOV(8,9) + 2*@VCOV(8,10) + 2*@VCOV(9,10); SET VCOV55= @VCOV(8,8) + @VCOV(9,9) +@VCOV(10,10) + @VCOV(11,11) + 2*@VCOV(8,9) + 2*@VCOV(8,10) + 2*@VCOV(8,11) + 2*@VCOV(9,10) + 2*@VCOV(9,11) + 2*@VCOV(10,11); MFORM(TYPE=GENERAL,NROW=5,NCOL=5) MIMP=0; SET MIMP(1,2)= (@COEF(8) @COEF(9))/SQRT[@VCOV(8,8) + @VCOV(9,9) 2*@VCOV(8,9)]; SET MIMP(1,3)= (@COEF(8) @COEF(10))/SQRT[@VCOV(8,8) + @VCOV(10,10) 2*@VCOV(8,10)]; 86

PAGE 87

SET MIMP(1,4)= (@COEF(8) @COEF(11))/SQRT[@VCOV(8,8) + @VCOV(11,11) 2*@VCOV(8,11)]; SET MIMP(1,5)= ( 2*@COEF(8) + @COEF(9) + @COEF(10) + @COEF(11))/SQRT(VCOV15); SET MIMP(2,3)= ( @COEF(9) @COEF(10) )/ SQRT[@VCOV(9,9) + @VCOV(10,10) 2*@VCOV(9,10)]; SET MIMP(2,4)= ( @COEF(9) @COEF(11) )/ SQRT[@VCOV(9,9) + @VCOV(11,11) 2*@VCOV(9,11)]; SET MIMP(2,5)= ( 2*@COEF(9) + @COEF(8) + @COEF(10) + @COEF(11) )/SQRT(VCOV25); SET MIMP(3,4)= ( @COEF(10) @COEF(11) )/ SQRT[@VCOV(10,10) + @VCOV(11,11) 2*@VCOV(10,11)]; SET MIMP(3,5)= ( 2*@COEF(10) + @COEF(8) + @COEF(9) + @COEF(11))/SQRT(VCOV35); SET MIMP(4,5)= ( 2*@COEF(11) + @COEF(8) + @COEF(9) + @COEF(10))/SQRT(VCOV45); SET MIMP(1,1) = @COEF(8)/SQRT(@VCOV(8,8)); SET MIMP(2,2) = @COEF(9)/SQRT(@VCOV(9,9)); SET MIMP(3,3) = @COEF(10)/SQRT(@VCOV(10,10)); SET MIMP(4,4) = @COEF(11)/SQRT(@VCOV(11,11)); SET MIMP(5,5) = -( @COEF(8) + @COEF(9) + @COEF(10) + @COEF(11) ) /SQRT(VCOV55); SET MIMP(2,1)= MIMP(1,2); SET MIMP(3,1)= MIMP(1,3); SET MIMP(4,1)= MIMP(1,4); SET MIMP(5,1)= MIMP(1,5); SET MIMP(3,2)= MIMP(2,3); SET MIMP(4,2)= MIMP(2,4); SET MIMP(5,2)= MIMP(2,5); SET MIMP(4,3)= MIMP(3,4); SET MIMP(5,3)= MIMP(3,5); SET MIMP(5,4)= MIMP(4,5); Print MIMP; end; 87

PAGE 88

REFERENCES American Peanut Council 2006. “U nited States Peanut Exports and Information about the American Peanut Council.” < http://www.peanutsusa.com > Retrieved July 2006. Chvosta, J., Thurman, W., Brown, B., & Rucker, R. 2002. “The End of Supply Controls: The Economic Effects of Recent Change in Fe deral Peanut Policy.” Paper prepared for presentation at the Southern Agricultural Economics Association Annual Meeting, Mobile, Alabama, February 2003. Dohlman, E., Hoffman, L., & Young, E. 2003. “U .S. Peanut Markets Adjust to Policy Reform.” Workshop on Agricultural Policy Re form and Adjustment, Imperial College, London. Eaton, J. & Tamura, A. 1994. “Bilateralism and Regionalism in Japanese and U.S. Trade and Direct Foreign Investment Patterns.” Journal of the Japanese and International Economies. 8(4):478-510. Fair, Ray C. 1978. “A Theory of Extramarital Affairs . ” The Journal of Political Economy . 86(1):45-61 FAOSTAT 2006. “Peanut Production , Imports and Exports.” < http://www.faostat.fao.org > Retrieved July 2006. Gujarati, D.N. 2003. Basic Econometrics , 4th. ed. New York: McGraw-Hill/Irwin. Long, J.S. 1997. Regression Models for Categorical and Limited Dependent Variables . Thousand Oaks, CA: Sage Publications. Quester, A.O. & Greene, W.H. 1982. “Divor ce Risk and Wives Labor Supply Behavior.” Social Science Quarterly . 63(1):16-27. Revoredo Giha, C.L., Ndolnyak, D.A., & Fletch er, S.M. 2005. “Contract Marketing in the U.S. after the 2002 Farm Act: The Case of Peanuts.” Discussion Paper Series: Environmental Economy and Policy Res earch, University of Cambridge. 88

PAGE 89

BIOGRAPHICAL SKETCH Julie Johnson was born, as Julie Davenport, January 24, 1981, in Jack sonville, Florida. Julie and her family later moved to Alachua, Florida, where she a nd her brother, Jason, attended school. Julie graduated from Sant a Fe High School in June of 1999. After graduating from high school, Julie attended Santa Fe Community College to earn an A.A. degree and then transferred to the University of Florida to earn a B.S. in food and resource economics. Upon completing her bachelor’s degree, Julie began graduate school, also in the Food and Resource Economics Department at the Universi ty of Florida. During her undergraduate and graduate career, Julie was very active in her church, First Baptist Church of Alachua. She worked with youth at church all seven years of college and had the opportunity to participate in many mission trips. These trips allowed her to tr avel all over the U.S. and to Japan and Mexico. Through these opportunities she was able to gain experience in working with people from diverse age groups, nationalities, ethnicities, social classes and cultu res throughout the United States and abroad. July 1, 2006, Julie married Gabriel Johnson a nd moved to Port St. Lucie where Gabe was living. Upon completion of her master’s degree, Julie plans to teach in the Port St. Lucie area. 89