Études rurales 180 | 2007Cafés et caféiers

The world behind the world coffee marketWim Pelupessy

Édition électroniqueURL : http://journals.openedition.org/etudesrurales/8564DOI : 10.4000/etudesrurales.8564ISSN : 1777-537X

ÉditeurÉditions de l’EHESS

Édition impriméeDate de publication : 30 novembre 2007Pagination : 187-212

Référence électroniqueWim Pelupessy, « The world behind the world coffee market », Études rurales [En ligne], 180 | 2007, misen ligne le 01 janvier 2007, consulté le 07 février 2020. URL : http://journals.openedition.org/etudesrurales/8564 ; DOI : 10.4000/etudesrurales.8564

© Tous droits réservés

Cet article est disponible en ligne à l’adresse :http://www.cairn.info/article.php?ID_REVUE=ETRU&ID_NUMPUBLIE=ETRU_180&ID_ARTICLE=ETRU_180_0187

The world behind the world coffee market

par Wim PELUPESSY

| Editions de l’EHESS | Études rurales2007/02 - 180ISSN 0014-2182 | pages 187 à 212

Pour citer cet article : — Pelupessy W., The world behind the world coffee market, Études rurales 2007/02, 180, p. 187-212.

Distribution électronique Cairn pour les Editions de l’EHESS.© Editions de l’EHESS. Tous droits réservés pour tous pays.La reproduction ou représentation de cet article, notamment par photocopie, n'est autorisée que dans les limites des conditions générales d'utilisation du site ou, le cas échéant, des conditions générales de la licence souscrite par votre établissement. Toute autre reproduction ou représentation, en tout ou partie, sous quelque forme et de quelque manière que ce soit, est interdite sauf accord préalable et écrit de l'éditeur, en dehors des cas prévus par la législation en vigueur en France. Il est précisé que son stockage dans une base de données est également interdit.

THE WORLD BEHINDTHE WORLD COFFEEMARKET

Wim Pelupessy

C OFFEE is the most important tradablecrop for 25 million smallholders in 60tropical countries. Coffee growing is

part of specific global commodity chains,which are border-crossing value-added crea-ting networks of producers, traders and serviceproviders, whose end result is the use of a fi-nished commodity [Gereffi, Korzeniewicz andKorzeniewicz 1994]. They process the coffeeberries picked manually into exportable greencoffee beans, which are transformed into aroasted, ground and packed final good powder,to be consumed as a brew by cup.

The crop suffers from market imbalances andhas a long history of private and public actionsto reach equilibrium without lasting success.After profound and persisting price crises in the1990s, there are now far-reaching proposals toreform the world coffee market with the imple-mentation of a long-term strategy to reducecoffee areas. This may cause welfare disastersamong coffee cultivators and mismatches withlong-run consumers preferences.

Introduction

The global commodity chain is a globalizationmetaphor to explain the income generation anddistribution by international production activi-ties and to design policies to improve the re-sults. A key issue in our methodology is theidentification and explanation of the gover-nance structure that controls the chain and de-termines in great part how resources and gainsare allocated, who may participate and underwhat conditions. Because of the increasingdemand-driven nature of the coffee chains,we use a chain reversal approach that startsfrom the consumer and goes through the dif-ferent nodes to raw material exploitation at thegrowers’ end.

Control is executed through market powerand coordination between firms in differentsegments or between external (as NGOs, go-vernments) and internal parties in the chain[Muradian and Pelupessy 2005: 2031]. Coor-dination is the exchange of extra-market infor-mation, capabilities and activities betweenparties (within or outside the chain) not linkedthrough ownership. This is meant to ensureparticular product specifications, perfor-mances, processes and logistics. When buyersin developed economies interact with suppliersfrom developing ones, coordination is neededto get reliable transactions in cases of highrisks, heterogeneous production conditions,technological backwardness and unstable fi-nancial systems [Hobbs and Young 2001].

Recent publications have given importantinsights into the structure and functioning ofcoffee chains [Talbot 2004; Daviron and Ponte2005], but the nature and impact of the

Études rurales, juillet-décembre 2007, 180 : 189-212

sequence of the imperfect or non-competitivemarkets in the chains should be more explicitlyconsidered.

The purpose of this article is to examine thenature and outcome of coffee market structuresin order to explain the analytics of chain go-vernance and its consequences for value crea-tion and distribution. This will be useful forpolicy design to reach a more balanced deve-lopment of the industry. Section 1 will pay at-tention to the structure and performance ofconsumers’ markets in developed economies.Section 2 treats the international markets ofgreen coffee, and the national ones wherecoffee berries are traded in tropical countries,will be discussed in section 3. The impact ofthe governance structure, public and privatecoordinations in the coffee chain will betreated in section 4, while a concluding sec-tion 5 will finalize the work.

Differentiated consumers’ markets

Today more than 7 million tons of green coffeeequivalents are consumed yearly: 26% in theproducing countries and 74% in the importingnon-producing ones. In the 2000s averageconsumption per head has reached 4.6 kilo-grams in the Western markets and a mere0.7 kilograms in coffee producing countries.1

The European Union (15) with 42% of worldconsumption is the biggest block, then followthe United States and Japan with respectively24 and 9%.

In most saturated and mature markets(> 5 kg per capita) trends are negative or stable(France, Germany, Netherlands, Sweden),while in Southern Europe (Italy) these are in-creasing. The emerging markets of Brazil and

Costa Rica are the only developing countrieswith similar coffee consumption per head,slightly higher than the United States or Japan.2

In saturated markets as Sweden the average is3.5 cups daily in 2005 or 6 cups a coffeedrinker a day (75% of the population over 15as in Norway).3

The only possibility for growth is to expandthe number of coffee drinkers. Coffee consump-tion is habit creating as tested for the UnitedStates with a rational addiction model [Olekalnsand Bardsley 1996]. It turned out that currentuse increases future desirability, which may ex-plain the low negative price elasticities of de-mand between 0.2 and 0.4. The long runelasticities are, in absolute terms higher than theshort run ones [Singh et al. 1977; Akyama andDuncan 1982; Vogelgang 1988; Bettendorf andVerboven 2000; Feuerstein 2001].

QUALITY ATTRIBUTES

In the main consumers’ countries income elas-ticities are smaller than 1, because of marketsaturation and other negatively impacting fac-tors as tea or soft drinks habits, increasingincome inequality and health concerns. Never-theless, final consumers’ markets have a

1. Foreign Agricultural Service, December 2006: “Tro-pical poducts: world markets and trade,” WashingtonDC.

2. Foreign Agricultural Service/United States Depart-ment of Agriculture, circular series 2004 and 2006: “Tro-pical poducts: world markets and trade,” WashingtonDC.

3. See European Coffee Federation 2006: “EuropeanCoffee Report.”

. . .190

Wim Pelupessy

dynamic nature because of the widespread firstwave of differentiations and the willingness ofconsumers to pay premiums for higher qualities.These were traditionally based on intrinsic orsensorial quality attributes, such as body, acidityand aroma [Boot 2002]. The characteristics arespecified by geographical location, altitude andperformance of the coffee cultivation, harves-ting, processing and blending practices.

Each country has its own taste preferencesand roasters adapt to these by offering coffeeblends or mélanges. There are great differencesin the composition of coffee varieties importedby consuming countries. The United States im-ports are for about 40% Other Milds mainlyfrom Central America and Mexico, because ofpolitical reasons and relatively low transportcosts. The blends in Sweden and Japan arehigher quality Colombian Milds intensivecomplemented with Other Milds (both > 40%),because of the sophisticated tastes of verywealthy consumers. Eastern European coun-tries, Spain, France and Italy use more than75% Robustas and Unwashed Arabica.4 Lowerper capita incomes, but also espresso popula-rity in Italy, France and Spain play a part in itand this may create market opportunities forlow quality coffee producers. The blending androasting intensity provide much of the parti-cular taste to a coffee brew.

