UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The competitive structure of the coffee industry and the impact of fair trade coffee Kojima, Hiroaki 2003

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_2003-0435.pdf [ 4.03MB ]
Metadata
JSON: 831-1.0090951.json
JSON-LD: 831-1.0090951-ld.json
RDF/XML (Pretty): 831-1.0090951-rdf.xml
RDF/JSON: 831-1.0090951-rdf.json
Turtle: 831-1.0090951-turtle.txt
N-Triples: 831-1.0090951-rdf-ntriples.txt
Original Record: 831-1.0090951-source.json
Full Text
831-1.0090951-fulltext.txt
Citation
831-1.0090951.ris

Full Text

THE_COMPETITIVE STRUCTURE OF THE COFFEE INDUSTRY A N D THE IMPACT OF FAIR TRADE COFFEE By HIROAKI KOJIMA Bachelor of Economics, Tokyo University, Japan, 1996 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE in THE F A C U L T Y OF G R A D U A T E STUDIES Agricultural Economics We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH C O L U M B I A July 2003 © Hiroaki Kojima, 2003 UBC Rare Books and Special Collections - Thesis Authorisation Form Page 1 of 1 In presenting t h i s . t h e s i s i n p a r t i a l f u l f i l m e n t of the requirements for an advanced degree at the Univ e r s i t y of B r i t i s h Columbia, I agree that the Library s h a l l make i t f r e e l y a v a i l a b l e f o r reference and study. I further agree that permission for extensive copying of th i s thesis f o r sch o l a r l y purposes may be granted by the head of my department or by his or her representatives. It i s understood that copying or pu b l i c a t i o n of t h i s thesis f o r f i n a n c i a l gain s h a l l not be allowed without my written permission. The U n i v e r s i t y of B r i t i s h Columbia , Vancouver, Canada http://www.library.ubc.ca/spcoll/thesauth.html 4/9/2003 ABSTRACT This thesis examines the current coffee market especially from the viewpoint of market structure. Though several previous literature suggest the presence of market power by large roasters or distributors in consuming countries, none of them showed any models or empirical analyses to prove this. Thus, the main research question of this thesis is to examine whether or not the distribution system in consuming countries has market power within the coffee market. To answer this question, the Farm-Retail Price model developed by Wohlgenant (1989) and Holloway (1991) (the Holloway model) was applied to the Canadian coffee sector. The advantage of this Holloway model is that linear restrictions imposed on the parameters in three-equation model can be used to test whether or not market sectors of agricultural commodities are competitive. The Canadian coffee market was picked up as an example because the coffee market inside Canada is not so large and it might be easier for several companies to obtain market power. Test results for market competitiveness, using Canadian data, provide evidence which implies that the Canadian coffee market might be perfectly competitive at the national level. Nevertheless, as regression results show a relatively poor fit to the data, these test results could be a rejection of the model itself. These test results indicate that the direct application of the Holloway model might have several problems in case of the Canadian coffee market. For example, the import price ii might need to be treated as exogenous and a time lag might be necessary to reach the market clearing condition. Considering that the Holloway model to examine perfect competition does not incorporate these elements in the model, the Holloway model seems to have certain limitations. Then, the findings of empirical analyses are used to examine the meaning and impact of FT coffee. According to several literature, FT coffee scheme appears to be based on the idea that large roasters in consuming countries are distorting the world coffee market. However, regression results weakly indicate that the Canadian coffee market is close to competitive. These findings suggest that FT coffee scheme may be built on an incorrect premise, at least in Canada. Since the willingness to pay for FT coffee seems higher than that for non-FT coffee, FT coffee is expected to have an effect to shift the retail demand curve rightward. It is considered that this effect leads to an increase in the equilibrium price and quantity. Nevertheless, as the market share of FT coffee among the total coffee is very small, this impact to increase coffee prices will be very limited. After all, the introduction of FT coffee cannot rectify main causes of the low and unstable world prices such as an oversupply of coffee and the cyclical fluctuation of the world coffee production. Considering also the market scale of FT coffee, it can be concluded that FT coffee does not have a large impact on the world coffee market. iii TABLE OF CONTENTS ABSTRACT ii TABLE OF CONTENTS iv LIST OF TABLES vi LIST OF FIGURES vii ACKNOWLEDGEMENTS viii 1. INTRODUCTION 1 1.1 P R O B L E M S T A T E M E N T A N D RESEARCH QUESTIONS 2 1.2 PURPOSE OF T H E STUDY 3 1.2 ORGANIZATION OF T H E STUDY 4 2. OVERVIEW OF THE CURRENT WORLD COFFEE M A R K E T AND THE EMERGENCE OF FAIR TRADE COFFEE 6 2.1 OVERVIEW OF T H E C U R R E N T W O R L D C O F F E E M A R K E T 6 2.2 HISTORY A N D E M E R G E N C E OF FAIR T R A D E C O F F E E 9 2.3 F L O CERTIFIED AND SELF-LABELED FAIR TRADE COFFEE 10 2.4 T H E F L O REGISTRATION STANDARD 11 2.5 A L L O C A T I O N OF T H E FAIR T R A D E PREMIUM 14 2.6 OVERVIEW OF T H E PRESENT FAIR T R A D E C O F F E E M A R K E T 15 2.6.1 Supply of Fair Trade Coffee 15 2.6.2 Demand for Fair Trade Coffee 16 2.7 SUMMARY 17 3. COFFEE M A R K E T STRUCTURE AND THE DISTRIBUTION OF MARGINS WITHIN THE COFFEE M A R K E T 18 3.1 C O F F E E M A R K E T STRUCTURE 18 3.1.1 Market Structure in Producing Countries 18 3.1.2 International Coffee Market 21 3.1.3 Market Structure in Consuming Countries 22 3.2 DIFFERENCE BETWEEN ARABICA AND ROBUSTA C O F F E E 24 3.3 DIFFERENCE OF M A R K E T STRUCTURE BETWEEN F T A N D N O N - F T C O F F E E 25 3.4 DISTRIBUTION OF MARGINS WITHIN T H E COFFEE M A R K E T 26 3.5 S U M M A R Y 28 4. EMPIRICAL TESTING OF THE PRESENCE OF MARKET POWER WITHIN CONSUMING COUNTRIES 29 4.1 EXAMINING M A R K E T POWER IN T H E DISTRIBUTION S Y T E M WITHIN CONSUMING COUNTRIES ...29 4.2 CONCEPTUAL M O D E L S TO E X A M I N E M A R K E T COMPETITIVENESS 31 4.3 M O D E L A N D M E T H O D O L O G Y 35 4.4 D A T A 41 4.5 RESULTS OF T H E REGRESSION 46 4.5.1 Results of the Unrestricted Holloway Model 46 4.5.2 Results of the Restricted Holloway Model 48 4.5.3 Tests for Competitiveness of the Market 49 4.6 S U M M A R Y 53 5. IMPACTS OF FAIR TRADE COFFEE ON THE COFFEE MARKET AND RELATED PROBLEMS 54 5.1 RESULTS OF T H E EMPIRICAL ANALYSIS AND F T C O F F E E 54 5.2 IMPACTS OF F T ON T H E C O F F E E M A R K E T IN CONSUMING COUNTRIES 55 i v 5.3 E F F E C T OF T H E ADOPTION OF THE M I N I M U M PRICE S C H E M E 5 7 5.4 S M A L L IMPACT OF F T C O F F E E ON T H E W O R L D C O F F E E M A R K E T 5 8 5.5 SUMMARY 5 9 6. S U M M A R Y A N D C O N C L U S I O N S 61 6.1 CONCLUSIONS A N D C O M M E N T S 61 6.2 LIMITATIONS OF THIS S T U D Y AND RECOMMENDATIONS FOR F U T U R E S T U D Y 6 3 R E F E R E N C E S 65 A P P E N D I X 1: N E C E S S A R Y A N D S U F F I C I E N T C O N D I T I O N S F O R P E R F E C T C O M P E T I T I O N O F T H E M A R K E T 70 A P P E N D I X 2: C H A N G E S I N E A C H V A R I A B L E 77 A P P E N D I X 3: W E I G H T S F O R M A R K E T I N G I N P U T S C O S T S 80 A P P E N D I X 4: R E G R E S S T I O N R E S U L T S U S I N G N O M I N A L V A L U E S 82 A P P E N D I X 5: R E G R E S S T I O N R E S U L T S U S I N G D E F L A T E D V A L U E S 91 v LIST OF TABLES T A B L E 2-1: MINIMUM PRICES OF FAIR TRADE COFFEE SET BY FLO 14 TABLE 2-2: PRODUCTION OF FT COFFEE AND PERCENTAGE OF TOTAL COFFEE PRODUCTION (1996/97) 15 T A B L E 2-3: SALES VOLUMES OF FT COFFEE PER YEAR (1997-2001) 16 T A B L E 3-1: M A R K E T SHARE OF THE INTERNATIONAL COFFEE TRADE COMPANIES (1998) 21 T A B L E 3-2: M A R K E T SHARE OF COFFEE ROASTING AND MANUFACTURING COMPANIES (1998) 22 TABLE 4-1: SIZE AND SIGNIFICANCE OF THE CANADIAN TEA AND COFFEE PROCESSING SECTOR 30 T A B L E 4-2: DEFINITIONS AND SUMMARY STATISTICS FOR THE VARIABLES IN THIS MODEL 45 T A B L E 4-3: INCOME ELASTICITY IN PREVIOUS PAPERS 46 T A B L E 4-4: ESTIMATES OF THE UNRESTRICTED HOLLOWAY MODEL 48 T A B L E 4-5: ESTIMATES OF THE RESTRICTED H O L L O W A Y MODEL 49 T A B L E 4-6: RESULTS OF F-TESTS TO EXAMINE MARKET COMPETITIVENESS 50 T A B L E CI: COSTS OF COFFEE AND TEA MANUFACTURING INDUSTRY IN THE US (1997) 81 T A B L E D-l: ESTIMATES OF THE UNRESTRICTED HOLLOWAY MODEL USING NOMINAL VALUES 82 T A B L E D-2: ESTIMATES OF THE RESTRICTED HOLLOWAY MODEL USING NOMINAL VALUES 86 TABLE D-3: RESULTS OF F-TESTS TO EXAMINE MARKET COMPETITIVENESS USING NOMINAL VALUES 89 T A B L E E - l : ESTIMATES OF THE UNRESTRICTED HOLLOWAY MODEL USING DEFLATED VALUES 91 T A B L E E-2: ESTIMATES OF THE RESTRICTED HOLLOWAY MODEL USING DEFLATED VALUES 95 T A B L E E-3: RESULTS OF F-TESTS TO EXAMINE MARKET COMPETITIVENESS USING DEFLATED VALUES 99 V I LIST OF FIGURES FIGURE 2-1: COFFEE MONTHLY PRICE FLUCTUATIONS (1982-2001) 7 FIGURE 3-1: COFFEE M A R K E T IN INDONESIA (SOUTH SUMATRA) 19 FIGURE 3-2: SHARE OF FARM-GATE PRICES AMONG EX-DOCK PRICES IN INDONESIA (ROBUSTA) 20 FIGURE 3-3: COFFEE MARKET STRUCTURE IN THE US 23 FIGURE 3-4: COMPARISON BETWEEN CONVENTIONAL AND FT COFFEE CHANNELS ....25 FIGURE 3-5: DISTRIBUTION OF RETAIL COFFEE PRICE (%) (1975-94) 26 FIGURE 4-1: STRUCTURE OF THE FARM-RETAIL PRICE MODEL 34 FIGURE 4-2: STRUCTURE OF MODEL USED IN THIS THESIS 34 FIGURE 4-3: MONTHLY PRICE FLUCTUATIONS OF THREE ARABICA COFFEE (1982-2001) 45 FIGURE 5-1: COFFEE M A R K E T AFTER THE INTRODUCTION OF FT COFFEE 56 FIGURE 5-2: EFFECT OF THE ADOPTION OF THE MINIMUM PRICE FOR FT COFFEE 57 FIGURE BI: PERCENT CHANGES IN MARKET MARGIN 77 FIGURE B2: PERCENT CHANGES IN RETAIL PRICE 77 FIGURE B3: PERCENT CHANGES IN IMPORT PRICE 78 FIGURE B4: PERCENT CHANGES IN MARKETING INPUT COSTS 78 FIGURE B5: PERCENT CHANGES IN RETAIL DEMAND SHIFTER 79 FIGURE B6: PERCENT CHANGES IN IMPORT SUPPLY 79 V l l ACKNOWLEDGEMENTS The completion of this thesis was supported by many people beside the author, and they all deserve recognition. Special thanks go to my supervisor, Dr. Richard Barichello, at the University of the British Columbia, who provided deep insight from the very early stage for choosing topics to the end of the research. Dr. Timothy Beatty gave me useful comments especially from the viewpoint of econometrics, and his comments caused me to rethink, rewrite, and upgrade this paper. Dr. Hiroshi Ohashi provided me many invaluable comments and assistance from different aspects. I also thank to Dr. Sumaila and Dr. Vercammen for their accepting to be an examiner and an examination chair of my Oral Examination Committee. My colleagues in the Department of Food Resource Economics supported me mentally and gave me many useful suggestions during these two years. Especially, I would like to thank to Sharan and Glenn, who checked my English grammar. The Japanese Ministry of Agriculture, Forestry and Fisheries gave me the chance to study in Canada and provided support for my stay in Canada. viii 1. INTRODUCTION The world coffee market is large — coffee was ranked as the second most important export commodity next to petroleum, in terms of export earnings for developing countries in 1998 (Akiyama et al. 2001). About five million metric tons of coffee was exported around the world, with an estimated value of 14 billion US dollars in 1998 (FAO 1998). For many coffee-producing countries, especially those in Sub-Saharan Africa and Central America, coffee brings not only foreign exchange but also an opportunity of employment in local regions (Akiyama et al. 2001). Coffee is produced in about 50 developing countries in Africa, Central and South America, and South East Asia, and is mainly consumed in developed countries such as the United States, Canada, European countries, and Japan. There exist lots of participants in the world coffee market from the local farmers in producing countries to consumers in consuming countries, and these participants are obtaining some benefits from the coffee trade. Considering the market scale of the world coffee trade and the large influence of coffee trade on many countries, examining deeply the market structure of the coffee trade has a significant meaning. Especially, to examine whether or not certain participants in the coffee market have market power seems important to understand the market structure. Recently, in this coffee market, a new kind of product called "Fair Trade (FT)" coffee has emerged with a sales volume that has been growing rapidly: 12.3% in 2001 (FLO 2002). 1 Although the growth rate of FT coffee is high, the market share of FT coffee in the total coffee market is presently small. 1.1 Problem Statement and Research Questions This thesis focuses on the world coffee market in terms of the market structure, and investigates the problems of the world coffee market in relation to the market structure. In particular, this thesis attempts to identify the parts of the coffee market that may have market power. Previous literature, such as Talbot (1997), Mendoza (2000), Ponte (2001), and Waridel (2002) mention the presence of market power by large coffee roasting companies, such as General Food and Nestle, and their undeniable influence on the world coffee market. Some of these articles show data that suggest the existence of this market power: Ponte (2001) reports that some large roasting companies have quite high market share in the world coffee roasting and manufacturing market. However, none of this literature shows any models or empirical analyses to prove the existence of this kind of market power. Considering the scale of the coffee market and many participants in the market, market power of large companies in consuming countries could have a significant influence. Thus, the main content of this thesis is to examine the presence of market power of large coffee roasting or distributing companies in consuming countries. Specifically, the following question is asked: 2 Do coffee roasters or distributors in consuming countries have market power within the coffee market? To answer this question, this thesis focuses on the Canadian coffee sector as an example of consuming country. Since the coffee market in Canada is not so large, it may be easier for several firms to have market power than in larger markets like the US or the EU. In addition, this thesis examines theoretically how the introduction of FT coffee can affect the coffee market to capture a new aspect of the current world coffee market. Especially, the impact of FT coffee on the coffee market and problems of FT coffee scheme will be analyzed. Here the following question arises: Does the introduction of Fair Trade Coffee have a large impact on the world coffee market? Though some previous literature such as Jones and Bayley (2000) and Ronchi (2002) report results of case studies in Tanzania and Costa Rica, respectively, and analyze impacts of FT coffee on some farmers' organizations and each farmer, the impact of FT coffee among the total coffee market has not been examined so far. Thus, in this thesis, the role and the impact of FT coffee in the world total coffee market will be analyzed. 1.2 Purpose of the Study The main purpose of this thesis will be to formulate a model and estimate empirically whether or not a coffee market in a consuming country is competitive to test for the 3 presence of market power by large roasters or distributors in consuming countries. The econometrical tool chosen to examine market competitiveness is the "Farm-Retail Price model" that is based on work by Gardner (1975) and developed by Wohlgenant (1989) and Holloway (1991). In order to analyze the existence of this kind of market power, it is necessary to describe the situation of current world coffee market, the structure and principal participants in the I world coffee market. Thus, this thesis also provides information about what happened and is occurring in the world coffee market recently, the trade channels of coffee, and each participant in the coffee market. As for another research question about the impact of FT coffee on the world coffee market, theoretical analysis based on some previous literature will be shown. By doing this analysis, this thesis intends to clarify the situation that is occurring in the current world coffee market and the possible situation that could arise from the introduction of FT coffee. 1.3 O r g a n i z a t i o n o f t h e S t u d y This thesis is organized in the following manner: Chapter 2 provides an overview of the current world coffee market especially in terms of the price trend. A brief history of FT coffee, standards for certified FT coffee, and the outlook for the present FT coffee market are also addressed in Chapter 2 to understand the new trend in the world coffee market. 4 Chapter 3 explores the current world coffee market, especially from the viewpoint of the market structure and the margins obtained by each stage in the market. The role and situation of each participant of the coffee trade are also described to better realize the market structure. Chapter 4 describes a testing of the presence of market power in the coffee distribution system in a coffee-consuming country, Canada. At first, conceptual models and methods are introduced that examine market competitiveness in agricultural commodity markets. These empirical models are then applied to the Canadian coffee sector by using Canadian data. Test results of this empirical analysis are reported at the last part in this Chapter. In Chapter 5, the theoretical impact of the introduction of FT coffee on the world coffee market and the meaning of FT coffee in the coffee market are examined combined with findings in Chapter 4. Several problems resulting from FT coffee scheme are also shown. Finally, in Chapter 6, a description of results, conclusions, and possibilities for future research are offered. 