<%BANNER%>

Effects of Demographic Variables on the Demand for Orange Juice

Permanent Link: http://ufdc.ufl.edu/UFE0022678/00001

Material Information

Title: Effects of Demographic Variables on the Demand for Orange Juice
Physical Description: 1 online resource (63 p.)
Language: english
Creator: Davis, Andrew
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2008

Subjects

Subjects / Keywords: Food and Resource Economics -- Dissertations, Academic -- UF
Genre: Food and Resource Economics thesis, M.S.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Effects of demographic variables on the demand for orange juice affects the state of Florida because Florida?s demographics are changing more rapidly than most other states. Florida is in the south where people with Hispanic heritage are settling in. This thesis identifies the changing demographics of the United States and also provides some conclusions about the difference in orange juice consumption between these demographics. The information provided should help companies in the orange juice industry to maximize their profits and improve efficiency.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Andrew Davis.
Thesis: Thesis (M.S.)--University of Florida, 2008.
Local: Adviser: Gunderson, Michael A.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2008
System ID: UFE0022678:00001

Permanent Link: http://ufdc.ufl.edu/UFE0022678/00001

Material Information

Title: Effects of Demographic Variables on the Demand for Orange Juice
Physical Description: 1 online resource (63 p.)
Language: english
Creator: Davis, Andrew
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2008

Subjects

Subjects / Keywords: Food and Resource Economics -- Dissertations, Academic -- UF
Genre: Food and Resource Economics thesis, M.S.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Effects of demographic variables on the demand for orange juice affects the state of Florida because Florida?s demographics are changing more rapidly than most other states. Florida is in the south where people with Hispanic heritage are settling in. This thesis identifies the changing demographics of the United States and also provides some conclusions about the difference in orange juice consumption between these demographics. The information provided should help companies in the orange juice industry to maximize their profits and improve efficiency.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Andrew Davis.
Thesis: Thesis (M.S.)--University of Florida, 2008.
Local: Adviser: Gunderson, Michael A.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2008
System ID: UFE0022678:00001


This item has the following downloads:


Full Text

PAGE 1

EFFECTS OF DEMOGRAPHIC VARIABLES ON THE DEMAND FOR ORANGE JUICE By ANDREW DAVIS A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2008 1

PAGE 2

2008 Andrew Davis 2

PAGE 3

ACKNOWLEDGMENTS The first people that I would like to thank are my parents. They taught me the importance of education and encouraged me throughout the whole process. Next, I thank Dr. House for helping me get the fellowship that allowed me to obtain my masters degree, and for all the editing help. Thanks also go to the USDA and the University of Florida Food and Resource Economics department which provided me with th e fellowship that made it possible for me to further my education. Without that fellowship I woul d not have been able to secure my masters degree. I also thank Dr. Gunderson for serving as my supervisory committee chair and making all the executive decisions. Lastly, I thank Dr Brown for working through the data set and explaining every detail so that I could easily understand it. I thank all of them for the help and encouragement. 3

PAGE 4

TABLE OF CONTENTS page ACKNOWLEDGMENTS...............................................................................................................3 LIST OF TABLES................................................................................................................. ..........5 LIST OF FIGURES.........................................................................................................................6 LIST OF ABBREVIATIONS.......................................................................................................... 7 ABSTRACT.....................................................................................................................................8 1 INTRODUCTION................................................................................................................. ...9 Introduction................................................................................................................... ............9 Demographic Background of the United States.....................................................................10 Problem Setting................................................................................................................ ......11 Research Question.............................................................................................................. ....13 Study Overview................................................................................................................. .....13 2 LITERATURE REVIEW.......................................................................................................21 Prior Findings of Demogr aphic Variables on Demand..........................................................21 Orange Juice Demand.............................................................................................................23 Florida Citrus Industry............................................................................................................24 Conclusion to Literature Review............................................................................................27 3 DATA......................................................................................................................... ............30 4 MODEL........................................................................................................................ ..........39 Seemingly Unrelated Regression............................................................................................40 Pooling Time Series a nd Cross-Sectional Data......................................................................41 5 RESULTS...................................................................................................................... .........43 6 CONCLUSION................................................................................................................... ....56 Conclusion..............................................................................................................................56 Implications and Topics for Additional Research..................................................................57 LIST OF REFERENCES...............................................................................................................61 BIOGRAPHICAL SKETCH.........................................................................................................63 4

PAGE 5

LIST OF TABLES Table page 1-1 Projected ethnicity of the United States............................................................................20 2-1 Industry purchases for Flor ida citrus fruit production......................................................28 2-2 Economic impacts of the Florida citrus industry, by industry group, 2003-04 season.....29 3-1 Sample means, all cities.................................................................................................. ..33 3-2 Sample means, by city..................................................................................................... .34 3-3 Standard deviations for sample means, by city.................................................................36 5-1 Parameter results...............................................................................................................48 5-2 Elasticities by demographic..............................................................................................4 9 5-3 Elasticities, by city..................................................................................................... .......50 6-1 Simulated impact on OJ demand of a 10% OJ Price discount and a 30 percentage point increase in the advertising variable...........................................................................59 5

PAGE 6

LIST OF FIGURES Figure page 1-1 OJ demand with respect to percent of th e U.S. population that is A) Asian B) Black C) Hispanic D) Advertising (ACNielsen)..........................................................................16 1-2 Percent of the U.S. population that is A) White B) Black C) Hispanic D) Asian (United States Census Bureau)..........................................................................................18 1-3 Predicted numerical ch ange in Hispanic population from 2000-2050 (United States Census Bureau)..................................................................................................................19 3-1 Map of 52 markets......................................................................................................... ...38 5-1 Income elasticity with respect to percent A) Black B) Asian C) Hispanic.......................52 5-2 Price elasticity with respect to percent A) Black B) Asian C) Hispanic..........................53 5-3 Cross price elasticity wi th respect to percent A) Bl ack B) Asian C) Hispanic.................54 5-4 Advertising elasticity wi th respect to percent A) Bl ack B) Asian C) Hispanic................55 6

PAGE 7

LIST OF ABBREVIATIONS FDOC Florida Department of Citrus FCOJ Frozen Concentrated Orange Juice NFC Not From Concentrate Orange Juice OJ Orange Juice OLS Ordinary Least Square RECON Refrigerated Orange Juice from Concentrate SUR Seemingly Unrelated Regression 7

PAGE 8

8 Abstract of Thesis Presen ted to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science EFFECTS OF DEMOGRAPHIC VARIABLES ON THE DEMAND FOR ORANGE JUICE By Andrew Davis December 2008 Chair: Michael Gunderson Major: Food and Resource Economics We investigated how the demand for orange ju ice is affected by consumer demographics. There are multiple demographic variables that may impact demand. Additionally, other variables affect orange juice demand including the own price of orange juice, the prices of other goods that are substitutes or complements, consumer income and advertising. Ethnicity is the demographic focused on in our study. Demographic variables ar e important in determining the tastes and preferences existing in different re gions. Data collected were weekly data over slightly more than a 2-year time frame. The seemingly unrelated regression method was used to examine the data. This project is intended to be be neficial to any firm, organization or individual interested in the demand for orange juice.

PAGE 9

CHAPTER 1 INTRODUCTION Introduction Orange juice (OJ) is a very important part of the economy in the state of Florida. Florida is by far the largest orange juic e producing state in the countr y producing over 90 percent of the orange juice consumed in this country (Binkley et al 2002). Other industries directly related to Florida OJ production are financial lenders, fer tilizer companies, ag ricultural chemical companies, producers of greenhouse and nursery products, producers of plastic pipes and fittings, other state and local government enterprises, and non education government (Spreen et al 2006). Previous research has been conducted to be tter understand the de mand for orange juice and other commodities. Some important factors in the demand for orange juice include price, tastes and preferences, prices of substitutes, purchasing power or consumer, advertising, and seasonality. Some variables that are may be overlooked are the de mographics of the consumers who are purchasing the orange juice. Prior studi es have found that demographic variables help predict changes in demand for agricultural commodities. One of the potentially important demographics is the ethnicity of the consumer who is buying orange juice. Demographics of the U.S. population with rega rds to ethnicity are changing; therefore it is importa nt to understand the impact of this factor on OJ consumption. However, this has yet to be determined. Figure 11 illustrates the relationship between percent of the U.S. population that is Asian, Black, Hispanic and the advertising va riable with the per capita demand for orange juice in 52 U.S. citie s. The figure representing advertising suggests that more advertising increases the demand for orange juice, although the relationship does not look extremely strong. Somewhat le ss clearly, the ethnicity graphs suggest that as the Asian and 9

PAGE 10

Black percentages of a citys popu lation increase, orange juice dema nd in that city may increase and decrease, respectively. The relationship between the percentage of Hispanic consumers and orange juice demand is unclear. A multiple regression analysis that includes all of the variables is used in this study to more accurately determin e the true effects of et hnicity on the demand for orange juice. Demographic Background of the United States The demographic background of the United States is changing with the minority population increasing in relative size. In fact, the percentage of the population that is White has been declining steadily since the early 1900s. Figure 1-2 is based on numbers provided by the United States Census Bureau. It provides the percent of the population that is White, Black, Hispanic, and Asian since 1970. Individuals in these growing minority populations may have different tastes and preferences for agricultural commodities than what has been observed historically. Gibson and Jung (2002) provide statistics on th e changing racial composition of the U.S. population. Their study analyzed data back to 1970, w ith statistics calculated every ten years. In 1970, 83.2 percent of the United States population was White. The same year 11.1 percent was Black, 4.7 percent Hispanic, and 0.7 percent Asian. In 1980 the make-up of the population shifted to 79.6 percent White, 11.7 Black, 6.4 pe rcent Hispanic and 1.5 percent Asian. These trends continued in 1990 when the White population decreased to 75.6 percent, percent Black rose to 12.1, percent Hispanic rose from 6.4 to 9.0, and the Asian population almost doubled to 2.9 percent of the population. The current percent of the population that is Black is 12, percent Hispanic is 12 and percent Asian is 4. This m eans that the Hispanic and Asian population have been trending upwards since 1970 (Gibson 2005). 10