A pyramid of blends gives each variety andquality its own place: a tastemaker, which is asmall but highly priced top of Colombian andhigh quality Other Milds (e.g. SHG or highaltitude coffee from Costa Rica); a largercomplementary part made of lower qualityOther Milds (lowland coffee from Nicaraguaand Costa Rica) and Unwashed Arabica

(Brazil); and the filler with an even greatershare of Unwashed Arabica and Robusta(Uganda, Vietnam). The precise compositionof a blend is a well-kept business secret andentry barrier. Substitution with cheaper varie-ties may take place as long as the taste of theparticular blend will not change notably [God-dard and Akiyama 1989: 148; Sellen and God-dard 1997: 134]. In the last fifteen years therehas been a tendency in the main importingcountries to substitute Colombian Milds for thelower priced Other Milds. The very popularRed Brand (Roodmerk) coffee of Sara Lee/Douwe Egberts with more than 40% of thetotal market in the Netherlands, uses in itsblends from one third till half part of Robusta,Unwashed Arabica or low quality Other Mildsin its blends [Hooghiemstra 1991: 16; Pelu-pessy and Van Tilburg 1994: 245]. Differen-tiation gives the Silver (about half Arabica)and Gold Brand (100% Arabica) with higherprices and much smaller market shares.

Other entry barriers are the transformationefficiency and roasting technology, high ad-vertising and other sunk costs (5-10% ofgross income), brand loyalty of customers,etc.5 To control the chain lead firms need toinvest high amounts of so-called sunk costs ininfrastructure, equipment, brand introduction

4. International Coffee Organization, 21-8-2004: “Price-elasticity of demand and coffee consumption in impor-ting countries,” p. 8.

5. 10 till 20% of the weight are lost in the transformationof green to roasted coffee beans. Efficient large-scaleroasters operate near the minimum percentage, whilesmall ones approximate the maximum limit.

. . .191

The world behind the world coffee market

and other fixed marketing costs. A recent fa-vourable technological innovation for largeroasters’ blends is the possibility to eliminatesome “bitterness” from the taste of Robustaand other low quality coffees. This may re-duce the costs of a mélange. Another is thedevelopment of consumers’ preparation ma-chines, which allow a better appreciation ofthe taste, aroma and freshness, like espressomachines, senseo, coffee pads, etc. [Muradianand Pelupessy 2005].

Mélanges are used by big and small roas-ters and traders to provide both high and lowqualities coffees. Nevertheless, the highconcentration ratios of roasters and retailersare also important to explain the blending prac-tices, which are instrumental to diversify sup-pliers and reduce costs and risks. In mostimporting countries, a few roasters control avery large part of the market with a fringe ofsmaller firms for the rest. For example, in theNetherlands Sara Lee/Douwe Egberts has amarket share of about 70%, Albert Heijn (Mar-velo) has another 12% and the rest of themarket is divided among some 20 smaller roas-ters. The largest four have a combined shareof more than 90% in this country. Four largemultinational companies provide more than60% of all coffee in the 25 main consumingcountries: Nestlé, Jacobs/Kraft General Foods,Sara Lee/Douwe Egberts and Procter &Gamble [Talbot 2004: 103-104]. These firmstogether with Starbucks, control 75-80% of theUS market.6

Let’s give a more detailed picture of themarket power of the earlier mentioned multi-national companies in the roasted and groundcoffee markets of 12 countries, representing

65% of world consumption in 2004-2005. Thesmall roasters participations are at a great dis-tance from the leaders. The multinationals arealso market leaders in several countries. SoNestlé leads in United Kingdom, Japan, Aus-tralia, Portugal and China. Kraft Foods isnumber one in the United States, France,Sweden and Canada. Sara Lee leads in the Ne-therlands, Belgium and Brazil. Tchibo has afirst place in Germany, Austria and the EasternEuropean countries Poland, Czech Republicand Hungary [Daviron and Ponte 2005].7 Onthe demand side, there are the individualconsumers, who are numerous with insignifi-cant market shares and low (negative) priceelasticities of demand because coffee is subjectto habit formation.

The second differentiation wave has to dowith the sustainability of the production pro-cess of coffee, which is an extrinsic or cre-dence attribute of the product. In 2003 thisincluded an estimated 1% of the worldwideexports. Ranked by quantities, the most impor-tant sustainable coffees had Organic, FairTrade, Utz Kapeh, Rainforest Alliance andshade-grown certifications [Daviron and Ponte2005: 167], sold under a diversity of brands.An organic production technology pretends tomaintain a sustainable agro-eco-system.Shade-grown coffee is supposed to preserve

6. UNCTAD, The coffee guide 2007: “The mainstreammarkets for coffee”, p. 6.

7. See also UNCTAD, The coffee guide 2007: “Nichemarkets, environment and social aspects,” “World coffeetrade,” “Futures markets,” “Quality control issues,” “Themainstream markets for coffee.”

. . .192

Wim Pelupessy

biodiversity and birds’ habitat, while fair tradeguarantees a fixed buying price for the coffee.Utz Kapeh and Rainforest Alliance are mix-tures of the earlier mentioned certificationswith weaker conditions, aiming at acceptanceby the mainstream blends.

In the following years all certificated sus-tainable coffees increased strongly. For 2005the International Coffee Organization (ICO)estimated that organic coffee exports grew toabout 0.9%8 of total exports, which was almostthe share of all sustainable categories togethertwo years ago. Since almost 70% of all organiccoffee exports were exported to the UnitedStates, Germany and Japan, market shares willbe considerably higher for these countries. Inthe same year Fair Trade coffee reached 0.7%of total exports, which were for 60% concen-trated to the United States, France and theUnited Kingdom.9 Utz Kapeh showed a per-centage of 0.9%, which makes 2.6% of worldexports for the three certifications together.

The volumes of certificated sustainabilitycoffees are still minor parts of the globalconsumption, despite their high growth rates asis shown for 2005. In absolute quantities mostorganic and Fair Trade coffees are consumedin the United States, but the percentages aresmall because of the extension of this market.These two certifications together vary from1.7% for the Japanese (only Organic) to 4.2%for the United Kingdom market.10 The figuresare still incomplete and Utz Kapeh certifica-tions are significant and also increasing. Noreliable information was available on Rain Fo-rest Alliance and shade-grown coffees, whichconsumption could be similar to the Utz Kapehfigures. The sustainability differentiation is

considered as part of the specialty coffee seg-ments, which cover 10% of the market. Origi-nally this term was used in the United Statesfor products sold in special coffee shops.Today it is a catchword for single origin, sus-tainable and unconventional (flavoured) cof-fees, high quality blends.11 Commercially thespecialty segment includes single origin highintrinsic value coffees for niche markets as Ja-maican Blue Mountain, Indonesian Luwak,premium brands of superior quality organiccoffee and natural Robustas, and mainstreamaverage quality. Small specialized roasters andtraders introduced the sustainability categoriesand are active in niche markets [Daviron andPonte 2005]. The big companies dominate thepremium brands and mainstream qualities ofspeciality coffees.

PRICES

In the differentiated coffee markets consumerspay price premiums for both intrinsic (sensorial)and extrinsic (non-sensorial) product attributes.

Retail prices of mainly standard blends arevery different among importing non-producing

8. UNCTAD estimations gave organic consumption as1% of the world total.

9. UNCTAD, The coffee guide 2007: “Niche markets,environment and social aspects,” pp. 12, 24.

10. International Coffee Organization, 19-4-2006: “Ef-fects of tariffs on the coffee trade,” and 25-9-2006:“Overview of the coffee market.”

11. UNCTAD, The coffee guide 2007: “Niche markets,environment and social aspects,” “World coffee trade,”“Futures markets,” “Quality control issues,” “The main-stream markets for coffee.”

. . .193

The world behind the world coffee market

countries. In a period of fifteen years (1990-2005) Japan had always the highest prices, Italywas second and Germany third. In 1990 theUnited States had the lowest average prices,which could be a consequence of the standard(low) quality blends consumption, withoutmany differentiations at that time. In the 2000sthis changed, as the cheapest roasted coffeeblends were then sold in France and in Swedenat prices also below those of the US. It is dif-ficult to find an explanation for the differencesand one may consider in the first place the pricesand qualities of the imported green coffee va-rieties. In the beginning of the 1990s Germany’simport package was for 70% composed by theexpensive Colombian Milds and Other Milds.12

However, the participation of these two varie-ties was considerably less in the imports ofJapan and Italy, which showed the highest re-tail prices of that time. The United States hadthe highest share of Other Milds and togetherwith Colombian Milds this reached a signifi-cant 60% of their imports. In the 2000s strongreductions of the import of Colombian Mildstook place in Germany with one of the higherroasted coffee price levels, as well as in theNetherlands and Sweden with moderate consu-mers’ prices. The latter strongly increased theimports of expensive Other Milds. Neverthe-less, consumers’ prices of Sweden showed adeclining trend to become one of the lowest in2005, despite their relatively expensive coffeebeans imports.

