Prepared by Smith Barney

Extremely low stocks of arabica coffees and expectations of smaller crops caused coffee prices to soar by over 200% during the first half of 1997. Despite the fact that coffee stocks have increased, causing prices to decline, they remain low relative to historical levels. The bullish price implications of the El Nino event, which is expected to reduce world coffee production in both the 1997-98 and 1998-99 seasons, and the likelihood that world coffee stocks will not increase appreciably in the 1997-98 crop year suggest limited downside price potential for the nearby futures from the $1.40 level. We favor buying the March futures at $1.30-1.35, risking 5¢.

For the 1997-98 (October- September) crop year, there may be little change in the overall world coffee balance. World production is expected to be between 100-103 million bags, with producing countries' domestic use about 25 million. Exports are likely to be around 75 million, for total use of 100 million bags. As a result ending stocks in 1997-98 are expected to be about the same as in 1996-97. This early in the season, there are still too many variables that will impact the final supply and usage numbers. In particular, the world coffee crop could end up less than 100 million bags if the weather from Mexico to Central America persists. A major variable is the 1997-98 Brazilian crop, which is estimated variously between 19 million and 28 million bags. In addition, Indonesia's crop is likely to decline by 1.0-2.0 million bags.

Coffee stocks increased since the lows of 1996. In part, the increase was due to the harvest of 1996-97 crop, and in part to the suddenly higher price and inverted market structure. Despite the fact coffee stocks have increased on an absolute basis, remain low relative to historical levels. Current deliver able supplies of CSCE (Coffee Sugar & Cocoa Exchange) coffee are under 400 contracts, while they have been over 4,000 contracts in the past. Thus, there is not a great deal of coffee around. One important question is whether stocks can build much more. Judging by coffee industry officials, there is very little coffee left to be exported from Mexico to Central America. Any increase in stocks at this point of the season will likely be marginal at best. Consider:

–the bullish implication that comes with the onset of El Nino, which is expected to reduce world coffee production in 1997-98, as well as in 1998-99; and

–a better chance that world stocks of coffee will not increase by any appreciable amount in 1997-98.

Trading Recommendation

Near term, we favor the long side of coffee. We would look to buy the March coffee futures between $1.30 and $1.35, risking 5¢. An initial objective would be $1.451.50, with the potential to move to $1.60-1.75. if the market does not begin to move higher soon, harvest activity in December and January could push prices back to the $1.20-1.30 area.

The July New York coffee futures contract bottomed on December 6, 1996, making a low at $1.00. The subsequent rally carried the market to $3.18 on May 29, 1997, a gain of 218% in 27 weeks–by any measure, a significant move higher. Since May, coffee prices have moved mostly lower. Before we address what will happen next, it is helpful to look at the causes of the bull market and the reasons for the subsequent decline, as these factors may provide clues as to where the market will head in the future.

At the outset, it should be noted that this bull market in coffee has been unique in the sense that the fundamental cause has been very tight stocks of arabica coffee in consuming nations. Historically, bull markets in coffee have been the direct result of weather developments in Brazil–specifically, cold weather and drought, which damage the crop that follows. The bull market of 1975 was caused by a Brazilian frost, the bull move of 1985 by a Brazilian drought, and the 1994 price explosion by a Brazilian frost in June, followed by another frost in July, which was followed by a drought–all of which did extensive damage to the subsequent 1995-96 crop.

While these types of fundamental events appeared to justify higher prices (at least until the full extent of the damage could be accurately assessed), the 1997 price rally was much different. It was spurred on by the fundamental fact that stocks of coffee (specifically, mild washed arabica coffee) were tight, especially when looked at in an historical context. This was the first significant bull market based predominantly on stocks considerations. What was interesting was that the shortage of coffee was confined to the high-quality mild washed arabicas, the basis of the CSCE coffee futures contract. This type of coffee is grown primarily in Colombia, Mexico and Central America, and was in short supply, as reflected by very high premiums for these coffees relative to Brazilian arabicas and African robustas. The price differential between these high-quality arabicas and the robustas (traded in London) became so large that the Association of Coffee Producing Countries (ACPC) decided in March to allow its members to increase their exports of arabicas to lower the arabica-robusta price differential.

Thus, we could argue that the rally was somewhat narrow in scope, given how much coffee was in short supply. One solution to the shortfall in these sought-after arabicas is to increase the blending with other, lower-quality arabicas and robustas. The problem here was that blending was already at levels that could not be changed very much, and there are limitations on changing blends. Therefore, the tightness in stocks was given a very bullish interpretation. In November 1996, the CSCE reported that certified warehouse stocks of coffee had dropped to less than 400 bags, a level that held for three weeks in a row. Roasters were aware more than anyone that stocks of these arabicas were tight, and they began to hedge their forward commitments by entering the long side of the market.