PAGE 11

The United States Census Bureau predicts th e demographics of Americas population from the year 2000 through 2050 (Table 1-1). The first tw o parts of the table illu strate the change in the total number of the population by race a nd ethnicity. The total populations by race are important, but the percent change for each demogr aphic group may be more revealing. The U.S. Census Bureau predicts that the population that is White (alone) will decline from 81 percent to 72.1 percent from the years 2000 through 2050. It pred icts the Black population to rise from 12.7 percent to 14.6 percent. Howeve r, the largest changes are seen in the Hispanic and Asian populations, expected to grow fr om 12.6 to 24.4 percent and 3.8 to 8.0 percent, respectively. With the number of Hispanic and Asian people in the U.S. are expected to double within 50 years, it becomes imperative to determine if different ethnicities have different tastes and preferences for orange juice These numbers pred ict that the percent of the population that is Asian and the percent of the popul ation that is Hispanic will double within 50 years. Figure 1-3 is a column chart that states th e predicted numerical change in Hispanic population from the year 2000-2050. This chart is highlighted because the Hispanic population in the state of Florida is growing faster than the majority of the states in the United States. These numbers represent the changing demographic landscape of the United States, and suggest the importance of figuring out how tast es and preferences for orange juice may be impacted by changes in ethnicity. Catering to ethnic group may become increasingly important to expand and/or maintain sales. Problem Setting The main problem that this study investigates is the fact that minority populations are rising quickly and the citrus indus try is not aware of the affect s of these changes on OJ demand. This study is important because it gives the orange juice industry an idea of where much of the 11

PAGE 12

demand for orange juice is coming. Additionally, it provides an opportunity to evaluate the prospects for expanding orange juice demand. Florida is the largest orange juice producer in the nation. It is also one of the most ethnically diverse states in America. The averag e city in Florida has a higher minority population than most cities in other regions. Different tastes and preferences are a result of the diversity of the state. With this difference in diversity co mes a larger effect on the Florida economy. This project aims to find the extent of the demographic effect. Many retail firms and businesses in the orange juice industry may be interested in the results of this research. The information in this study could be of extreme interest to the Florida Department of Citrus (FDOC). The FDOC is part of the state government and was established in 1935. The department was created because the Fl orida citrus industry requested to have a department of citrus to protect and enhance the quality and reputation of Florida citrus. The FDOC acts to protect the health and welfare and stabilize and protect the citrus industry in the state of Florida (Ward 2005). This in turn increase s the general health and welfare of the state. The FDOC carries out policy by conducting prog rams involving industry regulation, scientific research, market and economic research, advert ising, merchandising, consumer promotions, and public and industry relations (Ward 2005). The reason that this study would be of interest to the FDOC is that it has potentially important marketing implications. Over 80 percen t of the departments annual budget is spent on marketing activities. This is down from 87 pe rcent that occurred over the 1935-2004 period. The seven percent lost is now used for expenses related to mechanical harvesting. The FDOC advertises in magazines, infome rcials, TV, and on radio stations Advertising is also conducted in foreign markets. The FDOC concentrates on the United States, Canada, Europe, and Asia. The 12

PAGE 13

share of the marketing accounted for by mercha ndising was between 10 and 20 percent before the FDOC was downsized and eliminated a numbe r of field representatives who worked across the country merchandising Florida OJ. Currently, there is a small scale of merchandising representatives that have been rehired (Ward 2005). Research Question The primary objective of this project is to determine how ethnicity affects the demand for orange juice. The three demogra phic variables reviewed in this paper are th e percentages of the United States population that are Black, Hispanic and Asian. Examining these ethnic variables should assist in predicting the differences in de mand for orange juice across U.S. cities. The data used cover 120 weeks of the demand for orange ju ice and its substitutes in over 52 of the largest U.S. cities and their surrounding areas. This study estimates a model with different variables that effect orange juice demand. It concentrates on how the demand for orange juice is affected as the percentages of the populati on that are Black, Hispanic, and Asian change. The second objective is to measure the im pact of ethnicity on selected demand elasticities. This include s evaluating the impacts of the demo graphic variables on the responses of OJ demand to changes in the price of OJ, th e prices of substitutes, consumer income, and advertising. Information on how ethnicity inte racts with these basic demand factors may be useful in designing promotional programs to reach each demographic group specifically. Study Overview The next chapter of this study is the literature review which will discuss previous research related to the subjects addressed in this paper. The third chapter presents and analyzes the data used in the estimation. Chapter four discusses the theoretical model a nd the results of the 13

PAGE 14

estimation. Finally, chapter five provides conclusi ons, implications for the citrus industry, and some thoughts on additional oppor tunities for research. 14

PAGE 15

0 0.5 1 1.5 2 2.5 3 3.5 01234567 % AsianPer Capita OJ Demand8 A 0 0.5 1 1.5 2 2.5 3 3.5 05101520253035 % BlackPer Capita OJ Demand B 15

PAGE 16

0 0.5 1 1.5 2 2.5 3 3.5 05101520253035 % HispanicPer Capita OJ Demand C 0 0.5 1 1.5 2 2.5 3 3.5 50 60 70 80 90 100 % AdvertisingPer Capita OJ Demand D Figure 1-1 OJ demand with respect to percent of the U.S. population that is A) Asian B) Black C) Hispanic D) Advertising (ACNielsen) 16

PAGE 17

64.00% 66.00% 68.00% 70.00% 72.00% 74.00% 76.00% 78.00% 80.00% 82.00% 84.00% 86.00% 1970 1980 1990 2008 YearPercent White A 10.60% 10.80% 11.00% 11.20% 11.40% 11.60% 11.80% 12.00% 12.20% 1 YearPercent Black B 17

PAGE 18

0.00% 2.00% 4.00% 6.00% 8.00% 10.00% 12.00% 14.00% 1 YearPercent Hispanic C 0.00% 0.50% 1.00% 1.50% 2.00% 2.50% 3.00% 3.50% 4.00% 1 YearPercent Asian D Figure 1-2 Percent of the U.S. population that is A) White B) Black C) Hispanic D) Asian (United States Census Bureau) 18

PAGE 19

19 Numerical Change in Hispanic Population From 2000-20500 10,000 20,000 30,000 40,000 50,000 60,000 70,000 80,000 2000 20102010 20202020 20302030 20402040 20502000 2050 YearChange in Hispanic Population Figure 1-3 Predicted numerical change in Hispanic population from 2000-2050 (United States Census Bureau)

PAGE 20

Table 1-1 Projected ethnic ity of the United States Values in Thousands Population 2000 2010 2020 2030 2040 2050 Total 282,125 308,936 335,805 363,584 391,946 419,854 White alone 228,548 244,995 260,629 275,731 289,690 302,626 Black alone 35,818 40,454 45,365 50,442 55,876 61,361 Asian alone 10,684 14,241 17,988 22,580 27,992 33,430 All other races 7,075 9,246 11,822 14,831 18,388 22,437 Hispanic (of any race) 35,622 47,756 59,756 73,055 87,585 102,560 Population % 2000 2010 2020 2030 2040 2050 White alone 81.0 79.3 77.6 75.8 73.9 72.1 Black alone 12.7 13.1 13.5 13.9 14.3 14.6 Asian alone 3.8 4.6 5.4 6.2 7.1 8.0 All other races 2.5 3.0 3.5 4.1 4.7 5.3 Hispanic (of any race) 12.6 15.5 17.8 20.1 22.3 24.4 Numerical Change 2000 2010 2010 2020 2020 2030 2030 2040 2040 2050 2000 2050 Total 26,811 26,869 27,779 28,362 27,908 137,729 White alone 16,447 15,634 15,102 13,959 12,936 74,078 Black alone 4,636 4,911 5,077 5,434 5,485 25,543 Asian alone 3,557 3,747 4,592 5,412 5,438 22,746 All other races 2,171 2,576 3,009 3,557 4,049 15,362 Hispanic (of any race) 12,134 12,000 13,299 14,530 14,975 66,938 % Change 2000-2010 2010 2020 2020 2030 2030 2040 2040 2050 2000-2050 Total 9.5 8.7 8.3 7.8 7.1 48.8 White alone 7.2 6.4 5.8 5.1 4.5 32.4 Black alone 12.9 12.1 11.2 10.8 9.8 71.3 Asian alone 33.3 26.3 25.5 24.0 19.4 212.9 All other races 30.7 27.9 25.5 24.0 22.0 217.1 Hispanic (of any race) 34.1 25.1 22.3 19.9 17.1 187.9 Source: United States Census Bu reau, International Data Base 20

PAGE 21

CHAPTER 2 LITERATURE REVIEW There have been prior studies on the relationship between de mographic variables and the demand for agricultural commodities. The majority of these focused on specific demographic variables to concentrate on and developed a model which identified the effects those demographic variables had on that particular pr oduct. Studies on the effects that different variables have on the demand for orange juice are also common. Th is literature review starts with the prior findings of demographic variable s on demand and is followed by information is provided on the orange juice industry and OJ demand. The last subject reviewed is the mathematical model used in this study. Prior Findings of Demographic Variables on Demand Wilson (2005) identified the impacts of de mographic variables on meat demand. The demographics of this country have shifted thro ughout the last century. There is more cultural diversity in the United States today. Wilson c oncluded that age, residency, household size, education level, and percent of women in the labor force had impacts on the demand for meat (2005). The study concluded that the ow n price effects of meat dema nd are different from zero as expected. The cross-price effects for beef and poultry were also significan tly different from zero indicating that beef, pork, and poultry are substitutes for one a nother. The null hypothesis that demographic and health parameters were not si gnificant was rejected. Demand for beef, pork, and poultry may have been affected by these va riables. The Rotterdam Model was the base model for this study. It was chosen because it is consistent with demand theory, flexible, and most of all it is very useful in capturing non-price effects (Wilson 2005). 21