5 2. OVERVIEW OF THE CURRENT WORLD COFFEE MARKET AND THE EMERGENCE OF FAIR TRADE COFFEE 2.1 Overview of the Current World Coffee Market Coffee is produced in about 50 developing countries in the world, mainly by small growers who have less than five hectares of coffee trees (Waridel 2002). For these developing countries, coffee is one of the most important export commodities; for example, coffee was ranked as the second most important product next to oil in terms of export earnings in 1998 (Akiyama et al. 2001). Since 1962, the principal coffee exporting countries, which account for almost 90% of the world's total production, and importing countries, joined the International Coffee Agreement (ICA). Its main purpose was to increase and stabilize world coffee prices by imposing export quotas to member countries (Akiyama et al. 2001). The first ICA was agreed upon in 1962 and had been re-negotiated several times since then, at about six-year intervals. The ICA established a target range for the member market prices and each exporting country was assigned a percentage of the global export quota (Bohman and Jarvis 1996). This agreement, specifically, the quota system, worked relatively well to keep coffee market prices within the target range (Akiyama et al. 2001). The trend of coffee prices from 1982 to 2001 is shown in Figure 2-1. During the early 1980's when the quota was imposed, the world coffee prices remained almost constant. The quota regime was suspended in early 1986, following a severe drought in Brazil that 6 increased the world coffee prices. As the world prices peaked in the spring of 1986 and decreased thereafter, the quota was re-imposed in late 1987 (Talbot 1997). 350 r s j c n ^ ^ t D r ^ o O O i C D ' ^ c N < r o ^ L r ) t o r - - c o C J i O ' ~ C O C O G O O O a O C O O O C O O ) C > 0 ) O ) O ) O ) O ) O ) O ) C 7 > O O c i c n c n a > C T ) c r ) C J i c j > o i o i a > c T ) 0 > c T ) c r ) C T j O ) C T ) o o - » - Brazilian and Other Naturals - * - Colombian Mild Arabica Robusta Other Mild Arabica Figure 2-1: Coffee Monthly Price Fluctuations (1982-2001) (Source: ICO (2003a)) In July 1989, however, this quota system collapsed and has since been suspended. According to Akiyama, this is mainly because member countries could not agree on the method to control exports to nonmember countries and to distribute quotas among members. This suspension of export quotas urged many coffee exporting countries to export large amounts of coffee from their stocks. This dumping of coffee caused a sharp decline of world coffee prices in the market and of export revenues of these exporting countries in the early 1990's to almost half the level of their revenues in the early 1980's (Akiyama et al. 2001). Furthermore, since coffee is a tree crop and coffee trees take three 7 to five years after planting to start bearing coffee and take more two years to fully produce coffee, the world coffee production shows cyclical fluctuation (Talbot 1997). Thus, the increase of coffee prices in 1986 also might have contributed to the increase of coffee planting and the decline of coffee prices in the early 1990's. In response to the 1992 failure to negotiate the new ICA, some coffee-producing countries (28 countries in May 2002) formed the Association of Coffee Producing Countries (ACPC) in 1993, aiming to achieve a balance between the world supply and demand as well as raise and stabilize coffee prices (CAFE DE COSTA RICA 2003). At present, the ACPC is trying to restrict the supply of coffee by using a Coffee Retention Plan that requires members to withhold 20% of their exports from the market (Charveriat 2001). Nevertheless, since this agreement does not have punitive clauses and some of the major exporters, such as Vietnam and Mexico, are not joining, this association has only limited power (Akiyama et al. 2001). Figure 2-1 indicates an increase in coffee prices in the mid-1990's (from 1994 to 1997) and this increase was mainly caused by the decline of production in Brazil, the world's largest producer, owing to two frosts that occurred in that country. The Coffee Retention Plan by ACPC is also considered to have contributed to the increase of the world coffee prices since 1994 though it possessed the only restricted power (Talbot 1997). The production in Brazil recovered by the late-1990's and has again led to the decrease in the world price for coffee since 1998 (Akiyama et al. 2001). Also contributing to the 8 increased supply and the decline in prices has been the large increase in production in Vietnam over the 1990's (ICO 2003b). Although the quality of Vietnam coffee is generally low, the development of roasting technologies enabled the usage of cheaper low-quality coffee and increased the demand for Vietnamese coffee (Oxfam America 2002). Furthermore, new investments in the coffee region that reflects the higher coffee prices between 1994 and 1997 can be considered to be one cause of the world price decline since 1998 as there regularly exist cyclical fluctuations in the world production of coffee. Reflecting these developments, most coffee producers in developing countries have recently been suffering from low income levels caused by low and unstable coffee prices. As Figure 2-1 shows, the price of coffee has decreased since 1998 and coffee prices have fallen to their lowest levels in 30 years (ICO 2001a). Under these circumstances, FT coffee emerged in the world coffee market and has attracted a great deal of attention as an alternative trade practice to achieve benefits to coffee producers. 2.2 History and Emergence of Fair Trade Coffee The FT coffee movement originated from a variety of European initiatives in the 1960's that attempted to achieve a more equitable trade relationship between the worlds of the North and South (Raynolds 2000). Fair Trade Organisatie, in the Netherlands, was first to import fairly traded coffee from small farmers' cooperatives in Guatemala in 1973, and, 9 in 1988, the first Fair Trade mark for coffee was used in the Netherlands under the name of Max Havelaar (EFTA 1998). Following the Netherlands, three Fair Trade labels, namely, Transfair, Max Havelaar, and Fairtrade Mark were introduced in different countries in Europe (Raynolds 2000). In 1997, Fair Trade labelers formed an international umbrella group called the Fairtrade Labelling Organizations International (FLO) to harmonize different FT standards and to create a single market (Raynolds 2000). Currently, the FLO consists of 17 members in Europe, North America, and Japan with both the FLO and its members (referred to as the National Initiatives, NI) functioning as non-profit organizations. Examples of these NIs are Max Havellar Netherlands, TransFair USA, Fairtrade Foundation UK, etc. (FLO 2003). 2.3 FLO Certified and Self-Labeled Fair Trade Coffee Two kinds of FT coffee exist in the market — certified FT coffee by the FLO and self-labeling FT coffee. With certified FT coffee, producers and traders must be registered with the FLO or NI and comply with the standards for registration. Only FLO licensees, who sell final products to consumers, are authorized to use the FT label. These licensees need to pay a license fee to each country's NI, and the fees cover all of the FLO's certification and monitoring costs, as well as each NTs marketing expenses (FLO 2002a). As of June 2002, 274 producer organizations, representing almost 400 first-level producers' organizations and about 800,000 families of coffee producers and workers from over 40 countries in Africa, Asia, Latin America, and the Caribbean were certified 10 as FT-producing organizations by the FLO. Likewise, the number of registered traders with the FLO and authorized licensees from the NIs amounted to 236 and 416, respectively (FLO 2003 a). 2.4 The FLO Registration Standard The FLO sets a unified standard for each FT-labeled product, including coffee, and FLO standards not only require fair economic conditions but that certain social and environmental conditions exist, as well (FLO 2002b). For example, the Fairtrade Standards for Coffee are defined by the following three requirements: (1) Generic FT Standards for Small Farmer's Organization; (2) Product Specific Standards for Coffee; and (3) Trade Standards for Coffee. Standards (1) and (2) are for producers and their organizations, and standard (3) applies to traders. (1) Generic Fair Trade Standards for the Small Farmer's Organization Generic FT Standards for Small Farmer's Organization are applied to every kind of FT product, including coffee. These standards are broken down into Social Development, Economic Development, Environmental Development, and standards on Labour Conditions. The FLO Social Development standard requires that the majority of members of each organization need to be a small producer and must be managed democratically to achieve social improvement for the small producers. The Economic Development standard requires producers' organizations to manage the FT premium transparently and that these 11 organizations must have experience in the commercialization of products. The third standard, Environmental Development, requires that producers must comply with both national and international legislation related to the use of pesticides, protection of natural waters, and other ecosystems, etc. The last standard on Labour Conditions requires that forced labour or the abuse of child labour is not allowed. (2) Product Specific Standards for Coffee Product Specific Standards for Coffee applies only to coffee and can be also divided into Social Development, Economic Development, and Environmental Development. As for both the Social Development and Economic Development standards, no additional standards are specific to coffee, so that the same standards as described above in the Generic Fairtrade Standards, may apply. The coffee-specific Environmental Development standard requires that the majority of members of the producers' organizations need to grow FT coffee under shade trees (either forest trees or planted trees). Producers usually decide whether or not they will adopt shade-grown coffee by considering the demand for coffee, quality, or yields. Thus, it is not reasonable to assume that satisfying the FT coffee standard is the only reason for shifting to the production of shade-grown coffee. Moreover, even if producers adopt the shade-grown method, the switch-over costs are short-term and do not affect long-run operations for the shade-grown method. 12 (3) Trade Standards for Coffee Trade standards are mainly for importers as they define requirements for the pre-financing to farmers and the pricing of products. With pre-financing, importers are to make available up to 60% of the actual contract or estimated contract value on the request of the producers' organization. This condition enables the producers' organizations to buy coffee from their members. One of the purposes of FT coffee is to achieve "fair" prices to producers. Trade standards set the rules for coffee prices paid to producers' organizations by importers. These standards require that for prices over those for New York "C" (for Arabica coffee) and London "LCE" (for Robusta coffee), a fixed premium of 5 US cents per pound should be used for FT coffee and a premium of 15 US cents per pound for certified organic FT coffee. Moreover, these standards set minimum prices paid to producers' organizations (Table 2-1). Thus, if the prices for New York "C" (for Arabica) and London "LCE" (for Robusta) falls below these minimum prices, importers have to pay at least these minimum prices to the producers' organizations, and the difference between the minimum price and New York "C" or London "LCE" price then indicates the FT premium. Thus, FT premiums appear all the time, though the amount of premium differs depending on the world prices and types of coffee. The FLO estimates that the extra benefits for coffee producers in 2001 amounted to just short of US$ 30 million (FLO 2003a). 13 Table 2-1: Minimum Prices of Fair Trade Coffee set by FLO (US cents per pound FOB port of origin) Non-Organic Non-Organic Certified Organic Certified Organic Type of coffee Central America, Mexico, Africa, Asia South America, Caribbean Area Central America, Mexico, Africa, Asia South America, Caribbean Area Washed Arabica 126 124 141 139 Non-washed Arabica 120 120 135 135 Washed Robusta 110 110 125 125 Non-washed Robusta 106 106 121 121 (Source: F L O (2002b)) 2.5 Allocation of the Fair Trade Premium The FT premium described above is not paid to small producers directly, but compensates them indirectly through the producers' organizations. As required by the FLO coffee standard, small producers need to join organizations such as cooperatives to be registered by the FLO. Thus, even if higher prices are paid to producers' organizations, the FT premium is not necessarily fully distributed to small producers by the producers' organizations, but is held by them. According to the case study by Jones and Bayley (2000), the FT price premium was not passed on to producers, but was pooled at the level of the farmers' organizations and used to fund social infrastructure or market development activities. This is, in their analysis, partly because farmers produce both FT and non-FT products and the quantity of FT products represents only a small fraction of the total quantity produced. Likewise, in Costa Rica, producers' organizations (Coocafe) allocated 70% of FT premiums to producers (Ronchi 2002). 14 2.6 Overview of the Present Fair Trade Coffee Market 2.6.1 Supply of Fair Trade Coffee In 1996/97, FT coffee was estimated to be produced by more than 433,000 producers in 240 groups (in 2000, this number was updated to 321 producer groups that belong to 169 registered associations) from 20 countries, though mainly from Latin America and Africa (Rice 2001). Table 2-2: Production of FT Coffee and Percentage of Total Coffee Production (1996/97) Production of FT Total Coffee Percentage of FT Coffee (96/97) Production (96/97) coffee (%) Country (Metric tons) (Metric tons) T a n z a n i a 25,525 45,900 55.61 % U g a n d a 18,000 257,820 6.98 % M e x i c o 16,575 319,440 5.19% C o l o m b i a 7,046 652,560 1.08% P e r u 6,218 108,360 5.74 % Domin ican Republ ic 4,795 31,140 15.40% N i c a r a g u a 4,571 47,580 9.61 % Z a i r e 3,645 N.A. N.A. G u a t e m a l a 3,385 271,440 1.25% B r a z i l 3,045 1,659,840 0.18% C o s t a R i c a 2,097 127,560 1.64% B o l i v i a 1,217 7,980 15.26% H o n d u r a s 1,118 120,240 0.93 % V e n e z u e l a . 764 72,000 1.06% E l S a l v a d o r 337 152,040 0.22 % C a m e r o o n 200 85,920 0.23 % H a i t i 180 25,740 0.70 % Total Production of Total World Coffee Total FT Coffee/ FT Coffee Production Total World Coffee 98,718 6,156,720 1.60% (Source: Rice (2001) for Production of FT Coffee and ICO (2002) for Total Coffee Production) The major producers of FT-labeled coffee are Tanzania (25,525 metric tons/year), Uganda (18,000 metric tons/year), Mexico (16,575 metric tons/year), Colombia (7,046 15 metric tons/year), and Peru (6,218 metric tons/year) in 1996/97. Total production, which is assumed to include both FLO-certified labeled coffee and self-labeled coffee, amounted to 98,718 metric tons in 1996/97 (Rice 2001). As the total production of coffee in 1996/97 was 6,156,720 metric ton (ICO 2002), FT coffee production had a 1.6% share of the total coffee production (Table 2-2). 2.6.2 Demand for Fair Trade Coffee The sales volume of registered FT coffee from 1997-2001 is listed in Table 2-3 (FLO 2002). When looking at the FT coffee sales volume, the number is seen to have increased every year; especially in 2000 and 2001, when the growth rates were extremely high at 8.46% and 12.3%, respectively. When comparing FT coffee sales volume with total coffee imports, however, the market share of FT coffee is still at a very low level, in the range of 0.25% to 0.30%, from 1997 to 2001. Table 2-3: Sales Volumes of FT Coffee per year (1997-2001) 1997 1998 1999 2000 2001 FT Coffee Sales 11,543 11,664 11,819 12,818 14,398 Volume (Metric ton) (Metric ton) (Metric ton) (Metric ton) (Metric ton) Growth rate 1.04% 1.3% 8.46% 12.3% Total Coffee 4,487,888 4,570,675 4,752,912 4,892,628 4,827,774 Imports (Metric ton) (Metric ton) (Metric ton) (Metric ton) (Metric ton) FT sales/ Total 0.26% 0.26% 0.25% 0.26% 0.30% Imports (Source: FLO ( 2 0 0 2 ) for FT Coffee Sales Volume, and ICO ( 2 0 0 1 b ) for Total Coffee Imports) The market share of FT coffee is especially high in Europe with 3% of the coffee market in all of Europe, 2.5% in Holland, 5.0% in Switzerland, 1.0% in Germany (Transfair USA 2002a), and 0.2% in Canada in 2001 (Transfair Canada 2002). One reason for the 16 high market share of FT coffee in Europe seems to be because European countries have a relatively longer history of selling FT products and the market recognition rate is higher, compared with that in North America. For example, European countries began FT coffee sales more than 10 years ago, whereas, in the US, Transfair USA launched its FT label trade only four years ago in the spring of 1999 (Transfair USA 2002b). Galarraga and Markandya (2000) analyzed consumer demand for FT coffee in the UK coffee market. At first, they estimate how much is paid for the FT and organic features of coffee, using hedonic methods, and show that, under ceteris paribus conditions, the consumer's willingness to pay for FT and organic coffee is 11.26% higher than for non-FT and non-organic coffee in the UK. This finding about the consumer's higher willingness to pay is adopted when analyzing the impact of the introduction of FT coffee in Chapter 5. 2.7 Summary Especially, since the collapse of the export quota system, an oversupply of coffee to the market and cyclical fluctuations of the world coffee production have caused world coffee prices to be low and unstable. Under these circumstances, FT coffee emerged in the world coffee market. FT production had just a 1.6% share of the total production in 1996/97. As for the consumption of FT coffee, though the sales volume has been increasing, the market share still remains very small in the whole coffee market. 