OIL CROPS YEARBOOK October 23, 2000 October 2000, ERS-OCS-2000 Approved by the World Agricultural Outlook Board -------------------------------------------------------------------------------- OIL CROPS YEARBOOK is published annually by the Economic Research Service, U.S. Department of Agriculture, Washington, DC 20036-5831. This release contains only the text of the OIL CROPS YEARBOOK--tables and graphics are not included. Printed copies of this yearbook will be available from the ERS-NASS order desk. Call, toll-free, 1-800-999-6779 and ask for stock # ERS-OCS-2000, $21. ERS-NASS accepts MasterCard and Visa. -------------------------------------------------------------------------------- Summary U.S. farmers planted 73.7 million acres of soybeans in 1999, up from the 1998 record of 72.0 million. After mid-July, the Mid- Atlantic, Delta, and Southeast regions suffered from little rain and intense heat, cutting the national average soybean yield to 36.6 bushels per acre. The 1999 U.S. soybean crop was 2,654 million bushels on a harvested area of 72.4 million acres. The United States exported a record 973 million bushels in 1999/2000, surpassing the former 1981/82 record of 929 million bushels. With a smaller harvest and record exports, U.S. ending stocks of soybeans fell to 288 million bushels compared with the 1998/99 carryout of 348 million. Domestic soybean crushing declined from 1,590 million bushels in 1998/99 to 1,579 million in 1999/2000, as depressed oil prices perpetuated very narrow crush margins. The national soybean price averaged $4.65 per bushel in 1999/2000, which was the lowest since the 1972/73 average of $4.37 and well below the $5.26 national loan rate. For the 1999 crop, farmers received $2.1 billion worth of loan deficiency payments (LDP) on 2,319 million bushels of soybeans. Despite substantially lower import demand for soybean meal by the European Union and China, purchases by other importers in Asia and the Middle East supported U.S. exports at 7.3 million short tons, up modestly from 1998/99. The slowdown in domestic soybean crushing buoyed soybean meal prices in 1999/2000 to $168 per short ton, up from the 1998/99 average of $138.50. Higher meal prices, a contraction in hog production, and slower poultry output trimmed 1999/2000 domestic disappearance 0.7 percent to 30.45 million tons. Exports of U.S. soybean oil plunged to 1,375 million pounds in 1999/2000, from 2,371 million the previous year. A steady 3- percent expansion in domestic disappearance to 16,059 million pounds helped absorb some of the surplus. But the slump in exports caused 1999/2000 year ending soybean oil stocks to accumulate to 1,995 million pounds, nearly 30 percent more than the previous year. The global glut of vegetable oils depressed the 1999/2000 national average price of soybean oil to 15.6 cents per pound, down from 19.9 cents in 1998/99 and the lowest since 1986/87. Dry weather trimmed world soybean production 2 percent in 1999/2000 to 157.7 million metric tons. Global soybean imports grew 18 percent in 1999/2000 to 47.9 million tons. But global soybean meal consumption increased only 2 percent in 1999/2000, compared with 7 percent the previous season. With greater domestic production and moderate consumption growth, world soybean meal imports declined 0.3 million tons in 1999/2000 to 39.0 million. China’s soybean crush soared 18 percent in 1999/2000, which combined with a decline in domestic soybean output, swelled China’s soybean imports to 10.1 million tons, from 3.9 million in 1998/99. The resulting gain in China’s 1999/2000 domestic meal production more than replaced the 0.8- million-ton decline in soybean meal imports to 0.6 million tons. Global vegetable oil supplies for 1999/2000 increased 6 percent to a record 92.2 million tons, mostly from palm and rapeseed oils. World trade in soybean oil declined about 9 percent to 7.4 million tons. Increasing reliance on imports of soybeans and rapeseed shrank China’s soybean oil imports from 0.95 million to 0.6 million tons. Despite higher import tariffs, much lower world palm oil prices encouraged a 3-percent increase in Indian palm oil imports to 3.1 million tons in 1999/2000. Malaysia’s 1999/2000 exports surged 9 percent to 8.8 million tons, yet, stocks also rose to a record 1.3 million tons. A southern drought caused acreage abandonment and below-average U.S. cotton yields in 1999, which resulted in cottonseed production of 6.4 million short tons. Nevertheless, larger cottonseed supplies permitted 