Where Does Your Coffee Dollar Go?: The Division of Income and Surplus along the Coffee Commodity Chain
John M. Talbot
This article analyzes the division of the total income and surplus generated along the coffee commodity chain during the period 1971-1995. Until the late 1980s, coffee growers and producing states retained over a third of the total income and about half of the total surplus that was available. This was due in part to the collective actions of coffee-producing states, which led to the imposition of a regulatory regime involving export quotas, creating rents for the producing countries. By the late 1980s, coffee TNCs had consolidated their control over core markets, and began to use their market power to increase their shares of both income and surplus. This shift was greatly accelerated by the breakdown of the export quota regime in 1989. The article concludes that these results necessitate a reformulation of commodity chain analysis.
ne o f the central questions in the study o f d e v e l o p m e n t is, w h e r e does the surplus go? D e p e n d e n c y theorists h a v e argued that one of the f u n d a m e n t a l causes o f u n d e r d e v e l o p m e n t in the Third W o r l d has been the extraction o f surplus f r o m activities p e r f o r m e d in the periphery and its appropriation by capitalists b a s e d in the core (e.g., E m m a n u e l 1972, A m i n 1976). T h e " E C L A s c h o o l " and structuralist e c o n o m i s t s have argued that the declining terms o f trade for the p r i m a r y c o m m o d i t y exports of the Third W o r l d h a v e hindered d e v e l o p m e n t (e.g., P r e b i s c h 1962, Singer 1950). Neoclassical theorists h a v e responded that transfers o f surplus to the core have been overstated by dependentistas and structuralists; the real c a u s e o f
John M. Talbot received his doctorate in sociology from UC, Berkeley, this year. His dissertation is titled "Grounds for Agreement: The Political Economy of the Coffee Commodity Chain." Address correspondence to Department of Sociology, University of California-Berkeley, Berkeley, CA 94720. Studies in Comparative International Development, Spring 1997, Vol. 32, No. 1, 56-91.
underdevelopment has been dissipation of the surplus in Third World countries through rent-seeking and unproductive activities (e.g., Bhagwati 1982; Bohman, Jarvis, and Barichello 1996). More recent analyses by international political-economists have focused on the organization of global production systems by transnational corporations (TNCs); in industries such as apparel and autos. These studies suggest that TNCs organize production in order to maximize their control of surplus globally, and that this surplus may be directed to either the core or the Third World, depending on the global strategy of the TNC (e.g., Lee and Cason 1994, Bonacich et al. 1994). They also acknowledge that Third World states and firms can capture a share of this surplus, and that they may use it productively, or misuse it. Despite the centrality of this question of where the surplus goes, very few studies have thus far attempted to empirically determine the amount of surplus generated by various economic activities, and its division among economic actors. This article argues that a commodity chain approach is the most appropriate method for addressing this question, presents estimates of the division of the surplus along the coffee commodity chain, and analyzes how and why this distribution changes over time. The commodity chain approach has been developed by world-systems theorists (Hopkins and Wallerstein 1986; Gereffi and Korzeniewicz 1990, 1994). A commodity chain is defined as "a network of labor and production processes whose result is a finished commodity" (Hopkins and Wallerstein 1986, 159). The chain is conceptualized as a series of nodes, linked by various types of transactions (e.g., sales in a market or intrafirm transfers). "Each successive node within a commodity chain involves the acquisition or organization of inputs (e.g., raw materials or semifinished products), labor power (and its provisioning), transportation, distribution (via markets or transfers), and consumption" (Gereffi, Korzeniewicz, and Korzeniewicz 1994, 2). At each successive node, the commodity is transformed in some way; value is added to it, and profits are generated. We must trace the chains of production processes from finished commodities back to the various raw materials with which they begin, and sum up the profits realized at each stage or node, in order to get an accurate estimate of the total surplus. The chain concept has been used previously in a variety of political-economic analyses. Bargaining power analyses of mineral exports by Third World countries (e.g., Labys 1980, Girvan 1987) have conceptualized the extraction and processing of minerals as a chain of production processes. They have focused on the control by TNCs of key processing stages in the chain, which has given them the power to bargain highly favorable terms with mineral-possessing Third World states. Economists have studied the economic structure of vertically-linked markets as involving chains of transactions (e.g., Just, Hueth, and Schmitz 1982, chapter 9). And analysts of industrial organization such as Porter (1990) and Storper and Harrison (1991), have introduced concepts such as "value chains" or "production systems" to explain how firms make strategic decisions about the organization of production processes. But the commodity chain conceptualization is more general, and is capable of subsuming all of these somewhat more specialized variants of the chain concept (Gereffi 1994). As developed within the world-systems perspective, the commodity
Studies in Comparative International Devdopment / Spglng 1997
chain concept can be used to analyze how states, TNCs, and multilateral institutions shape economic globalization as they compete for shares of the surplus. World-systems theorists have proposed two conflicting approaches to the measurement of the distribution of benefits along a commodity chain. The HopkinsWallerstein (1994) position makes the surplus a central concept: "If one thinks of the entire chain as having a total amount of surplus value that has been appropriated, what is the division of this surplus among the boxes of the chain?" (p. 49). Thus their approach identifies "core-like boxes" (or nodes) in commodity chains based on the ability of firms ("units of production") controlling those boxes to appropriate a major portion of the total surplus generated along the chain. The Arrighi-Drangel (1986) position focuses on countries rather than chains. They argue that a country in which the core-like nodes of many different commodity chains are located will have a much higher aggregate income than one in which mainly peripheral-like nodes are located. Thus their approach classifies countries as core, peripheral, or semiperipheral based on their per capita GNP, as a measure of aggregate rewards. They regard the division of these rewards between factors of production as an internal political question that has no direct beating on a country's position in the stratification of the world-economy. "Whether or not the rewards of each class of factors of production (wages, rents, and profits), as opposed to aggregate rewards, are higher or lower in core or peripheral activities is a different issue" (p. 17, italics added). Each of these positions thus suggests a different measure of the distribution of benefits along commodity chains. The Hopkins-Wallerstein position is that the share of the total surplus generated is the key measure, while the Arrighi-Drangel position is that the share of total income generated is the key measure. If we are primarily interested in classifying countries as core, peripheral, or semiperipheral, as are Arrighi and Drangel, then total income is a sufficient indicator. But, in the analysis of particular commodity chains, how the total income accruing to each node is divided between wages, rents and profits becomes a central question. This suggests a two-stage approach that incorporates both measures: first, determine how the total income generated along the entire chain is divided between the nodes of the chain; and, second, determine how the total income accruing to each node is divided between wages, rents, and profits. The profits and rents can then be added up along the entire chain to determine who receives the largest share. Hopkins and Wallerstein (1994) link the division of the surplus among the different nodes of a chain to the degree of competition or monopolization that characterizes the activities at each node: "The first and in some ways the most important question is the degree to which the box is relatively monopolized by a small number of units of production, which is the same as asking the degree to which it is core-like and therefore the locus of a high rate of p r o f i t . . . " (p. 18, italics added). This assumption, that in order to measure the distribution of the surplus, the analyst need only identify which nodes of the chain are the most highly monopolized, and which are the most highly competitive, underlies most commodity chain analyses. Certainly this is a reasonable working hypothesis, but in the analysis of specific
commodity chains, we should make an attempt to measure the distribution of the surplus directly. First, it would provide an empirical test of this underlying assumption: a r e the degrees of monopolization correlated with the shares of the surplus across the different nodes of the commodity chain? If we measure the surplus distribution directly, we will probably find some commodity chains where this correlation is very close and others where it is not. This will point to other possible factors that determine how the surplus is distributed. Second, direct measures of the surplus would provide a comparative perspective. We could compare the distributions of the surplus across different types of commodity chains to see how the structure of the chain influences this distribution. Analysis here could focus, for instance, on trying to identify shifts in the "leading" commodity chains at different periods in the development of the world-economy. Third, and most important for the purposes of this analysis, this can also provide an historical perspective. If we can measure how the distribution of the surplus along a particular commodity chain changes over time, we can better understand the process by which this distribution is changed. When a commodity chain is significantly restructured, we can analyze how the surplus shifts among the new set of nodes. We may be able to confirm the key underlying assumption in a longitudinal manner, by finding that TNCs have created new nodes which they can monopolize, and that this is where the surplus has gone. Even when the overall structure of the chain remains the same, we may find that certain nodes have been completely reorganized in such a way as to shift shares of the surplus toward or away from these nodes. Or, as has been the case with coffee in some periods, both the overall structure of the chain and the organization of the individual nodes may remain constant, but political interventions in the market may shift the distribution of shares. The assumption that a high degree of monopolization insures control of a large share of the surplus overlooks the fact that rents can be obtained by other means besides monopolization of certain nodes of the chain by a small number of firms. For example, the Arrighi-Drangel approach leads to a few clear misclassifications: Saudi Arabia and Libya are classified as core countries in the 1975-1983 period. This obviously reflects the success of OPEC in capturing some of the surplus from the oil commodity chain. But in this case the surplus was appropriated by states which were able to assert some control over the flow of the commodity across their national borders. Barham, Bunker, and O'Hearn (1994) make a distinction between two different types of rents, which they call resource and strategic rents. Resource rents accrue to owners of relatively scarce resources. In the case of coffee, these scarce resources would be combinations of soil and climate that allow growers to produce coffee at lower cost than in most other coffee-growing regions, or that allow growers to produce particularly high grades of coffee. For instance, some higher altitude volcanic soils, as found in Colombia, Guatemala, or Costa Rica, are more conducive to production of the highest grade arabicas, which command higher prices on the world market. In other regions, for example, Brazil and Indonesia, coffee can be produced at a lower cost per unit of output than in most coffee-growing areas. In
