‘ " as ' i} " h: ut‘la‘. r. -4... \ ' JEWSE “3mm ‘F3' in? l ,u ‘A‘ui'u .LM.‘ -' .‘1-1 "" ‘ ‘. my.- “3: ”52““; L-“L'Ir. ,i‘. 3 5| 10‘ D: 7% s...‘ n 1 3 y 4, ‘ o'a-r'Ll-‘tw . L“. . ._., u 3‘9““ '- a? man‘s- .‘o a :55}: L-L- LI. v3? ,5... . m. A "333;? 1% ‘3, .. an?» “a... “I“... fl», .' *‘1 u-qucqu 4.. _, ‘ a??? ‘ , “a u} " ,2-v - 7.?" l1: Y flaw t 93;?" , a .x: 'A 9 it}: {'7’ THESE I2 2004. 9490170 523’ This is to certify that the thesis entitled AN ECONOMIC ANALYSIS OF THE MEXICAN DAIRY SECTOR AND PROSPECTS FOR US. DAIRY EXPORTS presented by MIGUEL ANGEL RAMIREZ has been accepted towards fulfillment of the requirements for the MS. degree in AGRICULTURAL ECONOMICS flflm I4 4% Major Professor’ s Si natu e 09/1 6/03 Date MSU is an Affirmative Action/Equal Opportunity Institution —.—.-.----------.-.---- .o-.-o-o-o--n-.-.-.—.-.-.-p-g-n-.-.-.----n-.-.-.-.-.-.--.---n- LIBRARY Michigan State University PLACE IN RETURN BOX to remove this checkout from your record. TO AVOID FINES return on or before date due. MAY BE RECALLED with earlier due date if requested. DATE DUE DATE DUE DATE DUE 6/01 c/CIRC/DateDuethS-pJ 5 AN ECONOMIC ANALYSIS OF THE MEXICAN DAIRY SECTOR AND PROSPECTS FOR US. DAIRY EXPORTS By Miguel Angel Ramirez A THESIS Submitted to Michigan State University In partial fulfillment of the requirements For the degree of MASTER OF SCIENCE Department of Agricultural Economics 2003 ABSTRACT AN ECONOMIC ANALYSIS OF THE MEXICAN DAIRY SECTOR AND PROSPECTS FOR US. DAIRY EXPORTS By Miguel Angel Ramirez Mexican dairy product demanded outgrew domestic quantity supplied during the 1990’s and the deficit has been filled with imports from the US, Europe, Oceania and South America. NAFTA enabled the US. to become the main dairy supplier to the Mexican market, but competition from other countries and Mexico itself has raised several questions about the prospects for US. exporters. Milk supply in Mexico was modeled using an aggregate profit function, the results suggest that milk supply is very own-price inelastic and sensitive to feed and capital cost. An assessment of dairy demand was performed considering income and population growth. Demand forecasts suggest that Mexico will have a milk deficit ranging from 4.06 to 6.10 million metric tons of milk equivalents in 2003 and 2011 respectively. Mexican dairy import behavior during the 1990’s was analyzed by estimating a Source Differentiated Almost Ideal Demand System and a marginal share analysis, the results suggest that Mexican importers differentiate dairy products by source of origin and US. will continue to be the main dairy supplier to Mexico. This is especially for cheese and other dairy products, as the US. faces no import tariffs for all dairy products, but milk powder, which faces a quota scheduled to disappear in 2008. Oceania, especially New Zealand, could potentially become the second largest supplier to the Mexican dairy market. C0pyright by MIGUEL ANGEL RAMIREZ 2003 ACKNOWLEDGEMENTS First and foremost, I would like to thank my professor and friend Dr. Christopher A. Wolf for his support and wise guidance during the duration of my grad studies, for suggesting the idea of this research and overall for encouraging me to do better. I would also like to thank my committee members Dr. Bob Myers and Dr. Kelly Raper for their patience and support during my first estimations, for their help and wise advice. Special thanks also deserve Mrs. Diane Lewis and Mr. Charles Timpko from the US. Dairy Export Council in Arlington, VA. For the firnding they provided me for this study and for sharing very valuable information for this research. My always friends at the US. Dairy Export Council in Mexico City, Mr. Rodrigo Fernandez and Mr. Larry Solberg for their support, advise, continuous interest in my studies and their unconditional friendship. I would like to acknowledge the valuable help from Dr. Don P. Blayney from the USDA-ERS for his advice and commitment helping me figure out the econometrics of this research. I would also like to thank the valuable opinions from Dr. Charles Nicholson from Cornell University for sharing his doctoral dissertation with me, Dr. Luis Garcia Hernandez from the Universidad Autonoma Metropolitana-Xochimilco in Mexico City for sharing with me his opinions about the Mexican dairy sector. I am certainly thankful to my family for their continuous support, my friends at MSU for their unconditional friendship and good times. Of course, my colleagues and friends from the Cook Hall, with whom I shared so many experiences, difficulties and good times during the completion of my studies, for providing me with an academic challenging but at the same time friendly and diverse atmosphere. Thank you all!!! iv TABLE OF CONTENTS fig; LIST OF TABLES viii LIST OF FIGURES x I. Introduction 1.1 Introduction ........................................................................... 1 1.2 Background ...................................... V ..................................... l 1.3 Research objectives .................................................................. 3 1.4 Study Outline ......................................................................... 4 II. The Mexican Dairy Sector 2.1 Introduction ........................................................................... 5 2.2 Government Programs for the dairy sector ....................................... 8 2.3 Production systems .................................................................. 9 2.3.1 The specialized system ........................................................... 11 2.3.2 The semi-specialized system .................................................... 11 2.3.3 The dual-purpose system ......................................................... 12 2.4 Dairy Marketing in Mexico ........................................................ 17 2.4.1 Formal Channels .................................................................. 17 2.4.2 Informal Channels ................................................................ 21 2.4.3 Final Distribution of Milk and Dairy Products ............................... 22 2.5 Dairy Consumption in Mexico .................................................... 23 2.6 Chapter Summary ................................................................... 30 III. Econometric Model of Mexico’s Aggregate Milk Supply and Input Demand IV. 3.1 Introduction ........................................................................... 3.2 Theoretical Considerations ........................................................ 3.3 Functional Form ..................................................................... 3.4 Data Definition and Sources ....................................................... 3.5 Estimation and Results ............................................................. 3.6 Technological Effects ............................................................... 3.7 Chapter Summary .............................. ‘ ..................................... An Econometric Model of the Mexican Dairy Imports 4.1 Introduction ........................................................................... 4.2 Model ................................................................................. 4.3 Data .................................................................................... 4.4 Estimation and Results ............................................................. 4.5 Implications .......................................................................... 4.6 Chapter Summary ................................................................... Implications for US. Exporters 5.1 Introduction .......................................................................... 5.2 Prospects for Increased Milk Production ........................................ 5.3 Prospects for Dairy Consumption ................................................ 5.4 Projections Implications ............................................................ 5.5 Prospects for US. Exporters ...................................................... 5.5.1 Prospects for Cheese ............................................................. 5.5.2 Prospects for Milk Powder ...................................................... 5.5.3 Prospects for Other Dairy Products ............................................ vi 31 32 33 37 41 49 51 53 59 62 64 69 76 77 77 80 83 85 86 87 88 5.6 Chapter Summary ................................................................... 90 VI. Summary and Conclusions 6.1 Summary and Conclusions ......................................................... 91 BIBLIOGRAPHY .............................................................................. 