As recently as the previous March, stocks at the CSCE had been as high as 150,000 bags, and have been over 1.0 million bags. Looked at in those terms, 300+ bags of coffee, was not very much at all. The decline in stocks had been fairly steady though there was little reaction in terms of prices. That appeared to have been due to expectations that the next Brazilian crop would be larger, as would the Mexican, Central American and Colombian crops. There was a feeling that the 1996-97 supply of these coffees would increase, which would enable stocks to be, rebuilt.

E.D. & F. Man, a commercial firm, reported that its survey of coffee stocks in consuming countries (Europe, Japan and the U.S.) showed that in December these stocks had fallen to 7.6 million bags, a record low supply. In 1993, world consumer stocks of coffee had been more than 18 million bags. Once again, the focus was on the U.S. and stocks there. The Green Coffee Association (GCA) reported that monthly stocks of coffee (which includes the CSCE stocks) actually made a multi-year low in October 1996 at 1.26 million bags. At the end of July 1997, they had risen back to 2.34 million, but that was still well below their level of more than 10 million bags, as recently as 1993.

With mild washed arabica coffee stocks tight and at minimal working levels, much of the market focus was on future supplies. While there had been optimism in late 1996 about the expected size of the coming coffee crops, in early December there were comments from officials in producing countries that the crops would not be as large as expected. Roaster expectations about a larger 1996-97 arabica coffee supply began to decline. Initial reports coming out of Brazil indicated that the 1997-98 coffee crop would likely be about 24 million bags, with some estimates reaching as low as 20 million. One reason cited was the biannual nature of production, with large crops followed by small ones, as the trees “rested.”

There was still no consensus as to the size of the 1996-97 Brazilian crop, which had been projected as high as 26-27 million bags, a large crop. At the same time, there was an increasing number of comments from Colombia to the effect that the 1996-97 crop would be smaller than expected, due to heavy rains during flowering. There were similar ideas for the area from Mexico down through Central America, where hurricanes had done some damage.

It appears that these comments about the possibility of smaller crops, along with the already tight stocks, led roasters to begin extending their coverage on the idea that prices had limited downside potential and more (possibly much more) upside potential. The relatively new “just-in-time” buying and inventory strategies, now widely employed by roasters and other holders of stocks, likely contributed to the initial increase in prices, as roaster stocks were probably at minimal levels.

With this background, it is not surprising that prices began to move higher, and the buying was supported by intense speculative interest. The 1994 bull market was remembered by speculators who were interested in seeing a repeat performance. As prices moved higher, there was new fundamental support. Colombia, the major exporter of the mild washed arabicas, had labor difficulties at export facilities, as well as with truckers bringing the crop to market. These problems reinforced the idea that Colombia perhaps was not that reliable a supplier of the needed high-quality arabica coffee. In Brazil, there were labor problems at a major coffee export facility. This was disconcerting, as Brazil is the largest exporter of arabica coffee.

As a result of these developments, the price structure of the market changed. The nearby contract increased its premium to the back months (becoming more inverted, or backward)–an indication of supply tightness and a signal to exporters and producers to sell their crop. That they did, as exports of coffee from Mexico and Central America surged to capture the high prices. That intense export activity resulted in a rebuilding of coffee stocks in consuming nations.

One result of high prices and a steeply inverted price structure is to call forth more of the commodity to the market in a shorter period of time; the market did this successfully. E.D. & F. Man reported that its survey of stocks of coffee in consuming nations showed an increase from the multi- year low of 7.6 million bags in December to 10 million bags in April, an increase of more than 30%. Similar increases in warehouse stocks were seen at the CSCE and by the Green Coffee Association. For its part, Man noted that the increase in consumer stocks was not that visible to the market, as it occurred at European ports–the result of heavy arrivals of arabica coffee from Central America and Mexico.

However, it could be expected that a doubling in the price of any commodity would result in some supply effect. Since the total stocks of a given commodity can never be accounted for completely, there was probably more coffee around than was indicated in the published data. In the case of the steeply inverted coffee market, the effect was to signal the market to provide it with supplies, which did take place. While production of arabica coffee from Mexico to Colombia was lower overall–a strong reason in itself for higher prices–the crop was harvested and sold quickly, and the result was a quick buildup in coffee stocks in the hands of consumers.

In late May 1997, coffee staged a spectacular three-day rally of almost 58¢, a speculative buying climax. There were no notable bullish events during this time. The May 29 high in the expired July contract was $3.18, and the September coffee reached $2.77, a backwardation of 41¢ per pound.

Since the May 29 contract and multi-year highs, coffee prices have been on the decline. A number of reasons can be cited for this. The most obvious was the fact that coffee stocks were rebuilt as a result of the backwardation, overall high prices and the accelerated export schedules. At the same time, the 1997-98 harvest occurred in Brazil. While initial reports from Brazil pointed to a crop of around 20.0 to 24.0 million bags, there were independent analysts who felt that the crop could be larger.