1999/2000 crushing to recover to 3.1 million tons from the previous year’s slide of 2.7 million. Continuing a trend that started in 1998/99, cottonseed feeding (at 3.5 million tons) exceeded crushing demand of 3.1 million tons. Cottonseed prices declined from the 1998/99 average of $129 per ton to $89. Despite relatively attractive potential returns for sunflowers in 1999, persistent rains during planting caused little change in U.S. acreage at 3.6 million acres. Wet weather during flowering promoted disease, reducing the national average sunflowerseed yield to 1,262 pounds per acre, the poorest since 1995. U.S. sunflowerseed production, at 4,342 million pounds, dropped 18 percent from the bumper 1998 harvest. With tighter seed supplies and very weak oil prices, sunflowerseed crushing in 1999/2000 declined 3 percent to 2,510 million pounds. Exports of sunflowerseed oil for 1999/2000 fell to 630 million pounds. U.S. farm prices averaged $7.53 per hundredweight, well below the $9.30 marketing loan rate and the lowest since 1986/87. Consequently, farmers received $120 million of loan deficiency payments on 80 percent of 1999 sunflowerseed production. U.S. Soybean Situation, 1999/2000 U.S. farmers planted 73.7 million acres of soybeans in 1999, up from the 1998 record of 72.0 million. The comparatively dim outlook for corn and wheat returns boosted soybean acreage 0.7 million acres higher than March intentions. Ten of the top 12 soybean producing States planted record acreage in 1999, absorbing more acres from corn and wheat. The largest prospective gains were in the western Corn Belt, including Nebraska (up 500,000 acres), South Dakota (up 450,000), and Iowa (up 400,000). Only Arkansas and several other southern States planted fewer soybeans than in 1998, where more land was switched into cotton and rice. Fewer winter wheat plantings also reduced the amount of soybean double cropping, especially in Missouri, Kentucky, and Tennessee. The other major factor for greater soybean planting was the marketing loan. In March, USDA announced that the 1999 marketing loan rates for soybeans, corn, and wheat would be set at $5.26, $1.89, and $2.58 per bushel, the same as last year. The marketing loan rate for minor oilseeds was set at 9.3 cents per pound. Under the 1996 farm legislation, these are the maximum rates allowed. Soybean cash prices were not expected to exceed the loan rates in 1999, so farmers could make up the difference through loan deficiency payments and marketing loan gains. Even if soybean cash prices had subsequently weakened, farmers were still assured of a better return by planting soybeans. Given relative production costs and expected yields, a ratio of 2.7-2.8 between the soybean marketing loan and December corn futures favored soybeans. Soybean yields also tend to suffer less under weather stress. The U.S. average variable costs of corn production are approximately double the cost of soybeans (about $81 per acre), so risk-averse farmers can better stretch their operating loans by planting more soybeans. In early July 1999, soybean crop conditions were generally quite good. Central Illinois spot prices reflected this by falling below $4.00 per bushel. But after mid-July, the Mid-Atlantic, Delta, and Southeast regions suffered from little rain and intense heat. As in 1998, weeks of record-setting summer heat and widening drought quickly deteriorated U.S. soybean crop conditions. The difference in 1999 was the inclusion of the Ohio River Valley among the most affected areas. Temperatures moderated in early August, but soil moisture in most areas was drawn down substantially by the time pod filling had begun. Late in August, a tropical storm alleviated dryness in eastern and southeastern parts of the country. This moisture may have helped pod filling, but poor pod counts and the advanced soybean development limited the yield improvement in these areas. Dry weather prevailed throughout the country in September and October, producing very favorable harvest conditions, but the dry summer cut the national average soybean yield to 36.6 bushels per acre. The protein content of U.S. soybean meal was also diminished by the adverse weather. Although conditions deteriorated in the East and South, high yields and large acreage yielded record or near record soybean harvests for Iowa, Nebraska, Minnesota, Wisconsin, Michigan, North Dakota, and South Dakota. The 1999 U.S. soybean crop was 2,654 million bushels on a harvested area of 72.4 million acres. Following