17 3. COFFEE MARKET STRUCTURE AND THE DISTRIBUTION OF MARGINS WITHIN THE COFFEE MARKET 3.1 Coffee Market Structure 3.1.1 Market Structure in Producing Countries When coffee trees become mature, coffee producers harvest coffee cherries from trees, mainly by hand, and sell them to local brokers, sometimes referred to as "coyotes." As small farmers cannot produce enough coffee to export directly to consuming countries, most of them have to sell their cherries to local brokers (Waridel 2002). Local brokers generally sell to processors who are often small local entrepreneurs, though in some areas, factories owned by multinational coffee companies are engaged in coffee processing (Waridel 2002). Usually, processors dry, clean, and polish coffee beans; grading and sorting beans first by size and then by density (Kraft 2003). Processors then sell coffee beans to exporters having the role to prepare the coffee to meet the precise quality demands of importers (Waridel 2002). In Figure 3-1, the coffee market structure in South Sumatra (Indonesia) is outlined as an example of a market in producing countries. Coffee farmers harvest and then partially dry and hull the coffee themselves, and then sell their product to agents or local brokers at local markets (ICO 1998). These middlemen (or brokers) then finish hulling the coffee and do a preliminary sort, before selling the coffee to sorting factories. In the sorting factories, coffee is dried further, cleaned, polished, graded, and sorted. After that, the 18 factories sell the coffee to exporters who check the coffee quality and negotiate with foreign buyers (McStocker 1987). Farmers - Harvest crop - Dry coffee partially - Hull coffee - Transport to village broker or factory - Sell coffee Agents - Buy from farmers on behalf of factory - Provide short-term credit - Transport from farmgate to factory Brokers - Purchase coffee - Hull any unhulled coffee - Preliminary sorting - Assembly and bulking - Arrange transport to sorting factory - Sell to sorting factory * 4 Sorting Factories - Obtain credit from banks to purchase crop - Dry crop further - Clean, polish, and grade - Colour-sorting by hand - M i x export grades according to world market prices - Bulk for exports - Hold stocks, i.e. assumes price risk - Sell coffee to exporters; process coffee for exporters or export coffee Exporters - Buy coffee from sorting factories - Check that coffee meets statutory standards of quality - Arrange transport, insurance, etc. - Negotiate with foreign purchaser - Hold stocks and assume price risk Figure 3-1: Coffee Market in Indonesia (South Sumatra) (Source: McStocker (1987)) In Indonesia, the relative importance of the middlemen has declined, since improved transport facilities have made it possible for small producers to take their coffee to 19 regional markets themselves (McStocker 1987). Furthermore, according to McStocker, exporters that try to obtain more consistent coffee quality tend to buy coffee directly from farmers, and this can be considered to have contributed to the exclusion of local middlemen from the market. Reflecting these situations, farmers obtain a relatively larger share of the profits from their products. This is shown in Figure 3-2, represented by the share of farm-gate prices among ex-dock prices that are seen to rise in the early-1990's. From 1994 until 1998, the farm-gate prices secured more than 70% of the share. These figures support the idea that the coffee market has become more competitive within Indonesia. The share of farm-gate prices among ex-dock prices varies depending on the coffee market situation in each coffee-producing country, and that in Indonesia is relatively higher compared with some other countries like Tanzania or Bolivia where the share ranges from 30% to 60% during the same period (ICO 2003a). 0.9 -| Y e a r Figure 3-2 Share of farm-gate prices among ex-dock prices in Indonesia (Robusta) (Source: ICO (2003a)) 20 3.1.2 International Coffee Market At the next stage, an international broker plays a role in accessing the world market. The world coffee trade occurs through two centers: the New York Exchange which brokers Arabica coffee, and the London Exchange which brokers Robusta coffee. International brokers buy and sell on commission (without owning or handling coffee), acting as intermediaries between the exporters in the producing countries and importers in the consuming countries (Waridel 2002). In the last two decades, according to Ponte (2001), international traders have restructured, and the coffee trader market has since become more concentrated. According to his data, the top two coffee traders, Neumann and Volcafe, have 29% of the total market share with the top five companies occupying 46% of the share (1998 data) (Table 3-1). However, these numbers do not point conclusively to the existence of much market power for any of these trading firms, unless cartel arguments are practiced. Table 3-1: Market Share of the International Coffee Trade Companies (1998) International Trade Companies Market Share Neumann 16% Top 2 Volcafe 13 % 29 % Cargill 6% Top 5 Esteve 6% 46% Aron •5% Man 4% Dreyfus 3 % Mitsubishi 3 % Others 44 % (Source: Ponte (2001)) 21 3.1.3 Market Structure in Consuming Countries Green coffee is supplied to the coffee industry in consuming countries by importers. Some medium and large roasters also import coffee directly (Rice and McLean 1999). At this stage in the market, a few multinational coffee roasters or manufacturers in consuming countries have a large market share. As Table 3-2 shows, in 1998, 49% of the world total market was occupied by two multinational companies: Philip Morris and Nestle (Ponte 2001). Table 3-2: Market Share of Coffee Roasting and Manufacturing Companies (1998) Roasting and Manufacturing Companies Market Share Philip Morris 25 % Top 2 Nestle 24 % 49 % Sara Lee 7 % Top 5 Procter and Gamble 7 % 69 % Tchibo 6 % Others 31 % (Source: Ponte (2001)) Also, many small-sized roasters are recently appearing, especially in the organic, shade-grown, and FT markets (Waridel 2002). These coffee roasters roast and package coffee and sell to distributors or retailers such as supermarkets, specialty coffee stores, and cafes or restaurants where the coffee is finally accessible to consumers. Again, large market share by several roasting or manufacturing companies does not necessarily mean the presence of market power of these large companies. Furthermore, even if their market power is proven by the cartel research, their market power only covers the roasting segment. So their influence on the final price may be limited. 22 Exporters, Small Farmer Co-ops, Large Estate Owners About 100 Importers (25 importers account for all imports of specialty coffee) Over 1,200 Roasters (Large) (l)Kraft (Maxwell House), (2)P&G (Folgers), (3)Nestle (Nescafe) Top 3; over 60% of total green bean in volume (Micro) More than half companies are microroasters (Retail) Supermarkets (70% of total sales in 1997) Specialty Coffee stores and cafes (15% of total sales in 1997) 2,500 Specialty stores 10,000 cafes at the end of 1999 (30% chains) Restaurants Consumers * Retail Share in 1998; Whole Bean 5.4%, Instant 8.2%, Roasted & Ground 77.4% and R & G Decaf 8.8% Figure 3-3: Coffee Market Structure in the US (Source: Rice and McLean (1999)) As an example of the coffee market in consuming countries, the US coffee market is represented in Figure 3-3. In the US, about 100 importers import green coffee from exporters and small farmers' cooperatives, and supply green coffee to roasters. Some large- or medium-sized roasters directly import green coffee (Rice and McLean 1999). 23 Almost 1,200 roasters exist in the market, with almost half of them being very small roasters who roast an average of 500, 60-kilogram bags per year. At the same time, very large roasters such as Kraft, Procter & Gamble, and Nestle have a market share that is more than 60% of total US market volume (Rice and McLean 1999). This figure indicates that several large roasters have large market share in the US and the situation in the US is similar to that of world total roasting and manufacturing market shown by Ponte (2001). Roasters usually sell their product to retail stores such as supermarkets, specialty coffee stores, and cafes and restaurants, or to consumers directly. 3.2 Difference between Arabica and Robusta Coffee Green coffee traded in the world coffee market is not an identical good. Two principal species of coffee traded are Arabica and Robusta coffee. Arabica coffee can be further divided into three categories: Colombian Mild Arabicas, Brazilian and Other Natural Arabicas, and Other Mild Arabicas. Some 70% of the world's total production is Arabica coffee; roughly 80% of all Arabica coffee comes from Latin America, while about 80% of all Robusta coffee are from African and Asian countries (Rice and McLean 1999). Arabica coffee is mainly used for roasted coffee and the Robusta coffee is often used for soluble coffee and inexpensive blending coffee (FAO 2001). According to Viani (2002), more than 80% of coffee is consumed as roasted and ground coffee, while a little less than 20% is used for the production of soluble coffee. Each of the major roasting companies, such as Nestle and P&G, make their own soluble coffee and also produce roasted coffee (Castle 2003). In contrast, small- or medium-sized 24 roasters are mainly concerned with roasted coffee and are not generally making soluble coffee. Thus, the possibility that large roasters or manufacturers have market power might be relatively higher in the soluble coffee market than the roasted coffee market though a careful investigation is necessary to obtain the conclusion. When we consider competitiveness of the coffee market, this difference of species, Arabica and Robusta, also needs to be taken into account. 3.3 Difference of Market Structure between FT and Non-FT Coffee (Conventional Coffee Channel) (Fair Trade Coffee Channel) 1. Small Producer I 2. Local Broker (Coyote) i 3. Transformer (Processor) i 4. Exporter 5. International Broker 6. Importer 4 7. Coffee Roaster I 8. Distributor (Wholesaler) i, 9. Retailer I 10. Consumer 1. Small Producer A 2. Cooperative (consisting of small producers) i 3. Importer I 4. Coffee Roaster i I 5. Retailer I 6. Consumer Figure 3-4: Comparison between Conventional and FT Coffee Channels (Source: Equiterre (2002)) One of the main characteristics of the FT coffee is that importers buy coffee directly from farmers' cooperatives in coffee-producing countries to exclude middlemen, such as local brokers, processors, and exporters, from the coffee channel and to lift the standard of 25 livings for the small producers in these countries (Zehner 2002). The FT coffee channel i compared to the conventional coffee route in Figure 3-4 where it is seen that the market channel of FT coffee is simpler than that of the conventional coffee. 3.4 Distribution of Margins within the Coffee Market Talbot (1997) divides the total income created in coffee commodity chains into four parts: (1) the total income of coffee growers ("Paid to growers"); (2) incomes of processors, traders, exporters, and so on, in producing countries ("Value added in producing countries"); (3) incomes of coffee manufacturers, wholesalers, retailers, etc., within coffee consuming countries ("Value added in consuming countries"); and (4) a residual category that includes, for example, shipping costs, income of shippers, and weight-loss relating to shipping ("Transportation costs and weight loss"). Figure 3-5 illustrates the share of each category within the total coffee income from 1975/76 to 1994/95. • Paid to growers • Value added in producing countries • Transportation costs and weight loss • Value added in consuming countries Figure 3-5: Distribution of Retail Coffee Price (%) (1975-94) (Source: Talbot (1997)) 2 6 According to Talbot, there exist two types of changes in the world coffee market that significantly affected the share of each category. One is cyclical fluctuation in the world coffee production. As was shown in Section 2.1, coffee is a tree crop and coffee trees need a long period to start bearing coffee after being planted. Thus, coffee prices fluctuated, and increases in the world prices shifted income to producing countries and decreases in the world prices brought benefits to consuming countries. The other type of change is the imposition of the export quotas by ICO. We can guess that the share in producing countries (sum of (1) and (2)) are generally larger when quotas were in effect than when quotas were suspended. For example, the suspension of the export quota regime in 1989 was coincided with the decrease in the share in producing countries during early 1990's. As may be seen, the value added in consuming countries has held the largest share through these periods though the share was fluctuating mainly reflecting cyclical fluctuation in coffee production and the imposition of export quota regime. Especially, the share in consuming countries continued to be over 70%. from 1989/90 to 1994/95. In contrast, since the late-1980's, the share in producing countries remained low and the level fell to below 20% from 1989/90 to 1993/94. The Coffee Retention Plan by ACPC and the decline of the Brazilian production because of the frost are considered to have contributed to the increase in the share in producing countries in both 1993/94 and 1994/95 (Talbot 1997). 27 Though large incomes of coffee manufactures, wholesalers, and retailers in consuming countries shown in Figure 3-5 do not necessarily prove the presence of market power in the distribution system in consuming countries, they suggest its possibility. Hence, market power of roasters or distributors in importing countries will be examined empirically by using models in Chapter 4. 3.5 Summary This chapter explores the current world coffee market from the viewpoint of market structure and the distribution of margins. Within the coffee market, many participants or middlemen exist in both producing and consuming countries. In some stages of the coffee market, such as the international market and the roasting and manufacturing market, a few companies have a high market share though large market share itself does not necessarily indicate the presence of market power by those companies. To better understand the coffee market structure, the difference between Arabica and Robusta coffee, and that between FT and non-FT coffee also should be considered. As for the distribution of margins generated in the coffee market, the share in producing countries is very low, while in consuming countries, it is quite high; the share of the value added in consuming countries remained over 70% from 1989/90 to 1994/95. 28 4. EMPIRICAL TESTING OF THE PRESENCE OF MARKET POWER WITHIN CONSUMING COUNTRIES 4.1 Examining market power in the Distribution System within Consuming Countries Though some previous literature mention that the large roasting or manufacturing companies are obtaining large margin by utilizing their market power (Talbot (1997), Mendoza (2000),. Ponte (2001), and Waridel (2002)), none of them empirically prove the presence of this kind of market power. Thus, in this chapter, the likelihood of large roasters or distributors distorting the coffee market is evaluated using a more formal empirical analysis. As discussed in Chapter 3, large roasters or distributors in consuming countries are obtaining a relatively larger margin compared with other participants in the market. According to Talbot (1997), the proportion of total income (the value of retail coffee sales) captured by manufacturers, wholesalers, and retailers in the coffee-consuming countries was 71.9%, and by producers, processors, and traders in producing countries was just 23.0% in 1994/1995. Considering the gap of incomes between producing and consuming countries, to examine the presence of market power of large roasters or distributors in consuming countries has a significant meaning. For this analysis, the Canadian coffee sector is used as an example of a consuming country. As Canada is a relatively small coffee-importing country and the coffee market inside Canada is not so large, it may be easier for several firms to have market power. 29 This is one of the reasons to use the Canadian coffee sector. Nevertheless, the result of this analysis may just apply to Canada and may not be appropriate to depict the coffee sector of other consuming countries. Other research will be necessary to examine the market situation of other countries and this is one of the limitations of this study. A few firms have a large share in the coffee roasting and manufacturing market inside Canada. The four-firm concentration ratio in the tea and coffee processing sector amounted to 87.9% in 1989 and this ratio seems relatively high (Agriculture and Agri-Food Canada 2003a) (Table 4-1). The value added in this tea and coffee sector was 33.5% of shipments in 1995 and this was equal to about 301.5 million Canadian dollars (Agriculture and Agri-Food Canada 2003a). These figures include not only roasted coffee but also soluble coffee and tea. Table 4-1: Size and Significance of the Canadian Tea and Coffee Processing Sector 4-Firm Concentration Shipments Value Added (*) ratio (in 1989) (in 1996) (in 1995) Tea and Coffee 87.9 (%) 0.9 ($ billion) 33.5 (% of shipments) sector * Value added is the value of the product sold by a firm or industry less the value of the materials purchased and used by the firm or industry to manufacture the products. (Source: Agriculture and Agri-Food Canada 2003a) A high concentration ratio by a few firms and a large value added in the processing sector does not necessarily indicate the presence of market power. According to Agriculture and Agri-Food Canada (2003b), the Canadian coffee market is considered as being competitive at the manufacturing stage, though this figure is not based on any empirical 30 analysis that we know of. Thus, the presence of market power by large coffee roasters or distributors will be examined using an empirical model in this chapter. 