PAGE 22

As age of the consumer increased the dema nd for beef increased while pork and poultry declined. An increase in ethnic diversity shifte d consumption to poultry and away from pork and beef. The percent of women with jobs and the e ducation level of the pop ulation had the largest impact on demand. As females enter the labor force demand for pork and poultry rises and the demand for beef declines. Poultry was the benefact or from the increased education level of the population. This result was consis tent with the speculation that the demand for poultry would rise when people have less time to cook. The poultry industry is fu rther along in the prepared meals category than the beef and pork industries (Wilson 2005). While meat is not the same as orange juice, there are some major points that are useful for the orange juice study. Overall, this study st ates that demographic variables do affect demand for agricultural commodities. Heien (1988) conducted research that focuse d on demographic effects on the demand for beef products. This study concluded that househol d size, region, tenancy, and ethnic origin were all statistically significant explanatory variable s. Employment status, shopper, and occupation were demographic effects that were statistically insignificant. The analysis also found evidence of strong own-price and cross-pr ice effects among meats. This paper was based on the Almost Ideal Demand System (AIDS) (Heien 1988). A study conducted by Blisard ( 2003) looked ahead to 2020 to determine the possible effects of demographics in the future. This st udy determined that the future population increase alone will increase spending on food in the U.S. by 26.3 percent. The demographic changes will increase per capita food purchases by 7 percent (Blisard 2003). The same study states that income disparities have a large effect on where families spend money on food. Higher income households spen d more money on food than lower income 22

PAGE 23

households. Food groups with the largest increase in consumption include food away from home, fruits, prepared foods, vegetables, and dairy. Also, people above the age of 74 will spend the most money on cereal and bakery goods, and frui t. Black households typically spend less than Whites and spend more on fish, meat, and eggs. The Northeastern United States spends the most on foods and the North Central sp ends the least (Blisard 2003). The change in the demographic profile of th e United States has connotations that will influence where, when, and what people will eat The recent trend has been that United States residents are becoming older, more educated, fina ncially better off, and more ethnically diverse (Blisard 2003). Fruit and fruit jui ces are substantially affected by this change in demographic variables. For every ten percent increase in income, household spe nding on fruits and fruit juices will rise 1.62 percent (Blisard 2003). Households with children age nine or under spend more money on fruit and fruit juices than households with older children. Orange Juice Demand There are an abundance of things that affect the demand for orange juice. A major interest exists on factors underlying demand increases or decreases for this commodity. Variables that possibly affect the demand for OJ include the pr ice of OJ, prices of substitutes, income, seasonality, demographics, and advertising. All of these are of interest, but advertising is a variable that the marketing firms can control. Th e role of advertising is increasing significantly especially in the agricultural commodities (W ard 1978). The effect that advertising has on demand is important for determining the amount of marketing firms c onduct. The following studies were conducted on factors th at change the demand for citrus. There are several different types of advertisi ng that can be used to increase the demand for orange juice. One form of advertising is c oupons. Brown and Lee found that the discounts and information provided by coupons increased the demand for FCOJ. Coupons offer a price 23

PAGE 24

reduction directly to the consumer. This make s it possible for manufact urers to reduce price where a demand problem exists (Brown 1985). Coupons influence the consumption of OJ through both advertising and pric e-discounting effects (Ward 1978). The success of advertising with coupons depends on multiple factors. A higher shelf price increases the likelihood that cons umers will redeem coupons. Demogr aphic variables also affect the use of coupons. Cents-off coupons are redeem ed by the White population more often than other races (Brown 1985). Coupons can stimulate consumers to increase their consumption. Effectiveness of coupons is significantly affected by age (Ward 1978). Brown and Lee also found that for each ten cent di scount (per capita) on a purchas e of FCOJ the per capita OJ purchase would increase by 2.4 to 3.5 ounces. High-income households are more likely to use cents-off coupons. The race variable was insignifi cant in this study (Brown 1985). Discounts and information provided by coupons increased de mand for FCOJ (Brown 1985). Coupons are an effective tool for informing the consumer about FCOJ (Ward 1978). Florida Citrus Industry The citrus industry was fairly small before th e 1940s. That is when frozen concentrated orange juice (FCOJ) was developed. That developm ent led to a significant amount of growth in the amount of citrus being produced At that time there was little to no brand identity in the industry (Ward 2005). The next year that there was a turning point in the industry was 1962. Th at is when a brutal freeze swept through Florida and cut the amount of processed oranges by more than 50 percent. This led to a lack of supply and in turn OJ prices skyroc keted. That also led Florida processors to turn to Brazil for another source of orange jui ce when the Floridian supply was unable to meet market demand. The addition of the Brazilian firms made the citrus market much more competitive (Ward 2005). 24

PAGE 25

In the early 1980s there were several seve re freezes. They occurred in 1981, 1982, 1983, and 1985. These freezes cut Floridas orange juice production in half. This was also the decade when new OJ brands emerged. Tropicana, Minut e Maid, and Proctor and Gambles Citrus Hill brand were beginning to battle fo r market share. Prices also increased dramatically during the time of the freezes (Ward 2005). The 1990s and 2000s brought some major brand advertising campaigns as more large firms entered the industry. The most recent change that may affect the industry is the growth of supercenters, particularly Wal-Mart Supercenters. They have become the largest retail grocer and have lower prices and lower margins. It is not cl ear how the growth of supercenters has impacted the citrus industry (Ward 2005). There are three major types of orange ju ice. These three types include frozen concentrated orange juice (FCO J), not-from-concentrate orange juice (NFC), and refrigerated orange juice from concentrate (RECON). Of the orange juice purchased in the United States, 98 to 99 percent falls in one of thes e three types. NFC is considered to be the highest quality of orange juice and is the fastest growing in gallons sold in the United States at this time (Brown 2000). Storage is a very important aspect in the orange juice industry because it is such a seasonal product. Consumption is somewhat seas onal with demand being at the peak during cold season. Production is very seasonal. FCOJ stocks are at the lowest in November and at a high in May. Frozen concentrate orange ju ice can be stored for over a y ear at appropriate temperature levels. NFC also lasts over a year when it is ei ther frozen or chilled (Binkley et al 2002). Prices for orange juice can be affected by income and location (Binkley et al 2002). The higher income areas in the United States tend to have higher OJ prices. Also, transportation 25

PAGE 26

theory would suggest that areas further away fr om Florida would have hi gher prices because of higher shipping costs. The Florida citrus industry is a large part of the Florida econo my, and also a huge supplier of exports to other states Nearly all (97%) of Floridas citrus products are consumed in other states or foreign countries. Table 2-1 is a table showing input purcha ses by the Florida ci trus industry. It demonstrates the total expenditures in each sector and the percentage of their revenue that stems from the Florida citrus industry. The total value of citrus fruit in 2003-04 was $1.778 billion. The total value of citrus juice was $3 billion, including $1.93 for canned ju ice and $1.08 billion for FCOJ. The direct output or sales revenue produced by the citrus indu stry in this season was $3.69 billion. The total output impact of the industry was $9.29 billi on. The indirect output im pacts resulting from purchases of inputs from othe r industry sectors were $1.93 bil lion. The induced output impacts resulting from consumer spending by employee households were $3.67 billion (Spreen et al 2006). The Florida citrus industry has an affect on ma ny other industries in th e state of Florida. The economic impacts of the Florida citrus industry, by industry group, for the 2003-04 seasons are demonstrated in Table 2-2. It shows that 20 different industry gr oups are significantly impacted by the Florida citrus industry. There ar e 76,336 jobs in the State of Florida as a result of the citrus industry, with 61,307 jobs due to th e processing sector and 15,029 jobs due to fresh fruit. Value added is a broad measure of tota l personal and business in come generated, and is equivalent to industry output less industry purchases. Total value added impacts were $4.87 billion. Labor income impacts amounted to $2.73 billion, which represents all wages and salary 26

PAGE 27

earnings by industry employees and proprietors income to busin ess owners. Indirect business tax impacts were $288 million, which includes most forms of local and state taxes, such as property tax, sales tax, water mana gement district levies, intang ible taxes, motor fuel and a vehicle tax, excise taxes, etc ., but does not include federal taxes (Spreen et al 2006). The economic impacts of the Florida citrus industry presented here for the 2003-04 season are consistent with those reported in a previous study for the 1999-00 season. The total economic impact of the Florida citrus industr y has declined by about 5% between the 1999-00 season to 2003-04 season (Spreen et al 2006). Conclusion to Literature Review The previous research on commodity demands and the impacts of demographics are the basis of this study. The gap in th e research that this study strives to fill is incorporating the effects of ethnicity in a demand model for oran ge juice. Many studies have focused on the effects of demographic variables on commodity demands, but demographic effects on orange juice demand have received less attention. Various variables that affect the demand for orange juice have been studied, but ethn icity variables have not usually been considered. All of these studies include the basic variab les (prices and income) in th e OJ demand equation. This study will include the same variables, but concentrate on the effect different ethnic variables have on demand. The seemingly unrelated regression method is the most appropriate for the time series and cross-section data that has be en provided. The orange juice industry is an important part of the economy. This study will provide additional information to the citrus industry. 27