Studies in Compar-ativeInterna~onal Development / Spring 1997
either case, growers may receive higher than average profits because of these resource rents. Barham, Bunker, and O'Hearn distinguish resource rents from what they call strategic rents. "Strategic rents accrue if, and only if, the resource holder or some other economic agent is able to push the price above the competitive p r i c e . . . " (1994, 18-20). This is the type of rent that was collected by OPEC from the mid-1970s through the early 1980s. In the case of coffee, strategic rents were created by the export quota system established under the ICAs. The quotas restricted coffee exports, thereby causing world market prices for green coffee to be higher than they would have been if all producing countries had exported their entire crops (minus domestic consumption). These strategic rents are another form of monopoly rents, but they are not obtained by monopolizing production at a particular node of the chain; rather they are obtained by regulating the flow of commodities between nodes. It thus seems necessary in the analysis of individual commodity chains to distinguish three types of rents: monopoly rents accruing to participants, usually TNCs, through monopolization of particular nodes; strategic rents accruing to participants, usually resource-owning states, through control of flows between nodes; and resource rents accruing to the owners of scarce resources. One recent study that has attempted to measure the amount of surplus, both profits and rents, and its distribution, is Schurman's (1993) study of the Chilean fishing industry. Schurman doesn't deal with the problems of measuring the distribution of the surplus along the entire commodity chain, because she restricts her analysis to only one segment of the chain. But this focus allows her to draw an important conclusion that is related to the Arrighi-Drangel position. Schurman concludes that the changes in the distribution of the surplus among the different participants in the fishing industry were caused by changes in their political and economic power as the market structure evolved. And she shows that under some conditions, workers and "artisanal" producers can capture a significant share of the surplus. Schurman also points out that what different groups do with the surplus they get makes a difference for the economic development of the region and for the future structure of the commodity chain. Contrary to another assumption, which seems to underly most commodity chain analyses, appropriation of a share of the surplus does not necessarily translate into capital accumulation. This article applies the two-stage approach outlined above to estimate the distribution of total income and of economic surplus along the coffee commodity chain. If surplus is being extracted from the periphery and transferred to the core, then world-systems and dependency theorists and structuralist economists all agree that the export of primary commodities in raw or semi-processed form from the periphery to be consumed in the core is one of the main channels for this transfer. Coffee has been one of the most important primary commodity exports of the Third World since World War II, second only to oil in total income earned by Third World exporters. Over 90 percent of the coffee exported by Third World producers is consumed in the core. In addition, the coffee commodity chain is a relatively simple one. Green coffee is a semi-processed raw material that is used to make only a few
final products--roasted, brewed, or instant coffee for final consumption. Very few other inputs are used in the growing or processing of green coffee or in its manufacture into final consumable forms. Therefore, the chain has very few side branches, unlike, for example, the complex networks of parts suppliers involved in the automobile commodity chain. The small number of directly linked nodes or stages makes it feasible to estimate the distribution of total income along the chain and the surplus at each stage. The coffee commodity chain is pictured in figure 1. For estimation purposes, it can be divided into three major stages: 1) the growing and initial processing of coffee on the farm; 2) processing up to the green coffee stage and exporting, by processors and exporters within the producing country, who are either local capitalists or state agencies; and 3) the importing of green coffee and the production and sale of roasted or instant coffee to consumers in the core. The first two stages of the chain are therefore under the control of individuals, firms, and agencies in the producing countries, and the third stage is under the control of TNCs based in the core. The basic measure of income used in this analysis is the retained value, the total income from coffee production that remains within the country. In the consuming countries, this is equivalent to the total value added (i.e., the retail price of coffee minus the costs of imported green coffee). 1 In the producing countries the retained value is divided into the total incomes of coffee growers and total value added (i.e., the export unit value minus the growers' price). 2 The measure of surplus is net profit, the price minus the costs of production, at each of these three stages: coffee growing, processing to the green coffee stage, and manufacturing and retailing of the final products. At the coffee-growing stage, some or all of these net profits may actually be resource rents, and at the second processing and export stage, some or all may be strategic rents. States in coffee-producing countries, through various policies to tax and regulate their coffee sectors, can capture most or all of these rents, or at least influence how they are divided among growers, exporters, and other intermediaries. At the final production and retailing stage, some or all of the net profits may be monopoly rents generated by TNCs' domination of consuming markets. The total surplus along the chain is appropriated by coffee growers, processors, exporters, importers, manufacturers, retailers, and by producing and consuming states through various forms of taxation. The remainder of this article presents the results of the two-stage estimation of the division of income and surplus along the coffee commodity chain. The next section presents an estimate of the gross distribution of income between coffee growers as a whole, other participants within producing countries as a whole, and participants in consuming countries as a whole. Consistent data over a relatively long period of time are available for this estimate, but it necessitates averaging together the diverse systems of production and distribution within the producing and the consuming countries. The diversity of these systems further suggests that, whatever the gross share of income, the allocation of this income to production costs, rents, and surplus probably varies widely across both the producing and consuming countries. Unfortunately, the data needed to make these more detailed, country-specific estimates
Coffee I m p o r t e r s / * tnd Traders ~ / '--.. Green
) ~ oJrPrivate I ,~ [
~ ~'/ Parchment
are spotty and inconsistent. Thus, the following two sections attempt to estimate how much surplus is retained in the major producing countries, and then in the major consuming countries. The fifth section uses these results to suggest a rough estimate of how the total surplus has been divided along the coffee commodity chain, and how this distribution has changed over time. It also compares coffee to some other primary commodity exports of the Third World. The last section argues that the results of this analysis necessitate a reformulation of commodity chain analysis. The Division of Income betweenGrowers, Producing Cauntries, and Consuming Countries The total income generated along the coffee commodity chain is equal to the total amount of money spent by consumers to purchase coffee products for final consumption. In order to make the results more intuitively comprehensible, this total can be represented by the average retail price of a typical pound of roasted and ground (R&G) coffee in the core markets. The percentage breakdown showing how this total income is distributed among the various participants at the different stages of the chain is the same using either basis. This total income per pound of coffee is divided into four parts: 1) the total income of coffee growers; 2) the incomes of processors, traders, exporters, and the state within producing countries; 3) the incomes of coffee manufacturers, wholesalers, retailers, and the state in consuming countries; and 4) a residual category including shipping costs, profits of shippers and financiers, and weight losses that occur during the shipping and roasting of the coffee. The data for this analysis were provided by the International Coffee Organization (ICO), a multilateral institution created by the International Coffee Agreements (ICAs), which began in 1962. Since the mid-1970s, the ICO has regularly published data on the average prices paid to coffee growers and the unit values of coffee exports for each coffee-exporting member country, and data on the unit values of imports and the average retail prices of coffee for each coffee-importing member country. 3 These data are used to estimate the division of income. The total income of coffee growers is indicated by the average price paid to growers, averaged over all coffee exporters. The average unit value of exports (fob) 4 of green coffee minus the average price paid to growers is used to estimate the income of all other participants in the producing countries. The income of all participants in the consuming countries is estimated by the retail price minus the cost of green coffee needed to produce a pound of roasted coffee (i.e., 1.19 pounds of green), based on the unit value of imports (cif). 5 The residual category includes this roasting weight loss plus the difference between the fob and cif prices. 6 Tables 1 and 2 present the results of this initial step in the analysis--the average retail price for a pound of R&G coffee is broken down into the shares remaining in the consuming countries; returning to the coffee growers; received by intermediaries, including the state, in producing countries; and the residual category. A further complication is introduced by the time periods for which the data are available. The
limiting factor here is the data on prices (both retail and growers'), which are available as calendar year averages for 1971-80, as monthly averages for the last month of each quarter for 1975-85, and as coffee year averages after 1985. To use the full range of data available, I constructed two overlapping sets of estimates, a calendar year series 1971-80 (table 1) and a coffee year series 1975-76, 1993-94
(table 2). 7