4.2 Conceptual Models to Examine Market Competitiveness Several papers examine the linkage from farm prices to retail prices and how the spread of these prices is affected by shifts in the retail demand, farm supply, and marketing costs. The model adopted in these papers is called the "Farm-Retail Price model", which is used not only to investigate how the price is transmitted from the farm level to the retail level, but also to determine whether or not the market for the specific commodity is competitive. The Farm-Retail Price model is based on work by Gardner (1975) that shows the theoretical framework for examining implications of simultaneous equilibria in three related markets: a retail product market, a farm product market, and a marketing inputs market. Gardner uses comparative static methods to display how the farm-retail price • spread changes when retail food demand, farm product supply, and the marketing input supply shifts under general competitive conditions. Wohlgenant (1989) develops a conceptual and an empirical framework for the retail-farm demand linkages, employing Gardner's framework. His conceptual model is based on reduced-form retail and farm price equations that estimate the food-marketing sector's supply and demand structure without having direct information about retail food products. 31 Holloway (1991) extends Gardners' model to an imperfectly competitive food market, by considering that Gardner's assumption of perfect competition limits the applicability of that model. At first, Holloway shows the conceptual framework for an imperfectly competitive market and then implements an empirical analysis by referring to the data and reduced-form retail and farm price equations from Wohlgenant (1989). Holloway develops the method to examine whether or not the market indicates competitive behavior by using these models. Later, in this chapter, his method is applied to analyze market competitiveness in the Canadian coffee sector. Similar research is carried out by Gordon and Hazledine (1996), who investigate the price linkage between farm and retail prices for eight agricultural commodities markets in Canada. Their work is based fundamentally on the work of Holloway (1991) and Wohlgenant (1989). At first, they apply the basic Holloway model to these agricultural markets. Since one of their concerns is the effect of imperfect competition after the farm gate, they examine competitiveness of these commodity markets by using the basic Holloway model. Results of their model indicate that the null hypothesis of competitive markets in each sector cannot be rejected. However, since each equation used in their model shows a generally poor fit to the data (low R and insignificant t-value), they conclude that the test results for competitive markets could be a rejection of the model itself (Gordon and Hazledine 1996). Again, employing the same model used by Holloway (1991) and Wohlgenant (1989), Zhao et al. (1998) look into the Australian beef industry. As much of the beef produced in 32 Australia is exported, they revise the basic model to account for the export component of the trade sector in the model. This research is another example that adopts the same model developed by Holloway and Wohlgenant to examine market competitiveness. In their results, each equation showed a generally good fit to the data (relatively high R and significant t-value of almost all variables). According to them, in a model without .a trade sector, there is strong evidence of noncompetitive behavior. However, in a model where the domestic and export markets are divided, results show that the domestic market is perfectly competitive while the export market is not (Zhao et al. 1998). One of the limitations of this model developed by Holloway and Wohlgenant seems to be that this model can just tell whether or not an agricultural commodity market is perfectly competitive. Even if results show the rejection of the null hypothesis of perfectly competitive market, this model cannot indicate the degree of competition. Thus, the rejection of the null of perfect competition does not necessarily mean that the market is either monopolistic or monopsonistic. Although previous empirical research by Wohlgenant (1989), Holloway (1991), Gordon and Hazledine (1996), and Zhao et al. (1998) analyze agricultural commodities such as beef, pork, chicken, milk, etc. in the US, Canada, and Australia, none of them focuses on coffee. Nevertheless, these models are not limited to the specific agricultural commodity market and can also be applied to the coffee sector. 33 Green coffee beans are imported to Canada by importers or large roasting companies and the green bean is then processed, roasted, and packaged within Canada. Finally, the coffee is sold to consumers by roasters or retailers. The coffee market structure inside Canada is similar to those for other agricultural commodities such as beef, pork, chicken, and milk, etc., which have been adopted in previous papers, in the sense that the coffee sector also contains a two-stage market — a raw green coffee market and a retail coffee market, resembling the input (farm) market and the retail market of other commodities. The structures of the Farm-Retail Price model and of the model used in this thesis are shown in Figures 4-1 and 4-2, respectively. Farm (Input Market) Farm Supply (Retail Market) Processor Retailer etc. Retail Product Consumer Marketing Inputs Figure 4-1: Structure of the Farm-Retail Price Model (Input Market) (Retail Market) Import Supply Retail Coffee Trader • Coffee Roaster Product Consumer Retailer etc. ^ I Marketing I Inputs Figure 4-2: Structure of Model used in this Thesis 34 4.3 Model and Methodology Basically, this thesis uses the models developed by Wohlgenant (1989) and Holloway (1991). Wohlgenant shows the conceptual model based on reduced-form specifications for retail and farm prices (Pr = retail price and Pf = farm price). At first, he puts the following form of the complete structural model for a particular commodity, assuming perfect competition in the retail and farm markets. Q/ = Dr(Pr,RD) (la) (retail demand, RD = Exogenous Retail Demand Shifter) Qr' = Z SV' (Pr, Pf, MC) (1 b) (retail supply, MC = Marketing Inputs Costs) Of" =2ZDf'(Pr,Pf,MC) (lc) (farm-level demand) Q/' = Predetermined (Id) (farm-level supply) In (lb) and (lc), "i" denotes an individual firm as the retail supply and farm-level demand are obtained as horizontal summation of each supply and demand function of individual firms (Wohlgenant 1989). The above equations (la-Id) may be rewritten as a two-equation system by assuming market clearing in both retail and farm markets. I 6V (Pr, Pf, MC) - Dr(Pr, RD) = 0' (2a) Qr-TDfi(Pr,Pf,MC) = 0 (2b) Totally differentiated, (2a) and (2b) may be expressed in the following elasticity form. ( £ r -e)-dln Pr+ • d In Pf = eno • d In RD - $MC • d In MC (3a) -4fr.d]nPr-4ff-d\nPf = $Mc-d\nMC-d\nQj (3b) 35 where E,n = the elasticity of retail supply with respect to retail price e = the elasticity of retail demand with respect to retail price Erf = the elasticity of retail supply with respect to farm price eRD = the elasticity of retail demand with respect to RD E,rMc = the elasticity of retail supply with respect to MC E/r = the elasticity of farm-level demand with respect to retail price Ejr = the elasticity of farm-level demand with respect to farm price EJMC = the elasticity of farm-level demand with respect to MC Then, by solving (3a) and (3b) for " d In P r " and " d In Pf ", the following reduced-form equations for retail and farm price can be obtained. d In P r = ArRD • d In RD + Aruc • d In MC + Arfd In Qy (4a) d\nPf- AfRD • d In RD + AJMC • dlnMC + Ajr • dlnQj (4b) where ArRD = -Ejf • eRD ID ArMC = [EjfE,rMC — E,rfEjMc) I D Arf=E,rflD AfRD = Efr • eRD I D AfMC = (-EfrE,rMC + (E,rr ~ e)EjMc) I D Aff = -{E,rr-e)ID D - -(E,rr - e)Eff + E,rfEfr Holloway (1991) specifies the three equation model (the Unrestricted Holloway Model) to measure changes in the marketing margin ratio (M ' =P r /P f ) , retail prices ( P r ) and farm (import) prices ( P f ) based on the above reduced-form equations obtained by Wohlgenant. The marketing margin ratio equation can be derived from the above retail and farm price equations (4a and 4b). The Unrestricted Holloway Model can be expressed as follows. 36 (Unrestricted Holloway Model) AlnM = /?mo + fimmclsAwMC + fimrdAlnRD + jSmqAlnQ + £l A In Pr = j3pro + fame A In MC + farrdA In RD + (3prqA XnQ + ei ^ (5) AXnPf = fipfo + j3pfmcAlnMC + jSpfrdAlnRD + j3pf<,A\nQ + £3 where MC = Marketing Inputs Costs RD = Exogenous Retail Demand Shifter Q = Farm Supply (Import Supply) si, si and£3 = random error terms Each variable is a first difference in logarithms, representing percentage change for small changes in variables and enables each coefficient to be interpreted as an elasticity. Moreover, the RD variable in equation (5) is determined as follows: Alni?Z) = __^AlnPr j + rjiyA In Y + A In Pop (6) •*j where T7jj = Cross-price elasticity Prj = Retail price of other good j rjjy = Income elasticity Y = Disposable Income per person Pop = Total consuming population This equation (6) is just used to generate data for the RD variable. To make the system of equations simple and to generate testable hypotheses about the competition, Holloway uses three assumptions, as commonly made in past studies (Zhao et al. 1998). The first assumption is that the farm supply (Q) is predetermined and is exogenous, and the second one is that the supply of marketing inputs is perfectly elastic (which indicates that MC is exogenous). The last one is that the retail demand shifter variable (RD) can be estimated as a linear combination of elasticities and value of an individual retail demand shifter, as shown in the above equation (5) (Zhao et al. 1998). 37 (Tests for Perfect Competition in the Commodity Market) Then, Holloway shows how to test for market competitiveness by using the above three equation model. He develops Gardner's retail-farm price model based on perfect competition and assumes a conjectural-variations oligopoly with endogenous entry of firms. Given an industry demand curve, an aggregate output of each firm, and the firms' conjecture function, the firm's first-order condition becomes as follows. Vr(l +9,/r/) = C(Pf,MC) (7) where Pr = Retail price of a food product 6i = Elasticity of industry output conjectured by firm i T] = Price elasticity of demand C(P/,MC) = Firm's marginal cost Pf = Farm price MC = Marketing inputs costs Two polar cases are obtained: perfectly competitive market, where 6i = 0; and monopoly market, where 0i = 1. In his analysis, Holloway assumes that each firm has identical technology and produces a homogeneous product. This assumption means that elasticity of industry output conj ectured by each firm is the same (6i=0j = 0). Since it is difficult to measure the value of 9 , Holloway derives a set of equivalent condition for 8 = 0, from the equilibrium model, that can be empirically tested (Zhao et al. 1998). As this derivation of equivalent condition for 0 = 0 requires many equations and formidable mathematical techniques, details are shown in Appendix 1. Only the main points are shown in the following several paragraphs. 38 Holloway focuses on Gardner's elasticities of the marketing margin ratio (M =Pr/Pf) with respect to three exogenous variables, RD, Q and MC. E M . R D - Ep R ,RD - Epf,RD (8) E M , Q - EpR,Q - Ep^Q (9) E M . M C = E p R > M C " Epf,MC (10) The first terms on the right-hand sides of (8)-(10) indicate elasticities of the retail price (Pr) with respect to RD, Q, and MC; the second terms represent respective effect of RD, Q, and MC on the farm price (Pf). From the equilibrium model, Holloway derives the expanded form of equations (8)-(10) in terms of all the parameters of the model including 6 (equations (21)-(23) in Appendixl). Next, he derives the testable hypotheses to examine whether the condition 6 = 0 holds or not by using these equations. He presents the following proposition: Necessary conditions for perfect competition in the food markets are: (i) E M . R D = - E M J Q (ii) Ep R ,RD = - EpR,Q, and (iii) E P f > R D = - Epf,Q. Since coefficients in the three-equation Unrestricted Holloway model (equation (5)) represent elasticities, these necessary conditions can be expressed by using coefficients in equation (5) (/3mrd = - f3mq, fipn-d = - pprq, and fip/rd = fipfq). Thus, necessary conditions imply that under perfect competition, the proportional impacts on the marketing margin ratio (M), the retail prices (Pr), and the farm prices (Pf), of changes in the retail demand and the farm supply have the same magnitude with opposite signs (Zhaoetal. 1998). 39 These necessary conditions can be easily proved by substituting 0 = 0 in equation (21)-(23) in Appendix 1; if 0 = 0 hold, then the above conditions (i)-(iii) hold. Hence, these conditions (i)-(iii) are necessary for perfect competition. Next, Holloway examines the sufficiency of each condition. He shows whether 0 = 0 holds if each of condition (i)-(iii) holds. According to his findings, if condition (i) and (iii) hold, then 0 becomes equal to zero. Thus, both conditions (i) and (iii) are sufficient for perfect competition. However, even if condition (ii) holds, 0 = 0 does not necessarily hold. In this case, not only is condition (ii) required, but (iv) E p r , M C ^0 (flprmc =£0 in equation (5)) is also required to work as the sufficient condition for# = 0 (Appendix 1 describes the details of these conditions for perfectly competitive market). Hence, Gordon and Hazledine (1996) and Zhao et al. (1998) calls the above condition (i)-(iii) as necessary conditions and (iv) as sufficient condition. The same notation as theirs will be adopted in this thesis. (Restricted Holloway Model) The advantage of the Holloway model is that linear restrictions imposed on the parameters in equation (5) can be used to test a null hypothesis to see whether or not the market sectors are competitive (Gordon and Hazledine 1996). To examine whether or not the above necessary and sufficient conditions hold, Holloway also presents the Restricted Holloway model; where the three equations in (5) are estimated to show variations in marketing ratio, retail price and import price with restrictions imposed so that 6jrd + Bjq = 40 0, j e m, P r and Pf. By comparing both the Unrestricted and Restricted Holloway models, the test for perfect competition in the industry can be implemented. Empirically, if any of the F-tests for the hypotheses about restrictions (/3jrd + /3jq = 0, j e m, P r and Pf) are significant, this can serve as at least one piece of evidence to show that the commodity market may not be competitive. In contrast, if none of the F-tests are statistically significant, and the test for the hypothesis that fiprmc = 0 is rejected, the empirical result indicates competitive behavior (Zhao et al. 1998). However, these test results do not indicate the degree of perfect competition at all, and this is a limitation of this test for market competitiveness. 4.4 Data The data set ranges from February 1995 to December 2000 on a monthly basis for the coffee market in Canada. In adopting the log-differenced value for each data, the data for January 1995 was used to generate the initial value. Definitions and summary statistics for each variable are listed in Table 4-2. The number of observations is 71 for the Unrestricted and Restricted Holloway model. Most of the data, such as for the retail prices of coffee, are from the CANSLM files of Statistics Canada. Appendix 2 provides a graphical representation of changes in each variable. Holloway (1991) and Zhao et al. (1998) estimate both the Unrestricted and Restricted models with prices in their nominal forms. This is because the normal homogeneity conditions for prices do not necessarily hold in the general imperfectly competitive model 41 (Zhao et al. 1998). Following their methods, nominal values are adopted to estimate equations in this thesis. However, as a comparison, models are also estimated with variables deflated by the consumer price index. As will be shown later in this chapter, the adoption of either nominal value or deflated value does not affect the main result of the estimation and tests for perfect competition. As described above, the data is transformed to represent the log differences (e.g. lnPt-lnPt-i), following previous work by Wohlgenant (1989), Holloway (1991), Gordon and Hazledine (1996), and Zhao et al. (1998). This transformed variable indicates the percentage change in the original variable, and this transformation enables the interpretation of each estimated parameter as an elasticity. Though Wohlgenant (1989), Holloway (1991), Gordon and Hazledine (1996), and Zhao et al. (1998) use the data for farm price and farm quantity, these data cannot be used in case of Canadian coffee sector since no green coffee is produced in Canada. Instead, the data for the import price (international price) and the quantity imported to the Canadian market are adopted. In the market, four types of coffee are present: 1) Colombian Mild Arabicas; 2) Brazilian and Other Natural Arabicas; 3) Other Mild Arabicas; and 4) Robustas. The first three types of Arabica coffee are mainly used for roasted coffee and the Robusta coffee is often used for instant coffee though Robusta is also a source of inexpensive blending with Arabica (Oxfam Canada 2003). Since the focus is on price changes for roasted coffee and 42 the data for Robusta coffee could not be divided into roasted coffee and instant coffee, the average international prices of the three types of Arabica coffee are used for the import prices, and the average prices of roasted coffee are used for the retail prices, given by the International Coffee Organization and the CANSLM files of Statistics Canada, respectively. In Figure 4-3, international prices of the three types of Arabica coffee show almost the same trend, thus, justifying the use of the average price of these three types. To generate marketing input costs (MC) variables, Gordon and Hazledine (1996) collected data for labor cost, containers and packaging cost, transportation cost, fuel and power cost, rent and storage cost, maintenance and repair cost, tax cost, service cost, utility cost, and office and other supplies cost. They obtain percentage change figures for each cost and finally, weight the percentage changes using the weighting scheme supplied by Agriculture Canada. In this thesis, a weighted scheme is similarly used. At first, the data for "labor cost" and "fuel and power" are log-differenced to show the percentage change figures; then these percentage change figures are weighted by using the weight obtained from the data for the coffee and tea manufacturing industry in the US (Appendix 3 provides details about acquiring these values). As data about the Canadian coffee industry is unobtainable, instead, US data are used here. In the process to obtain the exogenous retail demand shifter variable, RD (equation (6)), the income elasticity (r/iy), shown in previous papers, is used. For the Canadian case, Hassan and Johnson (1976) report an income elasticity of 0.1107 for coffee. This research is somewhat dated and the income elasticity of coffee may vary largely 43 depending on time and the country. Thus, the income elasticity of 0.1107 is used as a base value and sensitivity analysis is performed by changing values instead of adopting only one value. Since recent results for Canadian income elasticity are not available, the US income elasticities are used. Results from previous studies are reported in Table 4-3. In the US, income elasticity of coffee ranges from -0.19 to 0.53. By implementing sensitivity analysis, the values of income elasticity are increased and decreased from 0.1107 by 0.1 point within the range from -0.19 to 0.53. Thus, the income elasticity values adopted in this sensitivity analysis are -0.1893, -0.0893, 0.0107, 0.1107, 0.2107, 0.3107, 0.4107, and 0.5107. The cross-price elasticity (TJJJ) and price changes in substitutes or complements of coffee are ignored because of the lack of such data, though Wohlgenant (1989), Holloway (1991), and Gordon and Hazledine (1996) use both cross-price elasticity and income elasticity data obtained through a composite demand system. With regards to data for the disposable income per person (Y) and the total consuming population (Pop), constituting the retail demand shift variables, quarterly data is used instead of monthly data, which is unobtainable. To transform quarterly data to monthly data, the value is assumed to increase or decrease linearly each month. 