Horticultural & Tropical Products Division	Return to the H&TP Home Page

ESTIMATING AND FORECASTING CROPS AT USDA

National Coffee Association Annual Convention
March 11, 2000

Franklin E. Hokana
FOREIGN AGRICULTURAL SERVICE
HORTICULTURAL AND TROPICAL PRODUCTS DIVISION

Agricultural Production Estimates and Forecasts at USDA

For the 48 conterminous states, USDA employs a complex system of area and list frame sample surveys for use in forecasting and estimating crop production and livestock numbers. The USDA agency responsible for these estimates is the National Agricultural Statistics Service (NASS).

The principal methodology used by NASS for corn, soybean, and wheat production is the objective yield survey utilizing an area sampling frame. For estimating livestock numbers, NASS uses an enumerative survey that uses a list sample. The methodology is consistent for those states designated for these surveys and is summarized in Washington to obtain U.S. totals using state expansion factors.

For field crops such as soybeans, objective yield surveys are conducted during the growing season taken from plots two rows wide and 10 feet long that have been selected from an area frame sample. The total number of plots in any given state is determined by a planting intention’s report. Two plots are randomly selected within the boundaries of each field. These selected plots could appear any where within that field and thus are not designed as a measure of the yield for that particular field, but for the state. The number of plants for each plot is recorded as are other anatomies of the plant such as the number of lateral branches and pods throughout the growing season. Biological yields are estimated for each plot with an adjustment for harvesting loss to determine an actual yield.

NASS also uses random path methodology to estimate tree fruit production. In this procedure fruit along a path starting at the trunk and ending at a terminal branch is counted. At each forking of the branches along the path, random numbers are used to select one of the branches to continue the path. Measurements made at each forking, when combined with the fruit counted along the "random path", allow an unbiased estimate to be made of the total fruit on the tree, even though usually only about 5 percent of the actual fruit lies along the sampled path. A sample of the counted fruit can also be sized, allowing an estimate of production for the sampled tree. NASS uses this procedure to estimate production for almonds, walnuts, pistachios, tart cherries, and navel and valencia oranges.

For livestock estimates, there are two enumerative surveys which are conducted in June and December of each year. The name, derived from enumeration, describes the type of survey conducted, that is to obtain sample data by an actual count. Enumerators will visit livestock farms, record the number of livestock on a pre designed form. These results are later tabulated and expanded to state levels.

Each of the above surveys contains numerous elements for assuring accuracy. There are quality control measures for each survey that assure statistical accuracy and compliance.

World Supply and Demand Estimates and Forecasts

Both the Congress and public require supply and demand forecasts well in advance of the current crop or marketing year. Data on production, stocks, expected consumption, exports, and imports (PS&D) are all essential for future planning. Both foreign and domestic PS&D estimates and forecasts are reviewed by the World Agricultural Outlook Board (WAOB). The forecasts and estimates are processed through Interagency Commodity Estimates Committees. The USDA committee members, other than NASS mentioned above and their responsibilities are:

* Joint Agricultural Weather Facility (JAWF): Global weather and crop assessment.

* Economic Research Service (ERS): U.S. and foreign crop and livestock analysis.

* Farm Service Agency (FSA) U.S. Farm Programs

* Foreign Agricultural Service (FAS): Production, supply, and disposition (PS&D) of foreign crop and livestock production and marketing information.

The WAOB is the central body which coordinates analysis activities, providing independent review and clearance of the final product for consistency and objectivity. The objectives of the Board are to: assure a maximum complementarity of effort between agencies; eliminate and guard against duplication of effort; and achieve efficiencies inherent in joint efforts. In addition to the mission objectives it provides a common data base for all commodities that becomes uniform among all the USDA agencies. That is there is only one USDA official number.

The "FAS mission is to serve U.S. agriculture's international interests by expanding export opportunities for U.S. agricultural, fish, and forest products and promoting world food security."

"FAS takes the lead within the Department of Agriculture to protect and strengthen the long-term competitive position of U.S. agricultural, fish, and forest products in foreign markets. It accomplishes this goal through negotiating and monitoring trade agreements; market development, promotion, and outreach activities; country and commodity market intelligence collection, analysis, and dissemination; and international financial assistance programs.

The ERS provides economic analysis and intelligence on U.S. and global agriculture, food, natural resources, and rural America. The NASS mission is to serve the basic agricultural and rural data needs by objectively providing important, useable, and accurate statistical information and services for informed decision making. The FSA mission is to implement, monitor, and oversee U.S. farm programs.