As noted earlier, bull markets in coffee historically are not the result of tight stocks, but rather of cold-weather events (frost/freeze) and drought in Brazil. As the 1997-98 Brazilian crop harvest progressed (through the Southern Hemisphere winter), there was no sign of any cold weather. That seemed to spur speculative liquidation by weather bulls. In addition, the USDA surprised the market by estimating the 1997-98 Brazilian crop at 28 million bags, while the 1996-97 crop was raised to 27.5 million. All of this contributed to the decline. At this point, there are concerns about the size of the 1997-98 arabica crop, weather uncertainties due to El Nino, and the direction of stocks, all of which should stabilize prices.

F.O. Licht estimates world coffee production at 98.99 million bags in 1996-97, up from 87.88 million in 1995-96. The increase is due to the recovery of the Brazilian crop from the frost/drought damage of 1994, which affected the 1995-96 crop. If achieved, the 1996-97 world crop would be the largest since 1991-92, when world output was 102.6 million bags. Arabica coffee production was estimated at 65.5 million bags (two-thirds of world output), while robusta production was put at 33.5 million. In the record 1991-92 season, arabica production was 71% of the total, with robusta at 29%.

The overall increase in world coffee production in 1996-97 is deceiving. The Brazilian crop increased some 80%, while at the same time the production of mild arabica coffees in the important areas of Mexico, Central America and Colombia fell by 12%. It is this coffee that makes up most of the deliverable stocks at the CSCE, and a tight supply was the primary reason for the strong price, rally. With the 1996-97 season ending, the fundamental focus in terms of production will be on the 1997-98 crops. Some initial ideas on those crops are outlined below.

Brazil

Brazil is by far the world's largest producer of arabica coffee, and an important producer of robusta coffee in the state of Espirito Santo. Arabica is grown primarily in Parana, Sao Paulo and Minas Gerais. There are approximately 3.0-4.0 billion coffee trees in Brazil. The main harvest period is May-August, though early harvesting starts in April. Brazil has large production potential (the 1987 crop was more than 40 million bags), but also displays production variability due to the weather. Brazil is the one coffee crop that is most susceptible to frost/freeze damage, as happened in 1994. The May-August period is the coldest and driest time of the year, which favors the Brazilian method of harvesting. Brazilian coffee is not deliverable against the New York futures contract.

Because Brazil is such a large producer of arabica coffee relative to the market, any developments there become magnified and affect the entire market. Historically, bull markets in coffee are directly related to weather developments in Brazil during the late second quarter and first half of the third quarter. More than half the frosts in Brazil over the last 90 years or so have occurred after July 1, with the earliest on June 1. Smith Barney research indicates that of the 16 major frosts this century, four were in June, nine in July and three in August. This is what made the 1997 bull market unique: It was not the result of a frost in Brazil. The weather this year was good, with no threat of cold to the crop, though there was a great deal of rain early on, which could have affected the quality of the beans.

If there is a major production debate in the coffee market, it is over the size of the Brazilian coffee crop. With no central estimating agency like the USDA in Brazil, a variety of trade sources (including exporters) provide estimates of the crop. The range of estimates in a given year can be very wide. The 1995-96 crop, damaged by two consecutive frosts and a drought, has a range of estimates, with F.O. Licht at 14.8 million bags and the USDA at 16.8 million–a 14% difference.

Following that season, coffee production recovered. The higher prices in 1994 resulted in producers doing more maintenance. Trees were pruned, dead trees replaced and more trees planted. As a result, Brazilian production potential has increased. There is some biannual cyclicality in coffee production, in that large crops are followed by smaller crops as the trees rest, and then larger crops appear again. Of course, that cyclicality can be muted if there are more and more trees or if the trees become more productive.

The 1996-97 season reflected the impact of favorable weather and increased maintenance, as production is currently estimated at 28.5 million bags by Licht and 27.5 million by the USDA. Indeed, the initial estimates of the crop were lower and were gradually raised, based partly on Brazil's active export pace. Some of the latest increases in the estimates of the 1996-97 crop have occurred very recently and have contributed to the recent bearish tone. There have even been some comments that the 1996-97 crop was as large as 30 million bags.

Assuming the 1996-97 season output was 28 million bags, the next focus was the size of the 1997-98 crop. Once again, there is a wide range of crop estimates. Part of the current debate concerns the biannual cyclicality, where the large 1996-97 crop followed a small 1995-96 crop, which would imply a decline for the 1997-98 season. While this is not a bad assumption, improving productivity must be considered, i.e., better maintenance, spraying and fertilizer use, as well as more trees coming into a more productive part of the life cycle. All of these factors offset the biannual cyclicality. Add to this a favorable growing season, and the potential is there for another large crop.