PAGE 28

Table 2-1 Industry purchases for Fl orida citrus fruit production Sector Total expenditures Output Greenhouse and nursery products $21,589,880 1.22% Fertilizer mixing, manufacturing $87,820,996 4.97% Pesticides & agricultural chemicals $98,360,748 5.57% Plastic pipes & fittings $84,838,081 4.80% Financial lenders $195,331,379 11.06% Other state & local government enterprises $49,354,411 2.79% State & local government non-education $33,311,668 1.89% Total industry purchases $500,607,163 32.31% Source: (Spreen et al 2006) 28

PAGE 29

Table 2-2 Economic impacts of the Florida ci trus industry, by industry group, 2003-04 season Industry group Output Employme nt Value added Labor income Indirect business taxes -mil. $-jobs--------------Milli on $ ---------------Agriculture, forestry, fisheries & hunting 1,577.1 21,814 1,012.0 420.0 46.7 Mining 9.6 37 2.2 0.9 0.2 Utilities 88.3 163 60.3 18.5 8.8 Construction 478.4 4,281 205.4 168.6 2.5 Manufacturing 3,540.2 9,836 1,172.2 589.0 28.7 Wholesale trade 288.7 2,184 219.6 123.1 47.4 Transportation & warehousing 159.4 1,712 88.4 65.0 3.4 Retail trade 334.7 5,945 249.6 155.5 47.6 Information 131.2 538 61.0 32.1 5.2 Finance & insurance 418.8 2,496 264.6 130.4 8.5 Real estate & rental 224.7 1,537 149.9 39.4 23.5 Professional, scientific & technical services 288.1 2,808 172.0 144.4 2.8 Management of companies 92.5 578 55.2 42.5 0.9 Administrative & waste services 110.6 2,031 67.6 54.9 1.7 Educational services 32.9 684 19.1 18.6 0.4 Health & social services 368.6 4,897 228.2 199.7 2.5 Arts, entertainment, & recreation 43.5 741 27.7 19.0 2.9 Accommodations & food services 170.3 3,371 88.0 60.3 9.6 Other services 162.8 3,066 85.1 66.0 6.7 Government 768.5 7,616 645.1 383.6 37.6 Total 9,288.8 76,336 4,873.2 2,731.4 287.5 Source: (Spreen et al 2006) 29

PAGE 30

CHAPTER 3 DATA The data set used in this study is provided to the Florida Department of Citrus by ACNielsen. Figure 3-1 is a picture of the 52 different markets that are included in the ACNielsen data. The information that it provides is for 52 of the United States largest cities and their surrounding areas. New Orleans was one of the cities that the data were provided for, but after hurricane Katrina the demographic variables of New Orleans are still in question. In the model there are only 51 U.S. cities because of this. The data are collected weekly with 120 weeks starting with the week endi ng April 23, 2005 and ending with the week ending August 4, 2007. The data are for grocery stores with $2 milli on of annual business or greater. This does not include the Wal-Mart super stores. The data include the amount of total orange juice sold in gallons and dollars. Data on gallon and dollar sales for substitutes for orange juice are also provided. The substitute juices included in the data set are grapefruit juice, orange juice blends, grapefruit juice blends, grapefruit juice cocktails, orange juice drinks, and orange blend drinks. Advertisements are also accounted for in the information which is pr ovided. This variable is represented by the percentage of all commodity volume with a prominent orange juice feature. At the end of the data set is the demographic data. The income variable was found in the 2007 Market Scope: The Desktop Guide to Supermarket Share, by Trade Dimensions International, Inc. It is calculated as the num ber of households in the city times the citys disposable income per household (average house hold EBI) divided by the citys population. It tells us how much income the average househol d in each city has for all purchases. The percentages of the city which are Black, Asian, a nd Hispanic are listed next. These are the three ethnic variables that are availa ble and included in this study. 30

PAGE 31

Each city has different demographic and inco me variables. However, these variables are the same across the weeks for each particular ci ty. The effects of the income and ethnicity variables are determined from the differences in th e data across the cities. The effects of prices and promotional variables are based on the va riation in the data over time and cities. Sample means for all cities are shown in table 4-1. These means provide an information base for the orange juice market. Annual per cap ita orange juice consumption is 1.92 gallons with a standard deviation on .49. On average, a gallon of orange juice sells for $5.05 a gallon with a standard deviation of .71 and orange juice/drink sells fo r $4.10 with a standard deviation of .53. In the U.S. the average person in the workforce makes almost $19,900 per year with a standard deviation of $2,218.05. The average city in the U.S. has a Black population of 11.43 percent with a standard devi ation of 7.45. The average city has an Asian population of 2.37 percent with a standard deviat ion of 2.55. The average U.S. city has a Hispanic population of 8.9 percent with a standard deviat ion of 9.88. Of the stores in th e average city of the U.S., 86.74 percent of them have an advertising f eature with a standard deviation of 15.16. Sample means and standard deviations by c ity are shown in table 4-2 and table 4-3, respectively. Per capita OJ gallons consumed, the OJ price, the price of substitutes, income, percent Black, percent Asian, percent Hispanic, and percent advertising are provided for each city. The city with the highest per capita ora nge juice consumed is Boston, which consumes 3.13 gallons. The region with the lowe st per capita orange juice cons umed is West Texas, which consumes 0.90 gallons. The city with the highest price for OJ is San Francisco paying on average $6.27 per gallon. The city with the lowest pric e for OJ is Omaha paying $4.21 per gallon. San Francisco also has the highest substitute pr ice which is $5.42 per gallon. Birmingham has the lowest substitute price at $3.20 per gallon. San Fr ancisco has the highest per capita income at 31

PAGE 32

32 $26,564.01. West Texas has the lowest per capi ta income at $14,550.83. The city with the highest percentage of Black citizens is Memphis with 32.6 pe rcent of the population being African American. The city with the lowest percentage of Black population is Salt Lake City with just .7 percent of the popul ation being Black. The city with the highest percent Asian is San Francisco with 15.8 percent of th e population being of Asian descent. The lowest percent Asian is Miami with only .4 percent of the population being Asian. The city with the most Hispanic population is West Texas with 42.8 percent of the population being Hispanic. The lowest percent Hispanic is Pittsburgh with .5 percent of the population being Hispanic. Hartford/Connecticut has the highest percent advert ising at 97.64%. The lowest per cent advertising belongs to Birmingham at 60.33%.

PAGE 33

Table 3-1 Sample means, all cities. Variable Mean Std. Dev. Annual Per Capita OJ Ga. 1.92 0.49 OJ Price 5.05 0.71 Other Juice/Drink Price 4.10 0.53 Per Capita Income 19,899.09 2,218.05 % Black 11.43 7.45 % Asian 2.37 2.55 % Hispanic 8.90 9.88 % Features 86.74 15.16 33

PAGE 34

Table 3-2 Sample means, by city City Annual Per Capita OJ Ga. OJ Price Other Juice/Drk Price % Black % Asian % Hispanic % Features Albany 2.54 5.07 4.44 5.5 1.2 3.3 93.46 Atlanta 1.75 4.97 4.03 25 1.2 5.6 85.8 Baltimore 2.61 5.31 4.25 24.7 2.6 1.6 95.13 Birmingham 1.29 4.75 3.2 24.7 1.1 1.7 60.33 Boston 3.13 5.17 4.85 4.3 3 5.2 92.21 Buffalo/Rochester 2.74 4.82 4.19 9.3 1.5 3 80.9 Charlotte 1.79 5.23 3.83 18.6 1.3 4 93.66 Chicago 2.12 5.12 4.54 17.6 3.8 13.5 92.71 Cincinnati 2.28 4.61 3.94 11.4 1.3 0.9 85.04 Cleveland 2.03 5.04 3.89 12.3 0.4 1.8 89.48 Columbus 1.88 4.52 3.98 9.8 1 1.3 87.9 Dallas/Ft. Worth 1.28 4.85 3.76 13.2 1.7 17.1 84.28 Denver 1.96 5.25 4.8 4.4 1.5 14.5 88.74 Des Moines 1.96 4.64 3.57 2.1 2.5 2.5 69.42 Detroit 2.11 4.85 4.07 18.4 0.8 2.5 87.4 Grand Rapids 1.8 4.79 3.84 6.3 0.8 3.5 93.73 Hartford/New Haven 2.57 5.36 4.68 8.1 2.5 7.8 97.64 Houston 1.65 4.7 3.45 16.4 1.6 24.7 85.41 Indianapolis 1.78 4.82 4.01 8.5 0.7 2.1 87.55 Jacksonville 1.72 5.18 3.92 19.9 2.7 3.5 91.55 Kansas City 1.61 4.83 4.1 9.9 1.3 3.9 82.46 Las Vegas 1.74 5.07 3.78 8.6 3.3 17.5 92.59 Little Rock 1.21 4.66 3.97 14.5 1 2.4 72.87 Los Angeles 1.72 5.46 4.05 8.5 7.2 31.7 90.17 Louisville 1.69 4.56 3.92 9.3 1.3 1.6 84.73 Memphis 1.26 4.65 3.48 32.6 0.5 1.6 76.13 Miami 2.54 5.47 4.27 15.7 0.4 31 92.35 Milwaukee 2.22 4.93 3.88 10.8 0.7 4.7 94.22 Minneapolis 2.14 5.02 4.19 4 2.6 2.3 86.36 Nashville 1.73 4.71 3.86 11.1 1.1 1.7 78.11 New York 2.31 5.55 4.98 16.8 4.6 17 96.68 Oklahoma City/Tulsa 1.32 5.12 3.95 7.3 1.8 4 77.98 Omaha 1.87 4.21 3.44 4.8 1.2 3.9 72.72 Orlando 1.88 5.23 4.09 11.3 0.5 11.2 92.15 34