For coffee production estimates and forecasts and the PS&D data the World Board role is more simplified than what’s been described so far. Often times it may be only two or three people formally meeting, but a meeting is held for every coffee circular that FAS publically releases.

USDA Coffee Production Estimates and Forecasts

FAS’s methodology for estimating and forecasting coffee production and marketing is altogether different from how NASS prepares their estimates and projections. U.S. agricultural attaches and counselors based in U.S. embassies provide the major input for world green coffee production and trade reports. All forecasts and estimates are reviewed and approved by the WAOB. In addition to the attache reports, FAS makes considerable use of the information, data and analysis provided by the of the Crop Condition Assessment Branch (CCAB), NASS, and the National Oceanic and Atmospheric Administration in conjunction with USDA (JAWF). We also review reports from the International Coffee Organization (ICO), the Association of Coffee Producing Countries (ACPC), and the Food and Agricultural Organization of the United Nations (FAO) in developing our estimates. In addition, foreign Government forecasts/estimates along with news service reports, newsletters, and various U.S. trade journals are considered. On occasion, contacts with foreign embassies in Washington have been made for the purpose of gathering information. The coffee report, is analyzed in the Horticultural and Tropical Products (HTP) Analysis Branch.

Attache Reporting

FAS receives reports for the June coffee survey from 16 countries which account for about 75 percent of world total green coffee production. For the December report, five countries that account for about 55 percent of the world’s total coffee production have mandatory reporting. This does not include ‘Voluntary Reports’ which a Post will prepare if a significant change occurs in production or marketing. Some of the major producing countries for which FAS has no attache reporting are Ethiopia, Uganda, Papua New Guinea, Cameroon, and Nicaragua.

It is common for the Washington analyst to consult with the Post for clarification or review when more recent information appears after the report has been electronically filed. This is done through the agency’s Global Agriculture Information Network (GAIN) system. This system permits the reporting post to download the most recent official USDA estimates from a lotus spreadsheet file and submit any changes that it may wish to make. The reports are released to the public three working days after they are submitted to Washington. This transfers a greater burden of reporting to the Post from the Washington staff. More important, when this procedure was adopted, the report became available to the public much sooner than with prior processing methods.

Although reporting instructions are the same for each Post, preparation of the coffee report can vary considerably depending on the data and available resources. For most Posts, the attache relies on discussions with Government and trade officials, grower organizations, field travel, and other institutions such as banks and agricultural research stations.

FAS Circulars Reporting Coffee Production and Trade

World coffee production data are released by FAS in June and December each year in a Circular Series entitled, Tropical Products: World Markets and Trade. Production, supply, and distribution balances for 57 countries are included in these circulars. Numerous tables, apart from the coffee PS&D, are published that relate to consumption, prices, and stocks. A considerable number of tables are devoted to U.S. activity in terms of coffee imports, exports, and retail prices. In addition, information on other horticultural commodities and products are published, including cocoa updates.

In December, FAS provides a list of release dates for its circulars, and other reports, for the new year. These release dates, as well as the attache reports can be read on the FAS home page found at http://www.fas.usda.gov

USDA Review of Country Reporting

The World Agricultural Outlook Board and FAS are the agencies responsible for review and approval of individual country and world coffee production, supply and disposition estimates. The first forecast of coffee production is the most difficult, because of the limited knowledge available at the time when it is made. USDA usually publishes its first forecast for all countries in June.

Sources for data and information on the crop vary between countries, but generally stocks, trade, and consumption are obtained from official country sources, the International Coffee Organization , producers’ associations, and traders.

Major Producing Counties

It won’t be possible to review all of the 57 countries that are included in our circulars nor will we be able to discuss the World Coffee Supply and Distribution table that contains a great deal of information about these countries. Our discussion will attempt to focus on the production aspect and only for a few countries and one region. For ease of presentation, these first series of charts are tracking every other year from 19977/78 through 1999/2000. The first series of graphics will focus on the changes that have occurred in the production of coffee over the past 22 years. The first overhead will include the countries of Costa Rica, El Salvador, Honduras, Nicaragua and Panama that are combined into the region, Central America. It is shown along with Colombia’s and Brazil’s coffee production for comparison purposes used later in this presentation. The Central American countries increased production more than 50 percent during this interim or about 4.8 million bags. Honduras is producing about two and a half times more than it was in1977/78 while Guatemala, the largest producer in this region, has nearly doubled their production during this period.