There were two types of changes in the world market during the period considered here that might have altered the division of income between producing and consuming countries. One was fluctuation in the world production of coffee, caused primarily by bumper crops or natural disasters in Brazil. 8 Changes in world supplies of coffee cause price fluctuations exacerbated by the fact that coffee is a tree crop. Coffee trees take three to five years after planting to begin bearing coffee, and another two years or so to reach full production. Thus a frost in Brazil lowers world supplies for the next one to two years, before the trees recover, causing a prolonged period of shortage and high prices. The high prices encourage new planting, but the new trees do not have any immediate effect on prices, and there is a tendency toward overplanting. Beginning about five years after the frost, production from the new trees causes a glut on the market and falling prices. However, the trees continue to bear even when prices are low; growers cut back on their variable costs, such as maintenance and fertilizers, but generally continue to harvest the coffee. Thus prolonged periods of high prices tend to be followed by prolonged periods of low prices, until the reduced maintenance of the trees begins to reduce production. But once again, there is a tendency toward overcorrection, which reduces the production of coffee below worldwide demand, leading to a new cycle. A bumper crop in Brazil can also trigger this process in reverse, leading to a prolonged period of low prices followed by a prolonged period of high prices; however, bumper cropgenerated cycles tend to be less severe than frost-generated cycles (see Edwards and Parikh 1976, Ford 1978). Rapidly rising prices should shift income to producing countries, while falling prices should benefit consuming countries. The second type of change in the world market, which could have influenced the division of income, was in the regulatory regime governing the world market. Since 1962, there have been a series of International Coffee Agreements (ICAs) between producing and consuming states, which have, at times, imposed export quotas on producing countries. Quotas were in effect between 1962 and 1972, and again between 1980 and 1989. Between 1973 and 1980, and again after 1989, the quotas were suspended due to lack of agreement between producing and consuming states. Most econometric analyses of the world coffee market have concluded that the quotas increased world market prices above equilibrium levels (Edwards and Parikh 1976, Vogelvang 1988, Herrmann 1986, Akiyama and Varangis 1990, Bohman and Jarvis 1990). Thus, we would expect producers to get higher shares of the total income when quotas are in effect than when they are suspended. The data in table 1 begin just before the end of the first quota period in 1972. These data show no significant shift in retained value due to the end of the first quota regime. Through the early 1970s, a little over half of the total coffee income
"0 ~ C
Q ~ J
e o 9
~ 1 7 6
'i "1~~ ~
~r~ v v
Studies in Comparative Intamafional Development / Sprin~ 1 ~ 7
o ox 0
Ch v '0 r
oo '0 O~ OX r 0 o"1 v xO .-~ on 0
kO oo 0o l~ v oO v
oo Ox 0
00 r v
Ox Ox r r v 0 I~ 0
r 0 on r o"1 00 I~ O0 0
Go oO 0 C~ 0 Ox 00 ~t v t'q ',0 o'~ 0 ',0 0 0"1
% o o
0 v 0 0 ~ t~l ,-~
c~ ("4 c~ v r~ o
04 s I"-~'b v
v v 0 o,~ oo ~" 0o
~.~,~~,-~. ~oor~ "~ c~ ~:~t:~;:~ ,~ t
"~ ~ o ~
~ ~ o= ~ ~'-~, ~'-~,
9~ ' - ~
~ o ,~ ~ ~.~ o ~
O0 Ox ~1-
r.-,. ,--~ ('--4 v
... -~ 0
Studies in ComparativeInternational Development / Spring 1997
remained in the consuming countries. About a third of the total income returned to the producing countries, with growers receiving about two-thirds of that, or roughly twenty cents of each coffee dollar. However, with only two years' worth of data on the situation during this early quota period, it is difficult to place too much weight on this result. In addition, due to frosts in Brazil in 1969 and 1972, world market prices were rising during this period, possibly offsetting any income losses producers may have experienced from the end of the quotas. The devastating Brazilian frost of June 1975 reduced Brazil's production for crop year 1976-77 (April 1-March 30) to one-quarter of its pre-frost levels, producing record high prices that were the most significant feature of the 1973-80 non-quota period. World market prices shot up through 1976 and into 1977 as the effects of the frost on world supplies were felt. As expected, this price increase shifted income toward producers; consuming countries' share of total income fell below half, and producers' share exceeded 40 percent. Thus in coffee year 1976-77 and calendar year 1977, the share of income retained by producers exceeded that retained in consuming countries, but as prices began to fall after 1977, producer and consumer shares reverted to roughly their previous levels (tables 1 and 2). 9 Quotas were reinstated in 1980. Producers' and consumers' shares stabilized at about the same levels at which they had been in the non-quota period preceding the frost, contrary to expectations that the quotas would benefit producers. But world market prices were falling in 1980, as the new trees planted during the 1976--77 coffee boom began to produce, and the re-imposition of quotas may have forestalled an even deeper cut in producers' retained value. Further, shipping costs appear to have increased after 1980, which may have resulted from the second oil price shock in 1979 and the global recession of the early 1980s. This may also have reduced income to producers if the quotas had not been in effect, l~ Quotas were suspended once again in early 1986, following a severe drought in Brazil which lowered output and raised world market prices. Prices peaked in the spring of 1986 and declined gradually thereafter, leading to a reimposition of quotas in late 1987 near the beginning of the 1987-88 coffee year. The suspension of quotas in a rising market in 1986 coincides with a shift of more than 10 percent of total retained value from the producers to the consumers. But when the quotas were reinstated in a falling market in late 1987, this division of total income was maintained. Table 2 shows that TNCs were able to increase retail prices during 1985-86 in response to the increases in world market prices for green coffee; but when green coffee prices declined, retail prices did not. In addition, Akiyama and Varangis (1990) have shown that the quota system mitigated the effect of the drought on prices. Because of the quotas, producing countries were holding stockpiles of coffee, which could be released when the quotas were suspended, thus damping the price increase. They estimated that green coffee prices would have been 24 percent higher in 1986 had there been no quotas in the preceding period. Under those conditions, the drought may have produced a shift to producers similar to that observed in 1976-77; at least it may have offset some of the shift to consumers that occurred in 1986.
After the negotiations to renew the ICA broke down in 1989 and quotas were again suspended, world market prices crashed, falling by 50 percent within six months. Retail prices dipped slightly, and then actually increased in 1990--91 and 1991-92, while green coffee prices continued to decline, reaching historic lows in the summer of 1992 (Talbot 1995-96). Table 2 shows that the 1989 price crash coincided with an additional 10 percent shift in total income from producers to consumers. These two shifts, of roughly 20 percent of total coffee income over a period of less than five years, dwarf any changes observed over the previous fifteen years. While both shifts coincided with suspensions of the quota system, both shifts were probably also linked to the international restructuring of the coffee industry by the coffee TNCs during the 1980s. Through a series of mergers and buyouts, which accelerated through the 1980s, five TNCs came to dominate retail coffee sales in most major consuming markets, while six or seven large trading companies gained control over the majority of green coffee imports (Talbot 1995-96). The market power of the TNCs enabled them to maintain the level of retail prices of coffee while world market prices for green coffee were falling in 1987 and plummeting in 1989. In 1993, an attempt by producers to resurrect the ICA quotas failed, leading to the formation of a cartel, the Association of Coffee Producing Countries (ACPC), and a plan to withhold coffee from the world market beginning on October 1, 1993, the start of the 1993-94 coffee year. Then, in June and July 1994, two severe frosts struck the coffee-growing regions of Brazil, causing world market prices to more than double in a matter of weeks. Only preliminary data are available for the 1994-95 coffee year, but the data in table 2 suggest that these twin shocks may have caused a shift of as much as 10 percent of total income back toward the producers. The table also shows that the TNCs increased retail prices during 1994-95, and previous experience suggests that they are unlikely to decline much from this level. Meanwhile, green coffee prices were declining gradually through 1995 and into 1996, leading ACPC to establish a new retention plan for coffee year 1995-96. However, prices at the end of 1995 were still substantially higher that they had been in mid-1993, and compliance by producers with the retention plan was consequently less strict. Thus, the shift of income back toward the producers may be only temporary. These data indicate that changes in the regulatory regime interact in complex ways with the underlying price trends in the market. Changes in the regime (suspension or imposition of quotas) may affect the distribution of income mainly through their effects on the level of world market prices. Quotas tend to maintain world market prices above their equilibrium levels, and thereby transfer income to producing countries. Periods of shortage and rising prices shift income to producing countries, while overproduction and falling prices benefit consuming countries. Thus, the end of quotas in 1972, in a period of rising prices, and the imposition of quotas in 1980, in a period of falling prices, did not shift the distribution of income because the regime change and price change effects offset one another. Lifting quotas when prices are rising counteracts the shift of income to producers; imposing quotas when
Studies in Comparative International Development I Spring 1997
prices are failing counteracts the shift to consumers. When quotas were suspended in a falling market, however, as in 1989, both the regime change and price change effects favored consuming countries, and the shift in the distribution of income was significant. The 1986--87 shift of income from producers to consumers does not fit this pattern, because quotas were suspended in a rising market, and then reimposed in a failing market, but the distribution of income still shifted significantly toward consumers. Further, there is some evidence to suggest that the quotas damped the price increase during 1986 and prevented income from shifting toward producers. This shift suggests that the role of TNC market power also must be considered as a factor that can affect the distribution of income along commodity chains, because it allows them to maintain retail prices at the same level even as the prices of their raw material inputs fall. Finally, the 1993-95 shift suggested by the preliminary ICO data illustrates the effect of the imposition of export controls reinforcing the effect of a rising market, causing a significant shift of income to producers. Given the market power of the TNCs, it remains to be seen whether producers can hold onto this share, or whether the TNCs will regain their previous share by once again maintaining retail prices as green coffee prices slowly decline. This section has focused on the distribution of total coffee income between producing and consuming countries. The next question is, what are the costs of production at each stage of the coffee commodity chain, and, consequently, how much surplus is generated at each stage? In order to answer these questions, we must consider the specific systems of production and distribution in individual producing and consuming countries and the structures of costs and surpluses they generate. This task is tackled in the next two sections.