44 Table 4-2: Definitions and Summary Statistics for the Variables in this Model Variables Description Mean Standard Deviation Min Max A l n M Log-differenced market margin (margin = retail price/ import 0.0081 0.0971 -0.2642 0.2731 price) 0.0479 AlnP r Log-differenced average retail -0.0037 0.0186 -0.0454 price of coffee A l n P f Log-differenced average import -0.0118 0.0933 -0.2252 0.2891 price of coffee A l n M C Log-differenced marketing input costs 0.0027 0.0133 -0.0501 0.0466 AlnRD Log-differenced retail demand shifter (elasticity - 0.1107) 0.0011 0.0003 0.0006 0.0017 AlnQ Log-differenced import supply of coffee 0.0089 0.1454 -0.3451 0.3374 Figure 4-3: Monthly Price Fluctuations of Three Arabica Coffee (1982-2001) (Source: ICO (2003a)) 45 Table 4-3: Income Elasticity in Previous Papers Authors and Published Data Coverage Income Elasticity Country Year for Coffee Hassan and Johnson Cross Sectional Data 0.1107 Canada (1976) Huang et al. (1980) 1966-77 0.53 (Box-Cox function) 0.51 (Log-log function) US Heien and Pompelli 1977-78 -0.19 US (1989) Akiyama and Varangis 1974-84 0 US (1989) Maizels et al. (1997) 1979-90 0.46 us 4.5 Results of the Regression 4.5.1 Results of the Unrestricted Holloway Model At first, the Unrestricted Holloway Model (three equations in (5) without restriction) is estimated by applying Ordinary Least Square (OLS). All regression results in this thesis are estimated by using the software program, Stata. Results of estimating the Unrestricted Holloway model are reported in Table 4-4. As the results of sensitivity analysis by changing the value of income elasticity are almost the same, only the result for one representative case (income elasticity = 0.1107) is shown in Table 4-4. Other results are reported in Table DI of Appendix 4. Also, the estimation results obtained by using the deflated prices do not change the results very much, so only those obtained by adopting nominal prices are shown in this chapter. Estimation results using deflated prices are reported in Table El of Appendix 5. 46 Since the data used in this analysis is time-series data, the Durbin-Watson (DW) statistics are obtained to test the AR (1) serial correlation for each equation. According to the results for the DW statistics, we fail to reject the null hypothesis of no serial correlation against the alternative of a positive serial correlation at the 1% significance level in all three equations. Also, heteroskedasticity is tested using the White test; and the results show the presence of heteroskedasticity in some equations. To correct for heteroskedasticity, and obtain the robust standard errors, White's correction is applied to these estimations. As is clear from Table 4-4, each equation shows a generally poor fit to the data. The reported R 2 is low and most of t-values are insignificant. Only a few coefficients are significant; coefficients on import supply in both ratio and import price equations are significant at the 1% significance level. For example, a coefficient on import supply in the import price equation indicates that if import supply increases by 1 %, the import price of raw green coffee decreases by 0.3%. This is consistent with the prior expectation. In contrast, a coefficient on import supply in the ratio equation indicates that a 1% increase in import supply will lead to 0.3% increase in the marketing margin of processors or distributors. This means that the Canadian processors or distributors are obtaining a larger margin by utilizing lower green coffee prices when the supply of green coffee increases. 47 Table 4-4: Estimates of the Unrestricted Holloway Model (Income Elasticity = 0.1107, using White's Correction for Ratio and Import Price Equation) Ratio Equation, Retail Price Equation Import Price Equation Variables ( A l n M ) ( A l n P r ) ( A l n P f ) A In M C -0.6159 -0.2714 0.3444 (-0.71) (-1.59) (0.46) A l n R D 0.0843 1.4770 1.3927 (0.00) (0.18) (0.05) A l n Q 0.3054 0.0052 -0.3003 (3.26) (0.33) (-3.43) Constant 0.0070 -0.0046 -0.0115 (0.20) (-0.50) (-0.34) No. Observation 71 71 71 R-squared 0.2021 0.0367 0.2122 DW statistics 1.9661 1.7454 1.9575 AIC -1.954 -5.068 -2.046 * t-ratios in brackets, Robust standard errors obtained using White's correction for heteroskedasticity for Ratio and Import Price Equation 4.5.2 Results of the Restricted Holloway Model To contribute to the test for market competitiveness, the Restricted Holloway model is also estimated. The same procedure is applied as with the Unrestricted Holloway model and three equations in equation (5) are estimated with restrictions imposed so that Bjrd + (Sjq = 0, j e m, Pr and Pf. Again, as the results of sensitivity analysis by changing the value of income elasticity and those by adopting the deflated values do not affect the main content of the results, only a result for one representative case (income elasticity = 0.1107 and using nominal value) is shown here (Table 4-5). Other estimation results are reported in Table D2 of Appendix 4 and Table E2 of Appendix 5. 48 Results of the White test show strong evidence for heteroskedasticity though we cannot reject the null hypothesis of no serial correlation from the results of the D-W statistics. Hence, the White correction is applied to correct for the heteroskedasticity and to obtain robust standard errors in all three equations. Also in this case, each regression result reports a generally poor fit to the data. Imposing restrictions for a competitive market does not seem to improve the regression results much. The low R 2 indicates that very little of the variation in the dependent variable is explained by these equations. Again, only several coefficients are statistically significant. Table 4-5: Estimates of the Restricted Holloway Model (Income Elasticity = 0.1107, using White's Correction for all three equations) Ratio Equation Retail Price Equation Import Price Equation Variables ( A l n M ) ( A l n P r ) (AlnPf) A In M C -0.6158 -0.2711 0.3447 (-0.72) (-1.59) (0.46) A l n R D -0.3055 -0.0056 0.3000 (-3.28) (-0.30) (3.40) A l n Q 0.3055 0.0056 -0.3000 (3.28) (0.30) (-3.40) Constant 0.0074 -0.0030 -0.0104 (0.70) (-1.42) (-1.01) No. Observation 71 71 71 R-squared 0.2021 0.0362 0.2122 DW statistics 1.9660 1.7435 1.9577 AIC -1.983 -5.096 -2.074 * t-ratios in brackets, Robust standard errors obtained using White's correction for heteroskedasticity 4.5.3 Tests for Competitiveness of the Market Tests to examine perfect competition in the market are implemented using results of both the Unrestricted and Restricted Holloway models. As necessary and sufficient conditions 49 are j3jrd + Biq = 0 (j e m, Pr and Pf), and Bpvrnc J= 0, respectively, tests are made of the null hypotheses: (a) H 0 : Bird + Bk = 0, j e m, Pr and Pf, (b) H 0 : Bprmc= 0 The results of the F-test for the above null hypotheses (a) and (b) are reported in Table 4-6. Since results adopting both sensitivity analysis and deflated value do not change the conclusion, only one representative case (income elasticity is equal to 0.1107 and using nominal value) is reported here (see Table D3 of Appendix 4 and Table E3 of Appendix 5 for results using sensitivity analysis and deflated prices). Table 4-6: Results of F-tests to examine Market Competitiveness (a) H0:/3jrd + i3jq=0 : (b) H0:/3prmc=0 " j e m, Pr and Pf (Income elasticity = 0.1107) Ratio Equation 0.00 df= (1,67) Prob>F = 0.9901 Retail Price 0.03 df = (1,67) 2.54 df = (1,68) Prob > F = 0.8590 Prob > F = 0.1155 Import Price 0.00 df = (1,67) Prob >F = 0.9710 As for the necessary conditions, these results indicate that in all' three equations, the null hypothesis (a) (Ho: B^ + Bjq = 0, j 6 m, Pr and Pf) cannot be rejected even at the 10% significance level. F-statistics in all three equations are almost close to zero. This indicates that the necessary condition for perfect competition holds in this Canadian coffee case. As for the test for sufficient condition, we fail to reject the null hypothesis 50 (b) (H0: /3Prmc= 0) even at the 10% significance level in the retail price equation. Considering the sufficient condition ( j 3 p r m c ^ 0), the non-rejection of the null (b) indicates that the sufficient condition for perfect competitive market does not hold. Considering these test results of the null hypotheses, the Canadian coffee sector from input market to retail market seems to be close to competitive. This conclusion is consistent with the expectation by Agriculture and Agri-Food Canada. However, as the estimation results of both the Unrestricted and Restricted Holloway model report a relatively poor fit to the data, these test results for a perfectly competitive market could be a rejection of the model itself. Thus, this conclusion that the Canadian coffee sector may be perfectly competitive is relatively weak and other careful research might be necessary to obtain greater certainty regarding this conclusion. As is mentioned in Section 4-2, Gordon and Hazledine (1996) also get the similar results when they examined market competition by using the Canadian eight agricultural commodities: beef, pork, eggs, chicken, milk, cheese, butter, and ice cream. As each equation used in their model shows a generally poor fit to the data, they conclude that their test results for competitive markets might be a rejection of the model itself. They consider that these results are caused because one of the assumptions used in the Holloway model, the exogenous farm supply, may not be appropriate to any of their eight agricultural commodities. These commodities are either "marketing board commodities" whose output is set by quotas and the corresponding prices are generally set by cost of 51 production price formula, or "traded commodities" whose prices are set in the North American market (Gordon and Hazledine 1996). Thus, they point out that the farm price should be treated as exogenous instead of the farm supply for these commodities. This means that the farm price equation (the third one in equation (5)) may be incorrectly specified for their commodities. Hence, they try to replace the farm supply variable in the first and the second equations in equation (5) with the farm price variable. Gordon and Hazledine also think that since the data set used in their model is monthly data, this time period might be too short to allow for marketing clearing assumed in the equilibrium specification of equation (5). Then, they try to capture the lagged effect by incorporating the lagged value in equation (5). By incorporating these elements (the exogenous farm price and the lagged variables) in the model, they could improve their regression results; R 2 was clearly improved and many of t-values were also improved. However, their new models that treat the farm price as exogenous and that include the lagged values can not be used any longer to examine perfect competition of those commodities. In case of the Canadian coffee sector, the import price of coffee is the international price. As Canada is not a large importing country, it will be reasonable to assume that Canada is a price taker in the world coffee market; Canada imports almost two million metric tons of coffee and the share among the world total imports is just 3.2% in 2000 (ICO 2000). Hence, the import price might need to be treated as exogenous also in this model. Since the data used in this chapter are monthly, the fact that a lagged effect was not 52 incorporated in the model could be considered as one of the reasons of the poor fit of this model. As for the Canadian coffee market, the parameter estimates were improved by taking into account the exogenous import price and the lagged effects. These results justify the incorporation of exogenous farm price and lagged variables in the Farm-Retail Price model to examine certain commodities. Thus, results in this chapter and Gordon and Hazledine (1996) show the limitation of the Holloway model to examine.market competitiveness. 4.6 Summary A principal objective of this study is to examine whether or not the roasters or distributors inside Canada have market power. To test the presence of market power, the Farm-Retail Price model developed by Wohlgenant and Holloway is introduced and applied to the Canadian coffee sector. Test results for market competitiveness based on their model provide evidence to indicate that the Canadian coffee market might be close to competitive at the national level. However, as estimation results show a relatively poor fit to the data, these test results could be a rejection of the model itself. In case of the Canadian coffee sector, the import price might need to be treated as exogenous and a time lag might be necessary to reach the market clearing condition. Since the Holloway model to examine perfect competition does not incorporate these elements in the model, this Holloway model can be considered to have limitations. 53 5. IMPACTS OF FAIR TRADE COFFEE ON THE COFFEE MARKET AND RELATED PROBLEMS 5.1 Results of the Empirical Analysis and FT Coffee As was mentioned in Chapter 2, FT coffee has attracted great deal of attention as an alternative trade practice to achieve benefits to coffee farmers in producing countries mainly reflecting recent low and unstable world coffee prices. Waridel (2002) considers that large roasting companies in consuming countries have undeniable influence over the world coffee market and this kind of market power by large companies is one pf the causes of the present low and unstable coffee prices obtained by coffee producers. Oxfam America (2002) and Raynolds (2000) also insist that one of the purposes of FT coffee is to solve the inequality between producing countries and consuming countries since large roasters in consuming countries that have market power are obtaining large profits at the expense of lives of small coffee farmers in producing countries under the conventional trade. Thus, FT coffee trade seems to be based on the idea that large roasters or distributors in consuming countries are distorting the world coffee market and coffee producers are forced to have unreasonably small benefits. However, according to the findings shown in Chapter 4, regression results weakly indicate that the Canadian coffee market is close to competitive from importers to retailers. These findings suggest that FT coffee trade scheme may be built on an incorrect premise, at least in case of the Canadian coffee market though this may not apply to the coffee markets in other consuming countries. 54 5.2 Impacts of FT on the Coffee Market in Consuming Countries In Section 5.2 and 5.3, the coffee markets in consuming countries are assumed to be competitive to examine the impact of FT coffee on the coffee market reflecting the results of empirical analysis in Chapter 4. Hence, the analysis shown in Section 5-2 and 5-3 may only be applicable to consuming countries that have competitive markets. After the introduction of FT coffee, some part of the whole coffee market is considered to be shifted to the FT coffee market in consuming countries. In this section, this market is the focus of our attention. As was shown in Section 2.6, the willingness to pay for FT coffee by consumers can be considered to be higher than that for non-FT coffee. Galarraga and Markandya (2000) estimated how much was paid for the FT or organic characteristics of coffee in the UK market by using the hedonic model. Their results indicate that under ceteris paribus conditions, the presence of the FT or Organic label increases the coffee price by 11.26%. Moreover, according to consumer research by Transfair USA, among specialty coffee consumers, 50% of respondents said that they would pay US$ 1.00 or more per pound for FT coffee (Transfair USA 2002a). Reflecting on these situations in the market, it would be reasonable to assume that the demand curve for FT coffee would shift to the right from Do to D] in Figure 5-1. 55 Price So Pi Po N X Qo Qi Quantity Figure 5-1: Coffee Market after the Introduction of FT Coffee The new equilibrium price and quantity of FT coffee can be determined by Di and the supply curve for FT coffee, So- This new FT coffee price (Pi in Figure 5-1) would be higher than Po and the new equilibrium quantity (Qi in Figure 5-1) would be larger than Qo, reflecting the higher willingness to pay for FT coffee. In this sense, the introduction of FT coffee can be expected to somehow contribute to the increase of the retail coffee prices in consuming countries though more research is necessary to know how much higher prices will be actually achieved. However, considering the low market share of FT coffee in the whole coffee market (0.2% in Canada in 2001 (Transfair Canada 2002)), this impact of FT coffee will not be so large. . Even if prices become higher in the retail coffee market, this does not necessarily mean that coffee producers in producing countries can obtain higher prices. As is shown in Figure 3-4, many participants appear even in the FT coffee channel. These premiums 56 may be captured by some other participants in the market such as roasters, importers, and producers' organizations though coffee farmers also have possibilities to obtain some of these premiums. The distribution of FT coffee prices will need to be closely examined to investigate who is actually obtaining benefits from the FT coffee trade. 