In the next chart you can see how three individual countries’ coffee production has performed over the past 22 years. Mexican coffee production in 1977/78 was 3.3 million bags compared with the 1999/2000 forecast of 5.2 million for an increase of about 2.0 million bags. Indonesia produced 3.7 million bags in the mid-70's and is now expected to produce a crop of 7.5 million, or a 3.8 million bag increase. What’s even more impressive is tracking Vietnam’s coffee history of coffee production that was less than 500,000 bags through 1986/87 and then soaring to a 7.5 million bag forecast in 1999/2000.

Let’s turn our attention to Cote d’Ivoire and India as they seem to be heading in the same direction during the past few years. In 1977/78, India produced 2.0 million bags of coffee, but during 1999/2000 they are expected to produce 4.7 million bags. Cote d’Ivoire produced 3.1 million at the beginning of this series and is now forecast to produce 5.3 million bags.

In the Ethiopian and Ugandan chart its interesting to note that for Ethiopia, coffee production was in the 3.0 million bag range and today it has hardly increased at all to about 3.5 million bags. Ethiopia is often cited as the birthplace of coffee. On the other hand, Uganda produced 1.8 million bags in 1977/78 and today is expecting to produce 4.0 million bags. What does all this mean in view of changes that have occurred in world coffee production? For the Central American region and the seven countries shown, there are about 25 million bags more coffee on the market today than 22 years ago, and this does not include Colombia or Brazil, the world’s two largest producers. Further, it demonstrates that sometimes we tend to concentrate only on the large producers not always seeing the entire picture.

Colombia’s production in 1977/78 was 11.1 million compared with 10.8 in 1999/2000 while Brazil produced 17.1 million in 1977/78 compared with 26.5 today and a record 39.6 million bags in 1962/63. This chart hardly does justice to either country and we will take a further look at each. This is because of the yearly volatility of Brazilian coffee output and the change that is occurring in Colombia. No further comparisons are made in connection with the total added supplies, if any, as were made with the previous charts.

For Colombia and Brazil, note in this chart that two scales are USED and the year change beginning with 1989 and consecutive years after that. This chart attempts to depict the abrupt changes that occur year to year in these large producing countries since 1998/99. One can see that Colombia peaked in1991/92 at 18 million bags and Brazil in 1998/99 at 35.6 million. It can also be readily seen that Colombian coffee production has been reduced markedly since 1991/92.

Brazil

The dominance of Brazil in terms of the world’s coffee production and the volatility of its outturn is like no other country. Brazil’s share of world production was 33 percent in 1998/99 and 25 percent in 1999/2000. During the past five years its production reached a low of 16.8 million bags and a high of 35.6 million.

There are several reasons for periodic wide fluctuations in Brazilian coffee production: 1) On and off year production cycles 2) periods of high or low prices causing changes in crop inputs 3) weather related incidents, including drought and/or frost damage 4) cultural practices; pruning and tree removal, new plantings, new varieties, irrigation practices 5) and changes in the number of bearing trees, area, tree density per hectare, and cultural practices. The net effect over time has been a steady shift of the coffee tree population to the north. The shift in coffee production to the north has been caused by freezes which have either killed the coffee trees or set the trees back so that renovation through pruning was impractical. For future coffee production potential, this shift has had a positive effect with new varieties that begin to produce at a younger age, a significant increase in density per hectare, and an increase in irrigation. Further, the trees are not as vulnerable to frost or freezes because of the movement to warmer climates.

For 37 years, ending in May 1996, FAS conducted its coffee survey in Brazil by an agronomist working out of the agency’s Rio De Janeiro office. It was a thorough and consistent survey conducted at least twice and often four times each year. It would require about three weeks of field travel through 3,500 miles of Brazil’s coffee region. The crop survey ended by November when dehusking yields finalized the year’s outturn. During these field trips, the agronomist would observe new seedlings, tree removal, on/off year coffee tree cycle, vegetative growth, flowering, and plantation management. There was very little, if any, contact with the outside industry or trade sources. The focus of the survey was to look intently at the coffee trees and determine by appearance how the trees had fared in terms of potential productivity.

Today USDA prepares the Brazilian coffee report from Sao Paulo. Periodic and extensive travel is still made though the trips are fewer in number. FAS analysts travel extensively and now make contacts with the coffee industry and trade associations as well as other industry representatives. Forecasting and estimating Brazilian coffee production begins with field travel, during which analysts observe the vegetative growth and the general condition of trees, but with these contacts viewed as an important source for information. FAS analysts now discuss off/on year cycles, prices, tree numbers, area planted and harvested, yield, rainfall, and coffee crop potential.