PAGE 35

Table 3-2 Contd. Philadelphia 2.45 5.14 4.18 15.2 2.7 5 96.09 Phoenix 1.74 5.2 4.42 3 2.4 20.7 86.24 Pittsburgh 2.09 4.93 3.96 5.4 1 0.5 92.92 Portland 1.9 5.25 4.78 1.7 3.4 5.7 87.15 Raleigh/Durham 1.77 5.26 3.85 23.9 1.2 4.5 92.22 Richmond/Norfolk 1.89 5.2 3.93 27.4 1.7 2.1 91.53 Sacramento 1.63 5.62 4.57 5.8 6.9 17.1 92.03 St. Louis 1.84 5.13 4.09 15 1.3 1.1 88.97 Salt Lake City/Boise 1.69 4.72 4.05 0.7 1.8 7.5 90.99 San Antonio 2.02 4.28 3.23 5.9 1.6 35.5 93.07 San Diego 1.88 5.63 4.61 5.2 7 20.9 93.68 San Francisco 1.89 6.27 5.42 7.1 15.8 14.2 91.18 Seattle 1.91 5.44 4.83 4.1 6.3 4.4 89.25 Syracuse 2.55 4.94 4.36 4.4 1.1 2 85.13 Tampa 2.05 5.44 4.27 8.2 0.7 9.4 84.78 Washington D.C. 2.45 5.44 4.44 20.6 4.4 5.8 97.38 West Texas 0.9 5.13 3.53 3.6 0.8 42.8 60.58 35

PAGE 36

Table 3-3 Standard deviations for sample means, by city City Annual Per Capita OJ Ga. OJ Price Other Juice/Drk Price % Black % Asian % Hispanic % Features Albany 0.25 0.47 0.22 0 0 0 10.59 Atlanta 0.21 0.66 0.25 0 0 0 15.81 Baltimore 0.31 0.63 0.26 0 0 0 5.02 Birmingham 0.17 0.6 0.17 0 0 0 11.89 Boston 0.26 0.52 0.25 0 0 0 6.67 Buffalo/Rochester 0.34 0.47 0.31 0 0 0 20.08 Charlotte 0.23 0.6 0.2 0 0 0 11.74 Chicago 0.28 0.57 0.22 0 0 0 6.1 Cincinnati 0.32 0.61 0.29 0 0 0 22.16 Cleveland 0.27 0.54 0.24 0 0 0 6.58 Columbus 0.31 0.66 0.32 0 0 0 20.25 Dallas/Ft. Worth 0.15 0.54 0.21 0 0 0 8.33 Denver 0.22 0.59 0.36 0 0 0 19.32 Des Moines 0.23 0.62 0.16 0 0 0 14.1 Detroit 0.28 0.56 0.29 0 0 0 12.8 Grand Rapids 0.27 0.54 0.29 0 0 0 8.11 Hartford/New Haven 0.31 0.63 0.27 0 0 0 2.37 Houston 0.12 0.46 0.19 0 0 0 10.19 Indianapolis 0.24 0.54 0.29 0 0 0 13.08 Jacksonville 0.21 0.76 0.22 0 0 0 11.54 Kansas City 0.21 0.6 0.22 0 0 0 14.5 Las Vegas 0.2 0.68 0.37 0 0 0 10.31 Little Rock 0.19 0.55 0.33 0 0 0 15.76 Los Angeles 0.18 0.61 0.47 0 0 0 5.52 Louisville 0.24 0.63 0.33 0 0 0 20.25 Memphis 0.18 0.55 0.23 0 0 0 15.98 Miami 0.32 0.82 0.24 0 0 0 10.49 Milwaukee 0.23 0.53 0.2 0 0 0 4.61 Minneapolis 0.29 0.62 0.24 0 0 0 15 Nashville 0.25 0.62 0.26 0 0 0 16.77 New York 0.3 0.66 0.24 0 0 0 3.15 Oklahoma City/Tulsa 0.19 0.52 0.26 0 0 0 10.83 Omaha 0.27 0.56 0.35 0 0 0 13.94 Orlando 0.19 0.79 0.22 0 0 0 11.79 36

PAGE 37

37 Table 3-3 Contd. Philadelphia 0.29 0.59 0.27 0 0 0 2.82 Phoenix 0.22 0.7 0.41 0 0 0 9.81 Pittsburgh 0.29 0.61 0.22 0 0 0 7.47 Portland 0.22 0.66 0.32 0 0 0 12.47 Raleigh/Durham 0.22 0.59 0.19 0 0 0 13.88 Richmond/Norfolk 0.18 0.67 0.2 0 0 0 15.05 Sacramento 0.18 0.69 0.43 0 0 0 6.82 St. Louis 0.22 0.49 0.21 0 0 0 9.58 Salt Lake City/Boise 0.26 0.72 0.29 0 0 0 9.56 San Antonio 0.11 0.33 0.25 0 0 0 18 San Diego 0.19 0.61 0.5 0 0 0 6.42 San Francisco 0.23 0.67 0.26 0 0 0 5.46 Seattle 0.2 0.65 0.3 0 0 0 9.29 Syracuse 0.27 0.46 0.24 0 0 0 13.13 Tampa 0.27 0.74 0.26 0 0 0 10.9 Washington D.C. 0.27 0.6 0.23 0 0 0 5.06 West Texas 0.12 0.53 0.19 0 0 0 19.3

PAGE 38

38 Figure 3-1 Map of 52 markets

PAGE 39

CHAPTER 4 MODEL The seemingly unrelated regr ession method was used to analyze the demand for orange juice across cities. The model estimated by this method is discussed first, followed by some additional discussion on the estimation method. The dependent variable in the model for this pr oject is total orange ju ice gallons per capita. The explanatory variables are the price of ora nge juice, price of s ubstitutes, per capita buying income, percent Black, percent As ian, percent Hispanic, percent of stores with a prominent orange juice newspaper advertisement, display a nd/or price discount, and s easonality variables. Also included in the list of variables are twelve interaction terms. The interactions are between the OJ price and the three ethnicity variables, between the cross price and the three ethnicity variables, between income and the three ethnicity variables, and between advertising and the three ethnicity variables. The demand for orange juice by city and week can be written as Equation 1: Log(qit)=B0+B1*z1i+B2*z2i+B3*z3i+B4*log(xi)+B5*z1i*log(xi)+B6*z2i*log(xi)+B7*z3i*log(xi)+ B8*log(pit)+B9*z1i*log(pit)+B10*z2i*log(pit)+B11*z3i*log(pit)+B12*log(psit)+B13*z1i*log(psit)+B1 4*z2i*log(psit)+B15*z3i*log(psit)+B16*ait+B17*z1i*ait+B18*z2i*ait+B19*z3i*ait+B20*sint+B21*cost+eit, where the i and t subscripts stand for the city (i=1 -51) and week (t=1-120), respectively; q is the per capita OJ gallon sales; z1,z2, and z3 are the percent Black, Asian, and Hispanic, respectively; x is the income variable; p is the own price variab le; ps is the substitute price variable; a is the advertising variable; sint = sine(2*pi*t/52) and cost = cosine(2*pi*t/52) are the seasonality variables; and eit is the error term. The B s are coefficients to be estimated. The income, own price, and cross price el asticities for the demand equations can be written as 39

PAGE 40

Equation 2: Ex= B4+B5*z1i+B6*z2i+B7*z3i Equation 3: Ep= B8+B9*z1i+B10*z2i+B11*z3i Equation 4: Eps= B12+B13*z1i+B14*z2i+B15*z3i These equations indicate the percent change in orange juice demand for a one percent change in income, own price, and substitute price, respectively. The impact of the promotional variable on the demand for orange juice is Equation 5: Ea= B16+B17*z1i+B18*z2i+B19*z3i Since the advertising variable is a percentage it does not re present the exact advertising elasticity, but it is very similar. This equation re presents the percentage change in orange juice demand for a one percentage point ch ange in the advertising variable. The log form is used for the dependent variab le in the 52 equations. It is also the form used to specify the own price, cross price, and income. This specification directly provides the price, cross price, and inco me elasticities for orange ju ice demand as noted above. The advertising variable is repr esented by the percent of stores with a feature orange juice advertisement. The advertising numbers for th e cities are between zero and 100. The interaction terms with advertising are included to determine how the cities with different demographic variables respond to changes in advertising. The following subsections further discuss the seemingly unrelated regression method and related issues. Seemingly Unrelated Regression The Seemingly Unrelated Regression (SUR) model was developed by Arnold Zellner who first published it in 1962. In the SUR model th e residuals are uncorrelated over time but correlated across cross-section units. This type of correlation is common when some variables are omitted that are common to all equations. If there are a large number of cross-section units and the time series is short, seemingly unrelated regression is not a feasible model. It is essential to study whether there is a systema tic pattern in the resi duals. This reveals whether the residuals 40

PAGE 41

are auto-correlated and/or correlated across cros s-section units. After performing these steps the appropriate model will be revealed (Maddala 1977). A good example of when to use SUR is when one is estimating demand equations for a number of different consumption or investment goods or other cases that estimate a group of equations. Two conditions should be fulfilled in these cases; these two cond itions are that the set of explanatory variables should not be identica l for each commodity and there should be nonzero correlations between the error te rms in two or more equations. If those two conditions hold SUR will be more efficient than OLS. The gain in efficiency yielded by SUR over the OLS increases directly with the correlation between the errors from the different equations and inversely with the correlation between the different sets of explanatory variables (Johnston 1963). Kmenta (1986) agrees that SUR regressions ar e most appropriate when estimating multiple equations that may have correlated error terms. Th is is the case when there are demand functions for various commodities or producti on functions for different industries, with observations made over time or over some cross-sectional units (Kmenta 1986). Pooling Time Series and Cross-Sectional Data The type of data set used in th is study is one in which time series and cross-sectional data are pooled together. The 120 weeks of data per city are the time series part and the 52 targeted cities are the cross-sectional element. Many st udies have been conducte d studying this type of data. Terry Dielman wrote a pape r in 1983 that describes the diffe rent methods one can use to examine these statistics. This paper is used to i llustrate why seemingly unrelated regression is the best fit for our data. Dielman (1983) states that in econometrics literature a data base that provides a multivariate statistical history over time for each of a number of different units is described as a pooled cross-sectional and time series data base Survey statistics usually deal with cross41