The Distribution of the Surplusin the Producing Countries
In this section, estimates from previous studies of the costs of coffee production and processing in different producing countries are used to estimate the total surplus remaining in the producing countries and its division among the growers, the state, and others. A compilation of these results is presented in table 3. These studies used different sets of countries, different methods to decompose the export price into costs and surplus, and different methods of allocating the surplus among participants within each country. The task of making this data comparable across producing countries is made more difficult by the fact that their production systems vary widely. For instance, some growers sell fresh cherries directly to the processors, others wet-process the cherries into parchment coffee, and still others dry-process them. Intermediate processing of parchment or dried cherries into green coffee is sometimes performed by state agencies, sometimes by a few large exporters, and in other cases, by a myriad of small agents and traders. Nonetheless, table 3 presents the best estimates available of the amounts of surplus retained within producing countries, and of the distribution of this surplus. The variations between countries in the amount of surplus and its distribution, which are evident within each column of
table 3, are due largely to the differing production systems. But comparing data for each country across time, within the rows of the table, gives us some indication of how changes in the amount and distribution of the surplus are related to the shifts in the distribution of income analyzed above. Orlandi's (1978) estimates of the surplus in 1973, shown in the first column of table 3, are valuable because they are from the beginning of the period considered here, in the early 1970s, as the quotas that had been in effect under the first two ICAs were lifted. These estimates suggest that roughly half of the export price of green coffee was economic surplus that remained in the producing countries, and that the bulk of this surplus was appropriated by the states. The export prices shown here reflect differences in quality and world market prices of the three grades of arabicas produced by these countries. For Brazil, no data on costs of production were available. The state share, 39 percent of the export price, which can be considered as pure surplus, was collected through various forms of export taxation. Growers and exporters probably earned additional surplus, bringing the total surplus to about half of the export price. In Colombia, where the highest quality arabicas are produced, over 60 percent of the export price was surplus, and the quasi-state coffee agency, the Federacirn Nacional de Cafeteros (FNC), took over 60 percent of that. Still, coffee growers received a share of the total surplus equal to about a quarter of the export price. No data were available on the costs of private processors and exporters, who handled about 60 percent of Colombia's coffee in the early 1970s. Their gross income amounted to only about six cents per pound, shown here as the processing cost, so they received a negligible share of the surplus per pound. In E1 Salvador, the state's and growers' shares of the surplus amounted to only 30 percent of the export price, and the state took two-thirds of it. However, the gross income of processors accounted for 18 percent of the export price, much higher than in Brazil or Colombia, and a large part of this is almost certainly surplus also. In E1 Salvador' s plantation system, the processors are usually the largest plantation owners, who are sometimes exporters as well. Thus the total surplus was probably on the order of 45 percent of the export price, with the largest plantation owners taking a little over half of it, and the state taking most of the remainder. Based on the estimates from these three countries, we can say that, in the early 1970s, probably about half of the export prices of coffee from producing countries represented surplus. However, it is difficult to generalize from these rough estimates for three countries to the situation in all other producing countries. The rapid increase in retail prices of coffee following the devastating 1975 Brazilian frost created a flurry of interest in where consumers' coffee dollars were going, and spawned a number of studies attempting to explain why prices increased as much and as rapidly as they did. One of these studies was conducted by the US Government Accounting Office (GAO) for the House Agriculture Committee's Subcommittee on Domestic Marketing, Consumer Relations and Nutrition in 1977; this study attempted to estimate costs of production and profits in Orlandi's three countries as well as in Mexico and Crte d'Ivoire. These results are shown in the second column of table 3. As it turned out, the period on which this report focused was the
Studies in Cemparative International Development / Spring 1997
TABLE 3 Estimates of the Total Surplus Retained in Coffee-producing Countries, and Its Division between Growers, the State, and Others
Country Brazil (A) Export price Production cost Processing cost Total surplus Growers' share State share Others' share Colombia (A) Export price Production cost Processing cost Total surplus Growers' share State share Others' share El Salvador (A) Export price Produclion cost Processing cost Total surplus Growers' share Stale share Others' share Mexico (A) Export price Production cost Processing cost Total surplus Growers' share State share Others' share Costa Rica (A) Export price Production cost Processing cost Total stirplus Growers' share State share Others' share 119.7 49.9 18. I 51.7 (43%) 24.2 (47%) 27.5 (53%) 0 119.7 64.5 NA 55.2 (46%) 5.1 (9%) NA NA 315.0 31.0 I 8.0 266.0 (84%) 107.0 (40%) 136.0 (51%) 23.0 (9%) I 19.8 75.6 NA 44.2 (37%) 11.3 (26%) NA NA 58.0 26.6 13.8 17.6 (30%) 5.4 (31%) 12.2 (69%) NA 300.0 35.5 9.0 255.5 (85%) 146.5 (57%) 83.0 (32%) 26.0 (10%) 129.6 58.5 NA 71.1 (55%) 6.2 (9%) NA NA 69.3 20.2 5.9 43.2 (62%) 16.4 (38%) 26.8 (62%) 0 314.3 NA 2.6 NA NA 207.4 16.2 131.5 77. I 8.2 46.2 (35%) -10.0 55.2 1.0 126.5 73.2 NA 53.3 (42%) 2.2 (4%) NA NA 51.3 NA 3.3 NA NA 20.0 NA 369.11 40.0 1.5 327.5 (89%) 186.8 (57%) 101.3 (31%) 39.4 (11%) 115,7 54.4 6.8 54.5 (47%) 0 54.4 (100%) 0.I (0%) 97.5 54.8 NA 42.7 (44%) - 12.7 NA NA 1973 1977 ca. 1982 1987/88
TABLE 3 (continued)
Country Kenya (A) Export price Production cost Processing cost Total surplus
Growers' share Slate share others' share Rwanda (A) Export price Production cost Processing cost Total surplus Growers' share State share Others" ~hare 1973 1977 ca. 1982 132,4 70,8 24.9 36,7 ( 2 8 % )
Cameroon (R) Export price Production cost Processing cost Total surplus
Growers' share State share Others' share Cote d'lvoire (R) Export price Production cost Processing cost Total surplus Growers' share State share Others' share Indonesia (R) Export price Production cost Processing cost Total surplus Growers" ~hare State share Olhers' share
97.5 40.8 17,7 39.0 (40%) 5,0 (13%) 34.0 (87%) 0
94.0 88,5 NA 5.5 (6%). -20.5 NA NA
29t. I NA I 0,0 NA NA 240.9 0.4
86.2 408 13,2 32:2 (37%)_ 0,6 (2%) 31.6 (98%) 0
10t.8 ! 12.3 NA -10.5 -50.5 NA NA
79,4 36,3 36.3 6,8
86.7 69.0 NA 17.7 (20%) -9.0 NA NA
2,6 (38%) 4,0 (59%) 0,2 (3%)
Sources: 1973 data from Orlandi (1978); 1977 data from GAO (1977); 1982 data from de Graaf (1986); 1987-88 data on export price and price paid to growers provided by ICO; data on production cost from Landell Mills, as reported in Carta Cafetera (1990). (A) indicates a producer of arabica coffee; (R) indicates a producer of robusta coffee.
Studiesin ComparativeIn~rnationalDevdopment/ Spring 1997
acme of the coffee price boom, which followed the frost, around April 1977 (the ICO indicator price peaked at its all-time high of 314.96r per pound in this month). This means that these figures probably overestimate the amount of surplus that went to the producing countries. In addition, the GAO based their estimates on the New York (cif) prices of the coffees from most countries; this method apparently assigns most of the transport costs to the exporters' surplus, resulting in an overestimate of the surplus. Nonetheless, the data give us an approximation of the distribution of surplus within some major producing countries at the height of the boom. The overall picture from these GAO data, as compared to Orlandi's, is that the large price increases of 1976-77 greatly increased the total amount of surplus retained in the producing countries. Production costs increased somewhat between 1973 and 1977, but not nearly as fast as world market prices. The bulk of this surplus was divided between the coffee growers and the state, but their relative shares varied greatly across producing countries. In Brazil, the state's share of the surplus declined, which suggests that coffee growers reaped windfall profits during the boom. In Colombia, the total surplus increased, and the state's share alone amounted to two-thirds of the export price. The growers' share of the total surplus probably declined, but the absolute amount of surplus they received almost certainly increased. In El Salvador, there was a large increase in surplus, and most of it was apparently passed on to the growers. The total surplus in Mexico was similarly large, and the state took a little more than half of it. In Crte d'Ivoire, the state share alone, collected through its Caisse de Stabilisation amounted to almost 83 percent of the export price. These figures indicate that the coffee price boom increased surplus in producing countries from about half of the export price to about three-quarters of a much higher price. This increase can be considered a temporary resource rent created by the Brazilian frost. In some countries the state took most of this rent by holding down internal prices or increasing taxes; in other countries most of the increase was passed on to coffee growers. The most comprehensive published comparative analysis of production and processing costs in the producing countries was conducted in the early 1980s by J. de Graaf of the Royal Tropical Institute in Amsterdam. De Graaf (1986) analyzed production, processing, transport, and export costs for eight major producers: Brazil, Colombia, Costa Rica, Kenya, Rwanda, Cameroon, Crte d'Ivoire, and Indonesia. In coffee year 1981-82 these eight producers accounted for 62 percent of total ICO member exports. These data are summarized in column 3 of table 3. De Graaf's data show zero profits on average for Brazilian growers. He states that this is partly due to losses suffered by growers in Paran~i, where the 1975 frost occurred. It is reasonable to assume that profits per pound were relatively low, and de Graaf may have underestimated prices paid to growers. Using data from the ICO on prices paid to growers along with de Graaf' s estimated production costs yields a plausible estimate of 3.5 cents per pound. Since most Brazilian production comes from large diversified farms, low profit rates may be sufficient; the farms produce large volumes of coffee, and coffee income represents only one component of their total income. Almost half of Brazil's export price was surplus, and virtually all of it
went to the state. Since the growers probably made a small profit, the share of the state is overstated in these numbers, but it seems reasonable to conclude that the state (primarily the state coffee agency, the Instituto Brazileiro do Cafe (IBC)) did capture most of the surplus in the early 1980s. These figures refer only to coffee that is exported; they do not include income and profit generated by Brazil's sizable internal coffee market. The data show a negative profit for Colombian growers; again, using the ICO data on growers' price, there is a very small positive profit. Either way, it seems safe to conclude that most of the surplus (only 35 percent of the export price in Colombia vs. about half in Brazil) went to the quasi-state Federacidn Nacional de Cafeteros (FNC) in the early 1980s. This is consistent with the estimate for the late 1980s shown in column 4, but it must be considered in light of the services provided by the FNC to growers, which will be discussed in more detail later. The data for Costa Rica show a total surplus of about 45 percent of the export price, divided roughly equally between the growers and the state. A lower amount of surplus was retained in Kenya than in other arabica producing countries studied by de Graaf; only about a quarter of the export price was surplus. The data show growers receiving about two-thirds of this surplus. Rwanda is by far the smallest producer considered here; despite the fact that it is dependent on coffee for most of its export earnings, its production is an insignificant share of total world production. Table 3 shows an amount and distribution of surplus for Rwanda that is very similar to that of Costa Rica, another mild arabica producer. Surplus was divided primarily between the state and the growers, with the state getting a somewhat larger share. Cameroon is a producer of both robusta and arabica coffee; arabica production accounts for about 25 percent of total production. The figures shown here are for robusta production; the data show that 40 percent of the export price was surplus, and almost all of the surplus was retained by the state, through its marketing board. Crte d'Ivoire was the world's largest robusta producer in 1981-82, and presents an interesting case because it is also one of the world's major cocoa producers. All coffee transactions are under the control of the Caisse de Stabilisation, which annually publishes a schedule of rates and prices for various services and agents. The amount and distribution of surplus look very similar to those for Cameroon robustas; this makes sense since they have similar production systems. Indonesia is now the world's largest exporter of robustas (it was second behind Crte d'Ivoire in 1981-82); about 5 percent of its total production is arabica. Indonesia stands out as a low-profit producer--surplus represented only about 8 percent of the export price, well below all other cases analyzed here. Growers got a larger share of the surplus than in the other robusta producing countries, but not as large a share as arabica growers. The state still took a large share, but the surplus to be shared here was much less. What is probably the most comprehensive comparative study of production costs ever conducted is not available to the public. It was conducted by Landell Mills Commodity Studies (LM) of the UK, and is available only by special subscription; it is information which has itself become a commodity. A summary of the results
