The International Coffee Market

International Monetary Fund. Research Dept.
Published Date:
January 1964
  • ShareShare
Show Summary Details

STARTING IN THE LATE 1950’s, the international coffee market was characterized by large surpluses; exportable production exceeded exports year after year and unsold stocks continued to grow. As indicated in Table 1, by far the greater part of these stocks was accumulated in Brazil, where the supply of exportable coffee over the six crop years 1957/58–1962/63 appears to have exceeded exports by more than 50 million bags.

Table 1.Coffee: Exportable Production1and Exports, 1958–64(In millions of bags)
Total WorldBrazilAll Other
Calendar YearProductionExportsSurplusProduction2ExportsSurplusProduction2ExportsSurplus
Sources: U.S. Department of Agriculture, Foreign Agricultural Circular, various issues; Pan-American Coffee Bureau, Annual Coffee Statistics (New York), various issues.

Defined as total production minus estimated domestic consumption.

Production data are for crop years ended in the calendar year shown. Marketing begins during the second half of the calendar year; in some countries, like Brazil, it starts as early as July 1, and in other countries about October 1.

Sources: U.S. Department of Agriculture, Foreign Agricultural Circular, various issues; Pan-American Coffee Bureau, Annual Coffee Statistics (New York), various issues.

Defined as total production minus estimated domestic consumption.

Production data are for crop years ended in the calendar year shown. Marketing begins during the second half of the calendar year; in some countries, like Brazil, it starts as early as July 1, and in other countries about October 1.

The severe decline in prices which started early in 1957 and gained momentum in the following two years abated thereafter. From the end of 1959 to September 1963, the price of Brazilian coffee (Santos 4) declined at an annual average rate of about 2 per cent, compared with an average of 15 per cent over the preceding three years. Prices of mild varieties fell similarly, but those of African robustas showed a somewhat different pattern. Their décline came later and was more severe; recovery, however, started toward the end of 1961 while arabica prices were still weakening1 (Chart 1).

Chart 1.Prices of Coffee, New York, Quarterly, 1957–63, and Monthly, January 1963–June 1964

(In cents per pound; semilogarithmic scale)

The tapering off of the sharp decline in prices in spite of continuing excess production reflects, in part, restriction of supply under the terms of a series of agreements among producing countries, which gradually became more comprehensive and effective. The first one, known as the Mexico Agreement, dates back to October 1957. Set up by Brazil and Mexico and supported by 5 other Latin American countries, it was based on the retention of certain agreed quantities. It was followed one year later by a similar, though more comprehensive, agreement covering 15 Latin American countries. Later agreements were based on export quotas; the first one, which came into force in October 1959, included, in addition to the same group of Latin American countries, several African producers.

The initial failure of that agreement to halt the sharp decline of robusta prices is explained by the considerable expansion of output in 1959/60 and in the following season, and by the large proportion of production remaining outside the agreement. The subsequent adherence of the British territories to the agreement added strength to it, and late in 1960 robusta prices ceased to decline. By that time the margin between prices of robustas and Brazilian coffee (Santos 4) had increased to nearly 50 per cent of the latter, larger than at any time previously. The sharp upturn in robusta prices late in 1962 was mainly attributable to the increase in demand for this type of coffee in response to this widening margin and to the reduced supply as a result of smaller crops in 1961/62. To a certain extent, however, the movement also reflects the increasingly effective control of export supply by the Inter-African Coffee Organization (IACO), which was established late in 1960.

The present International Agreement, which became operative on October 1, 1963, is also based on export quotas, but it covers both importing and exporting countries. Though a reasonable balance between demand and supply at “equitable” prices is the first of its stated objectives, the Agreement has no provisions for specific price limits in either direction. Except for the very important inclusion of importing countries, the main features of the Agreement are not essentially different from those of its predecessors.

Recent Changes in the Market Situation

News of heavy frost in Paraná and drought in other coffee-producing regions of Brazil, which foreshadowed a considerable reduction in the 1964/65 crop, caused a drastic change in the situation. Demand from consuming countries responded sharply and led, in turn, to substantial advances in prices. Prices of “Brazils,” which had been weakening for a number of years, started to rise in late 1963; by March 1964, they were some 50 per cent above their September level. Prices of robustas showed a similar movement, but those of Colombian “milds” rose much less (Chart 1 and Table 2). Expectation of reduced supplies was also reflected in the behavior of the futures markets; trading in futures revived, and discounts on forward quotations turned into sizable premia (Table 3).