PAGE 42

sectional data while econometrics frequently uses time series. Data that describes a number of different individuals over time periods are pooled da ta. Pooled data describe each of a number of individuals which is common in cross-sectional data, but al so describe a single individual through time which mirror time series data (Dielman 1983). In summary, ordinary least squares (OLS) ar e the most common type of regression, but it is not always the most efficient. When multiple equations are to be estimated with explanatory variables differing across equati ons and the equation error terms being contemporaneously correlated, the OLS is not optimal. In this cas e, the seemingly unrelated regression is more appropriate. It is important th at the equations be related th rough cross-equation correlation of error terms. If there is no crossequation correlation of the errors there is no gain in efficiency. The SUR process can be considered as building a single model out of all the equations combined (Dielman 1983). 42

PAGE 43

CHAPTER 5 RESULTS Model (1) described in Chapter 4 was estima ted by the seemingly unrelated regression method. The SAS program for seem ingly unrelated regressions was used. Results are discussed in this chapter. The parameter estimates, pr ovided in Table 5-1, are discussed first. Most of the coefficients are st atistically significant at the = 0.10 level, with the exception of the second seasonality variable and th e interaction of advertising and percent Asian. The seasonality cosine variable is probably overs hadowed by the seasonality sine variable. That is not of particular concern because the sine vari able accounts for the seasonality and that is the main concern. The statistical significance prov ides evidence supporting the model. Examining the income, price, and cross pric e parameters gives us some background on the commodity. First, the parameters are exam ined under the assumption that there are no Black, Hispanic or Asian populati ons in a city, i.e., z1=0, z2=0 and z3=0. The coefficient for income is 0.17. This means that for every one percent increa se in income the demand for orange juice will rise 0.17 percent. The parameter estimate for price is -1.00. This illustrates that orange juice is a unit elastic commodity. For every one percent increase in price the demand for OJ will decline by one percent. The price for substitutes has a coefficient of 0.85. For every one percent increase in price of orange juice substitutes the de mand for orange juice rises 0.85 percent. The results for the three demographic variable s indicate that orange juice demand shifts across the ethnic groups studied. Consider the shifts in the intercept of equation (1) involving coefficients B0, B1, B2 and B3. .The cities with the highest pe rcentage of people who are Black and Hispanic have less demand for orange juice. The opposite is true for the cities with a large percentage of Asians. The coefficient for percen t Black and percent Hispanic are -0.58 and -0.32 respectively. This reveals that for every one perc ent increases in percent of the population that is 43

PAGE 44

Black the demand for orange ju ice decreases 0.58 pe rcent. For the Hispanic population it decreases 0.32 percent. As the Asian population rises one percent the demand for orange juice increases 0.40 percent. The last variable to examine is the advertising variable. Also, consider the case when there is no Black, Hispanic or As ian populations in a city. The co efficient for this variable is .0006 which means that if one store has a feature advertisement for orange juice and another one does not the store with the adver tisement will sell six percent mo re orange juice than the other store. That is a fairly substantial increase. The interaction terms examine the change in different coefficients when the demographic variables are changing. This allows us to see how the demographic variables impact the income, price and advertising responses of the model. This could be helpful in deciding what type of specific marketing or sales program to have in each city. The first variable scrutinized with an inte raction is income. As reported before the income elasticity is fairly low at 0.17. However, when interacted with percent Black it rises by 0.05 for each percentage increase in the Black po pulation. When interacted with percent Asian the elasticity drops 0.03 and percent Hispanic it increases 0.03, again for each percentage increase in the Asian and Hispanic populations, re spectively. These coeffici ents convey the fact that demand for orange juice is more elastic with respect to income for Blacks and Hispanics, but it is less for the Asian population. The next variable interacted with the demogr aphic variables is price. This exemplifies how the price elasticity changes as the population becomes more diverse. The results demonstrate that the Asian population is the most se nsitive to prices. The pr ice elasticity rises (in absolute value) by 0.01 for each percentage in crease in the Asian population. The Black and 44

PAGE 45

Hispanic population are less sensit ive to prices as el asticity with interact ion parameters of 0.012 and 0.009, respectively. Following the interactions to price are the cr oss price interactions. All three demographic variables have similar results when interacting w ith cross prices. As the cross prices rise the cross price elasticities become lo wer. This would suggest that the cities with a more diverse population are less sensitive to the price of orange juice s ubstitutes than other cities. The last variable which has an interaction is the advertising variable. As explained earlier, the advertising is a percentage which is between zero and 100. Examining the coefficients for the interactions between adver tising and demographic vari ables illustrates that advertising increases orange juice demand more with respec t to the Black and Hispanic populations. The opposite holds true with respect to the Asian population. One of the interesting parts of this study is th e ability to examine the changing elasticities for orange juice with respect to the three demogr aphic variables. Table 5-2 demonstrates how the own price, substitute price, income, and advertising elasticities differ for 27 different demographic levels. Three levels for each ethnicity variable, its mean, minimum, and maximum, are considered. .There is 27 different combin ations for the three demographic variables and three levels. Table 5-3 separate s the income, price, and cross-price elasticities by city and also lists the promotional impact in each city. Equa tions 2, 3, 4, and 5 represent the elasticities for income, price, cross price, and advertising. The important fact that these elasticities ex emplify is whether orange juice becomes more or less elastic as there is a hi gher or lower percentage of an ethnic group in the population. The interaction terms in the model address this issue, but the elasticity chart goes into more detail. 45

PAGE 46

Reviewing each variable with respect to the ma ximum, minimum, and m ean is the first step taken. This will show if there is a patt ern for the change in elasticities. Income is the first variable examined. Figure 5-1 illustrates income elasticities for the three demographic variables. The income elasticity for the mi nimums of all minorities is 0.21. The income elasticity for the mean is 1.04. The in come elasticity for the maximums is 2.93. This would suggest that as the percen t of the population that is not Wh ite rises the income elasticity for orange juice also rises. The lowest income elasticity listed is when the Black and Hispanic variables are at the minimum and the Asian variable is at the maximum. That elasticity is -0.27. The highest income elasticity is when the Black and Hispanic variables are at the maximum and the Asian variable is at the minimum. That el asticity is 3.42. These two numbers would suggest that the cities with a high Black and Hispanic population have the highes t income elasticities and vice versa. They also reveal that cities with more people of Asian descent would have lower income elasticities. The second elasticity taken is th at of orange juice own price. Figures 5-2 shows the price elasticities for each of the three demographic variables. The price elasticity for cities with the minimum amount of minorities is -0.99. The pric e elasticity for the mean is -0.80. The price elasticity for the maximum is -0.37. This would imply that as the minority population increases the price elasticity for orange juice would fall. Orange juice is the most inelastic when percent Black and Hispanic are at the maximum and percen t Asian is at the minimum. That value is 0.19. The highest price elasticity is when the percent Black and Hispanic variables are at the minimum and percent Asian is at the maximu m. That number is -1.17. These numbers would insinuate that the Asian populat ion is more price sensitive than the Black and Hispanic populations. 46

PAGE 47

The elasticity of orange jui ce substitute price is the th ird to examine. Figures 5-3 demonstrates the substitute price elasticitie s for each of the three demographic variables. The substitute price elasticity with the minimum am ount of the three minor ity variables is 0.82. The substitute price elasticity for the mean is 0.50. The substitute price elasticity for the maximum percent of minorities is -0.87. This trend indicates that as the percentage of the population that is a minority increases, the smaller the cross-price el asticity is. This elasticity switches from a positive to negative value; that is, from a substitute to a complementary relationship. The highest crossprice elasticity is 0.82, which as noted above occurs at the minimum for all three minorities. The lowest substitute price elasticity is -0.87, which occurs at the maximum for all three minorities. The final elasticity illustrated is for adve rtising. Figures 5-4 shows the advertising elasticities for the three demographi c variables. This may be the elasticity of the most interest to orange juice marketing firms. The advertisi ng elasticity for the minimum minorities is 0.07. The advertising elasticity for the mean ethnic leve ls is 0.17. The advertising elasticity for the maximum percent of minorities is 0. 37. This trend is a sign that adve rtising is more successful in cities with a high minority popul ation than other cities. The highest value for advertising elasticity is 0.42. This takes plac e when the Black and Hispanic variables are at the maximum and the Asian variable is at its minimum value. The lowest advertising elasticity is 0.02. This occurs when the Black and Hispanic variables are at their minimum and the Asian variable is at the maximum. The elasticities point out that advertising is more successful in the cities with higher Black and Hispanic populations. 47