94 vii 1&1; 2.1 2.2 2.3 2.4 2.5 ' 2.6 2.7 2.8 2.9 2.10 3.1 3.2 3.3 3.4 3.5 3.6 4.1 4.2 4.3 LIST OF TABLES Average Annual Production per Cow .............................................. Main Characteristics of Mexican Dairy Producing Systems .................... Variable Cost Structure per milk lb ................................................ Estimates of Milk Production Cost, 1998 ......................................... Milk Uses in the Processing Industry in 1998 .................................... Leading Pasteurized Firms in Mexico .............................................. The largest Mexican Dairy Companies ............................................ Dairy Apparent Consumption ...................................................... Distribution of Consumer Expenditures in Dairy Products in 2000 ...................................................................... Weekly Household Expenditures on Dairy Products by Income Decile ..................................................................... Variables and Definitions ............................................................ Supply Response Parameter Estimates ............................................ Hypothesis Tests of Own- and Cross-Price Parameters ......................... Parameter Estimates from the Restricted Model ................................. Restricted Model Equation Results ................................................ Own- and Cross-Price Elasticities Matrix ......................................... All Dairy Shares of Main Exporters to Mexico ................................... Main Exporters to Mexico Shares by Product .................................... Tariff Schedules ....................................................................... viii Page 10 14 15 16 19 20 21 27 28 29 4o 43 44 44 45 46 57 58 64 4.4 4.5 4.6 4.7 5.1 5.2 5.3 5.4 5.5 5.6 Estimated Results from the RSAIDS Model ...................................... RSAIDS Summary Statistics ........................................................ Hicksian Elasticities of Mexican Dairy Import Demand using the RSAIDS Model ................................................ Marshallian Elasticities of Mexican Dairy Import Demand using the RSAIDS Model ................................................ Mexico’s Dairy Sector Statistics, 1990-2001 ................................................. Mexico Income and Population Statistics, 1990-2001 ........................... Results for the Log-Log Demand Function for Total Dairy in Mexico ....... Mexico Milk Deficit Projections ................................................... Exchange Rate 1990-2001 ........................................................... Expenditure, Own-Price Elasticities and Marginal Shares for Imported Dairy Products ......................................................... ix 65 68 70 71 78 81 82 83 84 87 2.1 2.2 2.3 2.4 2.5 2.6 4.1 4.2 4.3 4.4 4.5 4.6 4.7 5.1 5.2 LIST OF FIGURES Total Milk Production and Consumption in Mexico ............................. Mexico Climatic Variation Zones .................................................................. Geographic Distribution of Mexico’s Milk Production ......................... Estimated Flow and Use of Milk in Mexico in 1998 ............................ Milk Marketing Channels ............................................................ Geographic Distribution of Dairy Household Consumption ................... Total Value of US. Exports to Mexico ............................................ Total Value of Mexico Dairy Imports ............................................. Milk Powder Imports ................................................................ Dairy Imports by Category ......................................................... Import Shares of Main Exporters to Mexico ...................................... Relation between Milk Powder Prices and Mexican Cheese Production ............................................................................. Relation between Milk Powder, Cheese and Other Dairy Imports ............ Nominal and Real Milk Price ($Peso/liter) ....................................... Model of Mexican Dairy Market ................................................... 10 13 13 19 23 25 54 54 55 56 57 74 74 79 79 CHAPTER 1 INTRODUCTION W2 Mexico has been the largest foreign destination of US. dairy products since 1990. Sales of US. dairy products to Mexico in 2002 reached $250 million according to the US. Dairy Export Council (U SDEC). In the ten years since the North America Free Trade Agreement (NAFTA) was implemented, tariff rates for all US. dairy exports to Mexico, with the exception of milk powder, were lowered to zero, endowing U.S. exporters with a competitive advantage in this market’. However, despite this advantage, US. companies continue to face fierce competition from the European Union, because of high subsidies, and New Zealand, because of low cost of production. The US. dairy industry has several questions regarding domestic production, consumption, and export prospects in the Mexican market. These questions can be answered with a better understanding of dairy supply and demand in Mexico. To answer these questions, this study examines the Mexican dairy industry and provides an outlook of the prospects for US. exports to Mexico, compared to other world-class competitors. Background Many policies created in the mid 1980’s were aimed at opening the economy and have changed the structure and performance of the Mexican dairy sector. In 1986, Mexico’s accession to the General Agreement on Tariffs and Trade (GATT) was a major step towards liberalization of the agricultural sector. Prior to 1989 and as a anti- IThe milk powder tariff will be lowered to zero in 2008. Currently, there is a non-tariff quota of 40,000 MT allocated to US. Imports exceeding the quota are subjected to a 139% tariff. Under the NAFTA guidelines the US. will be the only country capable of exporting non-tariff milk powder to Mexico in 2008. inflationary policy, milk price was fixed by the government and not allowed to rise despite increasing production costs. In order to encourage domestic milk production, the fixed price policy was removed in 1989, and as a result milk production increased from roughly 5,280 million liters in 1989 to 6,176 million liters in 1992 (a 16.96% increase). The North American Trade Free Agreement (NAFTA) in 1992 was the culmination of the liberalization of the government-controlled agricultural sector in Mexico. Mexico’s accession to the Organization for Economic Development (OECD) in 1994 required a further reduction in government involvement in the dairy sector. In 1995, the National Dairy Program was implemented and resulted on further price liberalization, technological modernization, and implementation of imported milk powder quotas. These changes in policy contributed to an evolving dairy market, and especially liberalized import requirements. As a result, annual milk production grth has averaged 5.38% during the 1990’s and in 2001 , Mexico produced 9,500 million liters. However, dairy production has failed to keep up with dairy consumption, which has grown 4% per year during this period, because of population and income increases. Despite a smaller growth in the consumption than in supply rate, consumption has been bigger than supply during the period of study, creating a milk deficit This milk deficit has been covered with imports, which represent around 20% of the total dairy consumption.2 The United States has been the major supplier to Mexico, because of proximity and trade preferences brought about by NAFTA. However, the Mexican market has 2 lrnports have ranged from 15.61% in 1991 to 31.93% of total dairy consumption in 1990. The total consumption covered with imports averaged 20% for the period 1990-2001. attracted other exporters such as the European Union, New Zealand and several South American countries. The Mexican dairy sector itself has evolved and some companies and dairy producers can produce milk efficiently in terms of quality and cost. As a result of these evolution in the sector, the US. Dairy Export Council reports that the product mix exported to Mexico has changed in the 1990’s and in fact, around 85% of the US. sales to Mexico are unsubsidized products such as cheese, whey and ice cream, while milk powder imports have decreased. Research Objectives The primary goal of this research is to understand Mexican dairy production and consumption, thus prospects for US. dairy exporters. To accomplish this objective three analyses are performed: 1. To understand Mexico’s milk supply response for the period 1990 to 2001. An econometric model of Mexico’s milk output supply and inputs demand is estimated for the dairy sector in Mexico. This model focuses on understanding total milk production, allowing a better assessment of the long-term prospects for increased milk production and self-sufficiency. 