harvest, low soybean prices and delayed exports from South America encouraged brisk U.S. export shipments, particularly to China, the European Union (EU), and Mexico. A recovery from the financial crisis also aided shipments to other Asian countries. The United States exported a record 973 million bushels in 1999/2000, surpassing the former 1981/82 record of 929 million bushels. With a smaller harvest and record exports, U.S. ending stocks of soybeans fell to 288 million bushels, compared with the 1998/99 carryout of 348 million. Soybean crushing for the first quarter of 1999/2000 was an all- time high. However, in spite of the low cost of soybeans and somewhat firmer meal prices, depressed oil prices perpetuated very narrow crush margins. Domestic soybean crushing subsequently weakened in the face of sluggish demand for both soybean meal and soybean oil exports. The competitive pressure from imports of Canadian rapeseed products also contributed to the contraction in U.S. soybean crushing. Several crushing plants closed permanently and many others operated well below capacity. Soybean crushing declined from 1,590 million bushels in 1998/99 to 1,579 million in 1999/2000. Cash soybean prices rallied to $5.40 per bushel by mid-May based on dry weather in South America and the United States and strong soybean exports. However, indications for record U.S. plantings and subsequent rains raised prospective 2000/01 output, and prices again receded. Soybean prices averaged $4.63 per bushel in 1999/2000, the lowest since the 1972/73 average of $4.37 and well below the $5.26 national loan rate. Unlike years prior to 1991, the soybean marketing assistance loan provides no floor for how low cash prices can fall below the loan rate. In October 1999, President Clinton signed the agricultural appropriations act for fiscal year 2000. Among the provisions to assist commodity producers was an effective payment of $438 million to oilseed producers based on each producer’s 1997 or 1998 production. U.S. oilseeds production in 1999 totaled 82.0 million metric tons, so the effective payment rate to soybean producers (on 72.2 million tons) was about 14.5 cents per bushel. More general provisions extended the discount of 30 percent on crop insurance premiums and the payment limitation for loan deficiency payments (LDP) and marketing loan gains was raised (for crop year 1999 only) from $75,000 to $150,000 per person. At average LDP rates for soybeans of about 91 cents per bushel, many mid-sized farms could have been affected by the lower payment limit in 1999. For the 1999 crop, farmers received $2.1 billion worth of LDPs on 2,319 million bushels of soybeans. Farmers also have the option to put soybeans under loan and can redeem the loan at a price below the loan rate. Farmers placed 284 million bushels of soybeans under loan in 1999, receiving $211 million in marketing loan gains. Together, the programs applied to 98 percent of the 1999 U.S. soybean harvest and supplemented the crop’s $12.2 billion market value. The strength of U.S. soybean exports also reflected a weakness in the exports of soybean meal and soybean oil. U.S. exports of soybean meal to the EU and China declined by 60 and 100 percent. Both markets relied more heavily on meal supplies from domestically crushed oilseeds, and rapeseed in particular. The slippage of the euro exchange rate against the U.S. dollar significantly raised costs for European importers in 1999/2000. However, meal purchases by other importers in Asia, the Middle East, and Latin America supported total U.S. exports at 7.3 million short tons, up slightly from 1998/99. The slowdown in domestic soybean crushing buoyed soybean meal prices in 1999/2000 to $168 per short ton, up from the 1998/99 average of $138.50. Higher meal prices, a contraction in hog production, and slower poultry output trimmed 1999/2000 domestic disappearance by 0.7 percent to 30.45 million tons. A relatively high oil extraction rate in 1999/2000 partially offset the reduction in soybean crush and soybean oil production. Despite bargain prices, U.S. soybean oil exports sagged from abundant world supplies of vegetable oil and much lower imports by China. Exports of U.S. soybean oil plunged to 1,375 million pounds in 1999/2000, from 2,371 million the previous year. A steady 3-percent expansion in domestic disappearance to 16,059 million pounds helped absorb some of the surplus. But the slump in exports caused 1999/2000 year ending soybean oil stocks to accumulate to 