The initial estimates of the 1997-98 crop, which came from various Brazilian sources, were clustered around 20-24 million bags, which would have to be considered positive for prices. There were even forecasts of a crop below 20 million bags. Yet there remained questions, given the fact that the weather during the development of the crop was reported to be very good. If productivity was increasing along with better weather and more trees, we could argue that the crop might, in fact, possibly increase and be larger than in 1996-97.

The USDA in mid-June issued an estimate of the crop of 28 million bags, marginally higher than the 1996-97 crop of 27.5 million. The USDA's former Brazilian crop analyst (now a private consultant) reportedly issued an estimate between 28.0 and 28.5 million bags. These higher-than- expected crop estimates contributed to the decline in prices but were questioned by Brazilian producers. Adding fuel to the fire of controversy, the USDA released the current Brazilian agricultural attache's May report on the crop, winch showed an estimate of 24 million bags.

However, it appears that the combination of excellent weather, increased maintenance and use of inputs, as well as new trees, will lead to improved production prospects. It should be noted that this will carry forward to future years, implying that Brazil is more likely to produce larger, not smaller, coffee crops in the coming seasons. No doubt the debate over the size of the 1997-98 Brazilian crop will continue, and there will be further estimates–probably until the start of the 1998-99 season in May of 1998.

Colombia

Having more market impact than the Brazilian crop this season were the prospects for the Colombian coffee crop. Like Brazil, there is no agreement or consensus on the size of the crop. Before the rally started last December, there had been talk in the trade that the 1996-97 crop would prove disappointing, the result of poor weather, the critical question was how disappointing. Colombian coffee is considered to be of the highest quality and is priced at a substantial premium to other arabica coffees. In part, the rally that began in December was due to rising expectations of a decline in Colombian coffee output. Added to that potentially very bullish development was the continued social unrest in the country.

At one point during the rally, there was a strike by port workers at Bueneventura, the major export facility for coffee, which delayed coffee destined for export. In addition, there was a nationwide truckers strike, important in that the truckers carry the coffee to the export facilities. The combination of these incidents, along with widespread political and social unrest, was unsettling to the trade, as it pointed out the vulnerability o Colombia as a reliable supplier of high-quality coffee.

The USDA estimated the 1996-97 Colombian crop at 10.3 million bags, a steep 20% decline from the 1995-96 crop. In recent years, Colombian coffee production has averaged about 13 million bags, though insect infestation from the Broca borer worm is a widening threat to the crop. F.O. Licht has estimated the 1996-97 Colombian crop at 9.6 million bags, which would represent 26% decline from its 1995-96 estimate. Based on data from Colombia, the National Coffee Fed estimates that the crop will end up being about 10.7 million bags.

With the smaller crop, the National Coffee Growers Federation was forced to use part of its stocks to help meet export obligations. Exports by Colombia were estimated at just over 11 million bags, up 3% from the previous year. For exports to increase with lower production means that stocks were depleted. Those stocks have been drawn down considerably. Thus, the prospects for the 1997-98 crop in Colombia will be an important fundamental consideration. The 1997-98 crop is currently estimated by the USDA at 11.3 million bags, up almost 10% from 1996-97. Officials from the federation have expressed the opinion that the new crop will be 12 million bags–though recent dry conditions are causing some concern–and that Colombia plans to export 10 million bags in 1997-98. The International Coffee Organization estimates the crop at 13 mullion.

Mexico

Mexico is another major producer/exporter of mild arabica coffee. Mexico is a major supplier of coffee to the U.S. and has been increasing its production. The crop, grown largely in the southernmost Chiapas, Vera Cruz and Puebla states, has shown some vulnerability to cold fronts moving down from the U.S. Still, it appears likely that production will trend higher in the future, as the Mexican Coffee Council has stated a goal to increase coffee output, and Mexico has the advantage of having the U.S. as its major market.

As with the other major producers, production estimates in Mexico have wide variability. The USDA estimates the 1996-97 crop at 5.6 million bags, up from 5.5 million in the prior season. In marked contrast, F.O. Licht estimates the 1996-97 Mexican crop at 5.15 million bags, which compares with 5.4 million in 1995-96. The decline was attributed to the biannual production cycle and a frost in December 1996.

For the 1997-98 season, the USDA estimates the Mexican coffee crop at 5.7 million bags, which is some 100,000 bags more than the estimate for 1996- 97. Mexican coffee producers have taken advantage of the high prices for coffee by harvesting and exporting the crop quickly, with 1996-97 exports estimated at 4.38 million bags. In the 1997-98 season, exports are projected to be 4.8 million. It is of some interest to note that domestic consumption in Mexico is declining. According to the USDA, rising retail prices, along with availability of other beverages, caused domestic use to decline by more than 8% in 1996-97, and that in turn freed up more coffee for the exporters.