2. To estimate import demand elasticities for diflerent exporters to Mexico in a way that accounts for differences in preferences for products from different sources of origin. This encompasses an analysis of the dairy import demand in Mexico during the period 1990-2002 by estimating a Restricted Source Differentiated Almost Ideal Demand System (RSDAIDS). 3. To analyze the prospects of US. exporters of dairy products to Mexico vis-c‘r-vis other exporters. By estimating forecasts based on models developed in l and 2, an analysis of the prospects for US. exporters to satisfy the import demand for dairy products in Mexico is performed. Study Outline This thesis is organized as follows. Chapter two deals with an analysis of the structure and conduct of the dairy sector in Mexico providing a historical background and information for the economic modeling efforts that follow. Chapter three presents the theoretical framework and estimation of a supply response model for the Mexican dairy sector following the work of Blayney and Mittelhammer (1990). Chapter four presents the analysis of the dairy import demand in Mexico by estimating a Source Differentiated Almost Ideal Demand System following the work of Yang and Koo (1994). Chapter five analyzes prospects for US. exporters drawing upon the results obtained from economic modeling and some opinions from specialists in Mexico. Finally, chapter six provides a summary of the results obtained and the conclusions of this research. CHAPTER TWO THE MEXICAN DAIRY SECTOR mgr; In 2001, Mexico’s economy was the ninth largest economy in the world with a GDP of $617,820 million USD. This number ranks Mexico as an upper-middle income country. Agriculture and related activities represented approximately 9% of the total GDP in 1990. However, by 2001, it represented only 3.75% of the total GDP, as the country’s economy has rapidly moved towards industry (27.5% of GDP) and services (68.8% of GDP) (Central Bank of Mexico). The objective of this chapter is to provide background information on the characteristics of Mexico’s dairy sector, which lays the groundwork for the economic analysis that follows. The first cattle in Mexico were introduced in the 16th century by Spanish settlers. The introduction of dairy cattle in Mexico was aimed to supply the urban settlements with milk and beef in the New World and the operations were located in close proximities to cities. By the beginning of the 1900’s, some improved breeds were brought into Mexico along with immigration into the western highlands giving rise to today’s second biggest milk production area in Jalisco State (Cervantes et al., 2001; Garcia Hernandez, 1996). In 1938, the Compafiia Nacional de Subsistencias Populares (CONASUPO), translated as National Company of Subsistence Commodities, was founded. CONASUPO’S goals were to increase farmer income and to provide dairy products to consumers at low cost (Garcia Hernandez, 1998). CONASUPO operated several sub- agencies that performed various marketing functions, such as storage, transportation, raw product assembly, wholesale and retail services. Among all the sub-agencies operated by CONASUPO, the most important was Leche Industrializada CONASUPO, LICONSA (Industrialized CONASUPO Milk). This agency was charged to dehydrate milk and support rural production and consumption. Among its most important roles was to guarantee a minimum farm-gate price to producers, Which sold to it, acting many times as the buyer of last resort (Garcia Hernandez, 1998; Nicholson, 1995). The dairy sector did not suffer radical transformation until the 1950’s. Mexico followed what is known as the “import substitution” development strategy, relying heavily on trade barriers to promote industrial development and relying more on the domestic rather than on the export market. This policy entailed a heavy dose of governmental regulation and accelerated the urbanization rate in Mexico but did not stress agriculture development. The growing urban population demanded more animal products than Mexican agriculture could produce with existing technologies and structure, leading to increasing food imports (Barham et al., 1994, Garcia Hernandez, 1996) By the beginning of the 1980’s, increased revenues from the oil operations together with increased government spending on social programs and agricultural subsidies led to a lack of sustained growth and high inflation. The fall in oil prices in the mid 1980’s along with a high external debt terminated the agricultural subsidy programs (Hallberg, 1992; Cranney, 1992). The Mexican government, struggling to maintain producer incentives while providing food to consumers at low prices, introduced a series of retail price controls on basic food commodities administered by government agencies. Favored urban consumers and retail price controls reduced producer incentives to expand dairy herds and milk production. (Nicholson, 1995; Hallberg, 1992). As a result, milk production declined, leading to increasing dairy imports. As the government budgetary pressures derived from the fall of oil prices, and as an anti-inflationary policy, policy-makers decided to import dairy products to meet the demand rather than increasing prices paid to milk producers. This decision was based on the availability of low cost dairy imports, resulting from heavy subsidies in industrialized countries, which had accumulated large stocks of dairy products offered for sale in international markets at low prices (Cranney, 1992). CONASUPO was in charge of procuring dairy imports, mainly skimmed milk powder (SMP) for LICONSA, which in turn reconstituted imported SMP and distributed the reconstituted milk through its own outlets, concentrated in rural areas. The SMP not employed by LICONSA social programs was auctioned to private industry at prices above the world price, because it worked as a monopoly extracting quasi-rents (Nicholson, 1995). CONASUPO was the single largest buyer of skimmed milk powder (SMP) in the world, its imports in the 1980’s represented between 15% and 21% of the world SMP supply (Garcia Hernandez, 1998). In 1986, Mexico’s accession to GATT (General Agreement on Tariffs and Trade) decreased CONASUPO’S role as dairy imports monopoly and several LICONSA facilities were sold to the private sector. This accession also required to transform import licenses into tariffs. The scheme of retail-controlled prices, aimed at providing accessible prices for consumers in urban areas, was changed to a policy known as “concerted prices” in 1988, which meant that consumer prices served as “base” prices, and the other prices in the system were determined based on traditional or “reasonable” margins negotiated by producers, processors, retailers and government in a regional basis. This policy change contributed to milk supply rebound of 10% annually in both 1990 and 1991 .1 Mexico’s accession to the OECD (Organization for Economic Development) in 1994 required increased market access and competition that further decreased CONASUPO’s role. As a result, the “concerted prices” scheme remained in effect until 1995 when price controls for all dairy products, except fluid milk because it was considered a basic food requirement, were liberalized only in specific states and then nationwide in 1998. In the same year, corruption scandals led to the total disappearance of CONASUPO. LICONSA continued and was transferred to the ministry of Social Welfare and Development (SEDESOL) with responsibility to produce reconstituted milk for lower-income people, but no longer regulating SMP imports (Garcia Hernandez, 1998; Nicholson, 1995). . Government Prgrams for the Dam Sector In order to increase milk production, some specific programs were developed by the Mexican government. The Milk Productivity Program was started in 1995 and specifically designed to support the specialized dairies. This program subsidized the acquisition of equipment and physical infrastructure, but has been very controversial, ' Milk production increased fiom roughly 5,280 million liters to 5.812 million liters from 1989 to 1990. In 1991, it increased to 6,176 million liters. because small dairy producers argue they have never been considered in the support programs aimed for the largest dairies operating under cooperative structures. The Grazing Establishment program