1,995 million pounds, nearly 30 percent more than the previous year. The global glut of vegetable oils depressed the 1999/2000 national average price of soybean oil to 15.6 cents per pound. The soybean oil price fell from 19.9 cents in 1998/99 and was the lowest since 1986/87. Oil prices rallied in March and April based on declining crush rates and very dry U.S. soil conditions prior to 2000 planting, but sank again by summer after moisture conditions had improved. World Oilseed and Protein Meal Situation World soybean production dipped 1 percent in 1999/2000, to 157.2 million metric tons. The decline was largely because of smaller U.S. output. Global soybean imports grew 18 percent in 1999/2000, to 47.9 million tons. China accounted for 80 percent of the expansion in world soybean trade. Global soybean meal consumption increased only 2 percent in 1999/2000, compared with 7 percent the previous season. With greater domestic production and moderate consumption growth, world soybean meal imports slipped 0.3 million tons in 1999/2000 to 39.0 million. After Brazil’s January 1999 currency devaluation, soybean prices in the local currency firmed, but relative prices generally favored more corn planting. In addition, the weakened financial situation of Brazil’s farmers, inflated production costs, and intense U.S. competition dampened the price enhancing effects of the devaluation. Dry conditions in Brazil slowed planting progress, which trimmed intended corn area. Consequently, Brazil’s soybean harvested area edged up to 13.4 million hectares, from 12.9 million in 1998/99. A series of rainshowers in late December substantially improved soybean-growing conditions in the northern half of Brazil. But drought persisted in the southernmost states of Rio Grande do Sul and Parana. Lower soybean yields were offset by higher area, resulting in a small increase in Brazilian production, from 31.3 million tons in 1998/99 to 32.0 million. Brazil usually exports several hundred thousand tons of soybeans during February-March, but such shipments were minimal in 2000 because of the delayed harvest. Yet, in subsequent months, strong demand by importers (particularly China) accelerated Brazil’s 1999/2000 soybean exports, which rose to 11.4 million tons, from 8.9 million the previous year. Following Brazil’s January 1999 devaluation, it became more profitable for producers in the center-west states of Brazil to export abroad than to supply domestic processors. Access to soybeans grown in the expansion areas of the center-west by Brazilian crushers (located mostly in the south) has been complicated by interstate value-added taxes. Such taxes reduce the values that these processors can pay for soybeans outside their own state than with foreign buyers. Drought had reduced local supplies in the southern states, so Brazil’s 1999/2000 soybean crush remained constant at the previous season’s 21.0 million tons. Limited supplies, growth in domestic use, and smaller world imports caused Brazilian soybean meal exports to slip 4 percent to 9.75 million tons in 1999/2000. With the still weak exchange rate, the relatively high domestic price (in local currency) compared with the export value (in dollars) meant that Brazilian meal consumers outbid more of their foreign counterparts. In December 1999, internal prices for soybean meal were 52 percent above a year earlier, compared with 38 percent higher soybean prices. By contrast, prices in dollar terms were up only 2 percent and down 21 percent, respectively. Exchange rates especially favored Brazil’s exports of domestically produced poultry, which have become the world’s third largest. Higher meat production helped domestic soybean meal consumption rise 4 percent to 6.9 million tons. Drought conditions also harmed Paraguay’s soybean growing regions, which border the most affected areas in Brazil. Paraguayan soybean production dropped to 2.9 million tons in 1999/2000, from 3.0 million the previous year, resulting in a reduction in exports to 2.2 million tons. For Argentina, smaller U.S. and Indian crops had somewhat strengthened the price outlook in late 1999, so soybean area edged up to 8.6 million hectares. An increase was possible because a 14-percent expansion of winter wheat area raised the proportion of double-cropped soybeans. Weakness of vegetable oil prices also discouraged sunflower planting in favor of soybeans. Argentina had quite dry planting conditions in late 1999, but soil moisture had substantially improved by February. Based