All the importing countries showed a similartrend in their retail prices: a strong increase inthe beginning 1990s and afterwards a decreasewith for some countries a partial recovery in2005. For others there was a continuous decline

since 1995, while Italy was the only countrythat ended in 2005 with higher prices than in1995. The breakthrough of speciality coffeesmay have increased the US price. Blendingpractices make it difficult to relate consumers’prices directly to those of the world market.Premiums for extrinsic characteristics of the se-cond wave of differentiation may give addi-tional price variations. However, their impactcould not be traced because of the reducedshares of certified coffees. Only Japan and Ger-many showed in 2005 high average retail prices,while the US, France and the UK belong to thegroup with the lower prices.

The market structures in Europe were cha-racterized by oligopolistic interdependency [Bet-tendorf and Verboven 2000; Feuerstein 2001].Prices for roasted coffee rose less than propor-tional with green coffee bean prices, becauseonly part of the marginal costs was accountedfor by green coffee, and a (small) reduction inmark-up took place. The concentration of marketpower by one or some roasters, does not excludecompetitive behavior by others.

A notable feature is the combination of pro-duct differentiation and brand loyalty for theposition of retailers in consumer markets. Pro-duct differentiation based on blends is gainingimportance with the strong increase of so-called gourmet and special coffees, which invalue have already captured 30% of the USmarket. In importing countries the demand isstrongly increasing for these higher quality and

12. International Coffee Organization, 21-8-2004: “Price-elasticity of demand and coffee consumption in importingcountries,” p. 8.

. . .194

Wim Pelupessy

other differentiated coffees [Fitter and Ka-plinsky 2001]. Competition among roastersoften takes place with non-price instrumentsas blends, advertising and bonus systems tomaintain brand loyalty, rather than with prices.As we have stated before, the geographicorigin plays a role in the differentiation of va-riety and quality.

Supermarkets and other retailers distribute70% of the coffee to final consumers.13 Theother 30% is consumed in restaurants, hotels,bars, institutions (work) and specialty shops asStarbucks and the Coffee Factory. The trendtowards high concentration in the roastingstage can also be observed in retailing, wherein most European countries the largest five su-permarket chains have a joint market share ofover 50%. Their weight vis-à-vis the roastersis considerable and increasing. Some retailersalso engage in backward integration with roas-ting and own brands as the case of the super-market chain Albert Heijn and roaster Marveloin the Netherlands. Nevertheless, the roastersare still powerful because of the definition ofthe brands by the specific blends.

One of the few studies on the coffee pricingstrategies of retailers is done for the US byP. Nelson and his colleagues [1992] assumingthere is monopolistic competition in the roas-ting industry with differentiation of productssupported by brand advertising. The studyfound that prices are influenced by the relativemarket share of the brand (+), times when andplaces where competition is unusually strong(–), advertising per unit of sales of competitors(–) and the prices of green coffee beans (+).Price is reduced and ads increased whencompetition becomes stronger.

An explanation for the positive observed re-lationship between market share and price couldnot be explained by consumers’ taste prefe-rences for the brand and ad-hoc evidence sug-gests that retailers cut their margins whenwholesale prices are increasing. The specialpreferences are therefore more likely to haveoriginated from the retailers than from the finalconsumers. It is possible that the retailers usethe roaster’s brand as a special offer, because itinduces extra-sales of other products. This retailstrategy may be worthwhile when there is alarge market share. Most other research consi-ders the retail trade only as intermediary thatbuys roasted coffee to sell it with a margin toconsumers, without particular own preferences.

Concluding, we may say that in the mainimporting countries it will be practically im-possible to increase the per capita consump-tion of coffee significantly due to marketsaturation and other factors. In these marketsthere may still be opportunities to augmentthe number of coffee drinkers. Emerging andunderdeveloped markets with less than 2 cupsper head a day, may offer more opportunities.However, in most developing countries coffeeconsumption is stagnating, with the exceptionof the largest domestic consumption marketof Brazil and the much smaller one of CostaRica.

The preservation and expansion of marketshares are therefore of utmost importance forroasters and retailers. Despite their high andstable concentration, both price competition

13. See European Coffee Federation 2005: “EuropeanCoffee Report.”

. . .195

The world behind the world coffee market

and markups are moderate. Large roasters andsupermarket chains are also entangled in astrong competition struggle, as in most of thefood industry. Non-price competition is moreimportant, because quality enhancement, pro-duct differentiation and innovation are themain trends. These trends are only partiallyvisible in the big differences among countriesin the varieties of imported green coffee. Mostof the upgrading for the consumer is done byblending, roasting intensity, packaging andbranding in the consuming countries, whichhave affected retail coffee prices.

Traditionally, all this had been related tothe intrinsic product attributes. Since the 1980sthe attention for extrinsic qualities as environ-mental and social friendliness, has acceleratedthe differentiation of consumers’ markets. Ad-vertising and large retailers strategies play cru-cial roles in this process. Large roasters couldmaintain positions by their blendings and cor-responding brands, while smaller ones coulddo so by specialization for niche markets. Theassumptions of coffee as a homogenouscommodity and the independence of the out-comes of vertically related markets in thechains must be reconsidered.

International markets

SPOT MARKETS

Because production takes place for almost100% in 57 tropical developing countries andconsumption for some 75% in 43 developedones, international trade is an essential linkageof the coffee chain. Here green coffee is mostlytraded and therefore the minor internationalflows of decaffeinated, roasted and ground, or

instant coffees will only indirectly be treatedin this article, as trade statistics are registeredin green coffee equivalents.

The major country importers of coffee arethe United States, Germany, Japan andFrance with shares of 25, 19, 8 and 6% res-pectively.14 Major exporters are Brazil (29%),Vietnam (15%), Colombia (11%) and Indo-nesia (7%). This implies the presence ofconcentration ratios (biggest four) of 0.58 forboth importing and producing countries. Butif we put the two trading blocs EU (25) andUS together, they will count for almost threequarters of all coffee imports. Mention shouldalso be made of re-exports by importing coun-tries, which have been increasing conti-nuously from about 19% of the total exportsin 1999 to 30% in 2005. Countries like Ger-many and the US rank fourth and seventh asexporters. A considerable part of their exportsare processed coffees. For 2004 and 2005 al-most all (93%) of the re-exports from the EU(25) to non-EU destinations were decaffei-nated, roasted and soluble coffees.15 This hasincreased the value added share of re-exportscompared to exports. The average prices ofthe coffee re-exports from the EU (25) tonon-EU destinations were 3.39 and 4.32euro/kg for respectively 2004 and 2005.

14. Imports and exports (bags) as an average percentageof world imports and exports in 2004 and 2005 of res-pectively ICO importing and exporting members, seewww.ico.org.

15. See European Coffee Federation 2006: “EuropeanCoffee Report,” p. 7. Most re-exports are intra-europeantrade.

. . .196

Wim Pelupessy

The prices were for the incoming coffees res-pectively 1.11 and 1.55 euro/kg for the sameyears.16

Market studies gave low-income elastici-ties of demand for high per capita consumingregions and relatively high values (> 1) forlow consumption economies, while price elas-ticities of demand have low (negative) valuesfor all. International prices are also sensitiveto changes of Brazil’s coffee production,which were often related to weather condi-tions. However, the declining share ofBrazil’s coffee and the changes of the locationof its cultivation have reduced these sourcesof volatility. L.S. Karp and J.M. Perloff[1993] demonstrated that even Brazil and Co-lombia together could only exert a limited de-gree of international market power. Largeinventories in consumers and producers coun-tries may also reduce the impact of big pro-duction fluctuations. The differences betweenthe available exportable coffee and consump-tion as indicators of the oversupply gap in theworld markets have declined in recent years.To estimate available supply, 50% of thecoffee stocks in producing countries were dis-counted as not immediately tradable and4 million bags from the inventories of impor-ting countries as iron ratios [Gilbert and Zant2001]. With the reduction of global over-supply the ICO price indicator has been in-creasing in the 2000s.