Studies in Comparative International Development / Spring 1997
has been published in Carta Cafetera (1990), the newsletter of the Colombian coffee exporters' association. The study covered twenty-one producing countries: eleven arabica producers, two robusta producers, and eight countries that produced both (although most are major producers of only one). The last column of table 3 presents data from this study only for countries included in one of the three previous studies reported here. Because the LM study focused only on production costs, these data were combined with ICO data on export prices and prices paid to growers to obtain estimates of the total surplus. The de Graaf and LM data provide a reasonably good basis for concluding that, during the 1980s, about one-third to onehalf of the export price of coffee represented surplus that was retained in the producing countries.ll Differences within this range depended mainly on the production systems, types and qualities of coffee grown, and the levels of world market prices for the different types and qualities. The LM data that have been published have no estimates of the state's share of the surplus, but even de Graaf's small set of countries suggests that this share varies considerably across producing countries, depending on the structure of the production system, and the state's role in it. What is interesting about these results is that, despite the large shift in income shares between producing and consuming countries in 1986-87 discussed in the previous section, there was apparently no decline in the amount of surplus retained in the producing countries. This brief sketch of the situation in the 1980s sets the stage for what happened after 1989. If the costs of production were more than half of the export price of coffee in the 1980s, and export prices fell by 50 percent during the 1989 price crash, then this crash wiped out the entire surplus. Growers could, and did, cut their production costs by cutting back on fertilizer, maintenance, and harvesting (picking less selectively or resorting to stripping, which lowered the quality of coffee they produced), but they could not match the fall in price. This is why growers worldwide were complaining by 1991-92 that the prices they were receiving were below their costs. And this despite the fact that the states had also cut back their coffee income and their roles in coffee processing and export, and in some cases were providing subsidies to growers out of income they received from other sources (note the sharp drop in value added in producing countries after 1989 in table 2). Coffee acreage declined in many countries in the early 1990s, and the coffee trees that remained were less well-maintained than they had been through the 1980s, creating conditions for another price increase. The 1993-94 rise in the world market price brought the income of producing countries back above their costs of production, but there is no data available to indicate the amount of surplus that returned to producing countries. The data in table 2 suggest that most of the 1993-94 increase was passed on to growers, but, when prices skyrocketed after the Brazilian frosts of 1994, states probably increased their shares. As in 1977, this was probably only a temporary increase in the surplus. These data show significant amounts of surplus appropriated by the producing states. The receipts of producing states must be evaluated not on the basis of their absolute levels, but based on the uses to which they have been put. For instance,
surplus extracted from coffee production in Zaire probably goes, as do all other forms of surplus, into feeding the archetypal predatory state (Evans 1989). This surplus must be judged differently from that extracted by the Colombian FNC and invested in improving the country's coffee quality and production infrastructure. The services provided by states or their agencies fall into three general categories: credit, extension, and R&D. Credit includes a variety of services such as lowinterest loans to pay costs before the harvest or advances of fertilizers and pesticides at subsidized rates. Agricultural extension services include education on cultivation techniques and proper application of fertilizers and pesticides, as well as provision of seeds and seedlings to grow high-yield or disease-resistant varieties. R&D is the most interesting of these categories, since most of the advances in new varieties and cultivation techniques have originated in the producing countries, Producing countries as diverse as Costa Rica and Uganda have developed their own high-yield varieties suited to their particular growing conditions; Colombia and other countries have developed varieties that are resistant to the diseases or pests that have caused particular problems for their coffee growers. Coffee depulping and hulling machines are generally produced in the producing countries, and the states may have also invested in R&D to improve the quality of coffee processing at these stages. It is difficult to generalize, but it seems clear that at least some of the surplus extracted by producing states was put to productive uses. The data on surplus analyzed in this section suggest that over most of the period considered here, the total surplus amounted to about 40-50 percent of the export price. There appear to have been two significant shifts. One was the coffee boom of 1976-77, during which about 75 percent of the export price may have been surplus. The second was the price crash of 1989, which drastically reduced this surplus; indeed, it became negative for some countries in 1990, and probably again in 1992. No data is yet available to indicate how much surplus the producing countries have been able to regain as a result of the increase in world market prices during 1993-94 and 1994-95, but it is unlikely that the surplus has regained its pre-1989 levels. This section of the analysis has focused on the amount of surplus that remains in the producing countries. But we know little about the distribution of the surplus along the entire coffee commodity chain until we also estimate the amount of surplus retained within the consuming countries. This is the subject of the next section.
Surplus in the Consuming Countries
On the consuming side of the market, the production system and the structure of the coffee commodity chain are more similar across countries than is the case for producing countries. TNC coffee importers bring in the green coffee and sell it to coffee manufacturers. They process and package the coffee and sell it to retailers who in turn sell to the final consumers. The difficulties encountered in the previous section, where estimates of the surplus might differ widely across countries with different production systems producing different types and qualities of coffee, are
Studies in ComparativeInternationalDevdopmeat / Spring 1997
TABLE 4 Estimate of Nestl6's Profit as a Percentage o f Retail Sales
1985 Operating Profit a Total Sales (Wholesale) I I% Retail margin b Est. Retail Sales Profit as % of Wholesale Profit as % of Retail Sales Green Coffee Costs as % of Retail Sales Advertising + Promotion as % of Retail Sales 3.4 40.4 48.3 231.7 28.6 260.3 20.9 18.6
1986 59.9 261.4 32.3 293.7 22.9 20.4
1987 77.8 256.2 31.7 287.9 30.4 27.0
1988 91.9 268.3 33.2 301,5 34.3 30.5
1989 99.6 276.7 34.2 310.9 36.0 32.0
39.9 3. I
Notes: aCalculated as return on capital employed X capital employed. bI assume a gross retail margin of 11 percent of the retail price. MMC states on p. 23 that the gross margin on Nestl6's most popular brand was under 10 percent, and of GFL's, less than 11 percent. But minor brands had margins close to double those on the major brands. Source: MMC (1991), Table 4.4, p. 52.
less o f a problem in the consuming countries. The intemationalization o f c o f f e e manufacturing has spread a similar system, and probably a similar distribution o f profits, across all consuming countries. The main c o m p o n e n t s in all countries should be the profits o f importers, roasters, and retailers. H o w e v e r , obtaining the data to estimate these components turns out to be much m o r e difficult than it was in the producing countries. First, the largest players in these markets are huge diversified TNCs, and it is almost impossible to sort out how much profit they make on their c o f f e e operations as opposed to their other product lines. Second, information on costs o f production can legally be considered a "trade secret," which does not have to be disclosed. T h e difficulty presented by this aspect o f c o f f e e manufacturing in the core is highlighted b y the 1991 report o f the U K M o n o p o l i e s and Mergers C o m m i s s i o n ( M M C ) on the prices o f instant coffee. The M M C was asked to investigate N e s t l r ' s pricing practices following the 1989 price crash, after which the retail prices o f instant c o f f e e generally remained at their pre-crash levels. In the UK, over 90 percent o f the c o f f e e c o n s u m e d is in instant form, and in 1990, NestI6 brands accounted for 56 percent o f the retail market. The M M C was asked to investigate whether N e s t l r ' s position in the market allowed it to make m o n o p o l y profits following the price crash. But
almost all of the data on Nestlr's profits are suppressed in the public version of the report, because it "would not be in the public interest to disclose [them]" (MMC 1991, iv). The only profit data left in this report are shown in table 4. The table shows profits as a percent of the total retail value increasing from a level of about 20 percent in 1985-86 to about 30 percent in 1988-89. It shows Nesflr's green coffee costs decreasing from about 40 percent of the retail price in 1985-86 to 23 percent in 1989. Since Nestl6 generally buys through importers, these costs already include the importers' profits. Thus, we can tentatively conclude that, in the late 1980s, surplus retained in the consuming countries was in the range of 30-40 percent of the retail price. The MMC report acknowledges that Nestlr's "profitability is considerably higher than that of industry in general or of other firms in its own industry," (p. 1; also see p. 75). But it finds no evidence that Nestlr's market position enabled it to gain an unfair advantage over its competition. What is considered to be its competition in this analysis is instructive: the second largest share in the market (25 percent) is held by General Foods Ltd., the Philip Morris/Kraft/General Foods UK subsidiary; the other two companies considered in detail are Lyons-Tetley and Brooke Bond, both major players in the world tea market, based in the UK. They produce or import instant coffee for sale in the UK market, and hold 8 percent and 6 percent shares of the market respectively. Thus all companies considered are large diversified TNCs who collectively benefit from their oligopoly position; together they account for 95 percent of the market. Although there are no "anti-competitive" discussions of pricing policy among these firms, none are needed; they all know what the other firms are doing and respond accordingly. In this situation, Nestl6 d o e s n ' t need to resort to "anti-competitive practices" and none of its TNC competitors are likely to complain about Nestlr' s higher profits, as long as theirs are sufficient. The data thus suggest a significant increase in Nestlr's profits beginning in about 1986. This coincides with the increase in value added in the consuming countries, which happened between coffee years 1985-86 and 1986-87, as shown in table 2. Although data for one TNC in one consuming market is not a good basis for generalization, it is tempting to conclude from this data that the TNCs, after a short lag, were able to increase retail prices to match the 1985-86 rise in green coffee prices, and that they were able to maintain these retail price levels through the late 1980s despite the fact that green coffee prices had fallen back to their early 1980s levels by mid-1987. When green coffee prices crashed in 1989, the TNCs were thus well positioned to continue to maintain their wholesale/retail prices, leading to the shift in value added toward the consuming countries that is shown in table 2, and to a further increase in TNC profits, which is only hinted at in the MMC report. An UNCTAD study (1984a) also contains some data on the distribution of coffee income during the period 1974-78 in one consuming country, the U.S. UNCTAD concludes that "gross margins from the activities of trading, processing, and distributing companies could be as high as one quarter of the retail price of coffee," (1984a:38). This would amount to roughly half of the retained value in the U.S.