Central America

The major Central American coffee producers are Costa Rica, El Salvador, Guatemala, Honduras and Nicaragua–all of which produce mild washed arabica coffees. Despite early ideas that overall coffee production would be lower, it appears that for the 1996-97 season, production was marginally above a year ago. According to F.0. Licht, these five producers produced about 11.9 million bags in the 1996-97 season, virtually from 1995-96. Production was higher in Costa Rica and Honduras, lower in Guatemala and Nicaragua, while El Salvador was about unchanged. Exports of coffee by these countries were marginally higher than the prior season.

Robusta Production

While arabica coffees are primarily grown in highland regions and are mild in taste, robustas are grown in lowlands and have a more pungent taste. The major producing areas are Asia and Africa, though Brazil also produces commercial amounts of robusta coffee. Among the major producers of robusta are Indonesia and Vietnam. Licht estimates that robusta production in 1996-97 was 33.47 million bags, up almost 13% from 1995- 96.

Indonesia is the largest robusta producer, and the third- largest coffee producer overall. Licht estimates production in 1996-97 (April-March) at 6.5 million bags compared to 6.3 million in 1995-96, while the USDA puts the 1996-97 crop at 7.6 million bags. The USDA estimates that 1997-98 Indonesian coffee production will be 6.8 million bags, a decline of 11%. There have been comments from industry officials in Indonesia that because of the very dry weather related to El Nino, production could conceivably fall by 30% in the 1997-98 season, and possibly by another 20% in 1998-99. That would imply a 1997-98 Indonesian crop of 5.3 million bags and a 1998-99 crop of 4.2 million. Given the fact that Indonesia is the third-largest coffee producer and the largest robusta producer, one conclusion is that El Nino will have a larger relative impact on the robusta market than the arabica.

Vietnam

Vietnam has been rapidly increasing robusta coffee production and is becoming a dominant player in the market. In the 1992-93 season, production was 2.25 million bags, according to the USDA, while in 1996-97 the crop was estimated at 5.0 million bags, and possibly higher. Industry sources in Vietnam put the current crop at 5.2 to 5.4 million bags. Recent comments from Vietnam estimate that the 1997-98 crop may even be more than 5.8 million bags, compared to 4.5 million in 1996-97–an increase of almost 30%. There is also a great deal of disagreement about the 1996-97 season, with F.O. Licht putting the crop at 5.15 million bags and the USDA at 4.22 million. It looks like Vietnam will be spared the effects of El Nino that Indonesia is now experiencing, and it is possible that Vietnam will export 5.0 million more bags of coffee than Indonesia.

World Coffee Production

World coffee production in 1996-97 was estimated by F.O. Licht at 99 million bags, up from 87.9 million the season before. The increase was due in large measure to the recovery of the Brazilian crop following the devastating frosts of 1994. Arabica production was estimated at 65.5 million bags compared to 58.1 million the year before, while robusta production was put at 33.5 million bags, up from 29.7 million the prior year. The USDA estimates the 1996-97 world coffee crop at 100.7 million bags, which compares with 89.2 million in 1995-96. Overall, there is a general agreement that world coffee production is about 100 million bags.

For the 1997-98 season, the USDA estimates world production at 103.7 million bags, some 3.0 million more than the current level. The USDA projects that there will be larger crops in Colombia and Brazil, Vietnam, India and Kenya, as well as Mexico and among the Central American producers. Production in 1997-98 will depend in part on weather. Already there has been dry weather in parts of Central America and Mexico, which could limit production. Colombia reports that conditions are drier than normal, and will have to be monitored.

El Nino effects will damage the crop in Indonesia. There are still questions about the size of the Brazilian crop, with estimates ranging between 19 and 28 million bags. Overall, because of the increasing possibility of poor weather, it looks like there is a better possibility that world coffee production in 1997-98 will be closer to the 1996-97 level of 100 million bags. If the Brazilian and Indonesian crops are smaller, which is a possibility, the USDA's estimate of world coffee production (almost 104 million bags) could be easily overstated by 4.0-6.0 million.

El Nino Considerations

The 1997-98 production season and the 1998-99 season will be different in that they will be affected by the El Nino weather event, which began around April 1997. While much has been written and said about this, the fact remains that it will adversely impact coffee production in both seasons. The key variables are the strength of the event and the duration. The current El Nino is very similar to the 1982-83 event, which was considered the strongest of the century. It is also likely to have impacts lasting from 12 to 18 months. The most salient effect will be to bring very dry weather to Southeast Asia–in particular Indonesia and its big robusta crop–which between this season and next may decline by 50%.