In real life the international coffee trade isconducted by a variety of private agents: dea-lers, brokers, specialized traders/exporters,and importers, large importing roasters andretailers of consuming countries. A strikingfeature of the international spot trade of coffee

is the concentration in the market. Only 7 tra-ding companies handled 56% of the globalimports in 1993 [Wheeler 1995]. More recentdata show 41.5% as the total share of the bigfive: Rothfos, E.D. & F. Man, Volcafé, Car-gill and Aron, while the top ten reaches 62.2%[Fitter and Kaplinsky 2001]. Volcafé boughtE.D. & F. Man and Cargill merged with Es-teve in the 2000s. Earlier Rothfos was takenover by Neumann which made the two largesttraders Neumann and Volcafé sharing 20% ofthe market and the top six with 50%, whilelater on the biggest three controlled 45% ofthe market [Daviron and Ponte 2005: 91, 93].As mentioned before, with the second diffe-rentiation wave, small traders and roastersemerged in the specialty segment trade.

J. Morisset [1998] is one of the few studiesthat deal explicitly with the role of the largetrading houses on spot markets. He found thatthe sources of the observed widening of thespread between international market andconsumer prices over a twenty-five years pe-riod were likely to be commodity-specific. Formany commodities as coffee, 6 or fewercompanies control about 70% of the interna-tional trade. These firms could hardly be onlyprice takers in the market. However, the mul-tinational concentration in the downstreambuyers market of big roasters and retailers be-came even stronger. This made the interna-tional trading houses loosing ground after the1980s; while the smaller and less specializedones were shaken out.

16. See European Coffee Federation 2006: “EuropeanCoffee Report,” pp. 7,9.

. . .197

The world behind the world coffee market

FUTURES MARKETS

The futures market is in the first place a fi-nancial market, in the sense that contracts arenearly always canceled by an opposite trans-action at the end of its duration. They are usedfor risk reduction and speculation and mayalso attract parties not involved in the coffeebusiness. In the London International Finan-cial Futures and Options Exchange (LIFFE)the volume traded was approximately fivetimes global Robusta coffee production in1997, while in the New York Board of Trade(NYBOT) this volume was over nine times theworld exports of all types of coffee. There wasan explosive increase of hedging and specu-lation over time in both New York andLondon. From the beginning 1990s to 2006the futures and options volume turnovers in-creased from 11 to 19 times that of grossworld imports of coffee.17

This indicates that a considerable amountof speculation is present in these markets. Fu-tures markets integrate into their prices allavailable information on (expected) supply anddemand and through arbitrage they may in-fluence the prevailing spot prices. The in-fluence of futures on spot markets makes itdifficult to explain international coffee pricefluctuations. Theoretically futures pricesshould approximate spot prices in the long runas contracts reach maturity and may providesome useful signals to facilitate decision-ma-king with respect to production, sale, purchaseand storage. If futures prices are unbiased es-timators of spot prices in the future, futuresmarkets may be used by developing countriesfor dealing with price risks.

C.L. Gilbert and C. Brunetti [1997] showamong others that the international price move-ments in 1993 were mainly due to the mereannouncement of the coffee retention schemeof the Association of Coffee Producing Coun-tries. The retention itself had fewer effects onfutures prices, which could indicate the impor-tance of expectations. In 1994 investors andspeculators from outside the coffee business,who diverted investment away from bonds andequity into commodities, caused price upheaval.The portfolio diversification away from equityand bonds into coffee was a very small amountin relation to the overall bond and equity ope-rations, but nevertheless it was a very signifi-cant sum of money for the coffee market.Because their profits were high due to the unex-pected frost in Brazil, coffee became a popularinvestment for a while, which was visible in theinterest that speculators took again in coffee in1995. The only difference was that the specu-lators now came less from outside, but morefrom commodity investment funds.

Although single actors cannot influencemarket outcomes, it is possible that a group ofspeculators with coordination can bid upprices, or when closing their positions all atthe same time, make the prices drop. Thisseems negative, but on the other hand, theiractions can also work in favour of coffee gro-wers. In fact, it is necessary on futures marketsthat there are parties who take opposite posi-tions, because otherwise no contracts can be

17. UNCTAD, The coffee guide 2007: “Quality controlissues,” “The mainstream markets for coffee,” onwww.intracen.org.

. . .198

Wim Pelupessy

concluded. Their functioning thus fits in thewider discussion on the positive or negativerole that speculation plays in markets. Surelyfutures markets may be strongly and erraticallyaffected by expectations and it is not clear ifparties concerned could affect them conti-nuously in a systematic way.

An interesting case was when at the end ofthe 1970s, Brazil, Colombia, Mexico and someCentral American countries, could success-fully raise their quotations on the futuresmarket through a joint private company Pan-café [Pelupessy 1993: 45]. It functioned as anordinary dealer, only bigger and over a longerperiod. Losses occurred because the accumu-lation of stocks was not followed by shortagesand the Commodity Futures Trading Commis-sion in the US changed their rules of the game.Brazil closed the initiative in exchange for thewillingness of large importing countries to ne-gotiate the fourth International Coffee Agree-ment. Another study showed how Côted’Ivoire could stabilize export incomes by fu-tures market operations [ibid.].

Finalizing this section on the internationalmarkets, it is clear that they are also veryconcentrated. However, the big internationaltrading houses have lost ground to roasters andretailers, who dominate the governance struc-ture of the coffee chains. Their relationshipwith international traders tends to be that ofunstable bilateral oligopolies. The impact offutures markets on spot prices makes thingsmore complicated, as (financial) speculatorsfrom other sectors have strongly increased par-ticipation. This makes market fundamentalsless relevant for determination of spot prices.Country behaviour based calculi were and are

not adequate to reach international market ba-lances, since only private companies can bechain actors and market parties.

Producers’ markets

The coffee demanded by final consumers indeveloped countries and traded and processedin the different stages of the chain, will be sup-plied by millions of smallholders operating onlocal markets in tropical countries. Buyers onthese markets are a much smaller number ofmiddlemen or intermediaries and processors.

COFFEE GROWERS

Usually coffee growers possess small plots, va-rying from on average 0.5 hectare coffee inEthiopia and Tanzania to 1.16 hectare in Indiaand more than 5 hectares in Nicaragua. Largeestates are scarce since internal economies ofscale are limited and land tenancy is traditio-nally fragmented for this crop. Normally, onlya minor part of the production is harvested onproperties of more than 10 hectares.

The grower’s income and production capa-city are the two key issues, which are normallyraised in the assessment of this end of thechain. They may provide arguments for public,NGO and other interventions in domestic mar-kets. Coffee shrubs become productive onlyafter three to five years; response to price mo-vements is therefore slow and price supplyelasticities are low. Most supply reactions areconsiderably less than proportional and in thefirst year not higher than 0.2, because in theshort run only the intensification of plot main-tenance and harvesting could raise production[Akiyama and Duncan 1981; Pollard and

. . .199

The world behind the world coffee market

Graham 1992; Pelupessy 1993; Branchi et al.1999]. For periods up to eight years themaximum estimated elasticity is 0.6.18 In thelong run the value of the elasticities mayincrease, such as in Vietnam, Brazil andColombia, where apparently there were possi-bilities to enlarge the agrarian frontier or tosubstitute other crops. The lack of smallhol-ders’ access to inputs and external servicesmay also cause low elasticities. Credits are res-tricted for small export commodity growers inliberalized countries.

Middlemen frequently operate between far-mers and processing plants or exporters, as inIndia and the “sabsabies” in Ethiopia. Theirrole is often subject to much debate: do theyabuse market power as is assumed by follo-wers of the different Fair Trade campaigns, oris their behaviour competitive, especially if therisk and transactions costs are calculated of theservices they provide to smallholders? Theseinclude the first collection of the harvested ber-ries, credit, transport and storage. Interventionby government agencies or parastatal monopo-lies as in Ethiopia, Ghana, Uganda (until 1990)and Vietnam had not always been favourableto growers.

Because of scale advantages, first proces-sing and hulling mills are far more concen-trated, which strengthens their market position.Therefore, minimum producer prices esta-blished by marketing boards or other institu-tions could be advantageous to small and lessefficient growers. The growers often bear therisk of fluctuating world market prices. In avery productive country such as Costa Ricathe impact of price fluctuations may haveenhanced efficiency, probably due to small-

holders’ good access to inputs, credit and ser-vices [Cardenas 1994]. Other important rea-sons are the cultivation on high altitudes andadequate first processing facilities, which re-sults in very good quality Other Milds coffeefor this country.