was designed to support the free-range dairy operations in need of pasture-based forage. This program subsidizes the acquisition of fences, pumping equipment, wells, and pasture seed. The Better Cattle Program was aimed to increase the milk and beef yield. It encourages the acquisition of improved breed animals, mainly imported from the US. and Canada, and artificial insemination technology to promote genetic improvement in dairy cattle (SAGARPA, 2001; Rabobank, 2000). Production Systems Diversity characterizes the production systems in Mexico. This section describes the different systems in Mexico. Milk production in Mexico has been steadily increasing (figure 2.1) but has been unable to keep up with the demand, so imports have had to complement that deficit. Despite of the relative large total volume of milk produced, the relative productivity per cow is low as compared to other countries (table 2.1). This low productivity can be understood in terms of the diversity that describes the milk production in Mexico, where there exists a wide dispersion between the smallest and biggest operations ranging from one or two cows to herds of 3000 or 5000 cows. In 2000, there were around 1.5 million dairy operations with an average of 16 cows per farm. For comparison purposes only, the US. dairy farm had on average 100 cows per farm and the European Union had 25 cows per farm for the same year. (Rabobank, 2000). Table 2.1 Average Annual Production per Cow (milk tons/cow/year) Country 1998 1999 2000 2001 2002 United States 7.80 8.06 8.26 8.23 8.43 Canada 6.82 6.92 7.15 7.43 7.50 EuropeanUnion 5.38 5.50 5.57 5.62 5.67 Australia 4.72 4.87 5.15 4.76 4.83 Argentina 3.78 4.12 4.00 3.88 3.57 Poland 3.58 3.48 3.58 3.80 3.87 NewZealand 3.56 3.37 3.67 3.70 3.71 Russian Federation 2.28 2.37 2.47 2.64 2.70 Others 1.99 2.09 2.15 2.26 2.32 Mexico 1.27 1.32 1.37 1.40 1.41 India 1.0] 1.01 1.01 1.01 1.01 Source: United States Department of Agriculture (USDA-F AS). Figure 2.1 Total Milk Production and Consumption in Mexico l 14 ————j IMilk Production 1 B Dairy Consumption l 124777 ,,. I, 10 ~ 8 _0__ 11:: 7,; 6 u Milk Million MT 1999 2000 2001 1990 1991 1992 1993 1994 1995 1996 1997 1998 Source: SAGARPA Explaining the diversity of production requires an understanding of the diverse climates in Mexico. Figure 2.2 shows the different climates where milk production takes place. Approximately two-thirds of the country is arid or semi-arid. There are three clearly differentiated production systems: the specialized, semi-specialized and dual- purpose systems (table 2.3). The specialized system Commercial or corporate farms have been increasing in size over the last 20 years. The average of cows per farm is often between 120 and 400, but some of them have thousands of cows. Their contribution to the total milk production is generally estimated to range from 50-61% (SAGARPA, 2001; Barham et al., 1994; Rabobank, 2000; Hernandez and Del Valle, 2000). These farms have the highest yields per animal (5 to 10 tons of milk) and are largely located in the Northern arid and semi-arid parts of the country. Dobson and Procter (2002) state that the shift in production location toward the Northern part of the country “may appear less than dramatic” and that it has been accelerated in recent years. However, these shifts, like the one experienced in the US. from the East and Upper Midwest to the West do not occur rapidly. In Mexico, this process started in the 1950’s, when Northern Mexico gave up cotton production and utilized the existing irrigation infrastructure to produce inputs for milk production. These farms are often organized as cooperatives and vertical integrated. The dairy herd is mainly Holstein confinement operations with state-of-the-art technology and production comparable to farms in California and Arizona. These farms feed their animals with concentrates and have specialized labor (SAGARPA, 2001; Rabobank, 2000) The semi-specialized svstem The semi-specialized farms tend to be in semi-confinement and free range grazing. Traditionally, this system is family-owned and operated with an average herd of five to 40 cows per farm and most are Holstein, Brown Swiss or crossbreeds. They 11 produce 25% of the total milk in Mexico (Rabobank, 2000). Their milk is marketed mostly through near-by processing companies and heavily relies on the many times under-compensated family labor. Most of the milking is performed by hand and many farms do not, have cooling tanks. The primary problems associated with this system are difficulties for the milk marketing process, low quality and lack of specialized infrastructure. Although, the degree of technology varies from farm to farm, these operations are located in the temperate and semi-arid areas of the country, because of grazing areas and favorable weather (FIRA, 2001). D_ual-Purpose svstem These operations are oriented primarily towards beef production and milk is a complementary part of their income. These farms are primarily located in the Mexican tropics and produce approximately 25% of the total milk in Mexico, but their participation is steadily decreasing. The herds are small, an average of 20 animals and most of the times are crossed-breeds of Zebu-Holstein. These cows are fed with native and improved pastures, which make the quality and production of milk very seasonal. The milk is marketed directly to the consumer and consumed raw most of the time. The milk is also utilized to produce artisan cheese varieties. These farms have the lowest production cost, but lack appropriate management and infrastructure to improve their quality (F IRA, 2001; Hallberg, 1992; SAGARPA, 2001; Hernandez & Del Valle, 2000). 12 Figure 2.2 Mexico Climatic Variation of Zones I Arid and semi-arid areas I Temperate areas I Tropical areas Source: SAGARPA (2001) Figure 2.3 Geographic Distribution of Mexico’s Milk Production I 400 up to 1.250 millions tons . I 100 up to 400 million tons ‘ 5 L E] 20 up to 100 million tons I E] 3 up to 20 million tons Source: SAGARPA The opinions and expectations about the growth in milk production in Mexico are almost as diverse as the production systems. This differentiation in production and technologies can explain the low average productivity per cow in Mexico. However, there is a consensus about the potential expansion of milk production in the specialized system. A summary of the main characteristics of the production systems is displayed in table 2.2. Milk production is concentrated mainly in the Northern states that allegedly lack the resource endowments, such as water and grain to produce milk. However, technology has enabled those areas to produce milk and has given them comparative advantages in terms of physical infrastructure (Garcia Hernandez et al, 1999; FIRA, 2001; Dobson and Proctor, 2002). Figure 2.3 displays the different milk producing regions in Mexico. Table 2.2. Main Characteristics of Mexican Dairy Producing Systems Characteristics Specialized Semi-specialized Dual Purpose Location North and North North and Central Southern and Central regions regions Southeastern rea'ons 1 Climate Arid and semi-arid Semi-arid and Tropical temperate Cows/Herd (Avg) 120-400 40 20 %Total herd 14% ‘ 23% 63% Breeds Holstein Holstein and/or Zebu crossed with Brown Swiss and Holstein and Swiss some crossing with Zebu Yield JDCI' cow 5 to 10 tons 2 to 4 tons 0.54 to 0.75 tons % Of Total Approx. 50% 25% 25% Production " These values vary by source and year, so they are only approximates. "Source: Rabobank (2000), F IRA (2001), SAGARPA (2001) A description of the production systems would be incomplete without including some estimates of the production costs. Due to the wide variety of production systems, 14 there is no singular estimate that precisely identifies the cost of producing milk. An accurate picture of the cost of production would be based on regional differences, different levels of technology and size. However, this information is not available and surveys of production costs are not common in Mexico (Cranney, 1992; Barham, 1994). Furthermore, the few estimates that are available are likely biased toward the commercial farms in Northern Mexico, which do not necessarily represent the majority of the operations. Previous studies (Cranney (1992); Hallberg (1992) and Nicholson (1995)) report estimates of production costs for the beginning of the 1990’s. Table 2.3 describes the cost structure per pound of milk produced. The conclusion is that feed costs represent approximately 70% of the production costs for milk in Mexico. This is the same value reported by Cranney (1992) and Nicholson (1995). Table 2.4 describes some production costs for specialized, semi-specialized and dual-purpose dairy farms in Mexico based on a survey carried out in 1998 by the Bank of Mexico Agribusiness Research Branch (FIRA). Following the discussion about the production systems in Mexico, the next section describes the Mexican dairy marketing sector. Table 2.3. Variable cost structure per milk lb. Irgrut Percentage Grains and concentrates 54 Forage 15 Labor and Cleaning operations 3 Financial cost 10 Other Expenses 18 I"Estimates for the specialized and semi-specialized systems Source: FIRA, 2001. 