The second area that can be affected consists of Central America, Mexico, Brazil and Colombia, which make up the most important region of arabica coffee production. The effect of El Nino there is to bring dry weather, and reports so far indicate that June and July were much drier than usual, with rainfall running some 10% to 20% less than the normal level. The dry weather is likely to reduce the 1997-98 crops to some extent, though the decline is not expected to be very large. For example, there have been comments from Costa Rica indicating that early-maturing coffee is actually maturing much slower than usual, which is attributed to El Nino. Colombia reports similar problems with dry conditions, though it is too early to tell if the crop will be damaged. There have been reports that the dry weather has reduced quality, a most important consideration in Colombian coffee. That, in turn, could reduce exports. The mild winter in Brazil is due to El Nino. Brazil experienced dry weather in July and August, but it started to rain in September and October. While the trees flowered, there have been questions about whether the rains will continue. Brazilian weather developments could turn out to be the most important factor in the coming months.

To quantify El Nino's possible effects, keeping in mind that not all changes are a direct result of it, we can look at the event historically. Looking at the production data, the USDA reported that in 1981-82, world coffee production was just over 98 million bags. In the very strong El Nino season of 1982-83, world coffee production fell to 81.9 million bags, a decline of 16%. In 1987-88, world coffee production rose, the result of a large Brazilian crop. In 1991-92, global production was 103.7 million bags, while in the 1992-93 El Nino season, production fell to 92.9 million bags, a drop of 10%. From this pattern, We can draw the general conclusion that world coffee production in 1997-98 and 1998-99 will be adversely impacted by El Nino-related factors, primarily drought. If we use the USDA 1997-98 production estimate of 104 million bags, that number could easily be reduced below 100 million by the weather effect.

Consumption Of Coffee

Consumption of coffee is a function of the retail price of coffee, prices of substitute beverages such as tea, disposable income and consumer tastes, as well as social factors (i.e., tradition). Factors like climate can also play a part, as per- capita consumption is generally higher in the colder climates. In the past, coffee consumption was assumed to be higher in the colder part of the season, though this has been disputed. What may be more significant is the fact that coffee now comes in an increasing variety of forms, including the very popular iced coffee, which could mean that warmer summers could in fact lead to higher rates of coffee use instead of lower. Specialty flavored coffees, high- quality coffees and coffee houses are enjoying a wave of popularity, which is increasing coffee use to some extent and introducing the beverage to new customers.

However, it is difficult to determine exactly how much coffee is being used, because stocks of coffee are held at various parts of the coffee- processing chain and these stocks are largely invisible. As such, coffee consumption is often estimated by looking at domestic use in producing countries and adding their exports to that figure. One very noticeable trend is an increase in coffee consumption in the producing countries themselves.

One area that has been of interest is the consumption of coffee by different income groups. To some extent, coffee can be considered a semi-luxury good. Among lower-income groups in developing countries, increases in coffee prices are likely to have a proportionately greater impact on the average food budget. At the same time, this effect may be mitigated to the extent that coffee has a unique flavor, mildly addictive properties, and is often consumed in a social setting such as a coffee house. As such, price increases may not reduce consumption on a proportionate basis. Various studies of the coffee market have found that coffee has a price elasticity of .2-.3 in higher-income countries, meaning that an increase of 1% in price results in a decrease in use of just .2% to .3%. In lower-income countries, it can be assumed that the elasticity is higher, perhaps much higher.

It can be expected that over time, the elasticity will in fact increase as more substitutes enter the market. Consumers will have more beverages to choose from in the future, and if coffee prices move higher, even less coffee will be consumed. Of particular concern to the coffee industry is the reaction of consumers to rapidly and suddenly increasing coffee prices–the fact that coffee prices increase so much in so short a time period. The recent rally saw futures prices for coffee increase more than 200% between early December 1996 and late May 1997. While only a portion of this increase was passed on at the retail level, the increase was still sudden. The other related concern is that consumers who switch to other beverages in a time of higher coffee prices will not return to coffee.

Following the explosive 1994 rally, caused by a double frost in Brazil along with a drought, the International Coffee Organization reported that its members in 1993 had per-capita coffee consumption of 4.84 kilograms, which fell to 4.66 kilograms in 1994, down 4%, and then to 4.52 kilograms in 1995, another 3% decline. This is in line with other estimates that put the decline in coffee use following the 1994 rally at between 6% and 8%. The decline can be attributed to the rapid increase in retail coffee prices. An analysis by Wheat First Butcher Singer indicated that U.S. per-capita consumption of coffee in 1993 was 25.3 gallons, which then fell to 23.4 gallons in 1994 (down 8%), while falling another 8% in 1995 to 21.5 gallons. This decline appears due in part to higher prices and in part to the declining long-term trend in U.S. coffee consumption.

Licht data indicate that in the 1992-93 season, coffee producers' exports were 77.8 million bags of coffee. This fell to 73.9 million in the 1993-94 season, and then to 65.9 million in 1994-95, down 11%. That would largely represent the impact of higher prices on coffee use. The USDA world export numbers show that in 1993-94, exports fell 1%, while in 1994-95 they were 65.9 million bags, down 9%.