In general, the market structures for berriesseems to be oligopsonistic. For each interme-diary there could be from 20 (Haiti) to 250 ormore growers (Côte d’Ivoire), while the moreconcentrated mills may serve from 100 inBrazil to 18.000 growers each, as in Côted’Ivoire. This means that growers are gene-rally price takers with a productive behaviourof either adaptive expectations19 or rational ex-pectations.20 However, only about a third ofthe variability in agricultural investment beha-viour could be related to market price move-ments [Branchi et al. 1999].

Foreign buyers with access to internationalfinance become very powerful when localcredit institutions are lacking or inefficient. Anexample is given by the emergence of verti-cally integrated exporters in Tanzania as sub-sidiaries of multinational companies, whichbecame the owners of all private processingmills in the country [Temu et al. 2001]. Butalso in this case a sufficient number of compe-titors and institutional improvements may res-trict oligopsonistic behaviour. The creation of

18. After full adaptation in the long run, elasticitiesmight be much higher (2-3) when new land is available.

19. See case studies of Costa Rica, Mexico and El Sal-vador in F. Jaramillo and T. Akiyama [1990].

20. See Jamaican case in S.K. Pollard and D.H. Graham[1992].

. . .200

Wim Pelupessy

producers’ cooperatives, increasing competi-tion by domestic buyers, falling supply andlarge-scale smuggling, as in Ethiopia, Uganda,and Bolivia, may also strengthen the marketposition of growers. Governments can also af-fect the outcomes by providing processing fa-cilities, setting minimum producer prices, orapplying other rules.

The liberalization of the market for coffeeberries and elimination of the government mar-keting boards may have resulted in higherprices for growers and capacity enhancing be-haviour. However, lack of access to inputs andprice uncertainty require institutional improve-ments to foster investment and a competitivemarket structure [Akiyama 2001; Temu et al.2001]. These improvements accounted for twothirds of the explanation of growers’ beha-viour. Since institutional change needs time,the higher long run price elasticities becomerelevant and in this way persisting price de-creases may have enduring negative effects onproduction capacity.

Considering the growers’ income by poundof coffee for five important growers’ countries(Brazil, Colombia, Guatemala, Indonesia,Vietnam) representing 60% of world produc-tion, it is interesting to note that these incomesare more coffee variety rather than country-specific.21 Internationally Robusta cultivatorsget significantly less income per pound thanArabica. The figures for Brazil and Indonesiaare eloquent and clearly differ according to va-riety. World market prices play an importantrole and possibly cultivators income is what isleft after deductions of fixed processing, trans-port and trade margins. In the 2000s Robus-ta growers’ income per pound increased

continuously, while Arabica dropped in thefirst one or two years, to assume increaseslater. These differences also appear for the twovarieties within the same country, while gene-rally the Robusta-Arabica growers’ incomegap tends to reduce in time.

PROCESSING

A second segment of the domestic coffee chainin producing countries is that of processed orgreen coffee production, which may be sold bythe mills to either exporters or domestic roas-ters. The coffee can be processed in both dryand wet ways. In the dry system the first dryingis done by the grower, to be delivered or storedas parchment coffee until the time is suitableto sell. This occurs frequently for the Robustavariety in for example Uganda, Vietnam andIndia. A great deal of Arabica in Ethiopia andNicaragua is still processed by the dry method,which is simple and cheap when sun drying isused. In Brazil most of the (natural) Arabicais also sun-dried. Under the wet method as ap-plied in for instance Central America, farmersdeliver the fresh berries to a processing plantwithin twenty-four hours of picking, which re-duces their market power. The berries are thentransformed; first into parchment and later intoexportable green coffee. In Colombia decen-tralized wet processing is in most cases inte-grated on the farm. Wet processing is morecapital-intensive than the dry one, but if cen-tralized it is also much more controllable and

21. See International Coffee Organization, 16-5-2005:“Overview of the coffee market;” 12-9-2005: “Obstaclesto consumption” and February 2007: “Letter from theexecutive director.”

. . .201

The world behind the world coffee market

often applied to obtain homogenous high qua-lity Arabica. Washed Arabica could receiveconsiderable premiums, as in Ethiopia, whereit varies between 20 and 100%. However, thereis a trade-off between this kind of quality in-creasing modernization and the environmen-tally friendlier dry method production, whichcan also be rewarding if sold as certified Or-ganic coffee.

Little is known about the marketing struc-tures between mills and exporters or roastersin coffee growing countries. Processors andexporters could also be vertically integrated asin Haiti, the vertically integrated (multina-tional) exporters in Tanzania and the coopera-tive mills in Costa Rica. Other exporters maybe domestic private companies or subsidiariesof foreign traders or governments agencies. InIndia and Tanzania auctions are playing an im-portant efficiency-enhancing role as marketmechanisms for exports [Akiyama 2001; Temuet al. 2001]. These countries and others likeUganda and Togo have shown a decreasingspread between growers and export prices, asa consequence of liberalization and improvedaccess of new actors to the local green coffeemarket. Export taxing had mostly been discon-tinued because of outward oriented strategiesand the international coffee crisis of the 1990s.

DOMESTIC CONSUMPTION

Domestic consumer markets of coffee growingcountries include about a quarter of worldconsumption, but represent the better opportu-nities for growth in the near future. Processorsusually sell lower quality green coffee to localroasters who generally manufacture single

variety ground coffee using outdated techno-logy of a dull and uneven roasting. Greencoffee traded for domestic consumption mayvary from between 25 and 50% of the harvestin Ethiopia and other local market orientedproducers as Sri Lanka, India and Bolivia. Itis 10% or less for the other typical export-oriented countries, with the exception of Brazilwhere about 30% of the production has a do-mestic destination. Consumption per capita isstill low in most producing countries, varyingbetween 2.6 kilograms in Nicaragua and1.5 kilogram annually in Ethiopia to 0.07 kilo-gram in Tanzania. This is considerably belowthe levels in developed countries, which mayreach 8 kilograms per capita or more.22

Today, in most cases the domestic trade forconsumption has been liberalized, but this hasnot dynamized local demand, despite thegrowth potential of underdeveloped markets.Low income per capita in the first place, tea-drinking habits and capacity shortage of roas-ting factories have negative effects. Aninteresting development has taken place inCosta Rica where falling world market pricesat the end of the 1980s and the booming touristsector have created higher domestic demand(4.34 kg/cap) and higher local prices for qua-lity coffee. It is to be expected that high qualityand product differentiation preferences fromdeveloped consumer markets will blow over toupper class markets in producing countries.Brazil with 5.01 kilograms/capita domestic

22. Foreign Agricultural Service/United States Depart-ment of Agriculture, circular series 2004 and 2006: “Tro-pical poducts: world markets and trade,” WashingtonDC.

. . .202

Wim Pelupessy

consumption is increasing above the stagnatedEU (25) average.

Policies such as forced delivery to the localmarket, retail price control and consumptiontaxes sometimes have negative effects on in-ternal prices for roasted coffee. For smallhol-ders, domestic consumption with lowerquantity and quality requirements may offerinteresting alternatives to exports. Price elas-ticities of demand (negative) are estimated as0.30 for Brazil and 0.379 for Haiti, while thelatter showed an income elasticity (positive) of0.804.23 Concentration of private exporters,state agencies, processing mills and domesticroasters may create bilateral oligopolies andcollusive actions in this market segment of pro-ducing countries, but the potential entry ofnewcomers stimulated by liberalization, mayrestrict the full and continuous excise ofmarket power.

Concluding this section it is clear that at thesupply side of the markets for coffee berriesthere are numerous smallholders with little orno market power. Supply elasticities are smallin the short run because of the specific natureand technology of the product, difficult accessto inputs and credit and scarcity of availableland. Middlemen, traders and processing millsare more concentrated, making it a buyersmarket, sometimes an unstable oligopsony.The elimination of state intervention in berriesand green coffee markets may have deliveredbetter prices to growers in some cases, but hasnever automatically improved the institutionsneeded for sustainability. An internationalcomparison of growers’ income in the 2000sindicates that differences are variety and lesscountry-specific. Robusta income per pound

has been growing continuously, while Arabicais still on top, but differences are decreasing.The domestic chain segments after processingare either vertically integrated or of a bilateraloligopoly nature. Consumer markets in produ-cing countries are still underdeveloped interms of quantity and quality, and are only forBrazil and Costa Rica in the emerging growthstage. Nevertheless, more attention should bepaid to developing markets as they have thereal potential to grow in the medium term andto become a feasible destination for small-holder production.