5.3 Effect of the Adoption of the Minimum Price Scheme Price P, / So Oversupply / = Q s - Q d ^ \ D , • Qd Q i Q s Quantity Figure 5-2: Effect of the Adoption of the Minimum Price for FT Coffee As explained in Section 2.4, the present FT standards set by the FLO contain a trade condition for the minimum prices paid to the producers' organizations, i.e., 126 US cents per pound for washed Arabica coffee. If importers of FT coffee are required to pay these minimum prices by producers' organizations, retail prices of FT coffee may be set at higher level than the market clearing price ( P i in Figure 5-1 and 5-2) by adding certain margin to these minimum prices by roasters or retailers. Suppose retail coffee prices are set at P m i n (Figure 5-2), then, oversupply of FT coffee ( Q s - Q d in Figure 5-2) will occur in 57 the market. Already, this kind of a tendency is appearing in the market. According to the report by Transfair USA, coffee producers sell an average of only 15 percent of their monthly products to FT coffee buyers (Zehner 2002). In this case, to keep the price of FT coffee at P m i n , some measures to restrict quantity would be necessary. One such measure would be for the government or some organization having enforcement power, to impose the producers' quota in aggregate at Q d . The allocation of a quota among countries, producers' cooperatives, and each of the producers may cause another serious problem, however, when considering the failure of the export quota system that was adopted by the ICA. At the producer level, a risk is present that a producer who are selling their coffee as FT coffee might not receive the FT coffee quota in the future (Zehner 2002). Moreover, even if the minimum price set by the FLO are paid to producers' organizations, whether or not farmers can enjoy higher prices will depend on the allocation of the FT premium by the producers' organizations. As pointed out in several case studies, such as those by Jones and Bayley (2000) and Ronchi (2002), not all the FT coffee premiums were passed on to producers, but some percentage of these premiums were kept by the producers' organizations. 5.4 Small Impact of FT Coffee on the World Coffee Market As was shown in Section 2.1, the world coffee prices are low and unstable recently. The main causes of these low and unstable world prices are an oversupply of coffee by coffee 58 producing countries such as Vietnam and the cyclical fluctuations of the world coffee production. The introduction of FT coffee cannot rectify these problems at all. Rather, the expectation for higher FT coffee prices by coffee producers might increase the production of coffee though more research is necessary to investigate the impact of FT coffee on the decision of each producer. Even if higher prices are achieved in a part of the total coffee market by selling FT coffee, reflecting the higher willingness to pay for FT coffee and the regulation set by the FLO, these impacts seem to remain very small/Because the market scale of FT coffee in the total coffee market is very small (the percentage of FT coffee sales among the total coffee import is just 0.3% in 2001) and the minimum price scheme set by the FLO is not sustainable in the long-run without some unfavorable measures for producers. 5.5 Summary FT coffee trade seems to be based on the idea that large roasters or distributors in consuming companies have market power. However, at least in Canada this kind of market power cannot be found and the FT coffee scheme might be built on an incorrect premise, although this may not be so in other consuming countries. It can be estimated that introducing FT coffee will have an effect on the demand side in coffee-consuming countries. The demand curve will shift to the right corresponding to a higher willingness to pay for FT coffee than for non-FT coffee. This effect leads to an increase in the equilibrium price and quantity. However, since the market share of FT 59 coffee is very small, this impact, of increasing coffee prices, will be very limited. Furthermore, even if higher prices are achieved, some premiums may be captured by other participants in the market and coffee producers in producing countries may not obtain large benefits. The minimum price scheme adopted by the FLO can achieve higher prices. However, this scheme will invite an oversupply of FT coffee. To make this situation sustainable, some measures such as the allocation of production quota or other quantity restrictions would be necessary. These measures will make the beneficial impact of FT coffee smaller. The introduction of FT coffee cannot rectify the main causes of the current low and unstable world prices. Considering its market scale and the fragile minimum price scheme, the impact of FT coffee on the world coffee market will be very limited. 60 6. S U M M A R Y A N D C O N C L U S I O N S 6.1 Conclusions and Comments This thesis examines the current coffee market especially from the viewpoint of market structure. Though several previous literature suggest the presence of market power by large roasters or distributors in consuming countries, none of them shows any models or empirical analyses to prove this. Thus, the main research question of this thesis is to examine whether or not the distribution system in consuming countries has market power within the coffee market. To answer this question, the Farm-Retail Price model developed by Wohlgenant and Holloway (the Holloway model) is applied to the Canadian coffee sector. The advantage of this Holloway model is that linear restrictions imposed on the parameters in three-equation model can be used to test whether or not market sectors of agricultural commodities are competitive. The Canadian coffee market is picked up as an example because the coffee market inside Canada is not so large and it might be easier for several companies to obtain market power. Test results for market competitiveness, using Canadian data, provide evidence which suggests that the Canadian coffee market might be close to competitive at the national level. Nevertheless, as regression results show a relatively poor fit to the data, these test results could be a rejection of the model itself. 61 These test results indicate that the direct application of the Holloway model might have several problems in case of the Canadian coffee market. For example, the import price might need to be treated as exogenous and a time lag might be necessary to reach the market clearing condition. Considering that the Holloway model to examine perfect competition does not incorporate these elements in the model, the Holloway model seems to have certain limitations. Then, the findings of empirical analyses are used to examine the meaning and impact of FT coffee. According to several literature, FT coffee scheme appears to be based on the idea that large roasters in consuming countries are distorting the world coffee market. However, regression results in Chapter 4 weakly indicate that the Canadian coffee market is close to competitive. These findings suggest that FT coffee scheme may be built on an incorrect premise, at least in Canada. Since the willingness to pay for FT coffee seems higher than that for non-FT coffee, FT coffee is expected to have an effect to shift the retail demand curve rightward. It is considered that this effect leads to an increase in the equilibrium price and quantity. Nevertheless, as the market share of FT coffee among the total coffee is very small, this impact to increase coffee prices will be very limited. Though one of the characteristics of the FLO certified FT coffee is the setting of the minimum price paid by importers to producers' organizations, this scheme may cause an oversupply of FT coffee in the future. To make this minimum price level sustainable, 62 some measures to restrict quantity such as the allocation of production quota would be necessary. These measures will make the introduction of FT coffee less beneficial. After all, the introduction of FT coffee cannot rectify main causes of the low and unstable world coffee prices such as an oversupply of coffee to the world market and the cyclical fluctuation of the world coffee production. Considering also its market scale, it can be concluded that FT coffee does not have a large impact on the world coffee market. 6.2 Limitations of this Study and Recommendations for Future Study In focusing on the Canadian coffee sector in Chapter 4, the conclusion is made that large roasters or distributors may not have market power in the coffee market though this result is based on weak evidence. This result may only be applicable to Canada and may not depict the total figures of other coffee-consuming countries. To more clearly understand whether of not market power of the large roasters or distributors exists, an additional investigation of the cases of other countries would be necessary. To do these analyses, the limitation of the direct application of the Holloway model also needs to be considered. Another model or framework might be appropriate for these examinations. Furthermore, the improvement of the basic Holloway model to take into account the exogenous farm price or a time lag might need to be pursued though it is not clear whether or not such models can exist. 63 As FT coffee began in earnest in the late-1990's, there is a scarcity of data on FT coffee and empirical analysis is not yet possible when examining the impact of FT coffee on the world coffee market. Instead, theoretical scenarios are described with reference to some previous studies. Further studies would be required to evaluate in greater depth the actual willingness to pay for FT coffee and the increase of equilibrium price. Furthermore, the distribution of FT coffee prices among each participants of the FT coffee channel will also need to be considered to understand who is getting benefits from the FT coffee trade. Last of all, the impact of FT coffee on the behavior of coffee producers can also be an interesting topic if data appropriate to test their decisions become available. 64 REFERENCES Agriculture and Agri-Food Canada (2003a), Table2: Sub-Sector Size and Significance in the Canadian Food and Beverage Processing Sector, 1988-96. Available: http://www.agr.gc.ca/food/industryinfo/processing/sector/table2_e.html (6 May 2003) Agriculture and Agri-Food Canada (2003b), The Canadian Tea and Coffee Industry, Food-Bureau, Sub-Sector Profiles. Available: http://www.agr.gc.ca/food/profiles/tea_e.html (6 May 2003) Akiyama, T. and P. N. Varangis (1989), Impact of the International Coffee Agreement's Export Quota System on the world's Coffee Market. World Bank Working Paper 148, World Bank, Washington, DC Akiyama, T., J. Baffes, D.F. Larson and P. Varangis (2001), Commodity Market Reforms: Lessons of Two Decades. The World Bank, Washington, D.C. Bohman, M. and L. Jarvis (1996), The International Coffee Agreement: A Tax on Coffee Producers and Consumers?, Working Paper #96-2, Department of Agricultural Economics, The University of British Columbia CAFE DE COSTA RICA (2003), Association of Coffee Producing Countries (ACPC). Available: http://www.icafe.go.cr/icafe/APPC_eng.htm (26 February 2003) Castle, T., J. (2003), Soluble Holds Extreme Niches within Market, Tea & Coffee Trade Journal. Available: http://www.teaandcoffee.net/1000/coffee.htm (17 June 2003) Charveriat, C. (2001), Bitter Coffee: How the Poor are Paying for the Slump in Coffee Prices. Oxfam [EFTA] European Fair Trade Association (1998). EFTA Fair Trade Yearbook 1997 Equiterre (2002). The Conventional Coffee route and the Alternative route. Available: http://www.equiterre.qc.ca/english/coffee/index.html (23 January 2003) 65 [FAO] Food and Agriculture Organization of the United Nations (1998). FAO Trade Yearbook Vol. 52, 1998 [FAO] Food and Agriculture Organization of the United Nations (2001). Policy for International Competitiveness of Agricultural Products of Viet Nam [FLO] Fairtrade Labelling Organization International (2002). Sales Volumes per Product per Year (in tones), 1997-2001 [FLO] Fairtrade Labelling Organization International (2002a). The Fairtrade Labelling Partners. Available http://www.fairtrade.net/en/how_does_flo_work.html (11 December 2002) [FLO] Fairtrade Labelling Organization International (2002b). Fairtrade Standards. Available http://www.fairtrade.net/standards.html (11 December 2002) [FLO] Fairtrade Labelling Organization International (2003). National Initiatives. Available: http://www.fairtrade.net/sites/contact/ni.htm (29 January 2003) [FLO] Fairtrade Labelling Organization International (2003a). Facts and Figures. Available http://www.fairtrade.net/sites/impact/facts.htm (29 January 2003) Galarraga, I. and A. Markandya (2000), The Use of Hedonic Methods to Evaluate the Economics of Eco-labelling. A Case Study for the United Kingdom, Paper for Royal Economic Society 2000 Conference. Gardner, B.L. (1975), The Farm-Retail Price Spread in a Competitive Food Industry. American Journal of Agricultural Economics, 57 (1975), pp. 399-409. Gordon, D. V. and T. Hazledine (1996), Modelling Farm-Retail Price Linkage for Eight Agricultural Commodities. Agriculture and Agri-Food Canada Policy Branch Economic and Policy Analysis Directorate, Technical Report # 1/96 Hassan, Z.A. and S.R. Johnson (1976), Consumer Demand for Major Foods in Canada. Agriculture Canada, Economics Branch Publication No.76/2, Apr. 1976 . 66 Heien, D. and G. Pompelli (1989), The Demand for Alcoholic Beverages: Economic and Demographic Effects. Southern Economic Journal, 55, pp. 759-770. Holloway, G. J. (1991), The Farm-Retail Price Spread in an Imperfectly Competitive Food Industry. American Journal of Agricultural Economics, 73 (1991), pp. 979-989. Huang, C. J., J. J. Siegfried and F. Zardoshty (1980), The Demand for Coffee in the United States, 1963-77. Quarterly Review of Economics and Business, 20(2), pp. 36-50. [ICO] International Coffee Organization (1998). Indonesia - September 1998. Available: http://www.ico.org/libser/indone.htm (3 December 2002). [ICO] International Coffee Organization (2000). Imports of non-members countries. Available: http://www.ico.org/frameset/traset.htm (13 June 2003) [ICO] International Coffee Organization (2001a), Review of the coffee market situation. September 2001, ICC 84-3, Economic Review. [ICO] International Coffee Organization (2001b). Imports of importing members. Available: http://www.ico.org/frameset/traset.htm (14 December 2002) [ICO] International Coffee Organization (2002). Supply data of exporting members. Available: http://www.ico.org/frameset/traset.htm (11 December 2002) [ICO] International Coffee Organization (2003a). Prices paid to growers in exporting Member countries (monthly) and ICO Indicator Prices (monthly averages). Available http://www.ico.org/frameset/traset.htm (8 April 2003) [ICO] International Coffee Organization (2003b). The Coffee Crisis Available: http://www.ico.org (15 July 2003) Jones, S. and B. Bayley (2000), Fair Trade: Overview, Impact, Challenges Study to Inform DFLD's Support to Fair Trade. Oxford, Oxford Policy Management. 67 Kraft (2003). Maxwell House Coffee Break: Steps to Coffee. Available: http://www.b-aftfoods.com/maxwellhouse/cschoolb.html (1 March, 2003) Maizels, A., Bacon, R. and G. Mavrotas (1997), Community Supply Management by Producing Countries: A Case-Study of the Tropical Beverage Crops. Clarendon Press, Oxford 1997 Mendoza, R. (2000), The Hierarchical Legacy in Coffee Commodity Chains. In Rural Development in Central America. Edited by Ruben, R. and J. Bastiaensen, New York, St. Martin's Press McStocker, R. (1987), The Indonesian Coffee Industry. Bulletin of Indonesian Economic Studies, Vol.23, No.l, April, pp. 40-69. Oxfam America (2002). Mugged: Poverty in your coffee cup Available: http://www.oxfamamerica.org/campaigncoffee/art3395.html (5 December 2002) Oxfam Canada (2003), Fair Trade Coffee Workshop. Available: http://www.oxfam.ca/campaigns/downloads/Coffeekit.pdf (6 March 2003) Ponte, S. (2001), The 'Latte Revolution'? Winners are Losers in the Restructuring of the Global Coffee Market Chain. Working Paper Sub-Series on Globalization and Economic Restructuring in Africa, CDR Working Paper 01.3 Raynolds, L.T. (2000), Re-embedding global agriculture: The international organic and fair trade movements. Agriculture and Human Values 17: 297-309 Rice, R. A. (2001), Noble Goals and Challenging Terrain: Organic and Fair Trade Coffee Movements in the Global Marketplace. Journal of Agricultural and Environmental Ethics 14: 39-66, 2001 Rice, P.D. and J. McLean (1999), Sustainable Coffee at the Crossroads. A White Paper Prepared For: The Consumer's Choice Council 68 Ronchi, L. (2002), The Impact of Fair Trade on Producers and Their Organizations: A Case Study with COOCAFE in Costa Rica. PRUS (Poverty Research Unit at Sussex) Working Paper No. 11 University of Sussex Talbot, J.M. (1997), Where does your coffee dollars go?: The Division of Income and Surplus along the Coffee Commodity Chain. Studies in Comparative International Development, Vol.32, No.l, pp. 56-91. Transfair Canada (2002). Facts and figures on fair trade in Canada. Available: http://www.transfair.ca/ftw/statistiques.html (11 December 2002). Transfair USA (2002a). Fair Trade Coffee: A History of Success and a Bright Future. Available: http://www.transfairusa.org/why/coffee.html (11 December 2002) Transfair USA (2002b). Fact Sheet. Available: http://www.transfairusa.org/update/facts.html (11 December 2002). Viani, R. (2002), Global Perspectives in Coffee Quality Improvement. AU Journal of Technology, volume 6(1), July 2002 Waridel, L. (2002), Coffee with Pleasure: Just Java and World Trade. Black Rose Books 2002, Montreal/ NewYork/London Wohlgenant, M.K. (1989), Demand for Farm Output in a Complete System of Demand Functions. American Journal of Agricultural Economics, 71 (1989), pp. 241-252. Zehner, D.C. (2002), An Economic Assessment of "Fair Trade" in Coffee. Chazen Web Journal of International Business Fall 2002. Zhao, X., G. R. Griffith and J. D. Mullen (1998), The Competitive Structure of the Australian Beef Industry: Accounting for Trade. Working Paper in Econometrics and Applied Statistics, The University of New England, Oct. 1998. 