What may be different in 1997-98 is the fact that consumers have become somewhat used to changes in coffee prices (given the recent 1994 increase) and tend to expect that what goes up will come back down again, based on past experience. Thus, consumption may not in fact decline very much. Also, it appears that roasters at the retail level are fully cognizant that consumers will in fact react to higher prices and will switch to other beverages, perhaps never to return to coffee. Therefore, they have likely kept retail price increases as low as possible. Also, any retail increases will not be as “sticky”–i.e., they will come back down more quickly.

F.O. Licht estimates world coffee consumption in 1996-97 at 102.4 million bags. Given the recent price increase, we would expect that consumption would decline in 1997-98. Other sources, such as the USDA, see 1997-98 consumption increasing. A reasonable projection would be for 1997-98 coffee consumption to be about 100 million bags.

If there is a critical question to be asked about coffee this year, it concerns the size of stocks. It was concern about tight stocks, along with expectations of smaller crops, that tripled prices early this year. There have been signs since October that coffee stocks are getting tighter. While new- crop coffee (from Mexico, Central America and Colombia) will not reach the market in any significant size until November-December, it is expected that arabica stocks will be tightest in the October- November period, right before the active harvesting.

Coffee stocks can be somewhat difficult to measure accurately, as they accumulate at various stages of the pipeline. Indeed, roasters may in fact be reluctant to reveal how much coffee they have for competitive reasons. Thus, gauging the amount of coffee on hand is difficult. One measure of help has come from the CSCE, which now releases the amount of daily stocks at its warehouses in New York, New Orleans, and Miami. This is arabica coffee that has been brought to the CSCE for grading and certification for delivery against the contract. The daily reporting provides one measure of the level of coffee stocks, though it can be somewhat misleading in that coffee may be intentionally moved to the CSCE, distorting the stocks picture.

The current market inversion, where the front position commands a steep premium to the back months, is in itself attracting a great deal of coffee to the CSCE for grading and certification. Some of it has failed to make the grade, and some of the coffee that has passed has been described as marginal. In addition to the normal amounts from Central America, Mexico and Colombia, there are increasing supplies from India, Uganda, Rwanda and Burundi. It should be noted that these stocks are at the CSCE and are not necessarily an accurate reflection of what roasters are holding.

At the beginning of the recent coffee rally, stocks of coffee at the CSCE had fallen to 321 bags. This amount did not constitute even one contract of coffee for deliverable purposes. In November 1996, when stocks were at their lowest, prices started to rally. The subsequent rally fulfilled one purpose of an organized commodity market–i.e., to bring more coffee to the market. That was indeed accomplished, though the amount of coffee was not large and deliveries have been light. Currently, stocks at the CSCE–after rising to more than 100,000 bags–are back under 40,000. On an absolute basis, the amount of coffee at the CSCE has in fact increased significantly from the November 1996 lows, but on a relative basis stocks are still quite low. Two years ago, CSCE stocks were over 1.0 million bags, or 4,000 contracts.

A different measure of U.S. coffee stocks comes from the Green Coffee Association. This measure of member warehouse stocks of green coffee reached a low of 1.26 million bags on October 31, 1996. While there has been an absolute increase in stocks since the recent lows, with the stocks most recently at 2.34 million bags, GCA stocks in the past have been much higher. In early 1993 GCA stocks were over 10 million bags.

Perhaps the most encompassing view of coffee stocks, comes from E.D. & F. Man, the London trade house, which tabulates stocks in consuming nations–i.e., Europe, the U.S. and Japan. The U.S. stocks figure would be represented by the GCA estimate. Man reported that at the market low in December, consumer stocks of coffee were a record low 7.5 million bags. By the end of August 1997, stocks had increased to 11.8 million. Man noted that this represented just over eight weeks of stocks in consumer hands at the end of August, which would provide roasters with a small measure of comfort as they approach peak roasting season.

Another measure of arabica coffee stocks is found in Colombia, the second-largest producer/exporter of coffee. The Colombia National Coffee Federation reported that at the end of June 1997, their inventory was 4.5 million bags, down more than 27% from the October 1996 total of 6.2 million. Colombian officials have indicated that because of the much smaller crop this season, they have had to rely on stocks to meet export commitments. They expect their stocks to be drawn down further ahead of the next harvest in the fall. Indications now are that the crop could be late, and if that is the case, National Coffee Federation stocks will be drawn down further, perhaps to under 2.0 million bags. There is some significance to this, in that the federation stocks acted as a buffer, allowing export commitments to be met. Going into the 1997-98 season, that buffer will be at minimal levels, and any shortfall in the Colombian crop would have even more bullish implications.