Table 2.Prices of Coffee, New York, Spot, Monthly Averages, September 1963–July 1964(In cents per pound)
Brazilian, Santos 433.037.544.946.349.848.647.346.946.6
Colombian milds (MAMS)39.439.345.045.749.948.549.348.649.3
Uganda, washed27.134.540.239.940.839.538.136.434.6
Source: Pan-American Coffee Bureau, Carta Semanal Mercado del Cafe (New York), various issues.
Source: Pan-American Coffee Bureau, Carta Semanal Mercado del Cafe (New York), various issues.
Table 3.Prices of Coffee, Futures, New York (Brazils) and London (Robusta), March 1963–July 19641
New York market (cents ver lb.)
Nearest month33.332.734.032.636.
10 months’ forward32.733.133.833.638.946.154.652.748.3
10 months’ forward as per cent of nearest month98.2101.299.4103.1108.1107.2104.9108.1104.5
London market (shillings per cwt.)
Nearest month210214191211255299329309273
10 months’ forward203212196209250299350325279
10 months’ forward as per cent of nearest month96.799.1102.999.198.0100.0106.4105.2102.4
Sources: George Gordon Paton & Company, Coffee Intelligence (New York), various issues, and The Financial Times (London), various issues.

Quotations at beginning of months.

Sources: George Gordon Paton & Company, Coffee Intelligence (New York), various issues, and The Financial Times (London), various issues.

Quotations at beginning of months.

As a result of the acceleration in purchases, world imports in the first four months of 1964 exceeded those in the same months of 1963 by 14 per cent, compared with increases of 2 per cent and 6 per cent in the corresponding periods of 1963 and 1962, respectively. Stocks in the United States by the end of May 1964 had increased to some 4.8 million bags, against 3.6 million a year earlier. Scattered data on stocks in some European countries indicate similar movements.

Quota restrictions do not seem to have hampered the flow of trade, though their mere existence and the delay in raising quotas in response to greater demand have at least temporarily added to the upward pressure on prices. Exports from both Brazil and Colombia in the first six months of the season actually exceeded even the increased quotas of these countries.

Although current production in 1964/65 was expected to fall far short of demand, the statistical position—the existence of large stocks, more than ample to supplement a series of short crops—did not suggest an impending shortage such as would explain the sharp reaction of demand and prices. However, there were serious doubts as to the condition and quality of these stocks, the largest part of which is held by the Brazilian Coffee Institute. A considerable proportion dates back to the years 1958 and 1959, mostly originating from Paraná; it has been reported that poor harvesting and processing of the respective bumper crops in that area greatly impaired the quality.2 Estimates varied as to the exportable quantity which might be obtained from the stocks held by the Institute; the Institute itself suggested a figure of some 22 million bags, less than half of the total held, and other estimates were even lower. Moreover, a strong preference, on various grounds, for current crop coffee prompted efforts to cover prospective requirements while new (1963/64) crop coffee was still available.

The buying wave subsided in the second quarter of 1964; prices weakened and exports showed a sharp decline. The relaxation of demand might be interpreted as resistance to the price advance, since prospects for the 1964/65 crop had not improved. Inventories in importing countries had been built up and permitted importers to adopt a “wait and see” attitude, at least for some months. The sharp decline of exports may have reflected, to a certain extent, the desire to bring exports into line with quotas under the Agreement, but efforts by exporting countries to maintain prices as demand abated seem to have played a major role. Brazil’s exports fell far short of those in any quarter for many years, and exports for the first nine months of the season were more than one half million bags below the quota for this period; in conjunction with the raising of the minimum registration basis to 43 cents a pound and increasing the loanable value of coffee, this strongly suggests that coffee was withheld in order to bolster the price. Measures to that effect were also taken by major exporting countries in Africa.

The first, and still very tentative, crop forecast by the U.S. Department of Agriculture (USDA), released late in June 1964, estimated total production in Brazil at 11 million bags, the lowest in several decades. Domestic consumption was estimated at 7 million bags, so that exportable production, defined by the Foreign Agricultural Service of the USDA as “total production minus estimated domestic consumption,” should amount to 4 million bags. In amplification of the very low figure thus derived, which has been used in the Department’s widely publicized forecast of exportable production, USDA has stressed that those who have evaluated the crop expect some 80 per cent (9 million bags) to be of “export quality”;3 this figure has been used in Table 4 for the purpose of Computing the shortfall to be expected.

Table 4.Coffee: Estimates of Total and Exportable Production Compared with Anticipated Demand, 1963/64 and 1964/65(In millions of bags)

Retained for


Exportable Production
Producing countries
Other Western Hemisphere20.84.616.216.2
Anticipated demand
For domestic use13.9
For export38.012.050.0
Producing countries
Other Western Hemisphere20.84.616.116.1
Anticipated demand
For domestic use13.9
For export38.012.050.0
Sources: U.S. Department of Agriculture (USDA), World Agricultural Production and Trade, June 1964, for revised data for 1963/64 and first estimate of 1964/65 crop. Data on production of arabica and robusta outside the Western Hemisphere and on demand for coffee are staff estimates supplemented by data supplied by Research Department, Pan-American Coffee Bureau.