Studies in Comparativelo.t.onmdona] Dcvdopmcat I Spring 1997
Since this period contains the 1976-77 coffee price boom, it is reasonable to assume that the surplus was about half of the retained value during the non-boom years, but substantially lower during the boom. UNCTAD derived some of this data from the Censuses and Annual Surveys of Manufactures, and it is possible to use these sources to put together a picture of how this distribution has changed over time. These data are shown in table 5. Since these data provide no estimate of retail value, it is difficult to compare it with the previous estimates. Nevertheless, the data show a declining trend in the cost of raw materials as a percentage of the wholesale value over time, from around 70 percent during the boom, to the low 60 percent range in the early 1980s, the high 50 percent range in the late 1980s, and finally falling below 50 percent after the 1989 price crash. These figures include some other inputs besides green coffee--in 1987, green coffee accounted for 80 percent of material input costs; most of the rest were for packaging materials. But they show a definite trend toward higher value added in the consuming countries from around 30 percent of wholesale in the late 1970s to around 50 percent in the early 1990s. Labor costs increased somewhat over this period as well, but they were a minor component of total costs. Other costs such as energy inputs also increased during this period, but they were also a minor cost---energy costs were about 1 percent material inputs in 1987. Therefore, increasing costs cannot account for all of the increased value added by manufacturing over this period, and there must have been a substantial increase in the surplus. The data in tables 4 and 5 thus are broadly consistent, and suggest a rather steady increase in the amount of surplus retained in the consuming countries. Taxes are a neglected component of the total surplus within the consuming countries. The fact that core states receive a share of the surplus is usually overlooked in discussions of the role of the state. These discussions tend to focus on the taxes taken by producing states as distorting factors in the world market, overlooking the fact that consuming states also raise revenues from taxes on coffee. And, unlike the situation in the producing countries, where some of these revenues are used by state coffee agencies to provide services to coffee growers, tax revenues in the consuming countries go directly into the general budget. No one questions whether core states use their shares of the surplus productively, or squander it, for instance, on bloated militaries. In fact, some consuming states extract more in taxes on coffee than some producing stZ'tes; this is particularly true in EEC countries such as UK and Germany, which have high VAT taxes. And it is particularly true for the post-1989 period, when most producing states cut their export taxes on coffee substantially. Overall, the data in this section suggest that the amount of surplus in the consuming countries has been steadily increasing over time. The two significant shifts in consuming country surplus seem to coincide with those noted in the previous section for producers: the 1976-77 coffee boom, and the 1989 price crash. During the boom, the amount of surplus probably dropped sharply, but then recovered and resumed its increase. After the price crash, it increased sharply, and we will have to wait to see whether this shift is permanent, or whether producing countries have been able to recover a share of the surplus since the 1993-94 price rise.
TABLE 5 Data on Cost of Raw Materials and Value Added for Coffee Manufacturing Establishments in the U.S., 1972--1991
Value of Shipments (Wholesale) Cost of Raw Materials Value Percent Manufncturing Value Added Value Percent Pay roll Value Percent
Source: 1987 Census of Manufactures, Industry Series--Miscellaneous Food and Kindred Products (U.S. Census Bureau, 1990) Table la-1, p. 2CI-8; 1988, 1989, 1990 and 1991 Annual Surveys of Manufactures, Industry Statistics, Table 2.
The Distribution of Benefits along the Commodity Chain
Pulling together the results presented in the three previous sections, we can summarize the changes in the distributions of income and surplus that have occurred over the past twenty years. In the early 1970s, a little more than half of the total income went to consuming countries, and about a third of it was surplus. About a third of total income went to producing countries, and about half of it was surplus. The result was that the total amount of surplus was roughly equally divided between
Studie~ in Comparative International Development I Spring 1997
the producing and consuming countries. The 1975 Brazilian frost and the subsequent coffee boom significantly shifted this distribution. Producing countries' share of the retained value briefly increased to more than half, and as much as threequarters of this may have been surplus. Consuming countries' income dropped to near 40 percent of the total, and as little as 20 percent of this may have been surplus. What was a relatively small shift in income thus apparently masked what was a massive, but very brief, redistribution of the surplus. For a period in 1976-77, the producing countries may have received as much as 80 percent of the total surplus available along the chain. This increase in surplus was due mainly to temporary resource rents created by the shortage of coffee. The situation seems to have quickly reverted to what it had been before the boom, and remained the same until 1986. Income in consuming countries was about half the total, with about a third, or maybe slightly more, being surplus; income in the producing countries was again about a third of the total, with a bit less than half of it being surplus. Thus the division of the surplus was again roughly equal, with the consuming countries probably having a slightly larger share than they had before the boom. The Brazilian drought of 1985, leading to price increases and the temporary suspension of quotas in 1986-87, appears to have triggered another significant shift. The income of consuming countries rose to about two-thirds of the total, while the income of producing countries fell to about a quarter. Surplus in the consuming countries rose to probably around 40 percent of income, while surplus in the producing countries seems to have held steady at about 40 percent of income. However, because the total amount of surplus increased, the net result was a shift in the distribution of the surplus: about 60 percent of the total surplus available along the chain now went to consumers and about 40 percent to producers. The next major shift recorded in these data was caused by the end of quotas and the price crash of 1989. Income in the consuming countries rose to about three-quarters of the total, while the income of producers fell to about 15 percent. Surplus in the consuming countries may have risen to about half of total income, while surplus in the producing countries fell to near, and sometimes below, zero. Thus, the bulk of the surplus, probably 80-90 percent, was controlled by TNCs in the consuming countries. The ACPC retention and the Brazilian frosts of 1993-94 shifted a large amount of income, and probably surplus as well, back to the producing countries. But how much of this income and surplus the producers will be able to retain in the longer run is an open question. On the one hand, the oligopoly and oligopsony positions of the major coffee TNCs (roasters and importers) put them in a good position to regain their shares; on the other, the producers have recently displayed an ability to cooperate in holding their coffee off the market, which helped to increase world market prices in 1993-94. From this summary of the results, it might seem that natural disasters in Brazil have as much influence on the distributions of income and surplus as any of the economic or political factors discussed above. This impression is correct to the extent that these disasters introduced exogenous shocks into the chain, to which all market participants had to respond. These shocks provided the opportunities for
much larger shifts than would have been likely under "normal" conditions. The impression is incorrect to the extent that it is impossible to know what would have happened in the absence of other political and economic changes. This is most clearly illustrated by the coffee boom and its aftermath. The 1975 frost raised prices and shifted income and surplus toward producers. Producers responded by planting more coffee, sowing the seeds of a production glut in the early 1980s. Other things being equal, this overproduction would have lowered income to producers and probably also lowered their share of the surplus. But this tendency was counteracted by the reimposition of quotas in 1980, which in effect helped producers to maintain the status quo that had prevailed before the frost. The 1985 Brazilian drought introduced another shock to which market participants had to respond, but in this case, unlike 1976-77, the result was a shift of both income and surplus toward consuming countries. The evidence suggests that this shift was caused primarily by TNCs using their market power to add income and surplus to the stages of the chain they controlled. The world market prices of green coffee and the absolute amount of surplus retained in the producing countries apparently remained at about the same levels through the 1980s, but the TNCs increased their prices in 1985-86, and maintained these prices after 1986 even as the costs of their raw material input declined. Finally, the 1989 price crash is the clearest example of a shift caused by a change in the regulatory regime governing the commodity chain. The end of the export quota system and the release of coffee stockpiles held by the producing countries put the TNCs in a very strong bargaining position. They were able to demand, and get, green coffee at bargain prices, while at the same time maintaining the levels of retail prices for their manufactured products, resulting in a huge increase in their shares of both income and surplus. Their oligopsony market positions and control of massive stockpiles of coffee enabled them to hold world market prices at depressed levels for a considerable period and to maintain their large share of income and surplus. As for the division of the income and surplus within the producing countries, it is clear that the largest shares go to the coffee growers and the state. The relative sizes of these shares vary greatly across producing countries and over time; in some cases, such as Brazil and Colombia, the state may take about half of the total income and an even larger share of the surplus. In other countries, such as Indonesia, the state may only take about 10 percent of the income and less than a third of the surplus, leaving the bulk of both for growers. Intermediate traders and processors, and exporters, get relatively small shares of the income and surplus. To put this into perspective, we also need to consider volumes. The estimates in this paper are all based on the production and export of a pound of green coffee. But a peasant grower producing one or two hundred pounds of coffee a year and earning a large share of the surplus per pound is still worse off than a large exporter who ships one or two hundred thousand bags (13-26 million pounds) per year, but earns only a very small profit per pound. In the consuming countries, the main beneficiaries are the TNC roasters, who get
Studies in ComparativeInternational Development / Spring 1997
the majority of the income and surplus and handle huge volumes of coffee. Some of this surplus is probably passed on to workers through higher wages, but this is undoubtedly a small share. The TNC coffee importers also earn a small share of the income, less than 10 percent of the total retail value, and a small share of the surplus. I have not been able to find any estimates of the importers' costs in order to make an estimate of the importers' share of the surplus. But, as for the exporters, the TNC importers handle such large volumes of coffee that they can do quite well on a very small profit per pound. Finally, some consuming states also take a share of the surplus through taxes, which, in the case of some European countries, can be quite substantial. Most of the surplus retained in the producing countries appears to have come from strategic rents created by the ICA export quota system. Resource rents created differences between the amounts of surplus retained in different producing countries, but the average level of the surplus across all countries estimated here was probably composed mainly of strategic rents. This is most clearly the case for the 1980-89 quota period, and was probably also the case in the earlier 1962-72 period. During 1972-80, when there were no quotas in effect, there was some collaboration among producing states to maintain world market price levels, but prices were probably maintained by shortages created by Brazilian frosts in 1969 and 1972, and then increased by the 1975 frost. Fortuitously for the producers, the strategic rents were converted into temporary resource rents which lasted for most of the non-quota period. These resource rents and producers' shares of the surplus probably would have declined significantly in the early 1980s, had the strategic rents not been recreated by the reimposition of quotas. And after 1989, strategic rents seem to have disappeared entirely; the 1993-94 price increases resulted from temporary resource rents. Most of the surplus retained in the consuming countries at the beginning of the period analyzed here seems to have consisted of the normal profits of the coffee TNCs, increasingly augmented by monopoly rents as the importing and manufacturing TNCs consolidated their dominance in the core consuming markets through the 1980s. In particular, the large shifts in the distribution of income and surplus in 1986 and 1989 seem to be largely due to increasing monopoly rents. It would be interesting to compare these results to the distributions of income and surplus for other primary commodity chains. Two series of studies conducted by UNCTAD in the mid-to-late 1970s did estimate the export prices of a number of commodities as a percentage of the retail prices of the final consumer products they were used to produce. Table 6 provides a summary of these results; they are only roughly comparable, and depend in part on the nature of the final consumption product used for estimation. Nonetheless, it is clear that coffee is at the upper end of the range of producing country retained values among the commodities included here. The 1976 UNCTAD study also included a detailed cost breakdown for the banana commodity chain. The data show that profits of producers and exporters in the producing countries were less than 2 percent of the final retail price, while the profits of TNCs and retailers were probably at least one-third of the retail price (50 percent gross margin, minus storage, distribution, and advertising costs, and taxes).