PAGE 48

Table 5-1 Parameter results Parameter Estimate Approx Std Err t Value Approx Pr > |t| Intercept -4.60131 0.1104 -41.66 <.0001 Percent Black (z1) -0.58528 0.00621 -94.3 <.0001 Percent Asian (z2) 0.405496 0.0166 24.35 <.0001 Percent Hispanic (z3) -0.32089 0.00552 -58.09 <.0001 Income (x) 0.170875 0.0116 14.76 <.0001 Interaction (Income and Black) 0.056983 0.000664 85.86 <.0001 Interaction (Income and Asian) -0.03165 0.00193 -16.4 <.0001 Interaction (Income and Hispanic) 0.032973 0.000603 54.65 <.0001 Price (p) -0.99938 0.0145 -68.87 <.0001 Interaction (Price and Black) 0.012473 0.000728 17.13 <.0001 Interaction (Price and Asian) -0.01169 0.00172 -6.8 <.0001 Interaction (Price and Hispanic) 0.009358 0.000545 17.16 <.0001 Cross Price (ps) 0.854982 0.0147 58.36 <.0001 Interaction (Cross Price and Black) -0.00665 0.000848 -7.84 <.0001 Interaction (Cross Price and Asian) -0.04661 0.00241 -19.31 <.0001 Interaction (Cross Price and Hispanic) -0.01817 0.000617 -29.45 <.0001 Advertising (a) 0.000657 0.000078 8.42 <.0001 Interaction (Advertising and Black) 0.000084 4.71E-06 17.82 <.0001 Interaction (Advertising and Asian) -0.00003 0.000021 -1.34 0.183 Interaction (Advertising and Hispanic) 0.00002 4.09E-06 4.91 <.0001 Seasonality (Sine) (w) -0.05454 0.00206 -26.45 <.0001 Seasonality (Cosine) (w2) 0.000098 0.00214 0.05 0.9637 48

PAGE 49

Table 5-2 Elasticities by demographic Black Asian Hispanic Income OJ Price Substitute Price Advertising % of City Population OJ Demand Elasticity 11.43 0.40 0.50 0.83 -0.85 0.75 0.16 11.43 0.40 8.90 1.10 -0.77 0.59 0.17 11.43 0.40 42.80 2.22 -0.46 -0.01 0.24 11.43 2.37 0.50 0.76 -0.87 0.65 0.15 11.43 2.37 8.90 1.04 -0.80 0.50 0.17 11.43 2.37 42.80 2.15 -0.48 -0.10 0.24 11.43 15.80 0.50 0.33 -1.03 0.03 0.11 11.43 15.80 8.90 0.61 -0.95 -0.11 0.13 11.43 15.80 42.80 1.73 -0.64 -0.73 0.19 0.70 0.40 0.50 0.21 -0.99 0.82 0.07 0.70 0.40 8.90 0.49 -0.91 0.67 0.08 0.70 0.40 42.80 1.60 -0.59 0.05 0.15 0.70 2.37 0.50 0.15 -1.01 0.73 0.06 0.70 2.37 8.90 0.42 -0.93 0.57 0.08 0.70 2.37 42.80 1.54 -0.61 -0.03 0.15 0.70 15.80 0.50 -0.27 -1.17 0.10 0.02 0.70 15.80 8.90 0.01 -1.09 -0.04 0.04 0.70 15.80 42.80 1.12 -0.77 -0.66 0.10 32.60 0.40 0.50 2.03 -0.59 0.61 0.33 32.60 0.40 8.90 2.30 -0.51 0.45 0.35 32.60 0.40 42.80 3.42 -0.19 -0.15 0.42 32.60 2.37 0.50 1.97 -0.61 0.51 0.33 32.60 2.37 8.90 2.24 -0.53 0.36 0.35 32.60 2.37 42.80 3.36 -0.21 -0.24 0.41 32.60 15.80 0.50 1.54 -0.77 -0.10 0.29 32.60 15.80 8.90 1.82 -0.69 -0.25 0.30 32.60 15.80 42.80 2.93 -0.37 -0.87 0.37 49

PAGE 50

Table 5-3 Elasticities, by city. City Income Elasticity Own-Price Elasticity Cross-Price Elasticity Promo Impact Albany 0.56 -0.91 0.70 11.8% Atlanta 1.75 -0.65 0.53 26.7% Baltimore 1.55 -0.71 0.54 28.0% Birmingham 1.60 -0.69 0.61 17.4% Boston 0.49 -0.93 0.59 10.3% Buffalo/Rochester 0.75 -0.87 0.67 12.6% Charlotte 1.32 -0.74 0.60 23.1% Chicago 1.50 -0.70 0.31 22.9% Cincinnati 0.81 -0.86 0.70 14.6% Cleveland 0.92 -0.83 0.72 16.7% Columbus 0.74 -0.87 0.72 14.0% Dallas/Ft. Worth 1.44 -0.69 0.38 18.7% Denver 0.85 -0.82 0.49 12.3% Des Moines 0.29 -0.97 0.67 6.1% Detroit 1.28 -0.75 0.65 21.2% Grand Rapids 0.62 -0.89 0.71 12.7% Hartford/ New Haven 0.81 -0.85 0.54 14.9% Houston 1.87 -0.58 0.22 23.1% Indianapolis 0.70 -0.88 0.72 13.2% Jacksonville 1.34 -0.75 0.53 22.9% Kansas City 0.82 -0.85 0.65 13.5% Las Vegas 1.14 -0.77 0.32 16.2% Little Rock 1.05 -0.81 0.67 14.7% Los Angeles 1.47 -0.68 -0.12 16.8% Louisville 0.71 -0.88 0.70 13.1% Memphis 2.07 -0.59 0.59 28.6% Miami 2.08 -0.52 0.17 26.5% Milwaukee 0.92 -0.83 0.66 16.8% Minneapolis 0.39 -0.95 0.66 9.0% Nashville 0.83 -0.85 0.70 13.3% New York 1.55 -0.69 0.22 23.6% Oklahoma City/Tulsa 0.66 -0.89 0.65 10.8% Omaha 0.54 -0.91 0.69 8.7% Orlando 1.17 -0.76 0.55 18.3% Philadelphia 1.12 -0.79 0.54 20.2% Phoenix 0.95 -0.79 0.34 11.6% Pittsburgh 0.46 -0.93 0.76 11.1% Portland 0.35 -0.96 0.58 7.7% Raleigh/Durham 1.65 -0.67 0.56 27.7% Richmond/Norfolk 1.75 -0.66 0.56 29.9% Sacramento 0.85 -0.85 0.18 12.2% St. Louis 1.02 -0.81 0.67 18.3% 50

PAGE 51

Table 5-3 Continued Salt Lake City/Boise 0.40 -0.94 0.62 8.2% San Antonio 1.63 -0.61 0.09 18.4% San Diego 0.94 -0.82 0.11 12.6% San Francisco 0.54 -0.97 -0.19 8.9% Seattle 0.35 -0.98 0.45 8.3% Syracuse 0.45 -0.93 0.73 9.7% Tampa 0.93 -0.81 0.59 13.9% Washington D.C. 1.40 -0.74 0.41 24.9% West Texas 1.77 -0.56 0.01 11.6% 51

PAGE 52

0 0.5 1 1.5 2 2.5 Max Mean Min Percent BlackIncome Elasticity A 0 0.2 0.4 0.6 0.8 1 1.2 Max Mean Min Percent AsianIncome Elasticity B 0 0.5 1 1.5 2 2.5 Max Mean Min Percent HispanicIncome Elasticity C Figure 5-1 Income elasticity with respect to percent A) Black B) Asian C) Hispanic 52

PAGE 53

-1 -0.9 -0.8 -0.7 -0.6 -0.5 -0.4 -0.3 -0.2 -0.1 0 Max Mean Min Percent BlackPrice Elasticity A -1.2 -1 -0.8 -0.6 -0.4 -0.2 0 Max Mean Min Percent AsianPrice Elasticity B -1 -0.9 -0.8 -0.7 -0.6 -0.5 -0.4 -0.3 -0.2 -0.1 0 Max Mean Min Percent HispanicPrice Elasticity C Figure 5-2 Price elasticity with respect to percent A) Black B) Asian C) Hispanic 53

PAGE 54

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 Max Mean Min Percent BlackCross Price Elasticity A -0.2 -0.1 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 Max Mean Min Percent AsianCross Price Elasticity B -0.2 -0.1 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 Max Mean Min Percent HispanicCross Price Elasticity C Figure 5-3 Cross price elasticity with respect to percent A) Black B) Asian C) Hispanic 54

PAGE 55

0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 Max Mean Min Percent BlackAdvertising Elasticity A 0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.2 Max Mean Min Percent AsianAdvertising Elasticity B 0 0.05 0.1 0.15 0.2 0.25 0.3 Max Mean Min Percent HispanicAdvertising Elasticity C Figure 5-4 Advertising el asticity with respect to percent A) Black B) Asian C) Hispanic 55

PAGE 56

CHAPTER 6 CONCLUSION Conclusion The results of this study should be important to orange juice marketing firms in the United States. As America becomes more diverse th e tastes and preferences in the country will change with the different demographics. As th e results demonstrate, demand for orange juice does change with the different demographic groups When there are a larger percentage of Black and Hispanic citizens in a city the intercept fo r the demand for orange ju ice tends to decline. When the percentage of Asians in the populat ion increases the intercept for the demand for orange juice rises. This exemplifies the different tastes and pref erences throughout these ethnicities. If orange juice firms keep the same marketing tactics, the demand for OJ will decline, considering just the inte rcept shifts and ignoring the impacts of the ethnic variables on the price, income and advertising effects. This is the result of a faster gr owth in the Hispanic population than that for the other demographic variables. The interactions between demographics and price and also between demographics and advertising should be of extreme interest. The pr ice interactions demonstrate which ethnicity is more sensitive to prices and vice versa. This repr esents an opportunity fo r orange juice retailers with respect to pricing their products through discounts, as well as possibly coupons. Prior studies involving coupons have c oncluded that marketing with coupons raises the demand for commodities. Since the Asian community responds to price changes the most, marketing with coupons might be most effective in places with a high Asian populat ion. The advertising interactions point out that adve rtising has a different effect on each ethnicity. This could help retailers decide how much to advertise in each city. 56