Domestic consumption estimate of USDA, except for Brazil in 1964/65, for which the figure of 2 million bags is the residual after deducting estimated volume of “exportable quality” from total production. It is assumed that domestic demand would be largely met by drawing on stocks.

Estimated “export quality.”

Sources: U.S. Department of Agriculture (USDA), World Agricultural Production and Trade, June 1964, for revised data for 1963/64 and first estimate of 1964/65 crop. Data on production of arabica and robusta outside the Western Hemisphere and on demand for coffee are staff estimates supplemented by data supplied by Research Department, Pan-American Coffee Bureau.

Domestic consumption estimate of USDA, except for Brazil in 1964/65, for which the figure of 2 million bags is the residual after deducting estimated volume of “exportable quality” from total production. It is assumed that domestic demand would be largely met by drawing on stocks.

Estimated “export quality.”

The data in Table 4 (last column) indicate that more than half of the 1964/65 deficit between total exportable supply and estimated demand for export might be met by using surpluses from the 1963/64 crop; however, this surplus consists largely of robusta, and substitution of the latter for arabica on such a large scale would be resisted by importers.4 A deficit in the current supply of arabica would have to be met by drawing on stocks of arabica, the largest part of which is held in Brazil. As pointed out earlier, only a fraction of these stocks—some 22 million bags according to the Brazilian Coffee Institute—is believed to be of exportable quality. After meeting the export deficit of nearly 10 million bags, some 12 million bags would be left. The estimate by the Institute does not indicate whether and how far stocks not of exportable quality could be used for domestic consumption. If part of the domestic deficit had also to be covered from exportable stocks, the carry-over would be even smaller. On the other hand, there are, according to various sources, some additional stocks in private hands which may supplement those held by the Institute. Since all the estimates used are highly conjectural, the residual figure of some 12 million bags can be considered as no more than a rough indication of the magnitude involved. In any event, there appears to be no problem in meeting the current deficit, though a substantial part of the coffee available would not be of recent vintage. Because of deterioration over time, however, it seems not unlikely that at least part of the remaining carry-over will not be in satisfactory condition for use in future contingencies.

Prospects for the next few crops—barring additional weather hazards—will depend largely on the degree to which trees have been damaged. While a severe frost may affect crops for several consecutive seasons, a light frost is followed by fairly rapid recovery.5 Early reports seemed to indicate severe and possibly protracted damage; however, more recent surveys have indicated a good recovery, which suggests that the 1965/66 crop output might approach that of 1963/64 and thus corne close to covering demand, with no (or only minor) drawing on stocks. The most critical period for possible frost damage is the third quarter of the year; if September has passed without frost in the main producing areas of Paraná, prospects for the following crop greatly improve. Uncertainty in this respect makes it particularly difficult to gauge probable market developments for, say, the next crop year. If prospects improve, importing countries, relying in part on larger inventories, may maintain their current cautious buying policy, and exporters may not be able to sell to the limit of their quotas at or close to current prices. However, even a minor turn for the worse may engender a renewed buying wave and a corresponding rise in prices.

Recent developments foreshadow major changes in conditions and prospects on the international coffee market. As indicated in Chart 2, Brazilian crops in the last few years have been subject to wide fluctuations which, in view of the country’s large share in the total, have been reflected in world output. Production has been increasing steadily in other areas, the advance being most pronounced for robusta in African countries. These developments raise a number of questions: What are the prospects for Brazilian crops and what would be the implications of continued fluctuations? How far might shortfalls in Brazil be met from expanded output and/or surpluses elsewhere? What are the implications of changed conditions for the International Agreement? No final answer to these questions can be attempted here; the following brief remarks are intended merely to point to the main factors that will have to be considered in reappraising prospects and policies concerning the coffee market.