TABLE 6 Export Price as a Percentage of the Retail Price of Final Consumption Products, for Some Major Primary Commodity Exports of the Third World
E x p o r t Price as % o f Retail Price Commodity Tin e Sisal d Copper e Coffee f G r o u n d n u t oil P h o s p h a t e rockg Tea h Sugar i JuteJ Cocoa k Bananas Oranges Iron ore I Alumina m Cotton Tobacco Bauxite m 3-5% 19-53% 21-53% 16-34% 12-34% 20-24% 29-32% 8-12% I 0- ! 4 % ! 3- | 6 % 4-8% 6% 2-3% 14% 19-42% 55-65% 18-50% 48-51% 45-47% 15-47% 1967-72 a 74 - 9 1 % 63-83% 54-58% 54-61% late 1970s b
Notes: a From UNCTAD (1976). Export prices refer to the average unit value of exports for all developing exporting countries, unless otherwise noted. b From Girvan (1987); UNCTAD (1984a, b, c, and d). These figures are often based on detailed studies of a few exporting countries. c Export price of tin metal as percent of price of straits tin. d Cif import price as percent of price of cloth, an overestimate because it includes shipping costs and insurance. e Export price of copper concentrates to price of copper wire. 1967-1972 data are for exports from Philippines and Bolivia only. f 1967-1972 low figure refers to retail price on the German market, high figure refers to retail price for roasted and ground coffee on the U.S. market. For late 1970s, lower figure refers to instant coffee, higher to roasted coffee on the US market. g Percent of the price of phosphoric acid. h Lower figure refers to tea bags; higher to loose tea. i Lower figure refers to sugar from the Caribbean sold in Japan; higher figure refers to sugar from Mauritius sold in UK. These figures may not be representative. J Raw jute export price as percent of price of jute yarn (higher) and jute fabric (lower). k Refers to prices of cocoa powder for export and for final consumption. The figures would be much lower for chocolate. l Export price as percent of price of steel bars and sheets. mExport price as percent of price of aluminum rods and sheets.
Studies in Comparative InternationalDevelopment/ Spring 1997
Thus, it seems plausible to infer that the coffee producing countries also retained larger shares of the total surplus than the producers of most of these other commodities.
If surplus was being extracted from the producers of primary commodities and transferred to capitalists in the core, then it appears that the coffee commodity chain was not a major mechanism for this transfer, at least from the early 1960s to the late 1980s. This analysis has indicated that during most of this period, roughly half of the total surplus generated along the entire chain was retained in the producing countries. After about 1986, however, there was a massive shift of surplus from the coffee producing countries to TNCs in the core, who used their market power to hold down the price of green coffee while inflating the price of coffee processed for final consumption. The main reason coffee producing countries fared so well for so long seems to have been the International Coffee Agreements, which were instituted in response to collective action by producing states attempting to increase their incomes from coffee exporting. The consuming states acquiesced in the establishment of the export quota system because it also suited their geopolitical aims--to provide economic assistance in somewhat disguised form to key Third World allies. The TNCs also acquiesced, because the ICAs stabilized supplies and prices of green coffee, and because they were still able to turn a handsome profit under the system. By the mid-1980s, however, neoliberal economics were becoming fashionable, the TNCs had consolidated control over the core markets, and the Soviet bloc was beginning to fall apart. These developments undermined the consensus among producing states, consuming states, and TNCs, leading to the collapse of the ICA and the huge transfer of surplus out of the producing countries. This is not to say that primary commodity production and exporting in general did not serve as a means of transferring surplus to the core. Indeed, coffee may have been unique in this respect. Coffee producers certainly made out better than the exporters of most other primary commodities, and the ICAs may have been the most successful commodity agreements negotiated in the post-war period. This is also not to say that no surplus was extracted from coffee growers, or from agricultural laborers in the producing countries. Some surplus certainly was extracted, but most of it went to the state and the largest growers, rather than out of the country. Even so, in most producing countries, coffee growing was more profitable than many altemative agricultural activities. The states were responsible for generating a large part of the surplus that remained in their countries, through the creation of strategic rents, and they kept it and used it for their own purposes. To be sure, many of these states squandered some of this surplus, but many also used it to provide services to coffee growers and rural areas, and to fund other development projects. In any case, the income from coffee exporting alone was not sufficient to serve as an engine of development, regardless of how wisely it was spent. These results suggest a number of modifications to commodity chain analysis as it
has been conceptualized thus far. First, these results cast doubt on the assumption that the bulk of the surplus generated along a commodity chain flows to the most highly monopolized nodes. The importing of green coffee and the manufacturing of roasted and instant coffee were more highly monopolized than coffee growing or coffee exporting throughout the period considered here. Yet for most of this period, coffee TNCs divided the total surplus available along the coffee chain roughly equally with coffee growers and producing states. Coffee growing in many countries approximates the ideal type of competitive small producers, yet coffee growers have at times realized high rates of profit and received relatively large shares of the total surplus. The number of coffee exporting countries increased during this period, and market shares became more dispersed, yet exporting states also received relatively large shares of the surplus. This confirms the need to measure the distribution of the surplus directly when analyzing particular commodity chains, rather than assuming that it is determined by the degree of monopolization. Second, these results point to the importance of an alternative way of generating and controlling shares of the surplus: creating strategic rents by regulating the flow of commodities between nodes of the chain. Acting collectively, the producing states developed regulatory regimes through the ICAs that enabled them to impose and collect these rents. The states then decided how much rent to keep for themselves and how much to return to the growers. The commodity chain approach has thus far focused primarily on TNCs, and has not fully incorporated the regulatory role of states. States can shift the distribution of income and surplus between nodes of a chain without directly controlling production at any node, and without necessarily changing the structure of the chain. Notes
This is a revised version of a paper presented at the annual meeting of the American Sociological Association, Washington, D.C., August 19-23, 1995. Portions of this analysis were also presented at the annual meeting of the Pacific Sociological Association, San Francisco, April 6--8, 1995. Thanks to Brian Wright for his comments on an earlier draft of this paper. My research at the International Coffee Organization was supported by an NSF Dissertation Improvement Grant. I am indebted to Mr. C. P. R. Dubois and his staff at the ICO for their assistance with my research there, in particular, Marianne Bradnock, Rebecca Adams, and Trevor Nash. Special thanks to R. S. Marks, ICO Statistician, who provided corrected and updated figures for table 2. 1. The retail price of coffee also includes the cost of other inputs, mainly packaging materials and energy, which, strictly speaking, would not b e included in value added. The production of these inputs are additional branches of the coffee chain, along which additional surplus is generated. Since these inputs are minor contributors to the total retail price of coffee, these branches are not traced here, and the costs of these inputs are included in the calculations presented here as part of the retained value, in order to simplify the analysis. 2. Again, the costs of other inputs are included in the retained value. But in this case, if these inputs are imported from core countries, including them in the retained value would cause an overestimate of the share of total income retained by producing countries. Fortunately, coffee is produced with very few imported inputs, primarily fertilizers and pesticides. Estimates by de Graaf (1986) suggest that imported inputs account for less than 10 percent of total robusta producing country income and less than 20 percent of total arabica producing country income. Therefore, the retained values presented here for coffee producing countries are somewhat overstated, but not by much. 3. To make the data on prices paid to growers comparable across countries, the ICO converts prices paid to
Studies in ComparativeInternational Development / Spring 1997
growers for all forms of coffee to their price equivalents per pound of processed green coffee (e.g., two pounds of dried cherries is equivalent to one pound of green coffee). To make the data on retail prices comparable across countries, the ICO converts the prices of all forms of coffee sold to their price equivalents for roasted and ground (R&G) coffee (e.g., one pound of instant is equivalent to about 2.2 pounds of R&G). In most of the major consuming countries, R&G accounts for the majority of coffee consumed; the major exception is UK, where 80-90 percent of the coffee consumed is instant. For most of the period under consideration, R&G can be considered to include sales of roasted whole beans. Beginning in the late 1980s, roasted whole bean specialty coffees sold at higher prices became a significant new market segment in the U.S.; for these later years, the ICO data reflect the average prices of supermarket brands. For the purpose of this analysis, the chain is assumed to end with the sale of roasted or instant coffee; the additional income and profits generated by the sale of brewed coffee in restaurants or cafes is not included. Free on board. In other words, this is the price of green coffee sitting in a ship at the port of export. Cost, insurance and freight. In other words, this price includes the cost of the green coffee, the cost of insuring it in transit between the docks in the producing countries and the warehouses in the consuming countries, and the shipping costs. Both of these costs could be considered as costs of doing business for coffee importers and manufacturers in consuming countries. They are separated out because they are relatively small components of the total price easily calculable from the basic data on unit values. The transport costs are not easily allocated between producing and consuming countries. The shipping lines, insurance companies, and banks involved are generally TNCs based in the core countries, but some of the larger exporters like Brazil and Colombia ship coffee on nationally-owned lines and obtain financing from their own banks. The weight loss is literally a deadweight loss, which does not really add to retained value in the consuming countries; by separating it out I estimate the retained value as the retail price of a pound of coffee minus the cost of enough green coffee to produce that pound. A final caveat to bear in mind is that it takes many months, at best, to move coffee from the tree to the cup. Grower prices are usually set by crop years, which also vary across countries, with most Southern Hemisphere producers (including Brazil and Indonesia) beginning April 1, while Northern Hemisphere producers (Colombia, Central America, and most African robusta producers) begin October I. It takes at least a few months to process this coffee into green beans ready for export, at which point, depending on whether quotas are in effect and on the size of the crop, it may be stored for as long as several years. Further, coffee which is sold and shipped to consumer markets may also be held in warehouses, as a part of working stocks or as a basis for speculation on the futures exchanges, for as long as a year, and occasionally even longer. In brief, although most of the coffee sold by growers in a given coffee year is processed and purchased for final consumption in that same year, a significant portion is not. This stored coffee, depending on world market price trends, may be a source of additional windfall profits, or losses, for exporters, importers, TNCs, or producing states who hold these stocks. There are no data available which would allow us to factor this Complication into the calculations. Brazil accounted for about one-third of total world production during the period considered here. Fluctuations in Brazilian production accounted for almost 70 percent of the variability in total world production over this period. Table 2 shows a consumers' share about 5 percent lower than in table 1. This appears to be due to a change in the retail price series in the ICO data for the U.S. Retail prices used in the calculations for table 1 are from the Market Research Corporation of America; those used for table 2 are from the Bureau of Labor Statistics of the U.S. Department of Labor. Changes in U.S. retail prices have a large influence on the weighted average because the U.S. accounted for about one-third of total world imports of coffee at this time. These data do not tell the full story of the division of income in the early 1980s, because of the two-tier market that opened after the quotas were reintroduced. The reimposition of the quotas, at a time when production was still expanding as a result of plantings during the coffee boom, left a number of producing countries with huge stockpiles of coffee that could not be exported to ICO member countries. Some of this coffee was offered at discounted prices to newer consuming countries which were not ICO members, and some of this coffee found its way to ICO member markets. Several analyses have concluded that producers still received higher prices overall under this situation than they would have without the quotas, but the data in table 2 probably overstate the total share of income retained by producers (Talbot 1995-96, Herrmann 1986, Akiyama and Varangis 1990, Bohman and Jarvis 1990). The LM data overestimate production costs for robusta producers, leading to negative estimates of total
surplus and growers' shares. This overestimate seems based on unreasonably high estimated processing costs, but the source of this error cannot be determined, since their methodology is not public. The editors of Carta Cafetera stated that the overestimates derived from errors in exchange rate conversion, but this does not explain why only the processing costs were affected.