15 Table 2.4. Estimates of Milk Production Cost, 1998 Specialized Traditional Dual Purpose Variable Total Top 25 % Total Top 25 % Total Top 25% Size and Investment Capacity (# cows) 400 400 20 19 55 58 Herd size of milking 252 312 11 11 - - cows Herd size capacity ratio 0.63 0.78 0.55 0.55 - - Total Investment per 2809 3178 _ 1544 2324 2297 1978 dairy cow (U SD) Production Milk per cow (lb) 15604.31 15992.33 7689.72 8496.62 1157.4 1298.5 Beef per cow (lb) - - - - 315.26 482.81 Economic variables Income per milk lb 0.1532 0.1532 0.1313 0.1681 0.1104 0.1072 (U SD) Income per beef lb - - - - 0.55 0.52 (U SD) Operation costs per lb 0.1090 0.0932 0.0918 0.0945 149* 101.2 (U SD) Financial costs per lb 0.007 0.008 0.0104 0.010 47.1“ 363* (USD) Labor units per 18 14 11.1 21.5 1.04 4.29 cow/year" Family labor units 1.16 0.59 35.5 43.6 3.45 5.5 Profitability per cow 686.6 957.5 303.5 624.3 133.1 249.4 (U SD) Returns to capital (%) 15 22 13 23 6 11 Qualitative Variables % Farms receive 80 83 24 38 - - technical assistance % Farms produce all of 41 50 42 50 - - the forage inputs % Farms integrated into 39 42 0 O - - commercialization operations " Operation and financial costs in a per cow basis (yearly average number of cows in the herd) “Labor units consist of 8 hours journeys Source: Bank of Mexico Agribusiness Development Branch (2001) 16 Daig marketing in Mexico The dairy industry in Mexico is the third most important activity in Mexico’s food sector. 'In 1999, it generated $4,300 million USD, roughly 13% of the Food Processing and Manufacturing GDP and employed around 50,000 workers (INEGI Manufacturing Census, 1999). There are two marketing channels in Mexico. The formal channel that involves the large dairy cooperatives, the dairy processing industry, and LICONSA, is similar to the US. dairy channels. The informal channel encompasses a vast number of small artisanal cheese producers and dairy processors. Figure 2.4, provides insight of how milk flows from the producer to the consumer and its estimated volumes. As contrasted with the US, much of the fluid raw milk (referred as “leche bronca” in Spanish) in Mexico is marketed through the informal channel. Different sources are concordant that approximately 30% of the milk is marketed in this way happening more often the tropics. The rest of the milk goes to the processing industry. Formal Channel The marketing process begins with the shipment of raw milk to a processing facility. In the specialized system, there are formal agreements for the recollection of milk in refrigerated tanker trucks. One interesting feature of this collection system is that large cooperatives often ship farm milk long distances to the processing facilities. Sometimes these distances are up to 1,000 miles from the farms in Northern Mexico to facilities outside Mexico City or Acapulco in the Pacific Coast. The final products are marketed by the cooperatives as their own brands, in a similar manner to the big cooperatives in the US. 17 Dairy farmers that are not members of the cooperatives usually have short-term arrangements with local and small dairy companies. The degree of technology varies accordingly to the size of the producer and the size of the processor in terms of cooling equipment and distribution. Farm milk is usually collected in milk cans or plastic containers, either by the farmer or an entrepreneur and delivered to the collection centers. Once in the collection center, the milk is weighted, filtered, chilled and often tested for fat and density2 until a tanker truck transports the milk to a central processing facility. Nestle and other large companies that acquire milk from the semi-specialized system have been installing more cooling tanks in the farms they have procured since 1994, the investment is done by the company and the farmers in different ways, for instance many times the processors set up the cooling tank for several farms and farmers receive a discounted price for their milk during an agreed time period to pay for the tank. As indicated in figure 2.4, approximately 80% of the total milk goes to private industry. However, LICONSA takes approximately 10-12% of the national production and the processing industry processes only 68% (Rabobank, 2000; F IRA, 2001). This quantity of milk is a considerable increase compared to the 58% of the milk processed by the industry in 1994. According to the Agribusiness Research Branch (F IRA), LICONSA’s and the informal channel participation are expected to decrease and leave more room to the processing industry, which often, has problems of idle capacity (Hernandez & Del Valle, 2000). Table 2.5 displays the different flows and uses of the milk processed by the industry in Mexico. 2 So much has been argued about milk adulteration with water in Mexico. It used to be a common practice, but most of the collection centers perform tests on density. Moreover, in the past, milk was paid in terms of volume and not quality, since the price of the milk was controlled. Today, milk is paid more on a quality basis with a premium if milk is cooled on-farm. l8 Figure 2.4. Estimated Flow and Use of Milk in Mexico in 1998 IVIilk Flow and Utilization I TotaIMilkAvailahility I I ' I National Production Dairy Imports 31% i 69% l | I . aw Milk and Dairy Industry uid and Milk ' anal Processing and 0069118001 P°Wd°f - . milk 1.7% 33% ro ducuon elaboration 19.62% 80.1% Lic onsa Welfare P \— Pasteurized and 7;?1ams UHT milk All of the other Industry Menard Consumption '— 24.9% I"Source: SAGARPA, 2001; Rabobank, 2000. “The percentages vary depending on the source, the values are only estimates Table 2.5. Milk Uses in the Processing Industry in 1998 Product % 0f Industry Value (million Milk dollars) Fluid milk 42% 1,788 Powder milk 15% 616.9 Cheese 1 1% 470.6 Yogurt 11% 453.4 Cream (Sour and others) 3% 131.5 Others 18% 771.6 ’Source: INEGI Manufacturing census 1998. 19 One characteristic of the formal channel is the degree of concentration, for instance the largest 17 companies in Mexico purchased around 50% of the total milk production and had 75% of the total dairy sales in 1999, suggesting that regional concentration and market power might be even greater in a regional basis. The fluid milk pasteurizing companies are the most important in Mexico employing 42% of the milk in the industry and generating 1,788 million dollars in sales (F IRA, 2001). These companies are organized in different business structures including completely integrated cooperatives, multinational companies such as Nestle, Dannon, Parmalat, Kraft Foods and New Zealand Dairy products, independent processing companies with farmer contracts, smaller-scale cooperatives and government companies such as LICONSA. In general, these companies market around 70% of the pasteurized milk in Mexico. Table 2.6, displays these companies and their respective market shares. The two biggest firms, Lala and Alpura are cooperatives completely integrated from production to retailing. Table 2.6. Leading Pasteurized Milk Firms in Mexico Company Market share percentage LALA 26 Alpura 15 Zaragoza 9 Sello Rojo 9 Boreal 6 San Marcos 6 Otros 29 Source: FIRA 2001. In terms of other manufactured dairy products, the structure is more fragmented. In the case of cheese, there are around 1,300 firms producing cheese, but most of them are small and artisanal ones. Large companies such as Nestle, some domestic companies 20 and Kraft Foods supply the biggest part of the production. These companies together supply around 65% of cheese production. Three firms, namely Dannon, Sigma Alimentos and Nestle supplied 60% of the total production of yogurt (INEGI, 1998; F IRA, 2001). This skewed distribution of the industry suggests that the Mexican dairy industry is an oligopsony with few firms leading the whole sector. Hernandez & Del Valle (2000) agree that the Mexican dairy industry is an oligopsony with a skewed technological distribution favoring the biggest firms and many times excluding the production sector. Table 2.7 shows the biggest dairy firms operating in Mexico. Nestle continues to be