Chart 2.Coffee Production: World and Brazil, Crop Years 1951/52-1964/651

(In millions of bags; semilogarithmic scale)

1 Figures for 1964/65 are estimated.

Prospects for Increased Variations in Output

During the last decade, the location of coffee production in Brazil has undergone a marked shift toward the State of Paraná, which in recent years has produced roughly half of the country’s output and more than 20 per cent of world output. The change in the percentage distribution by States has been as follows:

ParanáSão PauloOthers
1951/52–1954/55 average204535
1960/61–1963/64 average493021

The shift has been largely responsible for the wide fluctuations in Brazilian production during the last 5 years. Because of its location, the State of Paraná is particularly susceptible to damaging frosts; there have been five or six of varying severity in the course of the last 12 years. As shown in Chart 2, the growth in Paraná’s production during the 1950’s was interrupted by serious setbacks in 1953/54, 1954/55, and 1956/57; after reaching a peak in 1959/60, output declined in four of the following five seasons, twice as the result of frosts. In the early 1950’s fluctuations in Paraná’s output had a comparatively minor effect on Brazil’s total production; subsequently, however, they have increasingly affected the total. In the last few years, variations in output in other States have been more severe than formerly; and the sharp decline in the country’s 1964/65 output reflects, in addition to the reduction in Paraná, heavy losses caused by drought in São Paulo. However, some of the decline in the latter State may possibly be accounted for by the eradication of trees under the scheme of the Brazilian Coffee Institute.

Even though the cumulation of adverse factors which have been responsible for the current severe shortfall may not be repeated, the threat of considerable variations in crop output caused by periodic frost damage in Paraná will continue. Thus, if Brazil should wish to fill a quota of some 19–20 million bags a year, adequate buffer stocks would be required to supplement current production in years of shortfall, which are almost certain to recur. The size and duration of recent declines indicate that such stocks would have to be large. A rough indication may be provided by comparing total Brazilian requirements over the next few years—some 26–27 million bags (19–20 million for export plus 6–7 million for domestic consumption)—with the forecast of 11 million bags available from the current crop. The requirements to be covered from stocks would be 15–16 million bags. If, in addition, allowance were made for a further (though perhaps less severe) shortfall before the harvesting of one or two crops large enough to permit the replenishment of stocks, it would seem that stocks of at least 20 million (or, to be on the safe side, 25 million) bags would be needed. This would cover one or two additional, more moderate, shortfalls. Since frost damage is often protracted, provision has to be made for at least two consecutive low crops. If, instead of taking the deficit expected for 1964/65 as a starting point, the greatest cumulative shortfall from average production over the last 10 years is used to estimate the reserve stock required, a figure of 23 million bags is obtained.6 These estimates are, admittedly, highly conjectural, but they may serve as a rough indication of the order of magnitude which would be involved. In view of the preference for coffee of recent vintage and of the deterioration of stocks over time, frequent rotation of the stocks would be necessary, to make sure that, in case of need, they would be adequate not only in quantity but also in quality. Inasmuch as such buffer stock operations would become a permanent feature, their cost—initial investment, replenishment, maintenance of adequate storage facilities, and periodic rotation—would have to be considered as an integral part of the cost of production, regardless of whether the stocks were held by producers, exporters, or the Government.

Paraná is at present considered to be the lowest cost coffee-producing area within Brazil; whether and how far this would change if the extra cost of stockholding were added depends mainly on what the amount of the extra cost would be. No attempt can be made here to present an estimate; the aim is merely to emphasize the need for appraising the situation in this light.

In the absence of adequate stocks, and prior to a possible relocation of production, which at any rate would take a number of years, variations in Brazil’s output are likely to cause sharp fluctuations in prices over the next few years.

Substitution Between Different Types of Coffee

Perhaps the most remarkable development in the production of, and trade in, coffee during the last decade has been the considerable increase in the share of robusta, mainly from African sources but to a lesser extent also from Asia. Production of, and trade in, robusta more than doubled between the first half of the 1950’s and the early 1960’s; the share of robusta in the total coffee trade rose from 17 per cent to more than 23 per cent (Table 5). The main factor responsible for this advance was the steep rise in U.S. demand for robusta; in recent years, U.S. imports of this type of coffee amounted to more than 4 million bags, compared with less than 1 million in the early 1950’s (Table 6). This development has been closely linked to the rapid increase in the manufacture of solubles, for which the robusta type has proved to have some technical advantage. The proportion of robustas contained in solubles has been rising, which accounts in part for the fact that imports of robustas have been growing at a faster rate than the manufacture of solubles. It, nevertheless, is recognized that a superior quality of soluble coffee is obtained by using only Brazils or milds, and if the prices of these types should be sufficiently low compared with the price of robusta, the share of the latter in the production of solubles might be considerably reduced. For regular coffee, which absorbs over 80 per cent of total green coffee roasted, the use of robusta is limited because of its distinct flavor. However, the very low price of robusta, both in absolute terms and in relation to arabica, has resulted in some substitution for the latter.