AMIN, SAMIR 1976 Imperialism and unequal development. New York: Monthly Review Press. AKIYAMA, TAKAMASA, AND PANAYOTIS VARANGIS 1990 The impact of the international coffee agreement on producing countries. World Bank Economic Review 4(2): 157-73. ARRIGHI, G. AND J. DRANGEL 1986 The stratification of the world economy: An exploration of the semiperipheral zone. Review Journal of the Fernand Braudel Center 10, 1 (Summer): 9-74. BARHAM, BRADFORD, STEPHEN BUNKER, AND DENIS O'HEARN, EDS. 1994 States, firms, and raw materials: The world economy and ecology of aluminum. Madison: University of Wisconsin Press. BHAGWATI, JAGDISH 1982 Directly unproductive profit-seeking (DUP) activities. Journal of Political Economy 90, 5:988-1002. BOHMAN, MARY, AND LOVELL JARVIS 1990 The international coffee agreement: Economics of the non-member market. European Review of Agricultural Economics 17:99-118. BOHMAN, MARY, LOVELL JARVIS, AND RICHARD BARICHELLO 1996 Rent seeking and international commodity agreements: The case of coffee. Economic Development and Cultural Change 44, 2:379-404. BONACICH, EDNA, LUCIE CHENG, NORMA CHINCHILLA, NORA HAMILTON, AND PAUL ONG, EDS. 1994 Global production: The apparel industry in the Pacific rim. Philadelphia: Temple University Press. CARTA CAFETERA 1990 Comparaci6n de costos de producci6n de caf6 a nivel internacional. 1, 21 (Abril 15-30): 3-4. Bogotfi, Colombia. EDWARDS, R. AND A. PARIKH 1976 A stochastic policy simulation of the world coffee economy. American Journal of Agricultural Economics. May. EMMANUEL, ARGHIRI 1972 Unequal exchange: A study of the imperialism of trade. New York: Monthly Review Press. EVANS, PETER 1989 Predatory, developmental, and other apparatuses: A comparative political economy perspective on the third w o r d state. Sociological Forum 4, 4 (Dec.): 561-87. FORD, DEREK J. 1978 Commodity market modeling and the simulation of market intervention: The case of coffee. In F. Adams and S. Klein, eds., Stabilizing world commodity markets. Lexington, MA: Lexington Books. GEREFFI, GARY 1994 The organization of buyer-driven global commodity chains: How U.S. retailers shape overseas production networks. In Gereffi and Korzeniewicz, eds., (95-122). Commodity chains and global capitalism. Westport, CT: Praeger Publishers. GEREFFI, GARY, AND MIGUEL KORZENIEWlCZ, EDS. 1994 Commodity chains and global capitalism. Westport, CT: Praeger Publishers. GEREFFI, GARY, AND M1GUEL KORZENIEW1CZ 1990 Commodity chains and footwear exports in the semi-periphery. In W. Martin, ed., (4568). Semiperipheral states in the world-economy. Westport, CT: Greenwood Press.
Smdimin Comparath, Imu-~fion~ Devdopment/ Spring1997 e
GEREFFI, GARY, MIGUEL KORZENIEWICZ AND ROBERTO KORZENIEWICZ. 1994 Introduction: Global commodity chains. In Gerefti and Korzeniewicz, eds., (1-15). Commodity chains and global capitalism. Westport, CT: Praeger Publishers. GIRVAN, NORMAN. 1987 Transnational corporations and non-fuel primary commodities in developing countries. World Development 15, 5:713-40. DE GRAAF, J. 1986 The economics of coffee. Wageningen, Netherlands: Centre for Agricultural Publishing and Documentation. HERRMANN, ROLAND 1986 Free riders and the redistributive effects of international commodity agreements: The case of coffee. Journal of Policy Modeling 8, 4:597-621. HOPKINS, TERENCE AND IMMANUEL WALLERSTEIN 1994 Commodity chains: Construct and research. In Gereffi and Korzeniewicz, eds., (17-20). 1994 Conclusions about commodity chains. In Gereffi and Korzeniewicz, eds., (48-50). HOPKINS, TERENCE AND IMMANUEL WALLERSTEIN 1986 Commodity chains in the world-economy prior to 1800. Review Journal of the Fernand Brandel Center 10, 1 (Summer): 157-70. INTERNATIONAL COFFEE ORGANIZATION Various issues Quarterly Statistical Bulletin. London: ICO. JUST, RICHARD, DARRELL HUETH, AND ANDREW SCHMITZ Applied welfare economics and public policy. Englewood Cliffs, NJ: Prentice-Hall. 1982 LABYS, WALTER 1980 Market structure, bargaining power, and resource price formation. Lexington: Heath Lexington Books. LEE, NAEYOUNG, AND JEFFREY CASON 1994 Automobile commodity chains in the NICs: A comparison of South Korea, Mexico and Brazil. In Gereffi and Korzeniewicz, eds. ORLAND1, ALBERTO 1978 Prices and gains in the world coffee trade. CEPAL Review. First half of 1978: 161-97. PORTER, MICHAEL 1990 The competitive advantage of nations. New York: Free Press. PREBISCH, RAUL 1962 The economic development of Latin America and its principal problems. Economic Bulletin for Latin America, originally published 1949. SCHURMAN, R. 1993 Neoliberalism and natural resources: A case study of the Chilean fishing industry. Doctorate diss., University of Wisconsin-Madison. SINGER, HANS 1950 The distribution of gains between investing and borrowing countries, American Economic Review 40:473-85. STORPER, MICHAEL AND BENNETF HARRISON 1991 Flexibility, hierarchy, and regional development: The changing structure of industrial production systems and their forms of governance. Research Policy 20:407-22. TALBOT, JOHN M. 1995-96 Regulating the coffee commodity chain: Internationalization and the coffee cartel. Berkeley Journal of Sociology 40:113-49. UK MONOPOLIES AND MERGERS COMMISSION Soluble coffee: A report on the supply of soluble coffee for retail sale within the United 1991 Kingdom. London: HMSO. UNCTAD 1984a The processing and marketing of coffee: Areas for international co-operation. New York: UN. 1984b The processing and marketing of copper: Areas for international co-operation. New York: UN. 1984c The processing and marketing of phosphates: Areas for international co-operation. New York: UN.
The processing and marketing of tea: Areas for international co-operation. New York: UN. 1976 Commodities: Proportion between export prices and consumer prices of selected commodifies exported by developing countries, TD/1841Supp. 3 (January), US CENSUS BUREAU 1987 census of manufactures. Washington, D.C.: USGPO. 1990 Annual surveys of manufactures. Washington, D,C.: USGPO. Various years. US GOVERNMENT ACCOUNTING OFFICE 1977 Report to Congress by the Comptroller General, Coffee: Production and marketing systems. Washington D.C.: USGPO. VOGELVANG, BEN 1988 A quarterly econometric model of the world coffee economy. Amsterdam: Free University Press.
THE A special issue o f
This special issue of Social Research was occasioned by the publication last year of Avishai Margalit's important book, The Decent Society. The issue has been edited with the help o f both Avishai Margalit, of the Van Leer Jerusalem Institute, and Sidney Morgenbesser, of Columbia University. It is not meant simply as a discussion of'the book, but rather as a consideration of some of,the provocative issues that inhere in the idea of the "decent society" and are raised by the book. The contributors are H a r r y Frankfurt, Axel [Honneth, Steven Lukes, Philip Pettit, Anthony Quinton, A r t h u r 'Ripstein, Am61ie Rorty, Frederic Schick, and Avishai Margalit. We are pleased to offer this lively and thought-provoking issue,
SOCIAL RESEARCH 65 Fifth Avenue, NY, NY 10003 212-229-5776
 Please send me an invoice for a copy of the special issue on "The Decent Society." Copies are $7.50 each.  Please send me an invoice for a personal subscription to Social Research. 1 year $30 2 years $53 3 years $75
Name Street City State Zip
Social Research is a publication of the Graduate Faculty of the New School for Social Research. Arien Mack, Editor.