Governance and coordinationsFrom the assessment of the markets sequenceone may infer that the coffee chain could notbe seen as a randomly evolved network of bu-sinesses working with coffee. There are key ac-tors or leading forces who control and organizethe strategic parts of the chain. This is done forrent creation and appropriation, with price andnon-pricing behaviour based on market powerand coordination. Therefore we could not en-dorse the opinion of R. Fitter and R. Kaplinsky[2001: 78] that governance is “largely absent”in coffee chains. Coordination may be executedin a direct or indirect way. This would not bepossible if all coffee transactions take place inperfectly competitive markets, but this is defi-nitely not the case. Coffee markets are highlyorganized from within and outside the chain byprivate and public parties.

23. We may observe that Brazil and France have thesame consumption per head of 5.01 kilograms yearly andthe same negative demand elasticity of 0.30 as well.

. . .203

The world behind the world coffee market

PRIVATE

Leading actors have established rules to standar-dize and normalize market transactions and alsoapplied coordination in different chain segments.These are accepted by and legitimate for all chainactors and other parties involved. Grading andclassification is based on botanical variety, alti-tude or region, processing (wet or dry), bean sizeand shape and colour, number of defects, roastappearance, cup quality and density of the beans.24

This has led to the general acceptance of fourcommercial varieties (Colombian Milds, OtherMilds, Natural Brazilian, Robusta) and nume-rous subvarieties. International markets use theICO indicator price based on weights for Co-lombian Milds 13%, Other Milds 27%, Brazi-lian 25%, Robustas 35%.25

The number of bean defects which could becaused by cultivation, harvest or first processingpractices, determines the green coffee grading[Boot 2002: 98-104]. Cup-value is based on fra-grance and aroma, acidity, flavour, body, after-taste and the adjustment to increase or decreasethe aggregated five factors. The specialty orgourmet coffee that appeared in the 1980s in theUnited States uses thresholds for the same cri-teria: coffee beans should not contain any flavourdefect, minimal visible defects and cup-value ofat least 80% of the maximum [ibid.: 22].

Health concerns of consumers have increasedthe scope of control in developed countries, in-cluding the detection of the coffee mould ochra-toxin A, pesticide residues and hydrocarboncontamination by jute bags. The rules, practices,conditions and terminology of the internationalcoffee trade have been standardized by the Eu-ropean Coffee Federation (ECF) and the Green

Coffee Association of New York (GCA). The60 kg bag is used as unit of measurement, US $price quotations by pound, minimal traded quan-tity of 1 container (18 MT), green bean equiva-lents, etc. Most of these rules were introducedand applied long before the establishment of theliberalization strategies of the 1980s.

The governance structure by a small numberof big corporations preceded these and was nota consequence of the liberalization strategies assuggested by J.M. Talbot [2004] and by B. Da-viron and S. Ponte [2005]. Governments andICO interventions could have affected the in-come distribution in the coffee chain, but hadnever controlled its (private) governance. Theasymmetric income distribution in the chain tobe discussed later in this section did also pre-cede liberalization. What did come with libera-lization was the second differentiation wavewith the proliferation of sustainable coffee cer-tifications and corresponding private coordina-tions [Muradian and Pelupessy 2005]. It shouldalso be remembered that only value generatorscould become lead firms in the chain.

PUBLIC

The public coordinations or interventions thatcome from outside the coffee chain are impor-tant. They can take place on all levels and mayhave considerable effects on both performanceand outcome of the markets. A first interventioncategory to consider is taxes and tariffs in

24. UNCTAD, The coffee guide 2007: “World coffeetrade,” p. 3.

25. UNCTAD, The coffee guide 2007: “World coffeetrade,” p. 12.

. . .204

Wim Pelupessy

consuming countries, which are often neglectedin policy discussions. This may be unjustified,given the extensive and variable appropriation ofrents from the coffee chain by the importingcountries’ governments. Coffee exporting coun-tries of the Association of Coffee ProducingGroup (former French and British colonies inAfrica and the Caribbean) and the AndeanCommunity exporting to the EU are favoured bybilateral elimination of import duties. Andeancountries have also privileged trade relationswith the US (except Bolivia and Venezuela). Forprocessed coffees, Brazil is charged by the EU7.5% for roasted and 9% for decaffeinated andinstant coffees. Others as Vietnam, Indonesia andIndia have to pay from 2.6 to 3.1%. Mexico en-joys special advantages from the US and Canadawithin the NAFTA framework.

VAT rates of green coffee in some Europeancountries are much higher than the (former) exporttax rates in most producing ones. Examples areDenmark (25%), Norway (23%), Poland (22%),Austria (20%), Italy (20%), Slovakia (19%), andSweden (25% for out of home consumption). As acomparison, export taxes of producing countrieswere never more than 5% of the consumers price.There may also be additional rates on roastedcoffee. Germany raises actually a VAT rate of 7%with an excise tax of 2.19 euro/kg on roastedcoffee and 4.70 euro/kg on extracts. Some greencoffee imports to the EU are still charged with thecommon external tariff of 4%. Exchange rate mar-kets and policies may have indirect influence onthe coffee sector, because international trade incoffee is usually paid in US $.

VAT and tax rates in general have no strongeffects on coffee consumption, given the lowprice elasticities. In exporting countries there

are tariff and non-tariff measures to protect thelocal coffee industry. Examples are Brazil andColombia with import duties on processed cof-fees and Vietnam with 20% on green and 50%on roasted coffee.26 This may hamper furthergrowth or the development of blends to raisequality in the domestic markets.

Collusion in international spot markets mayrepresent a second category of intervention bystates, producers associations and others. The twomost important interventions of the past in thissense were those of the International CoffeeAgreement (which included both consuming andproducing countries) and the Association ofCoffee Producing Countries. The first brokedown in 1989 because of the persistent tendencyof overproduction, the problems of new quotadefinitions and free rider behaviour of members.The second was unsuccessful to reach clearagreements and monitoring mechanisms. Collu-sion among a (limited) number of countries mightgive other opportunities to increase the income ofproducing countries through market outcome.

E. Giraud-Heraud and his colleagues [1995]have investigated (game-theoretically) what theconsequences are of product differentiation inhigh and low quality coffees for the possibilitiesof collusive behaviour in the world market. Itis assumed that countries collude by setting quan-tities (Cournot game). A first result was that highlevels of product differentiation induce high in-centives to cheat on the collusion scheme for all,regardless of whether they produce the high or

26. International Coffee Organization, 12-9-2005: “Obs-tacles to consumption;” and 25-9-2006: “Overview of thecoffee market.”

. . .205

The world behind the world coffee market

low quality coffee. A second result was thatlarger product differentiation between twogroups of colluding producing countries en-hances the sustainability of each cartel, whenboth can set their own retention share. In contrast,when all countries with different qualities col-lude in one cartel with a uniform retention share,higher product differentiation increases the in-centive for the high-quality (Arabica) producersto cheat while reducing this for the low quality(Robusta) producers. In this case it seems wise tolet the higher quality producers set the retentionshare. The implication for the market dynamicsis that the high quality producers (should) havemore power in collusion negotiations. Marketingand advertising practices increase the degree ofproduct differentiation of coffee, which thereforeshould be taken more into account in interna-tional coordination schemes.

The conclusions are that rent appropriation bypublic intervention is mainly taking place down-stream of the coffee chain by importing coun-tries. Final consumption is charged by a numberof taxes and fees. It also seems to be difficult forproducing countries to obtain higher prices fortheir coffee by stable collusion in the interna-tional spot market. Free riders behaviour hasplayed an important role in the internationalcoordination efforts in the past and will compli-cate collusion in the future, given the increasingdifferentiation of coffee.

Even large exporting countries as Braziland Colombia have only restricted marketpower to collude [Karp and Perloff 1993]. Itappeared that the steady-state export prices ofthese two big producing countries were bet-ween 6 and 13% above price taking levels ofthe market. Maybe the producing countries are

simply too many, their interests too diverse,while the entry of new producing countries isnot too difficult and even stimulated by out-ward oriented development strategies, WorldBank and other international financial institu-tions loans (Vietnam).