the largest dairy company in Mexico and US. firms such as Kraft have not reached yet the sales magnitudes to be considered among the major players in the Mexican dairy sector. Table 2.7. The largest Mexican Dairy Companies C orryrany 1998 Sales (US millions) Ownership structure Nestle Mexico $1,650 Multinational (Switzerland) Grupo Lala $778 Cooperative Grupo Alpura $486 Cooperative Grupo Zaragoza $250 Private family-owned Lechera Guadalajara $171 Private family-owned Grupo Chilchota $169 Private family-owned Sigma Alimentos $120 Private public Gilsa $1 15 Cooperative Parrnalat México $80 Multinational (Italy) Dannon $70 Multinational (France) New Zealand Dairy $45 Multinational Board" I’Source: Rabobank (2000) l""‘The New Zealand Dairy Board changed its name to F onterra in 2000 Informal Channel The “inforrna ” channel is characterized by the large number of intermediaries. These “entrepreneurs” sell milk in the nearby towns or to small-scale processing plants, 21 where it is transformed into fresh cheese or other dairy products. The technology of this informal channel is rudimentary as is the milk quality. The milk marketed through this channel comes mostly from the dual-purpose farms and it is common in the tropics and small towns. Mistribution of Milk_and Dairy Products The last part of the marketing chain is also of interest. The largest companies have their own distribution systems at a national level, providing them direct delivery and negotiation power with the retailers. It is important to mention that none of the Mexican cooperatives have national presence; their presence is rather regional. Central wholesale markets and trading firms continue to play a key role in the distribution of dairy products (U SDEC, 2000). At the retail level, significant differences exist depending on the type of product. According to the US. Dairy Export Council (U SDEC), large retail chains account for only 25% of overall food sale, wholesale markets represent 25% and small regional retailers including mom and pop stores represent 40%. Food service companies handle the remaining 10%. However, given the perishable nature of the dairy products, their distribution is made mainly through small retailers and supermarkets3. Small retailers remain the most important distribution points for fluid milk, while 80% of processed dairy products are distributed through modern retail chains (U SDEC, 2000; Rabobank, 2000). Figure 2.5 summarizes the previous discussion on milk marketing in Mexico. 3 Fluid milk distribution through retailers was heavily supported by the government to better enforce the price control policy previously described. 22 Figure 2.5. Milk Marketing Channels Specialized fl Semimddiml — Dairy ”my,“ "*1? Comnialiutin Regional IRmnilkselfi commuter eonsmtiort ofm milk Processors and milk derivative industry Geranium | L; i iii 12:: |§ asteurized hflflh‘i CE! I 5: Other oration cream “17 butter and processors ”1' why , pmesso lupus I Wholesalers and retailersI-—' Final Cenumer " Source: Adapted from Rabobank (2000) and USDEC (2001) Dairy Consumption in Mexico This section describes the consumption patterns in Mexico and it is strongly linked to the milk marketing and production sections previously described. This section is aimed to give an insight into the dairy consumption patterns in Mexico. The dairy market in Mexico is not uniform. On one hand, there is milk that is basically bought by medium and high-income consumers and there is strong competition among local and international brands. On the other hand, the social programs provide milk to the poorer 23 strata of the society and subsidize its price. This difference is caused by the different production systems, which in turn determine differences in quality and price structure (Garcia Hernandez, 1998; Nicholson, 1995). Mexico is a vast country, with varied climates and huge differences that influence consumer preferences. Figure 2.6, shows the main milk consuming regions in Mexico. These differences are explained in terms of demographic, per capita income and distribution, urbanization rate and age distribution factors. The areas where most milk is consumed in Mexico are, not surprisingly, the ones that exhibit the fastest population growth and have the highest income levels located in Central and Northern Mexico (2000 Mexican Census). These results are consistent with previous estimations done by Gould & Kim (1998) and Nicholson (1995) who mentioned that the consumption of milk in those areas is a combination of a young, rapidly growing population and increasing urbanization.4 Sixty five percent of the total dairy production is consumed in or around Mexico City, Mexico’s largest city (USDA, 2001). An estimated 30% of the. milk in Mexico, mainly from the semi-specialized and dual-purpose systems, is consumed raw. This is perhaps one of the most important consumption patterns in Mexico. Some consumers reportedly regard this raw milk as higher quality than processed milk. What accounts for the high consumption of raw milk in Mexico? Nicholson (1995) suggests this phenomenon is in part a result from the restrictive price control policies of the 1980’s. He states these controls provided incentives for farmers and entrepreneurs to market their product themselves to achieve a higher price rather than selling it to the intermediaries or processors. ‘ According to the National Institute of Geography, lnforrnatics and Statistics (INEGI) approximately 80% of Mexico’s population lives in an urban area. 24 In the same way, milk processing was more focused on those products that were not controlled at the retail level, offering very few options on fluid milk availability to consumers. Quality of the pasteurized milk and the milk provided under the social programs was regarded as low quality,5 rooting the belief that milk purchased at the farm gate was “real and fresh.” In other areas of Mexico, the lack of processing and cooling infrastructure derived the consumption of raw milk as well. Although, the price controls no longer exist, it seems that consumer habits persisted. Figure 2.6. Geographic Distribution of Dairy Household Consumption N - 707. and up of households - 50707. of the households 30-507. of the households ‘Source: Done by the author with estimates from the INEGI 2000 and 1994 Household Income and Expenditure National Survey. "The percentages represent the % of households buying dairy products. 5 This quality perception has also implications in the way import products are regarded and their success into the Mexican market despite being higher-priced than their domestic counterparts. For further discussion, see the chapter on imports demand. 25 The milk and dairy products quality have increased in recent years. However, price controls created a mix of processing technologies nearly as diverse as production technologies used on farms. These technologies were characterized by the degree of substitutability of the dairy and in many cases non-dairy ingredients such as vegetable fat for butterfat. This created a functionality-based market for dairy ingredients such as caseins, whey powder, butter fat and other imported dairy dry ingredients that are still important today. To cite an example, what is called “imitation” cheese in the US, those cheeses made of milk powder or milk proteins recombined with other ingredients such as whey powder and vegetable fat, are very popular in Mexico. It is estimated that approximately 75% of the Mexican cheeses is “imitation”. This is legal in Mexico, as long as it is labeled as “analog” cheese. Considering this information, what factors generate the consumption of dairy products? Dairy consumption in Mexico entails the final demand by consumers and industrial demand (within and outside the dairy industry‘s). Determining the apparent consumption (as a proxy for final consumption) of dairy products in Mexico is a key step to understand the prospects for Mexico future dairy demand. Table 2.7 displays the apparent consumption of different dairy products. Milk consumption is the sum of milk production and all dairy imports, expressed in total solids milk equivalents, following the approach of Barham et a1. (1994) and the conversion factors from Selinsky et al. (1992). The numbers in table 2.8 reveal some interesting facts. First, domestic production has failed to outdo the total consumption, despite its increase. Second, the role of imports to meet the final demand for dairy has been somewhat constant over the years, regardless 6 Some dairy ingredients have uses in other industrial applications. However, the quantities consumed in other industries are very small in comparison to those consumed by the dairy industry. 