Stocks of coffee held by coffee-producing countries have also been declining. According to USDA data, in the 38-year period between 1960-61 and 1997-98, stocks of coffee-producing countries reached a high of almost 87 million bags in the 1965-66 season. At that point, stocks held by producers were greater than world coffee production, which was put at 67.5 million bags. Producer stocks fell to as low at 25.1 million bags in the 1978-79 season, only to build again to more than 50 million in 1988-89. Despite efforts by the Association of Coffee Producing Countries to restrict exports of coffee to the consuming nations, producer stocks have in fact fallen, and in 1997-98 are projected by the USDA to total only 20 million bags. With a USDA world coffee consumption estimate of 106 million bags, that would imply a producer stocks-to-world-use ratio of .19.

Total world stocks of coffee (producing and consuming nations) have been edging lower for some time. In the 1990-91 season, Licht estimated stocks at just over 55 million bags, with world production at 92.9 million. By 1996-97, global stocks had fallen to 33.5 million bags, with production of 99 million. While world production rose by less than 7%, world stocks fell by 39%. In the same period, the ending stocks-to-use ratio fell from .58 to .33.

To find an explanation for the 1996-97 bull market in coffee–particularly higher-quality arabica coffee–we need to examine production, exports and stocks. Looking at Mexico, Colombia and the five Central American producers, production in 1996-97 was 26.7 million bags, down 11% from the 1995-96 season, using F.O. Licht estimates. Almost all of the production decline was in Colombia, which produces the highest- quality arabica, coffee. The Colombian crop by itself was more than 25% smaller. Despite the decline in coffee output, exports did not decline, implying that coffee stocks were drawn down. Overall, 1996-97 season coffee exports are estimated at 26.7 million bags, up 1% from the previous season.

For the seven countries, stocks at the end of the 1995-96 season were estimated at 7.94 million bags, with three-fourths held by the Colombian National Coffee Federation. These stocks declined considerably, and by the end of the 1996-97 season (at the end of September) were expected to be only 3.27 million bags, down almost 50%. Looking ahead to the 1997- 98 season, which began in October, this drawdown in stocks of arabica coffee is contributing to a more positive price outlook.

Supply and demand estimates for the coffee market remain subject to a high degree of variability. For example, whether the 1997-98 Brazilian crop is 19 million bags, as the Brazilian National Coffee Council estimates, or 28 million bags (the USDA projection) will have a large impact on whether world stocks of coffee decline or increase. In addition, dry weather patterns over Central America and Colombia will result in reduced production potential for the 1997-98 new crops, though by how much is difficult to quantify at this time.

If we assume that 1996-97 world coffee production was 100 million bags, the 1997-98 season should see about the same–perhaps 98-102 million bags. While the USDA forecasts almost 104 million bags of world production, the Brazilian crop could well be lower, and the effect of the El Nino weather phenomenon worldwide is likely to shave 1.0-2.0 million bags off the projection.

In terms of coffee consumption, it does appear that there will be somewhat of a decline as a result of higher retail prices. However, prices are not the only determinant, and the increase in world population will mean more usage. The USDA estimated the 1996-97 season consumption at almost 108 million bags, while Licht puts it at 102 million. Expectations are that there will be a decline in coffee consumption of perhaps 5%, to maybe 97 million bags, though world coffee consumption could end up between 96 and 100 million bags. Hence, there should be a rough balance between the amount of coffee produced and consumed in the 1997-98 crop year.

Given the variables still at work in the market, like uncertainty about Brazil's crop and lingering El Nino effects, there is also greater variability in production prospects. If we assume a more bearish scenario (world production to exceed consumption by 3.0 million bags), then based on the 1996-97 Licht estimates, world stocks will increase to almost 37 million, marginally above the 1995-96 level and still well below recent world stocks. The stocks-to-use ratio would be .38, again marginally above 1995-96 but well below recent ratios. In that case, we'd expect coffee to trade between $1.20 and $1.60 for most of the season.

A second, neutral scenario would be for world production and consumption to be about the same. That implies that global coffee stocks would hold at about 34 million bags, and the stocks/use ratio would stay at about .33. It also implies that coffee prices would stay in a range of about $1.40-1.80 for most of the season.

The third scenario is bullish, and it assumes that world production is less than consumption. That could happen if the Brazilian crop is less than 24 million bags, if El Nino does further damage or in fact consumption turns out to be larger than has been assumed. In that event, world coffee stocks would fall, the stocks/use ratio would decline, and prices would be higher–perhaps in a range of $1.80-2.20.

November 1997Walter Spilka, Coffee Analyst, New York

Smith Barney

390 Greenwich Street, 1st Floor, New York, New York

Consensus National Futures and Financial On Line Index

Added to the WWW 12-12-97
Last updated on 12-12-97

Hosted by:
One Crossroads Place
610 West Maple Ave, Suite WWW
Independence, MO 64050
(816) 252-4080
sysop@kcmo.com

wmeubank@ocp.kcmo.com