However, international coffee trade ismostly done by private companies and govern-ments can at best try to influence their beha-viour or the market outcome. In an era ofdecreasing state intervention, market analysisshould be more concentrated on these compa-nies instead of states or associations of states.There may be considerable market power in thehands of international trading companies, largeroasters and other buyers in consuming coun-tries, which apply intra-chain coordinations.

INCOME DISTRIBUTION

The interaction of lead firms behaviour withprivate and public coordination efforts fromwithin and outside the chain results in a certainincome distribution.

The retail prices paid by final consumers givethe total value created in the chain for roasted,ground and packed coffee in US $/pound. If wechoose Sweden, US, Italy and Spain as importantcoffee buyers since 2000 from respectively Co-lombia, Guatemala, Brazil and Vietnam, we cannotice that consumers’ prices are quite divers, butin all cases producing countries receive a minorpart of these, which was also the case in earliertimes.27 As a result of increasing export prices,exporters and cultivators participations have

27. As shown in J.M. Talbot [2004: 166-169] who hadworked on the period 1971-2000.

. . .206

Wim Pelupessy

increased in the 2000s. For incomes of bothcategories one may clearly distinguish the Ara-bica Milds participations, from the lower classi-fied Brazilian and Robusta. In the beginning,Brazil’s export and growers shares were higherthan of Vietnam, but later the differences disap-peared and the post-harvest shares reduced forthe two cases. In the two Arabica Milds cases ofColombia and Guatemala post-harvest partici-pations increased, as did the growers shares. Inabsolute amounts the Colombian and Guate-malan post-harvest incomes more than doubledfrom around 10 to 25 US $/pound. Vietnamwent from 4 to less than 2 US $/pound, whilefor Brazil the reduction was from 10 to7 US $/pound. Government support to growerscould have played a role in the latter two coun-tries. For all considered cases one could ques-tion our earlier assumption of fixed rates forpost-harvest activities.

Previous studies have estimated the value dis-tribution within the chain rather straightforwardly.It is interesting to note that the income distributionresults of these microstudies are more or less inline with the previous data. The growers participa-tion varied between 10-20% of the consumers’price for roasted and ground coffees as a worldaverage [Fitter and Kaplinsky 2001: 73], between7 and 9% for Robusta exported from Uganda andTanzania to Italy [Daviron and Ponte 2005:207-211] and about 14% for Côte d’Ivoire toFrance [Pelupessy 1999: 127]. For Arabica it wasalmost 15% when exported from Costa Rica toGermany [ibid.] and 4% from Tanzania to Italy.Export value participations fluctuated from 9%for Arabica sold from Tanzania to Italy to 39% forthe world average. All other export participationswere lying between these two percentages.

From the data and discussion in section 1 itis clear that the leading sectors could becomposed by roasting and retailing companies.Their value shares varied on average for the roas-ters between 29 and 72% of the consumers’ priceof the 1990s. The share could increase when weconsider coffee bars and other (out of home) cus-tomers whose participations are strongly increa-sing in the main consuming countries.

Preliminary estimates of gross profit ratesalong the chain before the crisis of the 1990s andthe analysis of section 1, indicate that roasting andretailing are the most profitable segments of res-pectively 33 and 30% of the total profits in theCosta Rica to Netherlands chain. They have thehighest entry barriers of high concentration ratios,capital intensity, sunk costs, advertising and pro-tection rates for processed coffees in importingcountries. In the 1990s smallholders could oftennot recover their costs, while downstream bigcompanies’ profits were still rising considerably.Big roasting and retailing companies control thefinal consumer markets and the supply of the mostvalue generating differentiated products andbrands. Coffee re-exports from consumers’ coun-tries and brand alliances with coffee equipmentmanufacturers have enhanced this control.28

The stagnation in average consumption in deve-loped countries could partly be attributed to“made to measure” coffee equipment.29

28. E.g. Illy-expresso machines in Italy; DE-PhilipsSenseo coffeemakers in the Netherlands.

29. Centre for the Promotion of Imports from Develo-ping Countries, “Coffee, a survey of the Netherlands andother major markets in the European Union.” Report, DenHaag, 1997, p. 52.

. . .207

The world behind the world coffee market

Concluding remarks

The sustainability of the linkages betweencoffee growing in developing countries andfinal consumption in developed ones has beencontinuously under pressure. Market demandof the main consuming countries will not in-crease substantially and mismatches with chan-ging consumer preferences may causeadditional problems. In these countries only thenumber of coffee drinkers could be increased,while some opportunities are available in theemerging markets of Eastern Europe and theThird World. Easy access to coffee cultivationand trade has contributed to structural globaloversupply. Coffee chain governance and coor-dinations gave no durable solutions and couldnot prevent asymmetric income distributions.

Consumer markets are highly concentrated,but price competition seems to be moderate.Non-price competition is much more importantbecause consumers do prefer quality, differen-tiated and new products. Traditionally this hadbeen achieved with blending practices of mul-tinational roasters, which today cover 90% ofthe market. Quality differentiation was based onsensorial attributes in the first place, but sincethe 1980s extrinsic or non-sensorial ones as sus-tainability were added. Both kinds of differen-tiations are mainly realized by downstreamadding value activities near the final consumer,which increases the market power of big roas-ting and retailing companies. Big supermarketchains compete with the roasters, but the lattercould maintain their position by the blends andcorresponding brands. This strengthens their po-sition in the international markets, despite theapparent bilateral oligopoly relationship of

producers’ and consumers’ countries. Interna-tional trading houses also became competitors,but have lost ground to roasters and retailers.

In this way a few big (multinational) coffeecompanies had become the governance forcein most of the global coffee chains and executetheir control by operating in imperfect marketsand coordinating production and trade condi-tions. International price volatility is not onlycaused by spot market imbalances, but also bythe behaviour of speculators and hedgers in thefutures market. Concentration is also present inthe post-harvest stages in producing countries.However, the oligopsonistic structures are inthis case unstable in the long run because ofpotential entry of newcomers, often subsidia-ries of foreign companies with finance. This isquite different from the more stable oligopo-listic structures in consuming countries.

At the supply end of the chain are numeroussmallholders with little or no market power anddifficult access to inputs, with low supply elastici-ties as a consequence. In a period of increasingexport prices as in the 2000s, depending on thecoffee variety, growers receive between 10 and25% of the consumer price in the chain. Whenprices fall, the share of cultivators will be muchlower. In spite of liberalization, growers suffer theconsequences of fluctuating world market pricesand the asymmetries in power and income distri-bution of the local buyers’ markets.

Interventions are widely present in thecoffee chain. Lead companies in the first placeand other chain actors, apply coordinations toget reliable transactions. From outside thechain there are appropriations of high rents byimporting country governments and entrybarriers to protect their coffee companies.

. . .208

Wim Pelupessy

Attempts to increase export prices durably bycollusions of exporting and importing coun-tries or of producer countries only did not suc-ceed. Because of the increasing productdifferentiation there is a tendency to cheatingand overproduction, while even the marketpower of big coffee countries is very limited.

Rent seeking behaviour of lead firms innon-competitive markets has structured thecoffee chain market sequence. Principally bigmultinational roasters control the transmissionof material, value and information streamsalong the chain at strategic parts. This controlis obtained by market and extra-market coor-dination tools. High and very divergent finalconsumers’ prices are one of the outcomes ofthe dynamics of this structure. Growers’ in-comes at the other extreme of the chains aremuch lower and less differentiated when oneconsiders the geographical and socioeconomicdistances of producing countries. This asym-metry makes the cultivators segment the wea-kest link in the chain. If this segment collapses,the supply of the required category of coffee

could be put at risk, which may affect the sus-tainability of the whole chain.

A sustainable solution for the imbalances bet-ween raw material exploitation and finalconsumption could not be reached by interven-tions in individual segments of the coffee chain.Policies should aim at the enhancement of sys-temic efficiency of the chain, instead of pointefficiency in international markets. A verticallycoordinated effort to share long-term informa-tion should be made with the direct involvementof leading coffee companies, other chain actorsand stakeholders. In these efforts coffee institu-tions in producing countries should have an im-portant new role in providing the public goodsneeded to support growers and to adjust and dif-ferentiate their products in accordance with thetrends of distant consumers preferences.

But first and foremost, an earmarked inter-national taxation should be applied to the highsurpluses gained by downstream coffee compa-nies and retailers to support the necessary ad-justments to cultivate and market the coffeewanted by final consumers in each chain.