26 of the increase in production. The nature of the dairy consumption in Mexico requires an insight into the dairy purchasing expenditures. It is not a secret that income distribution is very uneven in Mexico. Nicholson (1995) reported that in 1992 an estimated 65% of the entire dairy was consumed by 40% of the population with the highest incomes. Raboka (2000) estimated that in 1999, 30% of the population consumed 50% of the total dairy products. This unequal distribution of the income has two important implications for the final demand for dairy products.’ Table 2.8. Dairy Apparent Consumption % 0f the % 0f the Domestic Milk All Dairy Total Dairy consumption consumption Production ImportsM Apparent supplied with supplied with (millions of (millions of Consumption domestic imported Year tons) tons) (millions of tons) production products 1990 5.81 2.73 8.54 68.07% 31.93% 1991 6.18 1.14 7.32 84.39% 15.61% 1992 6.38 2.45 8.83 72.25% 27.75% 1993 7.40 2.73 10.13 73.09% 26.91% 1994 7.32 2.29 9.61 76.17% 23.83% 1995 7.40 1.69 9.09 81.41% 18.59% 1996 7.59 1.91 9.50 79.86% 20.14% 1997 7.85 2.12 9.97 78.72% 21.28% 1998 8.32 2.02 10.34 80.45% 19.55% 1999 8.88 2.22 11.09 80.01% 19.99% 2000 9.31 2.31 11.62 80.12% 19.88% 2001 9.50 2.78 12.28 77.37% 22.63% ‘Source: Calculations made by the author with data from the Bank of Mexico, SAGARPA and USDA. "It considers fluid milk, milk powder, whey, cheese, yogurt, ice cream, condensed and evaporate milk and dairy desserts. First, low-income persons still consume dairy from subsidized social programs, mainly imported milk powder. So, as long as the social programs remain, there will be demand for fluid milk, even at the lowest income level that will be met with skimmed milk 27 powder imports and with domestic production7. Second, demand for other domestic and imported dairy foods will increase as income increases in Mexico. Considering the current marketing and production systems, it is expected that part of the final dairy demand will be in part supplied by imports.8 To corroborate the validity of these implications, data from the 2000 Household Income and Expenditure National Survey are shown in table 2.9 and 2.10. Table 2.9. Distribution of Consumer Expenditures in Dairy Products in 2000 Product Expenditure @ercentage) Fluid Milk 70 Cheese 1 1 Cream 4 Butter 1 Other 14 ‘Source: Calculations made by the author with INEGI’S data. The results in table 2.9 suggest that fluid milk accounts for the biggest expenditure in the dairy category. Table 2.10 suggests that the expenditures in dairy products increased in average 50% from the 1992 levels. The budget shares remained constant or with the exception of a decrease at the highest income levels, and an increase at the lowest income levels. These results suggest that as income increases in Mexico, dairy products consumption will increase but at different rates depending on the income level. Not surprisingly, Nicholson (1995) suggested that the biggest increase in consumption would come from the lower income strata. 7 Recently, LICONSA announced that some of its facilities would be procuring only domestic fresh milk for the social programs. a The discussion on marketing systems suggests that even if all the milk needed to meet the final demand, not all of it could be marketed to the final consumers due to the lack of contracts between producers and processors. 28 -I.III III. IliIl II III I I IIII IIII11III IIIIILI Iii .II 111 ii I111! III11.II. I II I1.1|II I -x _ . II I fig F. n. No. EoWINmF._..9_.95 Eng-:5. e.— 65 Wang... 3. 98:5 can... _ L - _-lil 11-.. 1? .I- 1- 1--1---..I.l - 1 I, .anoSu Gee—o _ _ . =_ _ E . c . <— .. <= 5: L _x Bow _ , . . _ 7 _ .. 406. h _ . L ‘ $33983 _ . L .— . . acacia... _ . . _ n 0305 . . fAm." Ambi upmd waAm AAbA. mo. Nu. mmbw. meE _ . mNAN SA. . w 403. w.cam.m,. R 9:20 _.\ . A..N A.mN A.mu. A.«A._ Amwu AQA. Amm“ mam. Nam m:um:a=c.m ‘ Hacmoiofi ~ _ mg iii-Ii I- ...... Ii. I-iI-iIliI L-- - - . ...-- u _ . ore—n. wwa Am. a. Aw.Ao__ mmm. .. 8.8? .3. _ m. mAAL mm mm. mm. um ._.o.m_ mane. . M L ._ . h 9.2me mNg Ab. Amm‘ Amm AQQ Am... A01: A. A” 9mm. New «moans. Ooficfigfi gm. 31.5 25.8. 5:. 98 m3... Ewe H1 0% BE ubgimmfiogfiwfiegifim $059.5. 2353 mega.- 3.165.. 583. gunman H _ # m «II I rlI . I 1.111.114. A _ L H 29 Summgy This chapter provided an insight on the Mexican dairy sector and the necessary background for the economic modeling efforts developed in the following chapters. It identified differences in production, marketing and consumption of dairy products in Mexico considering factors such as geographical distribution, urban growth, income, and consumer preferences. Different policies affecting the dairy sector, and their effects on the dairy sector were described. Regarding the marketing sector, it is characterized by an oligopolistic structure, where a reduced number of companies (17) concentrate more than 75% of the industry sales. Consumer preferences are shifting towards a more value- added trend bringing about consolidation within the industry and many times unequal power distribution along the marketing chain. 30 CHAPTER THREE ECONOMETRIC MODEL OF MEXICO’S AGGREGATE MILK SUPPLY AND INPUT DEMAND Introduction The Mexican dairy sector is characterized by a diversity of production and marketing systems. This diversity has characterized the behavior of milk production as “uneven” (Nicholson, 1995). The yearly rate of growth for the period 1980-1990 was on average -1.27% while for the period 1990-2001 the growth rate averaged 5.38%'. Several authors have argued that the growth of production was the result of “price liberalization” starting in 1989 (Nicholson, 1991; Cranney, 1992; Hallberg, 1992). Prior to 1989, as an anti-inflation policy, milk retail prices were fixed and not allowed to rise, placing a ceiling on the price paid to producers (Hallberg, 1992). These policies created a lack of price incentive to expand milk production and domestic production shortfalls led to increasing dairy imports. An aggregate econometric supply response model was estimated for the dairy sector in Mexico. This model was aimed at understanding the relations and effects on total milk production with respect to milk price and key production inputs. Technology is a crucial element of long-term shifts in agricultural supply. It follows that improvements in technology should be an important factor in determining the level of milk production in Mexico. ' Calculations done by the author with data from the Secretariat of Agriculture and Natural Resources (SAGARPA) 31 Theoretical considerations Following the approach of Blayney and Mittelhammer (1990), an aggregate profit function was defined for a given output, input prices and aggregate production relation. Farmers are assumed to maximize the present value of income over an infinite horizon with respect to inputs utilized to produce milk. The farm objective is first to maximize short-run profits with respect to variable inputs and then maximize the present value of long-run profits (Thij sen, 1992). According to Blayney and Mittlehammer (1990), the aggregate profit function must characterize the problem of maximizing the aggregate profit that can be generated by distributing production of output across industry firms. This can be represented as: (1) 7r(P,w)£-rqnaxPqu —ch(w,qj), ' ' " [=1 j=l where P is the output price, w is the vector of input prices, q, is the output of the jth firm, and Cj (W, q;) is the cost function of the jth firm. The optimality conditions for (1) require that each firm maximize its own profit at prices (P, w). The optimal output distribution across the firms is represented by the collection of continuously differentiable firm-level supply functions szqj (P, w), j=1, m. Substitution of the output supply functions into (1) yields: (2) 7r(P, w) = $qu (P, w) — C]. (w, q], (P, w))] = in], (p, w). j... 1.] Thus, the aggregate profit function can be represented as the sum of the individual finn- level profit functions (Blayney and Mittelhammer, 1990). By Hotelling’s Lemma, differentiating the profit function with respect to output price yields the supply function. Similarly, the negative partial derivative of the profit 32 function with respect to an input price yields the input demands: 57r(P, w) (3-1)—a—P— : Q(P,W)- (3.3—2w = ZxU.