PAGE 57

One marketing statistic that is very evident is that the Black population is more sensitive to a change in the advertising va riable than the other demogra phics studied. Another statistic involving the Black variable is th e interaction between percent Black and price. This relationship indicates that the African American population is less sens itive to a change in price. That is, in cities with a large Black populati on one can expect a relatively sma ll gain in orange juice sales by lowering price. On the other hand, advertising has a relatively large impact in cities with a large percentage of Blacks. The exact opposite of the prior paragraph is true for the Asian population. The amount of advertising changes the demand for orange juice less for Asians than other demographics. On the other hand, a slight rise in pri ce would cause orange juice dema nd to drop more significantly in places with a high Asian population than other ethni cities. This would suggest that pricing is more important in stores with more Asian tra ffic. Marketing by coupons or lower price would probably be more successful in these areas. The Hispanic population is of the most interest be cause that is the ethnicity that is currently growing the fastest. The variable representing Hispanics is very similar to the Black variable. Hispanics are less sensitive to a change in price and have a greater reaction to a change in the advertising variable. Although the demand change for th e Hispanic variable is not as large as that of the Black variable it is still a significant chan ge. The results of this study may be interesting to grocery retailers. It may be useful for the Fl orida Department of Citrus at national and regional levels. Retailers could use this info rmation on a store by store basis. Implications and Topics for Additional Research The majority of the retail industry knows th e primary demographics of each store they have open. These stores may vary in demographics throughout different citi es. If one part of a city has a higher African American population than another the fi rm could make more room for 57

PAGE 58

displays and use fewer coupons. On the other ha nd, in stores with a high population of Asians the store could offer more coupons for orange ju ice at the door. This w ould be a good advertising experiment for any firm trying to expand the demand for orange juice. This study might be useful for firms when d eciding about different marketing decisions. A couple steps should be taken in the future th at may have large impacts in regards to demographics and OJ demand. Both steps involve the data set. The data used in this thesis is accurate, but it does not include Wal-Mart super stor es. Sales in these giant grocery stores have been trending upwards and many different demo graphics shop at WalMart. The other step involves the advertising variable. This thesis conc ludes that the ethnicities respond to advertising differently. However, more detail could be useful with respect to advertising. The advertising variable is represented by the percentage of all commodity volum e with a prominent orange juice feature. The problem is that a prominent orange juice feature does not specify what type of advertising is being used. There is a large di fference between brand and generic advertising. Marketers need to know what type of advertisin g the different ethnici ties respond to, and also which is more successful. In conclusion, the demographics of a city do affect the amount of orange juice demanded. Those demographics also respond to changes in pr ice, cross price, income, and advertising at different levels. This information would suggest that different marketing strategies be put in place where different demographics reside. Growing the demand for orange juice is the main goal of the OJ industry and this in formation may be very helpful. 58

PAGE 59

Table 6-1 Simulated impact on OJ demand of a 10 % OJ Price discount and a 30 percentage point increase in the advertising variable City Impact of price discount Impact of price promo Total impact % Change in OJ Demand Albany 9.1 3.6 12.7 Atlanta 6.5 8.3 14.8 Baltimore 7.1 7.8 14.9 Birmingham 6.9 8.0 14.9 Boston 9.3 3.2 12.5 Buffalo/Rochester 8.7 4.4 13.1 Charlotte 7.4 6.7 14.1 Chicago 7.0 6.7 13.7 Cincinnati 8.6 4.8 13.4 Cleveland 8.3 5.2 13.5 Columbus 8.7 4.5 13.2 Dallas/Ft. Worth 6.9 6.1 13.0 Denver 8.2 3.9 12.1 Des Moines 9.7 2.6 12.3 Detroit 7.5 6.6 14.1 Grand Rapids 8.9 3.8 12.7 Hartford/New Haven 8.5 4.3 12.8 Houston 5.8 7.3 13.1 Indianapolis 8.8 4.3 13.0 Jacksonville 7.5 6.8 14.3 Kansas City 8.5 4.6 13.1 Las Vegas 7.7 4.9 12.5 Little Rock 8.1 5.6 13.7 Los Angeles 6.8 5.2 12.0 Louisville 8.8 4.3 13.1 Memphis 5.9 9.9 15.8 Miami 5.2 7.6 12.8 Milwaukee 8.3 4.9 13.2 Minneapolis 9.5 3.0 12.5 Nashville 8.5 4.8 13.3 New York 6.9 6.6 13.4 Oklahoma City/Tulsa 8.9 4.0 12.8 Omaha 9.1 3.4 12.6 Orlando 7.6 5.5 13.0 Philadelphia 7.9 5.8 13.7 Phoenix 7.9 3.8 11.8 Pittsburgh 9.3 3.4 12.8 Portland 9.6 2.6 12.2 Raleigh/Durham 6.7 7.9 14.7 Richmond/Norfolk 6.6 8.6 15.2 Sacramento 8.5 3.7 12.2 59

PAGE 60

Table 6-1 Continued Table 6-1 Continued St. Louis 8.1 5.7 13.8 Salt Lake City/Boise 9.4 2.6 12.0 San Antonio 6.1 5.5 11.6 San Diego 8.2 3.8 12.0 San Francisco 9.7 2.8 12.5 Seattle 9.8 2.7 12.5 Syracuse 9.3 3.2 12.6 Tampa 8.1 4.6 12.7 Washington D.C. 7.4 6.9 14.3 West Texas 5.6 5.4 11.0 60

PAGE 61

LIST OF REFERENCES Armah Jr., Bernard, Epperson, James. Export Dema nd for US Orange Juice: Impacts of US Export Promotion Programs. Agribusiness 13, no.1 (1997): 1-10. Binkley, James, Canning, Patrick, Dooley, Ryan, Eales, James. Consolidated Markets, Brand Competition, and Orange Juice Prices. Agriculture Information Bulletin 747-06 (2002). Blisard, Noel, Variyam, Jaychandran, Cromartie John. Food Expenditures by U.S. Households: Looking Ahead to 2020. Agricultural Economic Report 821(2003). Brown, Mark, Lee, Jonq-Ying. Coupon Redemption and the Demand for Frozen Concentrated Orange Juice: A Switching Regression Analysis. American Agricultural Economics Association (1985): 647-653. Brown, Mark. The Impact of Generic Advertising and the Free Rider Problem: A Look at Orange Juice Market and Imports. Agribusiness 12, no. 4 (1996): 309-316. Brown, Mark, Lee, Jonq-Ying. A Measurement of the Quality of Orange-Juice Consumption. Agrbusiness 16, no. 3(2000): 321-332. Dielman, Terry. Pooled Cross-Sectional and Time Series Data: A Survey of Current Statistical Methodology. The American Statistician 37 No. 2(1983): 111-122. Dooley, Ryan, Eales, James, Binkley, James (2000) The Demand for Nutritionally-Enhanced Varieties and Implications for Food Product Competition: The Case of Orange Juice. Tampa, Florida. Gibson, Campbell, Jung, Kay (2005) Historical Ce nsus Statistics on Population Totals by Race, 1790 to 1990, and by Hispanic Origin, 1970 to 1990, for Large Cities and Other Urban Places in the United States, ed. U.S. Census Bureau. Gould, Brian. Factors Affecting Demand for Fo od Items, University of Wisconsin-Madison. Heien, Dale, Pompelli, Greg. The Demand for Beef Products: Cross-Section Estimation of Demographic and Economic Effects. Western Journal of Agricultural Economics no. 13(1) (1988): 37-44. Johnston, J. Econometric Methods. 2 ed: McGraw-Hill, 1963. Kmenta, Jan. Elements of Econometrics 2 ed: Macmillan Publishing Company, 1986. Love, Leigh Ann, Sterns, James, Spreen, Thomas Wysocki, Allen (2006 ) Changing Patterns of Orange Juice Consumption in the Sout hern United States. Orlando, Florida. Maddala, G.S. Econometrics Edited by J.S. Dietrich and Michael Gardner: McGraw-Hill, 1977. Piggott, Nicholas. (2006) Consumer Price Formati on with Demographic Tran slating. Gold Coast, Australia. 61

PAGE 62

62 Spreen, Thomas, Barber Jr., R obert, Brown, Mark, Hodges, Alan, Malugen, Jordan, Mulkey, W. David, Muraro, Ronald, Norberg, Robert, Ra hmani, Mohammad, Roka, Fritz, Rouse, Robert. An Economic Assessment of the Fu ture Prospects for the Florida Citrus Industry. University of Florida, March 16, 2006. Ward, Ronald. Generic Promotions of Florida Citrus. University of Florida, April 8, 2005. Ward, Ronald, Davis, James. A Pooled Cross-Section Time Series Model of Coupon Promotions. American Journal for Agricultural Economics (1978).

PAGE 63

BIOGRAPHICAL SKETCH Andrew Davis is a part of the Food and Res ource Economics Master of Science program. Andrew was born in Fort Myers, Florida, and li ved there his whole life be fore college. Andrew attended Fort Myers High School and achieved ov er a 4.0 GPA. Andrew also was a part of the baseball team and NHS. After high school Andrew attended Lake Sumter Community College to play on the baseball team and took classes toward his Economics major. Andrew spent two years at Lake Sumter and was granted his AA degree. Next, he moved on to the University of Florida. He decided to major in food and resource economi cs and also played baseball for the team in 2005. Andrew graduated in 2006 with his bachelo rs degree in food and resource economics. Andrew was the graduate school organization treasurer his first y ear in graduate school. After graduation Andrew will accept a job from Edison National Bank in Fort Myers, Florida.