69 APPENDIX 1; Necessary and Sufficient Conditions for Perfect Competition of the market In this appendix, the necessary and sufficient conditions for the perfectly competitive market are shown, based on procedures from Holloway (1991). Similar notations as those used in his paper are used here. Suppose n firms face the following demand schedule. x = D(Px;N) (1) x — quantity of a food product, P x = price of a food product x and N = an exogenous variable which shifts the retail demand for a food product and x is an aggregate output of each firm i e {1,2, , n) n X = YJK (2) 1=1 Suppose each firm forms its conjecture x = k(xi), about an aggregate and its own output level, and the firm's first-order condition yields Px(\ + Oilrj) = MC() (3) 6i = (dki(xi) I dxi)(xi I x) = elasticity of industry output conjectured by firm i rj = (SD(-) / dPx)(Px I x) = price elasticity of demand MC(-) = firm's marginal cost and MC is defined over the price vector (Pa,Pb), which corresponds to a linear homogeneous technology in two variable inputs aj and bj. 70 aj = the firm's input quantity of a farm commodity with price P a bj = the firm's marketing service inputs with price P b As for equation (3), Qi = 0 holds when the market is perfectly competitive, and 6i = 1 holds when the market is monopolistic. Thus 6i e [0,1] In the following appendix, the assumption is made that the symmetric equilibria for each incumbent firm produces the same level of output (which also indicates that each firm possesses an identical technology and produces a homogenous product). Moreover, MC(-),Px,r/ and 6 become common to all firms, and the output of each firm is as follows: xi — XJ = x„ = xln Vz, j e {1,2, ,n) Then, firms are assumed to enter the industry until profits of firms become exactly zero. Pxxn-MC()xn-K = 0 (4) K = nonzero fixed costs From equation (3) and (4), the following may be obtained: 0 - -rjoh, where COK S KI PxXn denotes the share of fixed costs in food product revenue. The above relationship implies that the perfect competitive situation, 0 = 0, can be obtained when either demand is perfectly inelastic, 77 = 0, or the fixed costs are equal to 71 zero. In addition to above equations (l)-(4), the following six equations hold at the equilibrium. Equation (5) and (6) represent input demands for two factors; the farm commodity and the marketing input, respectively, and from the symmetric assumption, equations (7) and (8) are obtained. Moreover, equation (9) and (10) indicate inverse-supply relations for the farm commodity and the marketing input, respectively. a„ = dMC(-)x„/dPa (5) bn = dMC(-)xn/dPb (6) a = nan (7) b = nbn (8) Pa = h(a;W) (9) Pb = g(b;T) (10) W = an exogenous shifter of farm commodity supply T = an exogenous shifter of marketing inputs Then, allowing for above equations (1)-(10) to be displaced by changes in exogenous variables, the equilibrating adjustments may be solved in each of the endogenous variables. To obtain these effects, properties of the dual unit cost function are used, and express the displaced system of equations in percent-change terms as follows: X*^J]PX*+J]NN* (11) 72 x* = n*+Xn* (12) [(// 1(6 + 77)) - (61(6 + rj))yP,]Px * -(61(6 + rj))^N * +(61(6 + rj))(\ + y™)xn * = COaPa * +CObPb * Px * -(6 I lj)Xn * ~((6 + 7]) I ri)(DaPa * ~((6 + 7]) I T])G)bPb* - 0 (14) an* = (DbGPb * -CObCjPa * +Xn * (15) bn* = OJa CiPa * -<X>a (jPb * + Xn * (16) a* = n*+an* (17) b* = n*+bn* (18) Pa* = (l/ea)a*+ewW* (19) Pb* = (\/eb)b*+eTT* (20) Asterisk (*) indicates proportional changes (x* = Axf x). TJN = elasticity of £>(•) with respect to N coa and m = cost shares a = the elasticity of substitution between farm commodities and marketing inputs ea and eb = input supply elasticities ew = the elasticity of h(-) with respect to W er = the elasticity of g(-) with respect to T YJ, j e {Px, N, Xn} represent elasticities derived from second-order differentiation of the demand and conjectural-variations functions with respect to Px, N and xn yPx = (d2D(-) I dPx2)(Px /(SD(-) / dPx)) YN = (d2D(-) I dPxdN)(N /(dD(-) I dPx)) yx„ = (d2K(-)/dxn2)(dK(-)/dxn)) 73 Next, Holloway (1991) introduces assumptions that are useful in the empirical analysis. The first assumption is that retail demand shifter N variable is expressed in percent-change terms as follows: N* = exjPj * +exyY * +Pop * je {1,2, ,m;j*m) j Pj = prices of other products exj = cross-price elasticities of other products Y - per capita income exy = income elasticity of demand Pop = the total consuming population This assumption and the specification of the demand shifter means that both TJN and yn are equal to one. The second assumption is that the supply of the farm commodity is exogenous. Holloway (1991) uses results of Hausman's test implemented by Wohlgenant (1989) that shows that the supply of farm commodities is predetermined over the annual periodicity of the data. This assumption implies that we can delete equation (19) and treat a* as being exogenous. The third assumption means that the supply of marketing inputs to the food industry is perfectly elastic, which leads to the deletion of equation (20) and enables the price of marketing inputs, P b * to be considered as exogenous. By combining the above assumptions with equations (11)-(20), Holloway (1991) presents the elasticities of the retail-farm price ratio, R = Px/Pa, with respect to three exogenous variables, N, a and Pb. Each of the elasticities are as follows: 74 ER, N = [(2 + yxn)coa + (01(0 + T]))o-a)b] IO - [(2 + yXn)(r/ 1(6 + rj)) - (61(0 + rj))ypx + (07]/(O + r]))]/O ( 2 1 ) ER, a = -[(2 + yxn)C0a\IO + [(2 + yxn)(r\ 1(0 + rj)) - (01(0 + Jj))yPx] I <D (22) ER, k = [(2 + yxn)cjCQb] IO - [(2 + 74,1)77 + [(771(0 +1]))(2 + yxn) - (01(0 + J]))ypx]o-]cob I <D (23) where O = [(rj 1(0 + rj))(2 + yxn) - (01(0 + r]))ypx](JCOb - r]C0a(2 + yxn) The first terms on the right-hand sides of the above equations (21)-(23) indicate impacts of N, a, and P b on P x , namely EP*, N , EPX, a and EPX, Pb, respectively. Likewise, the second terms represent effects of N, a, and P b on P a , namely £>„, N , EP,, a and EP„, />*. (Proposition for the Perfect Competitive Market) Holloway (1991) presents the following proposition: Necessary and "almost" sufficient conditions for perfect competition in the retail food markets are:(i) EP«, N - -EP*. a, (ii) EPX, N = -EPx, a and (iii) ER, N = -ER, a. (The Proof of Necessary Conditions for the above Proposition) The proof of the above proposition is as follows; the necessity of (i)-(iii) can be easily proved by substituting 0 = 0 in equation (21)-(23). Thus conditions (i)-(iii) are necessary for perfect competition. (The Proof of Sufficient Conditions for the above Proposition) Next, Holloway (1991) shows the sufficiency of each condition. By putting F(-) = (2 + yxn)(t] 1(0 + 77)) - (01(0 + rf))yp* in the second terms of the right-hand 75 sides of both equation (21) and (22), EP„ N = F(-) /O + (0t] 1(0 + 77)) / Q and - EP„, a = F(-) IO may be obtained; with a result indicating that EP„, N = -Epa, a is sufficient for #77 = 0, namely 0 = 0 or 77 = 0, or both. As shown above, however, from equations (3) and (4), 0 = 0 can be obtained when demand is perfectly inelastic, namely when 77 = 0 holds. Hence the condition (i) is sufficient for perfect competition. Likewise, by defining J(-) = (2 + yxn)o)a and using the first terms of the right-hand sides of equation (21) and (22), the result is EP,, N = J(-) /O + (0/(0 + T]))acob IO and - EP,, a = J(-) IO. Since co ^ 0 from the assumption, either 0 or cr, or both will equal zero. Although the empirical analysis by Wohlgenant (1989) suggests that a * 0, the possibility also exists that a is equal to zero. When a = 0 holds, another restriction appears; that is EPX, P* = 0, as derived from equation (23). Therefore, not only is condition (ii) required, but a * 0, namely Epx,pb * 0 is also required to work as the sufficient condition forr? = 0. Thus, Holloway calls only condition (ii) as being "almost sufficient" for competition. Moreover, condition (iii) can be proven to be "almost sufficient" when considering the following: ER, Z = EP,, Z - EP„, Z Z e {N, a, Pb} 76 APPENDIX 2: Changes in Each Variable This Appendix presents graphical representations of changes in each variable used in the empirical analysis to examine the Canadian coffee sector in Chapter 4. • 3 H .1 H 2 oH - 1 H - 3 H 1 9 9 5 m 2 2 0 0 0 m 1 2 Month Figure BI: Percent Changes in Market Margin (AlnM) - . 0 5 H 1 9 9 5 m 2 2 0 0 0 m 1 2 Month Figure B2: Percent Changes in Retail Price (AlnPr) -3 H 1 9 9 5 m 2 2 0 0 0 m 1 2 Month Figure B3: Percent Changes in Import Price (AlnP f) . 0 0 2 H . 0 0 1 5 -.001 -. 0 0 0 5 H 1 9 9 5 m 2 2 0 0 0 m 1 2 Month Figure B5: Percent Changes in Retail Demand Shifter (A In RD (elasticity = 0.1107)) •4H !_! n ! , !— 1 9 9 5 m 2 2 0 0 0 m 1 2 Month Figure B6: Percent Changes in Import Supply ( A l n Q ) 79 APPENDIX 3: Weights for Marketing Inputs Costs To obtain the marketing inputs costs variables, a weighting scheme is used (adopted by Gordon and Hazledine, 1996). First, the data for "labor cost" (CANSEvI#L57740) and "fuel and power" (CANSIM#E13225) are log-differenced to show the percentage change figures; then these percentage change figures are weighted by using the weight listed below. Input Data Weight Labor Cost 0.625 Fuel and Power 0.127 Next, how these weights are obtained is shown, with reference to the data for the coffee and tea manufacturing industry in the US. According to the "1997 Economic Census -Manufacturing Industry Series - Coffee and Tea Manufacturing," costs for all establishments in the coffee and tea manufacturing industry in 1997 are according to Table CI. The cost of material consumed and cost of resale, which relates to other input, is subtracted from the total costs, to give the total marketing inputs costs: US$ 406,907,000. By dividing "Production worker wages" by the total marketing input costs yields the weight for labor cost (0.625). Likewise, the sum of "Cost of fuels" and "Cost of purchased electricity" are divided by the total marketing input costs to give the weight (0.127) for fuel and power. 80 Table C l : Costs of Coffee and Tea Manufacturing Industry in the US (1997) Types of Costs (inUS$ 1,000) Production workers wages 254,360 Total Cost of Materials Cost of materials consumed Cost of resales Cost of fuels Cost of purchased electricity Cost of contract work 4,396,045 4,142,110 188,816 21,611 30,163 13,345 Total Rental Payments 30,727 Cost of purchased services for the repair of buildings and other structures 4,415 Cost of purchased services for the repair of machinery and equipments 21,053 Cost of purchased communications services 5,919 Cost of purchased legal services 3,687 Cost of purchased accounting and bookkeeping services 1,633 Cost of purchased advertising services 16,163 Cost of purchased software and other data processing services 2,015 Cost of purchased refuse removal (including hazardous waste) services 1,816 (Source: 1997 Economic Census - Manufacturing Industry Series) 81 APPENDIX 4: Regression Results using Nominal Values Table D1-D3 report results of each regression in Chapter 4. Sensitivity analysis is adopted to examine the impact of income elasticity by changing it from -0.1893 to 0.5107. Income elasticities used in this sensitivity analysis are -0.1893, -0.0893, 0.0107, 0.1107, 0.2107, 0.3107, 0.4107, and 0.5107 referring to previous researches. Table DI: Estimates of the Unrestricted Holloway Model using Nominal Values 1) Income Elasticity = -0.1893 Ratio Equation Retail Price Equation Import Price Equation Variables (AlnM) (AlnP r ) (AlnP f) A l n M C -0.5599 -0.2688 0.2911 (-0.69) (-1.57) (0.38) AlnRD 15.3630 0.6254 -14.7376 (0.77) (0.15) (-0.78) AlnQ 0.2951 0.0052 -0.2899 (3.93) (0.32) (-4.05) Constant 0.0020 -0.0032 -0.0052 (0.16) (-1.20) (-0.44) No. Observation 71 71 71 R-squared 0.2092 0.0365 0.2193 DW statistics 1.9930' 1.7445 1.9847 AIC -1.963 -5.068 -2.055 * t-ratios in brackets 2) Income Elasticity = -0.0893 Ratio Equation Retail Price Equation Import Price Equation Variables (AlnM) (AlnP r ) (AlnP f) A l n M C -0.5625 -0.2683 0.2943 (-0.69) (-1.56) (0.38) AlnRD 22.6391 1.1988 -21.4403 (0.77) (0.19) (-0.76) AlnQ 0.2933 0.0049 -0.2883 (3.88) (0.31) (-4.00) Constant -0.0059 -0.0037 0.0023 (-0.30) (-0.87) (0.12) No. Observation 71 71 71 R-squared 0.2091 0.0367 0.2190 DW statistics 1.9947 1.7453 1.9858 AIC -1.963 -5.068 -2.055 * t-ratios in brackets 82 3) Income Elasticity = 0.0107 Ratio Equation Retail Price Equation Import Price Equation Variables (AlnM) (AlnP r ) (AlnP f) A l n M C -0.5890 -0.2687 0.3203 (-0.73) (-1.57) (0.41) AlnRD 25.1066 2.2490 -22.8576 (0.57) (0.24) (-0.54) AlnQ 0.2954 0.0047 -0.2907 (3.89) (0.29) (-4.00) Constant -0.0135 -0.0048 0.0087 (-0.36) (-0.61) (0.24) No. Observation 71 71 71 R-squared 0.2060 0.0370 0.2157 DW statistics 1.9857 1.7467 1.9756 AIC -1.959 -5.068 -2.050 * t-ratios in brackets 4) Income Elasticity = 0.1107 (using White's Correction for Ratio and Import Price Equation) Ratio Equation Retail Price Equation Import Price Equation Variables (AlnM) (AlnPr) (AlnP f) A l n M C -0.6159 -0.2714 0.3444 (-0.71) (-1.59) (0.46) AlnRD 0.0843 1.4770 1.3927 (0.00) (0.18) (0.05) AlnQ 0.3054 0.0052 -0.3003 (3.26) (0.33) (-3.43) Constant 0.0070 -0.0046 -0.0115 (0.20) . (-0.50) (-0.34) No. Observation 71 71 71 R-squared 0.2021 0.0367 0.2122 DW statistics 1.9661 1.7454 1.9575 AIC -1.954 -5.068 -2.046 * t-ratios in brackets, Robust standard errors obtained using White's correction for heteroskedasticity for Ratio and Import Price Equation 83 5) Income Elasticity = 0.2107 (using White's Correction for all three equations) Ratio Equation Retail Price Equation Import Price Equation Variables ( A l n M ) ( A l n P r ) ( A l n P f ) A l n M C -0.6032 -0.2718 0.3314 (-0.69) (-1.58) (0.43) A l n R D -8.4102 0.4866 8.8968 (-0.36) (0.08) (0.41) A l n Q 0.3065 0.0055 -0.3010 (3.30) (0.29) (-3.43) Constant 0.0181 -0.0036 -0.0217 (0.55) (-0.43) (-0.69) No. Observation 71 71 71 R-squared 0.2034 0.0363 0.2138 DW statistics 1.9682 1.7440 1.9608 AIC -1.956 -5.068 -2.048 * t-ratios in brackets, Robust standard errors obtained using White's correction for heteroskedasticity 6) Income Elasticity = 0.3107 (using White's Correction for Ratio and Retail Price Equation) Ratio Equation Retail Price Equation Import Price Equation Variables ( A l n M ) ( A l n P r ) ( A l n P f ) A l n M C -0.5932 -0.2716 0.3216 (-0.67) (-1.57) (0.42) A l n R D -8.1348 0.1689 8.3037 (-0.46) (0.04) (0.49) A l n Q 0.3054 0.0056 -0.2998 (3.29) (0.30) (-4.24) Constant 0.0197 -0.0033 -0.0230 (0.66) (-0.46) (-0.81) No. Observation 71 71 71 R-squared 0.2046 0.0362 0.2150 DW statistics 1.9722 1.7437 1.9650 AIC -1.958 -5.068 -2.050 * t-ratios in brackets, Robust standard errors obtained using White's correction for heteroskedasticity for Ratio and Retail Price Equation 84 7) Income Elasticity = 0.4107 (using White's Correction for Ratio and Retail Price Equation) Ratio Equation Retail Price Equation Import Price Equation Variables (AlnM) (AlnP r ) (AlnP f) A l n M C -0.5875 •-0.2713 0.3162 (-0.66) (-1.57) (0.41) AlnRD -6.9546 0.0587 7.0133 (-0.51) (0.02) (0.55) AlnQ 0.3044 0.0056 -0.2988 (3.28) (0.30) (-4.23) Constant 0.0196 -0.0031 -0.0227 (0.71) (-0.49) (-0.90) No. Observation 71 71 71 R-squared 0.2053 0.0362 0.2157 DW statistics 1.9748 1.7436 1.9676 AIC -1.958 -5.068 -2.050 * t-ratios in brackets, Robust standard errors obtained using White's correction for heteroskedasticity for Ratio and Retail Price Equation 8) Income Elasticity = 0.5107 (using White's Correction for Ratio and Retail Price Equation) Ratio Equation Retail Price Equation Import Price Equation Variables (AlnM) (AlnP r ) (AlnPf) A l n M C -0.5841 -0.2712 0.3130 (-0.66) (-1.56) (0.40) AlnRD -5.9225 0.0135 5.9360 (-0.53) (0.01) (0.58) AlnQ 0.3038 0.0056 -0.2982 (3.28) (0.30) (-4.22) Constant 0.0192 -0.0030 -0.0222 (0.74) (-0.52) (-0.95) No. Observation 71 71 71 R-squared 0.2057 0.0362 0.2162 DW statistics 1.9764 1.7435 1.9693 AIC -1.959 -5.068 -2.051 * t-ratios in brackets, Robust standard errors obtained using White's correction for heteroskedasticity for Ratio and Retail Price Equation 85 The following Table D-2 shows the results of the estimation of the Restricted Holloway Model in Chapter 4. Table D-2: Estimates of the Restricted Holloway Model using Nominal Values 1) Income Elasticity = -0.1893 (using White's Correction for all three equations) Ratio Equation Retail Price Equation Import Price Equation Variables ( A l n M ) ( A l n P r ) ( A l n P f ) A l n M C -0.6165 -0.2711 0.3454 (-0.72) (-1.59) (0.47) A l n R D -0.3055 -0.0056 0.2999 (-3.27) (-0.30) (3.39) A l n Q 0.3055 0.0056 -0.2999 (3.27) (0.30) (-3.39) Constant 0.0071 -0.0030 -0.0101 (0.67) (-1.43) (-0.99) N o . Observation • 71 71 71 R-squared 0.2018 0.0362 0.2119 D W statistics 1.9649 1.7435 1.9566 A I C -1.982 -5.096 -2.074 * t-ratios in brackets, Robust standard errors obtained using White's correction for heteroskedasticity 2) Income Elasticity = -0.0893 (using White's Correction for all three equations) Ratio Equation Retail Price Equation Import Price Equation Variables ( A l n M ) ( A l n P r ) ( A l n P f ) A l n M C -0.6163 -0.2711 0.3452 (-0.72) (-1.59) (0.47) A l n R D -0.3055 -0.0056 0.2999 (-3.27) (-0.30) (3.39) A l n Q 0.3055 0.0056 -0.2999 (3.27) (0.30) (-3.39) Constant 0.0072 -0.0030 -0.0102 (0.68) (-1.43) (-1.00) N o . Observation 71 71 71 R-squared 0.2019 0.0362 0.2120 D W statistics 1.965 1.7435 1.9570 A I C -1.982 -5.096 -2.074 * t-ratios in brackets, Robust standard errors obtained using White's correction for heteroskedasticity 86 3) Income Elasticity = 0.0107 (using White's Correction for all three equations) Ratio Equation Retail Price Equation Import Price Equation Variables ( A l n M ) ( A l n P r ) ( A l n P f ) A l n M C -0.6160 -0.2711 0.3449 (-0.72) (-1.59) (0.46) A l n R D -0.3055 -0.0056 0.2999 (-3.27) (-0.30) (3.39) A l n Q 0.3055 0.0056 -0.2999 (3.27) (0.30) (-3.39) Constant 0.0073 -0.0030 -0.0103 (0.69) (-1.42) (-1.00) No. Observation 71 71 71 R-squared 0.2020 0.0362 0.2121 D W statistics 1.9656 1.7435 1.9574 AIC -1.982 -5.096 -2.074 * t-ratios in brackets, Robust standard errors obtained using White's correction for heteroskedasticity 4) Income Elasticity = 0.1107 (using White's Correction for all three equations) Ratio Equation Retail Price Equation Import Price Equation Variables ( A l n M ) ( A l n P r ) ( A l n P f ) A l n M C -0.6158 -0.2711 0.3447 (-0.72) (-1.59) (0.46) A l n R D -0.3055 -0.0056 0.3000 (-3.28) (-0.30) (3.40) A l n Q 0.3055 0.0056 -0.3000 (3.28) (0.30) (-3.40) Constant 0.0074 -0.0030 -0.0104 (0.70) (-1.42) (-1.01) No. Observation 71 71 71 R-squared 0.2021 0.0362 0.2122 DW statistics 1.9660 1.7435 1.9577 AIC -1.983 -5.096 -2.074 * t-ratios in brackets, Robust standard errors obtained using White's correction for heteroskedasticity 5) Income Elasticity = 0.2107 (using White's Correction for all three equations) Ratio Equation Retail Price Equation Import Price Equation Variables ( A l n M ) ( A l n P r ) (AlnPf) A l n M C -0.6155 -0.2711 0.3444 (-0.72) (-1.59) (0.46) A l n R D -0.3056 -0.0056 0.3000 (-3.28) (-0.30) (3.40) A l n Q 0.3056 0.0056 -0.3000 (3.28) (0.30) (-3.40) Constant 0.0074 -0.0030 -0.0104 (0.70) (-1.42) (-1.02) No. Observation 71 71 71 R-squared 0.2022 0.0362 0.2123 D W statistics 1.9663 1.7435 1.9580 AIC -1.983 -5.096 -2.074 * t-ratios in brackets, Robust standard errors obtained using White's correction for heteroskedasticity 87 6) Income Elasticity = 0.3107 (using White's Correction for all three equations) Ratio Equation Retail Price Equation Import Price Equation Variables ( A l n M ) ( A l n P r ) ( A l n P f ) A l n M C -0.6153 -0.2711 0.3442 (-0.72) (-1.59) (0.46) A l n R D -0.3056 -0.0056 0.3000 (-3.28) (-0.30) (3.40) A l n Q 0.3056 0.0056 -0.3000 (3.28) (0.30) (-3.40) Constant 0.0075 -0.0030 -0.0105 (0.71) (-1.42) (-1.03) No. Observation 71 71 71 R-squared 0.2023 0.0362 0.2124 DW statistics 1.9666 1.7435 1.9584 AIC -1.983 -5.096 -2.074 * t-ratios in brackets, Robust standard errors obtained using White's correction for heteroskedasticity 7) Income Elasticity = 0.4107 (using White's Correction for all three equations) Ratio Equation Retail Price Equation Import Price Equation Variables ( A l n M ) ( A l n P r ) ( A l n P f ) A l n M C -0.6150 -0.2711 0.3439 (-0.72) (-1.59) (0.46) A l n R D -0.3056 -0.0056 0.3000 . (-3.28) (-0.30) (3.40) A l n Q 0.3056 0.0056 -0.3000 (3.28) (0.30) (-3.40) Constant 0.0076 -0.0030 -0.0106 (0.72) (-1.42) (-1.04) No. Observation 71 71 71 R-squared 0.2024 0.0362 0.2125 DW statistics 1.9669 1.7435 1.9587 AIC -1.983 -5.096 -2.074 * t-ratios in brackets, Robust standard errors obtained using White's correction for heteroskedasticity 8) Income Elasticity = 0.5107 (using White's Correction for all three equations) Ratio Equation Retail Price Equation Import Price Equation Variables ( A l n M ) ( A l n P r ) (AlnPf) A l n M C -0.6147 -0.2711 0.3437 (-0.72) (-1.59) (0.46) A l n R D -0.3056 -0.0056 0.3001 (-3.28) (-0.30) (3.40) A l n Q 0.3056 0.0056 -0.3001 (3.28) (0.30) (-3.40) Constant 0.0077 -0.0030 -0.0106 (0.73) (-1.42) (-1.04) No. Observation 71 71 71 R-squared 0.2025 0.0362 0.2126 DW statistics 1.9673 1.7435 1.9591 AIC -1.983 -5.096 -2.075 * t-ratios in brackets, Robust standard errors obtained using White's correction for heteroskedasticity 88 Table D-3: Results of F-tests to examine Market Competitiveness using Nominal Values H 0 : /3jrd + /3j q — 0 H 0 : 3pnm: - 0 j 6 m, P r and P f 1) Income elasticity = -0.1893 Ratio Equation 0 . 6 2 df = ( 1 , 6 7 ) Prob > F =  0 . 4 3 2 5 Retail Price 0 . 0 2 df = ( 1 , 6 7 ) 2 . 5 4 df= ( 1 , 6 8 ) Prob > F = 0 . 8 8 1 1 Prob > F = 0 . 1 1 5 6 Import Price 0 . 6 3 df = ( 1 , 6 7 ) Prob > F = 0 . 4 3 0 3 2) Income elasticity = -0.0893 Ratio (PyPf) 0 . 6 1 df= ( 1 , 6 7 ) Equation Prob > F = 0 . 4 3 9 3 Retail Price (Pr) 0 . 0 4 df = ( 1 , 6 7 ) 2 . 5 4 df= ( 1 , 6 8 ) Equation Prob > F = 0 . 8 4 7 6 Prob > F = 0 . 1 1 5 6 Import Price (Pf) 0 . 6 0 df = ( 1 , 6 7 ) Equation Prob > F = 0 . 4 4 3 0 3) Income elasticity = 0.0107 Ratio (Pr/Pf) 0 . 3 3 df = ( 1 , 6 7 ) Equation Prob > F = 0 . 5 6 5 2 Retail Price (Pr) 0 . 0 6 df = ( 1 , 6 7 ) 2 . 5 4 df = ( 1 , 6 8 ) Equation Prob > F = 0 . 8 0 8 9 Prob > F = 0 . 1 1 5 6 Import Price (Pf) 0 . 3 0 df = ( 1 , 6 7 ) Equation Prob > F = 0 . 5 8 3 2 4 ) Income elasticity = 0.1107 Ratio Equation 0 . 0 0 df = ( 1 , 6 7 ) Prob > F = 0 . 9 9 0 1 Retail Price 0 . 0 3 df = ( 1 , 6 7 ) 2 . 5 4 df = ( 1 , 6 8 ) Prob > F = 0 . 8 5 9 0 Prob > F = 0 . 1 1 5 5 Import Price 0 . 0 0 df= ( 1 , 6 7 ) Prob > F = 0 . 9 7 1 0 5) Income elasticity = 0.2107 Ratio (PyPf) 0 . 1 2 df = ( 1 , 6 7 ) Equation Prob > F = 0 . 7 3 1 5 Retail Price (Pr) 0 . 0 1 df= ( 1 , 6 7 ) 2 . 5 4 df = ( 1 , 6 8 ) Equation Prob>F = 0 . 9 3 6 0 Prob > F = 0 . 1 1 5 5 Import Price (Pf) 0 . 1 5 df = ( 1 , 6 7 ) Equation Prob > F = 0 . 6 9 5 5 89 6) I n c o m e e las t i c i t y = 0.3107 Ratio (pyPf) 0.20 df= (1,67) Equation Prob > F = 0.6577 Retail Price (Pr) 0.00 df = (1,67) 2.54 df= (1,68) Equation Prob > F = 0.9677 Prob > F = 0.1155 Import Price (Pf) 0.22 df = (1,67) Equation Prob>F = 0.6386 7) I n c o m e e las t i c i t y = 0.4107 Ratio (P/Pf) 0.23 df= (1,67) Equation Prob > F = 0.6296 Retail Price (Pr) 0.00 df = (1,67) 2.54 df= (1,68) Equation Prob > F = 0.9842 Prob > F = 0.1155 Import Price (Pf) 0.28 df = (1,67) Equation Prob > F = 0.6016 8) I n c o m e e las t i c i t y = 0.5107 Ratio (FyPf) 0.25 df= (1,67) Equation Prob > F = 0.6165 Retail Price (Pr) 0.00 df = (1,67) 2.54 df = (1,68) Equation Prob > F = 0.9941 Prob > F = 0.1155 Import Price (Pf) 0.30 df = (1,67) Equation Prob > F = 0.5833 90 APPENDIX 5: Regression Results using Deflated Values Table E1-E3 report results of regression in Chapter 4 by using the deflated prices and costs instead of nominal prices and costs. Sensitivity analysis is also adopted to examine the impact of income elasticity by changing it from -0.1893 to 0.5107. Table E l : Estimates of the Unrestricted Holloway Model using Deflated Values 1) Income Elasticity = -0.1893 Ratio Equation Retail Price Equation Import Price Equation Variables (AlnM) (AlnP r) (AlnPf) A l n M C -0.5375 -0.2788 0.2587 (-0.66) (-1.61) (0.33) AlnRD 2.3701 -1.8050 -4.1751 (0.15) (-0.55) (-0.28) AlnQ 0.3002 0.0043 -0.2959 (3.94) (0.27) (-4.07) Constant 0.0048 -0.0037 -0.0085 (0.34) (-1.23) (-0.63) No. Observation 71 71 71 R-squared 0.2009 0.0397 0.2163 DW statistics 1.9675 1.6984 1.9626 AIC -1.953 -5.056 -2.050 * t-ratios in brackets 2) Income Elasticity = -0.0893 Ratio Equation Retail Price Equation Import Price Equation Variables (AlnM) (AlnP r ) (AlnPf) A l n M C -0.5258 -0.2763 0.2495 (-0.64) (-1.59) (0.32) AlnRD 8.0950 -1.9711 -10.0661 (0.32) (-0.36) (-0.41) AlnQ 0.2967 0.0037 -0.2930 (3.87) (0.23) (-4.01) Constant 0.0005 -0.0034 -0.0039 (0.02) (-0.77) (-0.20) No. Observation 71 71 71 R-squared 0.2018 0.0372 0.2173 DW statistics 1.9750 1.6941 1.9688 AIC -1.954 -5.053 -2.052 * t-ratios in brackets 91 3) Income Elasticity = 0.0107 Ratio Equation Retail Price Equation Import Price Equation Variables (AlnM) (AlnP r ) (AlnP f) A l n M C -0.5171 -0.2676 0.2495 (-0.63) (-1.54) (0.32) AlnRD 28.2563 3.0390 -25.2173 (0.64) (0.33) (-0.60) AlnQ 0.2927 0.0010 -0.2917 (3.88) (0.06) (-4.06) Constant -0.0164 -0.0073 0.0091 (-0.45) (-0.93) (0.26) No. Observation 71 71 71 R-squared 0.2055 0.0369 0.2196 DW statistics 1.9827 1.6994 1.9687 AIC -1.959 -5.053 -2.054 * t-ratios in brackets 4) Income Elasticity = 0.1107 (using White's Correction for Ratio and Import Price Equation) Ratio Equation Retail Price Equation Import Price Equation Variables (AlnM) (AlnP r ) (AlnP f) A l n M C -0.5557 -0.2738 0.2819 (-0.64) (-1.59) (0.37) AlnRD 14.2983 5.1637 -9^1346 (0.73) (0.84) (-0.49) AlnQ 0.3040 0.0024 -0.3015 (3.25) (0.16) (-3.42) Constant -0.0066 -0.0094 -0.0029 (-0.34) (-1.59) (-0.16) No. Observation 71 71 71 R-squared 0.2035 0.0455 0.2166 DW statistics 1.9621 1.7173 1.9484 AIC -1.956 -5.062 -2.051 * t-ratios in brackets, Robust standard errors obtained using White's correction for heteroskedasticity for Ratio and Import Price Equation 92 5) Income Elasticity = 0.2107 (using White's Correction for Ratio and Import Price Equation) Ratio Equation Retail Price Equation Import Price Equation Variables ( A l n M ) ( A l n P r ) ( A l n P f ) A l n M C -0.5598 -0.2775 0.2823 (-0.64)' (-1.61) (0.37) A l n R D 5.4851 3.0327 -2.4523 (0.44) (0.85) (-0.21) A l n Q 0.3058 0.0036 -0.3022 (3.25) (0.23) (-3.41) Constant 0.0008 -0.0078 -0.0086 (0.05) (-1.88) (-0.57) N o . Observation 71 71 71 R-squared 0.2019 0.0457 0.2156 D W statistics 1.9572 1.7165 1.9471 A I C -1.954 -5.062 -2.049 * t-ratios in brackets, Robust standard errors obtained using White's correction for heteroskedasticity for Ratio and Import Price Equation 6) Income Elasticity = 0.3107 (using White's Correction for Ratio and Import Price Equation) Ratio Equation Retail Price Equation Import Price Equation Variables ( A l n M ) ( A l n P r ) ( A l n P f ) A l n M C -0.5588 -0.2786 0.2802 • (-0.64) (-1.61) (0.37) A l n R D 2.8806 2.0251 -0.8555 (0.32) (0.83) (-0.10) A l n Q 0.3058 0.0040 -0.3018 (3.24) (0.25) (-3.40) Constant 0.0031 -0.0070 -0.0101 (0.22) (-2.03) (-0.75) N o . Observation 71 71 71 R-squared 0.2014 0.0452 0.2154 D W statistics 1.9568 1.7144 1.9481 A I C -1.953 -5.062 -2.049 * t-ratios in brackets, Robust standard errors obtained using White's correction for heteroskedasticity for Ratio and Import Price Equation 93 7) Income Elasticity = 0.4107 (using White's Correction for Ratio and Import Price Equation) Ratio Equation Retail Price Equation Import Price Equation Variables (AlnM) (AlnP r ) (AlnP f) A l n M C -0.5578 -0.2791 0.2787 (-0.64) (-1.61) (0.37) AlnRD 1.8214 1.5003 -0.3211 (0.26) (0.82) (-0.05) AlnQ 0.3056 0.0041 -0.3014 (3.24) (0.26) (-3.39) Constant 0.0041 -0.0066 -0.0107 (0.31) (-2.11) (-0.84) No. Observation 71 71 71 R-squared 0.2011 0.0448 0.2153 DW statistics 1.9569 1.7130 1.9490 AIC -1.953 -5.061 -2.049 * t-ratios in brackets, Robust standard errors obtained using White's correction for heteroskedasticity for Ratio and Import Price Equation 8) Income Elasticity = 0.5107 (using White's Correction for Ratio Equation) Ratio Equation Retail Price Equation Import Price Equation Variables (AlnM) (AlnP r ) (AlnPf) A l n M C -0.5569 -0.2793 0.2776 (-0.63) (-1.62) (0.36) AlnRD 1.2853 1.1861 -0.0992 (0.23) (0.80) (-0.01) AlnQ 0.3054 0.0042 -0.3012 (3.23) (0.27) (-4.22) Constant 0.0046 -0.0063 -0.0109 (0.37) (-2.17) (-0.83) No. Observation 71 71 71 R-squared 0.2010 0.0446 0.2153 . DW statistics 1.9572 1.7121 1.9497 AIC -1.953 -5.061 -2.049 * t-ratios in brackets, Robust standard errors obtained using White's correction for heteroskedasticity for Ratio Equation 94 Table E-2: Estimates of the Restricted Holloway Model using Deflated Values 1) Income Elasticity = -0.1893 (using White's Correction for Ratio and Import Price Equation) Ratio Equation Retail Price Equation Import Price Equation Variables (AlnM) (AlnP r) (AlnP f) A l n M C -0.5490 -0.2710 0.2780 (-0.63) (-1.57) (0.37) AlnRD -0.3034 -0.0022 0.3013 (-3.26) (-0.14) (3.43) AlnQ 0.3034 0.0022 -0.3013 (3.26) (0.14) (-3.43) Constant 0.0064 -0.0048 -0.0113 (0.61) (-2.16) (-1.11) No. Observation 71 71 71 R-squared 0.2005 0.0354 0.2152 DW statistics 1.9605 1.6944 1.9490 AIC -1.981 -5.080 -2.077 * t-ratios in brackets, Robust standard errors obtained using White's correction for heteroskedasticity for Ratio and Import Price Equation 2) Income Elasticity = -0.0893 (using White's Correction for Ratio and Import Price Equation) Ratio Equation Retail Price Equation Import Price Equation Variables (AlnM) (AlnP r) (AlnP f) A l n M C -0.5485 -0.2710 0.2776 (-0.63) (-1.57) (0.37) AlnRD -0.3033 -0.0022 0.3011 (-3.26) (-0.14) (3.43) AlnQ 0.3033 0.0022 -0.3011 (3.26) (0.14) (-3.43) Constant 0.0065 -0.0048 -0.0113 (0.61) (-2.16) (-1.11) No. Observation 71 • 71 71 R-squared 0.2005 0.0354 0.2152 DW statistics 1.9607 1.6943 1.9494 AIC -1.981 -5.080 -2.077 * t-ratios in brackets, Robust standard errors obtained using White's correction for heteroskedasticity for Ratio and Import Price Equation 95 3) Income Elasticity = 0.0107 (using White's Correction for Ratio and Import Price Equation) Ratio Equation Retail Price Equation Import Price Equation Variables (AlnM) (AlnP r ) (AlnPf) A l n M C -0.5480 -0.2709 0.2771 (-0.63) (-1.57) (0.37) AlnRD -0.3032 -0.0021 0.3010 (-3.26) (-0.14) (3.43) AlnQ 0.302 0.0021 -0.3010 (3.26) (0.14) (-3.43) Constant 0.0065 -0.0048 -0.0113 (0.61) (-2.16) (-1.12) No. Observation 71 71 71 R-squared 0.2005 0.0353 0.2152 DW statistics 1.9610 1.6943 1.9497 AIC -1.981 -5.080 -2.077 * t-ratios in brackets, Robust standard errors obtained using White's correction for heteroskedasticity for Ratio and Import Price Equation 4) Income Elasticity = 0.1107 (using White's Correction for Ratio and Import Price Equation) Ratio Equation Retail Price Equation Import Price Equation Variables (AlnM) (AlnPr) (AlnP f) A l n M C -0.5475 -0.2709 0.2766 (-0.63) (-1.57) (0.37) AlnRD -0.3030 -0.0021 0.3009 (-3.26) (-0.14) (3.43) AlnQ 0.3030 0.0021 -0.3009 (3.26) (0.14) (-3.43) Constant 0.0065 -0.0048 -0.0113 (0.62) (-2.16) (-1.12) No. Observation 71 71 71 R-squared 0.2005 0.0353 0.2152 DW statistics 1.9612 1.6942 1.9501 AIC -1.981 -5.080 -2.077 * t-ratios in brackets, Robust standard errors obtained using White's correction for heteroskedasticity for Ratio and Import Price Equation 96 5) Income Elasticity = 0.2107 (using White's Correction for Ratio and Import Price Equation) Ratio Equation Retail Price Equation Import Price Equation Variables (AlnM) (AlnP r ) (AlnP f) A l n M C -0.5470 -0.2708 0.2762 (-0.63) (-1.57) (0.37) AlnRD -0.3029 -0.0021 0.3008 (-3.26) (-0.13) (3.43) AlnQ 0.3029 0.0021 -0.3008 (3.26) (0.13) (-3.43) Constant 0.0065 -0.0048 -0.0114 (0.62) (-2.16) (-1.12) No. Observation 71 71 71 R-squared 0.2005 0.0353 0.2152 DW statistics 1.9614 1.6942 1.9504 AIC -1.980 -5.080 -2.077 * t-ratios in brackets, Robust standard errors obtained using White's correction for heteroskedasticity for Ratio and Import Price Equation 6) Income Elasticity = 0.3107 (using White's Correction for Ratio and Import Price Equation) Ratio Equation Retail Price Equation Import Price Equation Variables (AlnM) (AlnP r ) (AlnPf) A l n M C -0.5464 -0.2708 0.2757 (-0.63) (-1.57) (0.37) AlnRD -0.3027 -0.0021 0.3007 (-3.26) (-0.13) (3.43) AlnQ 0.3027 0.0021 -0.3007 (3.26) (0.13) (-3.43) Constant 0.0066 -0.0048 -0.0114 (0.62) (-2.16) (-1.12) No. Observation 71 71 71 R-squared 0.2004 0.0353 0.2153 DW statistics 1.9616 1.6941 1.9507 AIC -1.980 -5.080 -2.077 * t-ratios in brackets, Robust standard errors obtained using White's correction for heteroskedasticity for Ratio and Import Price Equation 97 7) Income Elasticity = 0.4107 (using White's Correction for Ratio and Import Price Equation) Ratio Equation Retail Price Equation Import Price Equation Variables (AlnM) (AlnP r ) (AlnPf) AlnMC -0.5459 -0.2707 0.2752 (-0.63) (-1.57) (0.37) AlnRD -0.3026 -0.0020 0.3005 (-3.26) (-0.13) (3.43) AlnQ 0.3026 0.0020 -0.3005 (3.26) (0.13) (-3.43) Constant 0.0066 -0.0048 -0.0114 (0.63) (-2.16) (-1.13) No. Observation 71 71 71 R-squared 0.2004 0.0353 0.2153 DW statistics 1.9618 1.6941 1.9511 AIC -1.980 -5.080 -2.077 * t-ratios in brackets, Robust standard errors obtained using White's correction for heteroskedasticity for Ratio and Import Price Equation 8) Income Elasticity = 0.5107 (using White's Correction for Ratio and Import Price Equation) Ratio Equation Retail Price Equation Import Price Equation Variables (AlnM) (AlnP r ) (AlnPf) A l n M C -0.5454 -0.2707 0.2747 (-0.63) (-1.57) (0.37) AlnRD -0.3024 -0.0020 0.3004 (-3.26) (-0.13) (3.43) AlnQ 0.3024 0.0020 -0.3004 (3.26) (0.13) (-3.43) Constant 0.0066 -0.0048 -0.0114 (0.63) (-2.16) (-1.13) No. Observation 71 71 71 R-squared 0.2004 0.0353 0.2153 DW statistics 1.9621 1.6940 1.9514 AIC -1.980 -5.080 -2.077 * t-ratios in brackets, Robust standard errors obtained using White's correction for heteroskedasticity for Ratio and Import Price Equation 98 Table E-3: Results of F-tests to examine Market Competitiveness using Deflated Values H 0 : $ r d + /3jq - 0 H 0 : /3prmC - 0 j e m, P r and P f 1) Income elasticity = -0.1893 Ratio Equation 0.03 df = (1,67) Prob > F = 0.8632 Retail Price 0.30 df = (1,67) 2.48 df= (1,68) Prob > F = 0.5840 Prob > F = 0.1201 Import Price 0.09 df = (1,67) Prob > F = 0.7621 2) Income elasticity = -0.0893 Ratio (Pr/Pf) 0.11 df = (1,67) Equation Prob > F = 0.7446 Retail Price (Pr) 0.13 df = (1,67) 2.48 df = (1,68) Equation Prob > F = 0.7190 Prob > F = 0.1202 Import Price (Pf) 0.18 df= (1,67) Equation Prob > F = 0.6729 3) Income elasticity = 0.0107 Ratio (P/Pf) 0.42 df = (1,67) Equation Prob > F = 0.5168 Retail Price (Pr) 0.11 df = (1,67) 2.48 df= (1,68) Equation Prob > F = 0.7454 Prob > F = 0.1203 Import Price (Pf) 0.37 df = (1,67) Equation Prob > F = 0.5433 4) Income elasticity = 0.1107 Ratio Equation 0.55 df = (1,67) Prob > F = 0.4610 Retail Price 0.71 df = (1,67) 2.47 df = (1,68) Prob > F = 0.4024 Prob > F = 0.1203 Import Price 0.25 df= (1,67) Prob > F = 0.6155 5) Income elasticity = 0.2107 Ratio (Pr/Pf) 0.22 df= (1,67) Equation Prob > F = 0.6415 Retail Price (Pr) 0.73 df = (1,67) 2.47 df= (1,68) Equation Prob > F = 0.3966 Prob > F = 0.1204 Import Price (Pf) 0.05 df = (1,67) Equation Prob > F = 0.8164 99 6) Income elasticity = 0.3107 Ratio (P,/P f) 0.13 df= (1,67) Equation Prob > F = 0.7229 Retail Price (P r) 0.69 df = (1,67) 2.47 df= (1,68) Equation Prob > F = 0.4080 Prob > F = 0.1204 Import Price (P f) 0.02 df = (1,67) Equation Prob > F = 0.8926 7) Income elasticity = 0.4107 Ratio (PyPf) 0.09 df= (1,67) Equation Prob > F = 0.7609 Retail Price (P r) 0.67 df= (1,67) 2.47 df = (1,68) Equation Prob > F = 0.4168 Prob > F = 0.1205 Import Price (Pf) 0.01 df = (1,67) Equation Prob > F = 0.9256 8) Income elasticity = 0.5107 Ratio (Pp/Pf) 0.08 df = (1,67) Equation Prob > F = 0.7805 Retail Price (P r) 0.65 df= (1,67) 2.47 df = (1,68) Equation Prob > F = 0.4231 Prob > F = 0.1205 Import Price (P f) 0.00 df = (1,67) Equation Prob > F = 0.9522 100 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0090951/manifest

Comment

Related Items