Bibliography

Akiyama, T. — 2001, “Coffee market liberalization since1990. Commodity market reforms”, World Bank.Akiyama, T. and R.C. Duncan — 1982, “Analysis ofthe world coffee market”, World Bank Staff Commo-dity Working Paper 7.Bettendorf, L. and F. Verboven — 2000, “Incompletetransmission of coffee bean prices: evidence from The

Netherlands”, European Review of Agricultural Eco-nomics 27 (1): 1-16.Boot, W. — 2002, “National policies to manage qua-lity and quantity of coffee in Central America”. Wa-shington, Inter-American Development Bank.Branchi, M., A. Gabriele and V. Spiezia — “Tradi-tional agricultural exports. External dependency and

. . .209

The world behind the world coffee market

domestic price policies”. UNCTAD/OSG/DP/140, Fe-bruary 1999.Cárdenas, M. — 1994, “Stabilisation and redistribu-tion of coffee revenues: a political economy model ofcommodity marketing boards”, Journal of Develop-ment Economics 44: 351-380.Daviron, B. and S. Ponte — 2005, The coffee paradox.Global markets, commodity trade and the elusive pro-miseofdevelopment. LondonandNewYork,ZedBooks.Feuerstein, S. — 2001, “Do coffee roasters benefitfrom high prices of green coffee?”, InternationalJournal of Industrial Organization 20: 89-118.Fitter, R. and R. Kaplinsky — 2001, “Who gains fromproduct rents as the coffee market becomes more dif-ferentiated?”, IDS Bulletin 32 (3): 69-82.Gereffi, G., M. Korzeniewicz and R.P. Korzeniewicz— 1994, “Introduction: global commodity chains”, inG. Gereffi and M. Korzeniewicz eds., Commoditychains and global capitalism. London, Praeger: 1-14.Gilbert, C.L. and C. Brunetti — 1997, “Speculation,hedging and volatility in the coffee market,1993-1996”. Queen Mary and Westfield College,University of London.Gilbert, G.L. and W. Zant — 2001, “Restoring ba-lance by diversion in the world coffee market”. Re-port. Amsterdam.Giraud-Heraud, E., C. le Mouël and V. Requillart —1995, “Competition and collusion on the world coffeemarket”, European Review of Agricultural Economics22: 336-353.Goddard, E.W. and T. Akiyama — 1989, “UnitedStates demand of coffee imports”, Agricultural Eco-nomics 3: 147-159.Hobbes, J.E. and L.M. Young — 2001, “Vertical lin-kages in agrifood supply chains in Canada and theUnited States”. Canada, Strategic Policy Branch forAgriculture and Agrifood.Hooghiemstra, E. — 1991, “Market research for the Fede-ration of Coffee Corporations on gourmet coffees and stan-dard coffees in the Netherlands”. MA thesis, Deventer.Jaramillo, F. and T. Akiyama — 1990, “Policy res-ponses to the collapse of world coffee prices. Thecases of Costa Rica, Mexico and El Salvador.” Confi-dential Report. World Bank, Washington DC.

Karp, L.S. and J.M. Perloff — 1993, “A dynamicmodel of oligopoly in the coffee export market”, Ame-rican Journal of Agricultural Economics 75: 448-457.Morisset, J. — 1998, “Unfair trade? The increasinggap between world and domestic prices in commo-dity markets during the past twenty-five years”, WorldBank Economic Review 12 (31): 503-526.Muradian, R. and W. Pelupessy — 2005, “Governingthe coffee chain: the role of voluntary regulatory sys-tems”, World Development 33 (12): 2029-2044.Nelson, P., J. Siegfried and J. Howell — 1992, “A si-multaneous equations model of coffee brand pricingand advertising”, The Review of Economics and Sta-tistics 74 (1): 54-63.Olekalns, N. and P. Bardsley — 1996, “Rational addic-tion to caffeine: an analysis of coffee consumption”,Journal of Political Economy 104 (5): 1100-1104.Pelupessy, W. — 1993, El mercado mundial del café.San José, DEI. — 1999, “Coffee in Côte d’Ivoire andCosta Rica: national and global aspects of competiti-veness”, in H. Van der Laan, T. Dijkstra and A. VanTilburg eds., Agricultural marketing in tropical Africa.Ashgate, Aldershot: 109-130.Pelupessy, W. and E. Van Tilburg — 1994, “El mercadosolidario del café y el pequeno productor en Centroa-mérica”, in M. Samper ed., Crisis y perspectivas del cafélatinoamericano. San José, ICAFE-UNA: 239-259.Pollard, S.K. and D.H. Graham — 1992, “Rationalexpectations and output supply evidence from thesugar cane and coffee industries in Jamaica”, Socialand Economic Studies 41(1): 89-101.Sellen, D. and E. Goddard — 1997, “Weak separa-bility in coffee demand systems”, European Reviewof Agricultural Economics: 133-144.Singh, S., J. de Vries, J.C.L Hulley and P. Yeung — 1977,“Coffee, tea, and cocoa: market prospects and develop-ment lending”. World Bank Staff Occasional Paper 22.Talbot, J.M. — 2004, Grounds for agreement. Thepolitical economy of the coffee commodity chain.Oxford, Rowman and Littlefield Publishers Inc.Temu, A.A., A. Winter-Nelson and P. García — 2001,“Market liberalization, vertical integration and pricebehaviour in Tanzania’s coffee auction”, Develop-ment Policy Review 19 (2): 205-222.

. . .210

Wim Pelupessy

Vogelvang, B. — 1988, A quarterly econometricmodel of the world coffee economy. Amsterdam, FreeUniversity Press.

Wheeler, M. — 1995, “Coffee to 2000: a market un-tamed”. The Economist Intelligence Unit ResearchReport M 223.

RésuméWim Pelupessy, Le monde derrière le marché mondialdu caféLes liens entre la production du café dans les pays envoie de développement et sa consommation dans les paysdéveloppés ont été l’objet de tensions fréquentes. Nousrecourons ici à une approche qui rend compte, de manièreglobale, de la chaîne production-consommation pour ana-lyser ces problèmes et proposer des solutions. Les désé-quilibres sur les marchés internationaux de crédits spotrésultent de l’accès facile à la production et au commercedu café, des inconvénients climatiques dont souffrent lesgrands pays producteurs, ainsi que des décalages quiexistent entre l’offre et le goût des consommateurs. Lerecours aux fonds spéculatifs et aux contrats à terme surdevises accroît plus encore la volatilité des prix. Lesgrands torréfacteurs et détaillants internationaux contrô-lent le transfert des denrées, les flux de valeur et d’in-formation dans les secteurs stratégiques de la chaînegrâce au pouvoir qu’ils exercent sur le marché et auxoutils de coordination dont ils disposent. Il en résulte desprix élevés pour les consommateurs et très bas pour lesproducteurs. La structure de gouvernance qui régit ac-tuellement le système caféier n’offre ainsi aucune solu-tion durable aux déséquilibres du marché et des revenus.Il faut donc donner plus de poids à la coordination ver-ticale en partageant l’information, en réformant les ins-titutions qui régissent l’industrie dans les paysproducteurs et en introduisant des mécanismes de redis-tribution des surplus, susceptibles de renforcer les mail-lons les plus faibles.

Mots clésmarché mondial du café, chaîne production-consomma-tion, gouvernance

AbstractWim Pelupessy, The world behind the world coffeemarketLinkages between coffee growing in developing coun-tries and final consumption in developed ones have beenfrequently under pressure. We apply a global commoditychain framework to understand these problems and topropose possible solutions. Imbalances on internationalspot markets are related to the easy access to coffee pro-duction and trade, climatic drawbacks in large producingcountries and supply mismatches with changing consu-mers’ preferences. Hedgers and speculators in futuresmarkets cause additional price volatilities. Big multina-tional roasters and retailers control the transmission ofmaterial, value and information streams in the strategicparts of the chain by their market power and extra-marketcoordination tools. The outcomes are high consumers’prices and very low growers’ prices. A sustainable solu-tion for both the market imbalances and the asymmetricincome distribution could not be reached by the presentgovernance structure of the coffee chains. There is a needfor the strengthening of vertical coordination by sharinginformation, reorientation of coffee institutions in produ-cing countries and introduction of intra-chain surplus re-distribution to support the weakest links.

Keywordscoffee, global commodity chain, chain governance

. . .211

The world behind the world coffee market