(P,w) = X,(P,w), i = 1,..n, where Q (P, w) is the aggregate milk supply, Xij (P, w) is demand for input i by firm j, and X is aggregate input demand. Functional Form Since theory does not suggest a specific functional form, applied economists are faced with the important task of selecting a functional form for the aggregate profit fimction. Resulting problems can be reduced by specifying a flexible functional form, which can be viewed as (usually second-order) approximations to a general fitnctional form. Following the work of Blayney and Mittelhammer (1990), the Box-Cox form was selected for this study. According to Genc and Bairam (1998, p. 55), the Box-Cox form is popular because it allows unrestricted substitution among inputs and satisfies the homogeneity restrictions imposed by economic theory. The Box-Cox form requires a non-linear transformation of the dependent variable (Y) in a functional relationship: Y (A) = F (X), where Y is the vector of original observations of the dependent variable transformed by the parameter A, and X is a matrix of independent variables. There are two popular forms of the Box-Cox transformation (Genc and Bairam): (4a) Y“’ = 53—22—11,; at 0; 4b ( ) lnY,A=O. 33 A generalized Box-Cox transformation of the dual profit function was specified as (Blayney, 1988): 7rP,w@—l 194—1 w’I—l ( ) =m , ). o 2. )1 (5) for M0 and ®¢O. The parameters 6) and K are the transformation parameters and h denotes a general functional relationship. By defining a vector of transformed prices Z and a form for h (Z) that includes all of the own price and cross price relationships in the following manner: <6)h(2)=(P/1".W;1), the following equation is obtained (Blayney): (7) 7r(P, w) = [1 + oh(2)]"° h(Z) = a, + a'Z + 0.52'AZ, where a0 is a constant, a’ and A the row vector and a symmetric matrix of price effect parameters denoted by: a' = (a P , aw) (8) A = [app aw] ’ awp aw where the subscripts P and w denote output price and input price respectively. Economic theory requires a profit function to be homogenous of degree one (Genc and Bahram, 1998), by substituting equation (6) into (7), one obtains: 7r(P,W)=(1+G)[ao+ap(P1 "1)+a,(WA '1)+ (9) P‘—1, W‘—l, P‘—1 W‘—1 V, 1/2a,,,,( 2. )+1/2a...( )+a,,.( .1 )( )1 )1) . 34 By assuming D0, and imposing the following restrictions: 1+ 90,: 69/1 (a,+ (1,.) (10) aP=1/21(aPP+aPW) aW=1/21(aw+am,) any = -1/2(aPP+aWW ) One can obtain the homogeneous function (Blayney): 1 P“ + —1—a,,,,,P‘W‘ + ——21 W“)]"". (11)7rP,W = G) , q ( ) [ (2,112,, )1- 210W Multiplying both P and w by a positive constant k results in: l (kP)“ + %am,(kP)‘(kW)‘ (kP,KW) = [(~)( 2 (12) 2/1 a", +5%awI, where t is the technological variable (a trend variable starting in 1988=1 and ending in 2001:14). A form g (P, W, t) that accommodates the notion of technological change is given by: 35 (15) g(P,W,t) = 2(1“ + 2y, 111(12, )), where Rj is either the price of the output or the price of an input. The parameters of the aggregate profit function are the functional parameter A, the neutral technology measure 1:, the vector of biased technology effects 7, and A, an N x N symmetric matrix of parameters associated with own- and cross-price relationships. Linear homogeneity of the aggregate profit function is imposed by: (16) ix=0, where yj’s are the elements of gamma. The profit function was not estimated directly. Instead, Hotteling’s lemma was applied, which resulted in a system of equations representing aggregate output (milk) supply and the aggregate input demands (Blayney, 1990. p. 865-866). Q :(l-1(ZvAZ))l/2/l er(r+y'R) (%) + N e,(,+r'R)(/1—l (ZvAZ))(l/2/t)-l 2:1 (alan-l + ZaleA-le_l , (17) t "=2 X, = -[(A-I(Z'AZ))l/u et(r+y'R)(h)+ W I(r+y'R) -l 21 1—1 A 1—1 2. e (’1 .(al+l,l+1W‘ + au+lwt p + Zai+l,}+le W1 )]' 1*! The 30’s denote cross-price parameters. Algebraic manipulation of the above equations allows the level of profit,7t to be substituted into (17) and the matrix A to be removed from the expressions. In particular, the supply and input demand equation can be transformed into: 36 N 77’ - r+v' - — (18) Q:”(_Fl_)+fl_l 21821.1( , R) .(aHPZIl 1+ZaUPl le4), J=2 X. = n(—W' ) + rz'H’lemmr'R) ~(a W”1 + a pr’1 ' P , 1.1+] 1 1+1,1+l 1-1 )1 + zanlJHWj W1 )' 1:! These are the forms that are estimated and the results are reported below. Qua Definition and SOurces Data quality and availability on the Mexican dairy sector are major problems when estimating a supply response function for the sector. As the years have passed by, more data on the dairy sector have become available. This model incorporates the best and most complete information available to date. Annual data were available for the period 1988 to 2001, a total of 14 observations. This period of study is important to analyze, because it includes the implementation of the milk price liberalization in 1989, the 1993 North America Free Trade Agreement (NAFTA), as well as the economic and financial crisis of 1995. The price of milk was defined as the national average of price paid for 1,000 liters of milk in 32 Mexican States collected from the Secretariat of Agriculture and Natural Resources (SAGARPA). The quantity of milk was defined as the aggregate milk production in 32 Mexican States collected by SAGARPA, defined in thousand millions of liters. The feed price and quantity variables deserve special attention. Two feeds were considered, concentrates and forage. Actual feed quantities were expressed in millions of 37 tons and collected from the General Cattle Coordination Branch of SAGARPA. 2 The feed price was defined as the average of concentrates and forage weighted by their respective quantities. Concentrate prices was defined as FOB Gulf Ports corn and oilseed paste prices reported by the US. Department of Agriculture, because most of the grains and feeds in Mexico were imported from the US The forage price was represented by the price paid for a metric ton of hay in Northern Mexico reported by SAGARPA. The cow price was the price of dairy replacement cows imported into Mexico as reported by the Food and Agriculture Organization (FAQ). The dairy herd population is expressed in millions of head and collected from the SAGARPA. The price of capital was defined as the commercial lending interest rate reported by the Central Bank of Mexico. The proxy for quantity of capital was the amount of milk machines imported into Mexico as reported by the F A0. This variable is a proxy given the lack of data on capital investments such as buildings, parlors, cooling tanks or others in Mexican dairy farms. The price of labor was defined as the minimum wage hourly remuneration reported by the Labor Secretariat in Mexico and the US. Department of Labor. The quantity of labor in Mexico’s dairy farms was estimated from the Mexican Census Department (INEGI), the Labor Secretariat and the SAGARPA, defined as millions of hours per year. This estimate included paid labor hired for dairy operations, but did not include the owner and unpaid (underpaid) family labor devoted to the farm. This labor estimate is the only estimate for Mexican dairy farms. 2 Over the period considered the average ration consisted of 40% concentrates (35% grains and 5% oilseeds) and 60% forage. 38 The equations in (18) require substituting for a profit proxy. There have been very few studies of the cost of production and management of Mexican dairy farms (Nicholson, 1995; Cranney, 1992; Barham, 1994; FIRA, 1998). All of the existing studies are biased towards the largest farms in the Northern part of the country and there are no studies or time series data on small-scale farms available. Furthermore, dairy farming profits are not positive in every period in Mexico or anywhere else, but the specification of the equations to be estimated in (18)require the profit proxy to be positive and greater than zero, so that the Box-Cox form does not yield imaginary or zero values that can cause numerical problems in the estimation.3 To overcome these problems, a study performed by the Agricultural and Food Policy Center in Texas A&M University was used. This study titled “Mexican Representative Dairy Farms 1998 Economic Outlook” describes the profitability, measured as the difference between the total cash receipts and the total cash expenses of different dairy farms in Central and ‘ Northern Mexico for the period 1988-1998 and makes projections up to 2002. According to Ochoa et a1. (1998), the following assumptions were made for the profit estimation: 0 All farms surveyed are of moderate scale to be representative of full-time commercial farming operations in the study area. - Herd size was held constant for all farms over the 1997-2002 planning horizon. o All prices were converted to US. dollars. 0 Minimum family living withdrawals for family dairies were assumed at a base rate ranging from 17 to 34 percent of gross receipts accordingly to the panel suggestions. 3 The term [121'] yields imaginary numbers if 1! is negative and 21-1