Table 5.Coffee: Exportable Production and Exports, by Main Areas






(million bags)increase(million bags)increase
Other Western Hemisphere5.99.0545.88.444
Asia and Oceania0.72.32200.61.5150
Western Hemisphere26.241.05725.931.923
Africa, Asia, and Oceania1.22.81331.12.5127
Africa, Asia, and Oceania5.213.31555.210.9110
Sources: Pan-American Coffee Bureau, Annual Coffee Statistics (New York), various issues, and data from Research Department, Pan-American Coffee Bureau.
Sources: Pan-American Coffee Bureau, Annual Coffee Statistics (New York), various issues, and data from Research Department, Pan-American Coffee Bureau.
Table 6.Coffee: U.S. Imports and Production of Solubles(In millions of bags)
Imports of Arabica













1951–53 average20.619.910.
1954–56 average19.317.
Sources: Pan-American Coffee Bureau, Annual Coffee Statistics (New York), various issues, and data from Research Department, Pan-American Coffee Bureau.

As indicated by the amount of green coffee roasted for this purpose.

Sources: Pan-American Coffee Bureau, Annual Coffee Statistics (New York), various issues, and data from Research Department, Pan-American Coffee Bureau.

As indicated by the amount of green coffee roasted for this purpose.

Correlation analysis, examining the changes in U.S. imports of robusta relative to imports of Brazils, in response to changes in the relative prices of these two types during the period 1953–63, shows a substitution elasticity of −0.98;7 however, the trend factor, reflecting largely the rise in the manufacture of solubles, appears to have been stronger than changes in relative prices in bringing about the growth of imports.8 The ratio of imports of robusta to those of Brazils increased from about 10 per cent in the first half of the 1950’s to over 40 per cent in the early 1960’s. However, there are reasons for doubting that the strong upward trend in U.S. imports of robusta will continue. Production of solubles, after nearly a decade of rapid growth, has shown signs of leveling off; moreover, the ratio of robusta imports to the production of solubles in the last few years has been so high that, even if the advance in the latter should be resumed, imports of robusta could not be expected to rise more than in proportion. Thus, to bring about a further shift in favor of robusta in the United States, reliance would have to be placed primarily on a lowering of prices of this type, relative to prices of other types. If the 1963 import ratio of 44 per cent is taken as a starting point, and if the substitution elasticity obtained, i.e., close to −1.0, is used, the price of robusta relative to Brazils would have to decline by some 30 per cent in order to induce the substitution of 1 million bags of robusta imports for the equivalent amount of Brazils. If no change in the current price of Brazils (Santos 4) in New York (about 45 cents a pound) is assumed, a reduction in the price of robusta (now about 37 cents a pound) to about 26 cents would be required. A similar computation indicates that reduction of robusta to 18 cents, the lowest price in recent years—again on the assumption that the price of Brazils is unchanged at 45 cents—would induce substitution of somewhat more than 2 million bags.

In an examination of the European market, where in recent years roughly half of the world total of robusta imports has been absorbed, it is convenient to consider France separately since, owing mainly to severe restrictions on imports, the share of robustas that is imported and consumed there is far above the world average and considerably higher than that in any other European country (Table 7). There is little room for further substitution; on the contrary, the recent freeing of coffee trade in France is likely to result in at least some shift in favor of arabica imports. The rest of Europe has increasingly turned to using a larger proportion of arabica (Table 7). This movement appears to reflect mainly the gradual reduction of restrictions and, in some instances, a lowering of import duties. Correlation analysis indicates that changes in relative prices have not played a significant role; nor has there been a discernible trend. In the United Kingdom, the shares both of solubles in consumption and of robusta in imports are large. With that exception, however, solubles constitute a relatively minor fraction of European coffee consumption. Any significant expansion in the consumption of solubles might raise the proportion of robusta imports. The common tariff for the six members of the European Common Market which became effective on June 1—free entry of coffee from affiliated African countries and a 9.6 per cent ad valorem duty for coffee from all other sources—might induce some shift in favor of robustas. Judging, however, from the lack of response to the sharp rise in the margin between robusta and arabica prices in 1959 and 1960, the effect of the preference, which will result in a comparatively minor increase of the price margin, should not be great.

Table 7.Coffee: Imports by Main Areas(In millions of bags)
TotalArabicaRobustaRobusta as Percentage of Total
United States19.723.618.819.80.93.8516
Other Europe7.315.35.512.21.83.02720
Rest of World2.
Source: Pan-American Coffee Bureau (New York), various publications.

Subdivision by types of coffee has been estimated on the basis of sources of imports.

Source: Pan-American Coffee Bureau (New York), various publications.

Subdivision by types of coffee has been estimated on the basis of sources of imports.

These considerations suggest that the possibility of substituting robusta for Brazilian coffee, as a device for offsetting temporary deficiencies, is limited. It would entail a sharp relative reduction in the price of robusta, which would certainly not be favored by the producing and exporting countries.