mostly on the larger area harvested, Argentine soybean production increased from 20.0 million tons to 20.7 million. Wet harvest conditions added to delays in marketing the crop. While Brazil’s domestic market for soybean products expanded, Argentina’s small domestic sector necessarily makes it heavily dependent on exports. Despite a smaller U.S. soybean harvest and a larger domestic crop, lackluster growth in international demand for soybean products (particularly by China) stifled Argentine soybean crushing. Poor margins made it hard for Argentine oilseed processors, like their American counterparts, to operate to capacity, which had expanded substantially in the mid-1990’s. After several years of large gains, Argentina’s 1999/2000 soybean crush fell 0.2 million tons to 17.3 million. Consequently, Argentine soybean meal exports increased only slightly to 13.65 million tons. Yet, stronger world demand for whole soybeans instead of soybean products swelled Argentine soybean exports from 3.2 million in 1998/99 to a record 4.6 million tons in 1999/2000. Thus, Argentina’s end of September soybean stocks tightened compared with large levels of the previous 2 years. A major constraint on world trade for the major soybean meal exporters was the remarkable reversal of China’s buying patterns. Weaker prices diminished the incentive for soybean sowing in China, which reduced 1999 output to 14.3 million tons, from 15.2 million in 1998, and expanded its import requirements. China’s policies shifted toward greater use of domestic oilseed crushing capacity versus imports of protein meal and vegetable oil. By some estimates, Chinese crushing capacity expanded 25 percent from 1998, or as much as 3 million metric tons annually. Vegetable oil prices within China have been more than double world prices, which are maintained by strict import quotas on vegetable oils. The relative absence of import barriers on oilseeds, however, provided a substantial advantage to domestic crushers in producing highly valued vegetable oil. The recovery in consumption of protein meal and vegetable oil from the slowdown of 1998/99 also boosted soybean demand. Domestic crushing soared 18 percent in 1999/2000, which combined with the decline in domestic soybean output, swelled China’s soybean imports to 10.1 million tons from 3.9 million in 1998/99. This turned China into the world’s single largest soybean importing nation. Virtually all of China’s soybean imports are crushed. The resulting 18-percent gain in 1999/2000 domestic meal production exceeded the 1-million-ton decline in soybean meal imports to 0.6 million tons. A large expansion in domestic rapeseed crushing also minimized China’s import requirements for soybean meal and soybean oil. The soybean meal imports that did arrive were encouraged for a short period by a State Bureau of Taxation decision in June to exempt oilseed meals from the 13-percent value-added tax. However, other Government ministries did not sanction the policy change, and it was never officially implemented. For India, import controls and strong consumer oil demand make the domestic market resistant to low worldwide oilseed prices. The Government also guaranteed farmers a higher soybean support price. Although the monsoon season started off well in June 1999, the early rains faltered before planting was complete. Soybeans in India are entirely dependent on the summer monsoon, as irrigation is seldom used. The 1999 soybean harvest was 5.2 million tons, down from 6.0 million the previous year. The smaller crop and low prices slowed market deliveries of soybeans. India does not import oilseeds for crushing to cover the deficit, but has instead imported large volumes of low-priced vegetable oil. The squeeze on post-harvest crush margins had discouraged strong soybean price bids and slowed processing, but crushers benefitted from a January-May 2000 rally in soybean meal prices. Smaller supplies available for crushing cut 1999/2000 Indian soybean meal exports to 2.35 million tons (the smallest in 5 years) from 2.8 million in 1998/99. Without substantial soybean meal imports by China, the surplus of South American supplies were diverted into EU and Asian markets at a very favorable price. But with bumper EU rapeseed harvests, margins for EU soybean crushers were squeezed very tight. Exchange rates also influenced the composition of Western European imports. From January 1999 to January 2000, Rotterdam prices (in dollar terms) were down 6 percent for soybeans and up 19 