In Washington, USDA begins following the progress of an upcoming crop soon after the previous crop is harvested. Analysts from FAS and WAOB begin monitoring the upcoming crop by JAWF weekly weather reports, daily news reports, coffee prices, and satellite imagery.

In the handout is a chart that contains a history of Brazilian coffee statistics dating back to 1970/71. This chart shows the coffee producing regions in Brazil and the percent of total production by state based on the past five years that JAWF has updated. It includes a coffee crop calender for most of the coffee grown in Brazil. The map depicts where coffee is grown and not the volume of coffee produced by each state.

In the chart that follows, titled Brazil: Environmental Vegetation Index (EVI) comparisons for October, 1999, 1998, and 1997. The EVI legend is located on the lower right hand side of the chart. The EVI is derived from the Advanced Very High Resolution Radiometer (AVHRR) instrument on board the National Oceanic and Atmospheric Administration (NOAA) series of polar orbiting satellites. The states are highlighted and these satellite images were all taken near the end of October. Note the orange and red coloring in SW Minas Gerais compared to the two previous years and when we discuss the latest attache report, reference will be made to this chart.

In the last two charts, there is a satellite image for the month of February 2000, 1999, and 1998 also with the states of Brazil highlighted. The February 2000 year is an image taken from a satellite near the first of March and relates to the upcoming 2000/2001 upcoming coffee crop, while the 1999 imagery relates to the 1999/2000 season, and the 1998 year refers to the 1998/99 year. A color code appears at the bottom of the page describing the vegetative index conditions at the time of the image.

In the last chart for Brazil, for November-December 1999 and January-February, 2000 this reflects the vegetative conditions in late 1999 and shows the effect of the rainfall that fell in Brazil after December 10. It has become apparent that vegetation has responded to rains that began falling nearly three months ago. Some of the areas that appear marginal in January-February 2000 are the result of ripening fields or fields of other crops that have been harvested.

As mentioned earlier, the attache in Sao Paulo released the first forecast this week for the 2000/2001 year. Excerpted comments from this report said coffee trees in the southwestern and center-western part of the state of Minais Gerais were stressed by the drought. The leaf drop rate in this region was high, varying from about 30 to 80 percent of the plant except for irrigated areas. He said older trees appeared generally more affected and less productive than younger trees, which were showing greener leaves. Fruit set was fair, but above average fruit drop was noted. Southeastern Minas Gerais is expected to produce a good crop in 2000/2001.

In Espirito Santo, both the center-north where the robusta is grown and in the south where arabica is grown, production is expected to be higher than last year, reflecting favorable weather and good crop management.

In the state of Sao Paulo, the forecast of 2.5 million bags is down one million bags from the 1999/2000 crop. Coffee trees were stressed by the 1999 drought especially close to the Minas Gerais border and, you can clearly see this on the chart. The leaf drop rate is high except for irrigated plantings. Fruit set is fair, but above average fruit drop has resulted because of heavy rainfall in late December and early January.

For the state of Parana, the forecast for 2000/01 is 2.2 million bags, a drop of 15 per cent from last season. Coffee trees were partially affected by below average temperatures in September 1999 and the drought that followed. The drought impact was at a lower extent than in the state of Sao Paulo and southwestern and the center-western part of Minas Gerais.

For other producing states, the coffee production forecast is up slightly from last year.

Now we can compare some of the state forecasts and previous year’s estimates with what we have learned. First, we know that 2000/2001 is an on/year cycle with some early estimates of 40 million bags. Secondly, the accumulated rainfall through the crucial flowering period October-November for most of these states was well below last year and normal. Minas Gerais is the largest producing state with three distinct areas. Coffee production estimates and the attache report for 2000/01 for the three regions in Minas, Espirito Santo, Sao Paulo, and Parana are as follows:

1/ Bahia and Rondonia

Because the year 2000/01 is an on-year in the coffee production cycle and early estimates were near the 40-million bag mark, a compariosn will be made not only to last year, but also to the year 1998/99 when 35.6 million bags were produced.

Graphs for presentation can be viewed or downloaded in freelance or pdf format.

For questions regarding this presentation please contact Frank Hokana at (202)720-0875 or at hokana@fas.usda.gov

Last modified: Wednesday, July 21, 2004