Mild coffees and Brazils are widely interchangeable; however, the use of a larger proportion of the former yields a superior blend. Thus, as one might expect, the elasticity of substitution is considerably higher between milds and Brazils than between robusta and Brazils. Correlations for the U.S. market, covering again the period 1953–63, gives a substitution elasticity of −1.6 for Colombia milds; the value for Central American coffees (Guatemala, El Salvador, and Mexico combined) relative to Brazils is still higher, −2.8. These figures indicate that the prices of milds, relative to those of Brazils, would have to decline much less than those of robusta in order to induce their substitution for Brazils.9 Correlations with respect to the European market proved inconclusive, but there has been a strong upward trend in imports of milds, both in absolute and in relative terms.

Although mild coffees are much closer substitutes for Brazils than are those of the robusta type, the scope for their substitution in the short run is limited by physical availability. Their production has risen much less than the production of robusta, and—because of the longer gestation period of arabica, compared with robusta, trees—expansion of the output of milds would take more time.

Implications for the International Coffee Agreement

The changed conditions on the coffee market suggest the need to examine certain aspects of the International Agreement and to consider whether and how far modifications might seem desirable. Three specific points come to mind: the provisions of the Agreement on prices, quotas, and stocks. Though all of them are interrelated in some way, the first two are so closely connected that they will be discussed together.

Prices and quotas

The Agreement contains no specific provisions for prices. Although “keeping prices at equitable levels” is one of the stated objectives, no actual range defining what would be considered as equitable is spelled out. At the time that the Agreement was negotiated (the third quarter of 1962), prevailing conditions suggested that crops far in excess of consumption, and further additions to already burdensome stocks, would continue to exert pressure on prices. It was thought that export restrictions under the Agreement would prevent a further sharp decline in prices, but it was not expected that either restrictions or market factors (or a combination of the two) would, during the lifetime of the Agreement, cause a substantial rise in prices. There seemed to be no need, therefore, to seek agreement on a specific price range and, in any event, there was little if any chance of reaching agreement at that time. Lacking specific price provisions, the Agreement also lacks the provision—which had been foreseen in other agreements of a similar nature—for automatic relaxations of restrictions on quotas if prices increase beyond a certain limit. Under the Agreement, increases in quotas are possible but not mandatory; according to Article 32, the Council may review the quota situation and may vary the percentages of basic export quotas, if market conditions so require. A proposal for a revision was made at a Council meeting in November 1963, after prices had started to advance, but the proposal was not carried. Early in 1964, when the advance had gained momentum, urgent requests by the United States, backed by other importing countries, resulted in an upward adjustment of quotas. The timing of price movements10 and the fact that exports (particularly from Brazil and Colombia) in the six-month period October 1963-March 1964 were in excess of (adjusted) quotas make it seem unlikely, however, that quota restrictions have been a significant factor in causing prices to rise, though the delay in their revision may temporarily have added to the upward pressure. However, when demand abated and prices showed signs of weakening, exporting countries (notably Brazil) adopted measures designed to maintain prices close to their recent peaks. The measures taken were mainly to withhold supplies from the market; as mentioned above (p. 372), Brazil’s exports in the April-June quarter of 1964 were exceptionally low. Similar measures had been adopted on earlier occasions, but, in the absence of quota restrictions, consumers were usually able to shift to other sources of supply. Quota restrictions greatly limit such opportunities and enhance the scope for individual exporting countries to implement independent price supports by cutting supplies to less than their quotas. If carried out with moderation this may not be objectionable per se, but it may arouse resistance in importing countries and reduce their willingness to participate in an Agreement which offers unrestrained latitude to exporters but affords no protection to consumers.

These reflections suggest that certain modifications in the Agreement may be worth considering. One way of assuring supplies to importers and markets to exporters would be through a mutual obligation that a certain quantity would be supplied at a price not higher than an agreed maximum and purchased at a price not lower than an agreed minimum. This, however, would necessitate agreement on both quantity and price limits, would complicate the administration of the Agreement, and would be likely to meet with considerable difficulties. A simpler way, requiring only minor modifications, would involve linking quotas with prices. As mentioned before (p. 383), provisions to that effect were contained in other agreements, calling for quota adjustments, up or down, when prices reached certain high or low levels (and failed to fall or rise, as the case may be, within a certain predetermined period, say two weeks). If a stipulated “maximum” price were reached or exceeded, quotas would be suspended altogether. Obviously, if the price increase reflected a genuine shortage, such suspension would not (and should not) per se bring the price down; it would only permit importing countries to shift freely between sources of supply, and thus would greatly reduce the temptation to drive prices higher still by withholding sales.


Perhaps most important in the light of developments since September 1963 are the provisions in the Agreement concerning the implementation of, and the addition to, existing provisions on stocks. Article 51 provides that “the Council shall take measures to ascertain world coffee stocks, pursuant to Systems which it shall establish, and taking into account the following points: quantity, countries of origin, location, quality, and condition.” Such information, valuable at any time, is of the utmost significance when the current supply has to be supplemented by drawing on stocks. The prospects for greater variability of crops, and for increasing reliance on stocks in the future, make it essential to have such information currently on hand.

Furthermore, it may well prove necessary to extend the present provisions in order to make it mandatory for countries with major fluctuations in output to maintain stocks of adequate size and quality, so that they can give reasonable assurances of being able to fill their quotas.


Elasticities of Substitution

Elasticities of substitution for robusta and Brazils, as well as for milds and Brazils, were computed on the basis of U.S. imports of coffee and relative prices on the New York market over the years 1953–63. Equations were estimated by using the least-squares method. A number of different alternatives were tried: the equations were computed in natural numbers, with and without trend, in first differences, in logarithms, and in deviations from trend. The price data used were lagged and unlagged. The three equations listed below are those which give the most satisfactory statistical explanation of the dependent variable. All three are in logarithmic form; hence, the price coefficients represent the elasticities. In equations 1 and 2, current prices are used; prices in equation 3 are lagged by one year.

Equation 1: Colombian coffee

Equation 2: Other Latin coffees

Equation 8: Robusta

List of variables

QB = U.S. coffee imports: from Brazil.

QC = U.S. coffee imports: from Colombia.

QOL = U.S. coffee imports: from El Salvador, Guatemala, and Mexico.

QR = U.S. coffee imports: robusta from all sources.

PB = N.Y. spot price, annual average, Brazil, Santos 4.

PC = N.Y. spot price, annual average, Colombian mild (MAMS).

POL = Average of N.Y. spot prices for coffees of El Salvador (central standard), Guatemala (prime washed), and Mexico (prime washed).

PR = Average of N.Y. spot prices for coffees of Ivory Coast (superior), Portuguese West Africa (Ambriz No. 2 AA), and Uganda (native standard), weighted by U.S. imports of the respective types for the years 1953–63.

t = Straight line trend.

The standard errors of the coefficients are given in parentheses beneath the respective figures. R¯2 is the coefficient of determination, adjusted for degrees of freedom. s¯ is the standard error of estimate (root-mean-square error), corrected for the degrees of freedom, as a percentage of the sample mean of the dependent variable, d represents the Durbin-Watson test statistic for serial correlation, the asterisk in the Colombian equation indicating that the test is inconclusive. r¯2, given for equation 3 only, is the partial coefficient of determination, corrected for degrees of freedom.

Le marché international du café


La perspective d’une récolte déficitaire de café au Brésil, conséquence de conditions climatiques défavorables, a amené un essor considérable de la demande provenant des pays consommateurs et un relèvement des prix vers la fin de 1963 et pendant les premiers mois de 1964. Les facteurs principaux incitant les importateurs—malgré les stocks très importants accumulés dans les pays producteurs, stocks qui dépassent largement le total des exportations normalement enregistrées au cours d’une année—à accroître leurs achats et à augmenter leurs inventaires sont examinés dans l’optique de révolution de ces dernières années. A part la perspective d’une récolte en 1964/65 insuffisante pour faire face à la demande prévue, le plus important de ces facteurs est l’incertitude qui règne quant à la qualité et même au volume des stocks actuels. En outre, les torréfacteurs montrent une préférence marquée pour le café de la récolte courante ou d’une récolte très récente, car l’emmagasinage prolongé semble causer une certaine détérioration, au moins légère. Finalement, les contingentements et l’adaptation tardive des contingents à l’augmentation de la demande ont accentué, au moins temporairement, la tendance à la hausse des prix.

Les auteurs font ressortir l’incidence d’expériences récentes sur l’évolution future. La concentration de la production brésilienne dans l’état de Paraná, où des gelées périodiques causent des variations considérables d’une récolte à l’autre, indique la nécessité de maintenir d’importants stocks régulateurs permettant de faire face à des insuffisances temporaires. Le coût du maintien et du renouvellement de ces stocks s’ajouterait au coût de la production dans l’état de Paraná, considéré actuellement comme une des régions où le coût de production est le plus faible. Comme la production de cafés africains du type robusta a beaucoup augmenté au cours des récentes années, la possibilité de remplacer le café brésilien par du robusta est examinée, et semble être assez limitée.

En conclusion, cette étude suggère certaines modifications possibles de l’Accord international du café, qui paraissent s’imposer à la lumière des événements récents.

El mercado internacional del café


La perspectiva de que se produjera una cosecha escasa de café en el Brasil como resultado de las vicisitudes del tiempo, provocó un brote subito considerable de la demanda de los países consumidores y un alza de los precios a fines de 1963 y primeros meses de 1964. En este artículo se analizan, contra el fondo de los acontecimientos de los años recientes, los factores principales que indujeron a los importadores a intensificar sus compras y aumentar sus existencias, no obstante las muy grandes cantidades de que disponían los países productores, las cuales excedían en mucho las exportaciones totales de un año. El motivo principal, aparte de que se vislumbrara que la cosecha de 1964–65 no alcanzaría a satisfacer la demanda que se esperaba, es la incertidumbre en cuanto a la calidad y aun en cuanto al volumen de las existencias actuales. Además, los torrefactores tienen decidida preferencia por el café de la cosecha corriente o de las muy recientes, debido a que parece ser inevitable que el almace-naje prolongado ocasione por lo menos cierto deterioro. Por último, las restricciones de las cuotas y la demora en ajustar éstas al aumento de la demanda han intensificado, siquiera temporalmente, la presión ascendente de los precios.

El artículo advierte la trascendencia que las experiencias actuales podrân tener en el mañana. La concentración de la produccion del Brasil en Paraná, donde las heladas periódicas causan notables variaciones en las cosechas, sugiere la necesidad de mantener grandes existencias reguladoras con que hacer frente a cualquier escasez temporal. El costo de la conservacion y rotacion de semejantes existencias aumentaria el costo de produccion en Paraná, region que actualmente se considera como una de las de más bajo costo. Dado que la producción de los cafés africanos del género robusta ha aumentado grandemente en años recientes, se analiza la posibilidad de sustituir mediante este tipo al café del Brasil, pero se encuentra que esa posibilidad es limitada.

El artículo concluye con sugerencias sobre posibles modificaciones del Convenio Internacional del Café, las cuales parecen ser necesarias en razón de los recientes sucesos.

Miss Lovasy, Advisor in the Research and Statistics Department, was formerly with the Economic and Financial Department of the League of Nations. She is the author of a memorandum on “International Cartels,” published by the United Nations, and of several articles in economic journals.

Miss Boissonneault, economist in the Special Studies Division, is a graduate of the George Washington University, Washington, D.C.

There are three main types of coffee: “milds” grown mainly in Central and South America other than Brazil and, in small quantities, in certain areas of East Africa; “Brazils” grown in Brazil; and “robustas,” which constitute some 80 per cent of African and Asian production. “Milds” and “Brazils” come from the same variety of tree, the “arabica”; the superior quality of the “milds” results from differences in climate (including altitude), cultivation, and preparation. The third type cornes from a different, sturdier tree, bearing a coarse fruit. On account of its particular flavor, which makes it less suitable for blending, robusta is considered inferior to arabica and sells at lower prices.

Govêrno do Estado do Paraná, Secretaria de Agricultura, O Paraná e a Economia Cafeeira, 1963. See also J. W. F. Rowe, The World’s Coffee (London, 1963).

U.S. Department of Agriculture, World Agricultural Production and Trade, Statistical Report, June 1964, pp. 14–15.

For a further discussion of the question of substitution, see pages 378–82, below.

The same applies to the effects of drought. The sharp cut in the 1964/65 crop was caused by both drought and frost: drought in São Paulo, which reduced production from 11 million to 2 million bags, and frost in Paraná where output, after a decline from 16.5 million to 10 million bags in the previous season, was lowered to 5.5 million bags. However, damage to trees in both States seems to have been less severe than was originally believed.

A group of Brazilian experts have suggested, on the basis of the large fluctuations experienced and expected, the need for reserve stocks equal to one year’s total requirements (exports and domestic consumption). See O Paraná e a Economia Cafeeira, 1968 (cited in footnote 2, p. 372, above), p. 1/21.

It would have been preferable to analyze substitution on each of the two markets—regular and soluble coffees—separately, but the data available do not permit this separation.

This is confirmed by the “partial coefficients of détermination,” which indicate that the trend factor would explain 73 per cent—or, alternatively, the price factor 45 per cent—of the variation not otherwise explained. (For equations, see Appendix.)

As indicated in Table 6, U.S. imports of arabica from sources other than countries in the Western Hemisphere—mainly from Ethiopia—have greatly increased in recent years. Computation of a substitution elasticity has not been attempted because recent changes in price differentials have reflected, to a large extent, improvements in the quality of Ethiopian coffee.

There was no acceleration of the increase in prices immediately following the November meeting when the proposal to raise quotas was defeated. Nor was there an immediate slackening of the advance when quotas were adjusted in a special session of the Council on February 10–12, 1964. Prices continued to rise through the first week of March, reached a plateau thereafter, and did not weaken before mid-April.

Other Resources Citing This Publication