percent for soybean meal. In terms of the euro, however, import costs to European buyers for both products rose 9 percent and 38 percent. Therefore, crushing of imported soybeans became comparatively more attractive than importing soybean meal. Margins for EU soybean crushers were not entirely favorable, though, as vegetable oil prices dropped by one-third. EU soybean imports of 15.75 million tons were down 1.0 million in 1999/2000 from the previous season. Declining soybean meal consumption marginally reduced EU meal imports to 19.8 million tons in 1999/2000. Similarly in Japan, higher imports of rapeseed weakened domestic crush margins for soybeans. Japan’s 1999/2000 soybean imports edged up to 4.85 million tons, from 4.81 million in 1998/99. In addition, soybean import demand by the countries most impacted by the Asian financial crisis in 1998 continued to recover. A worldwide boom in rapeseed planting lifted global production to 42.6 million tons in 1999, from 35.9 million in 1998. While a 7- percent expansion in harvested area accounted for much of the trend, excellent growing conditions worldwide also swelled yields. Global rapeseed exports soared by 21 percent in 1999/2000 to 11.0 million metric tons. Despite the robust 16-percent growth in world consumption, rapeseed use still trailed output gains. Consequently, global rapeseed stocks doubled in 1999/2000 and prices plunged. Rapeseed harvested area in the European Union reached a record high 3.6 million hectares. However, part of the increase was from a doubling of industrial rapeseed planted on set aside land and the remainder would displace sunflower and soybean area. Therefore, the rapeseed expansion did not push total EU oilseeds area above the limit agreed to in the 1992 U.S.-EU Blair House Agreement. The EU rapeseed harvest produced a bumper crop of 11.3 million tons. Rapeseed production in Germany and France was higher at 4.2 million and 4.4 million tons. Together the two countries typically account for three-fourths of EU rapeseed output. Farmers in Eastern Europe also produced excellent rapeseed harvests, including Poland, the Czech Republic, and Hungary. Newly added crush capacity absorbed Poland’s output gains, while a large Czech surplus led the Government to reduce restrictions on rapeseed exports. China’s rapeseed production bounced back from disappointing 1998 yields to a 1999 bumper harvest of 10.1 million tons. Despite greater domestic supplies, China’s rapeseed imports from Canada, Europe, and Australia were very brisk. Greater requirements for vegetable oil than protein meal encouraged greater rapeseed use at the expense of soybeans. For 20 years, China has been the world’s leading rapeseed crusher, based on its large domestic output. But in 1999/2000, China (which imported virtually no rapeseed just 2 years earlier) surpassed Japan as the world’s largest importer. China’s rapeseed imports swelled from 2.15 million tons in 1998/99 to 3.7 million, accounting for three- fourths of the 1.85-million-ton expansion in world trade. Combining the record domestic harvest with massive rapeseed imports, China even became an exporter of surplus rapeseed meal production to countries like South Korea. In India, rapeseed is the largest source of domestically produced vegetable oil. India uses its own entire domestic crop and does not import any seed for crushing. Persistent drought and poor irrigation supplies in northwestern India again curtailed the country’s 1999/2000 rapeseed harvest. But yields were still somewhat better than 1998/99, raising 1999/2000 production to 5.5 million tons, from 4.9 million. Canada’s seeding intentions survey indicated that farmers intended to increase canola planting 3 percent in 1999. Actual canola plantings might have been even larger, but extreme wetness stalled spring seeding throughout much of the Canadian prairies. Farmers in southern Saskatchewan could not finish before the mid- June cutoff date, which was compensated by a large acreage seeded with good short season varieties. Warm summer temperatures rapidly matured the crop and later than normal freezing weather allowed most of the crop to reach maturity. Thus, Canada harvested a record canola crop of 8.8 million tons in 1999. Robust imports by China negligibly raised Canadian exports to 3.9 million tons in 1999/2000. Yet, stronger competition from European and Australian producers and a shortage of railcars for canola slowed Canadian exports. Some foreign buyers were also deterred by an abnormally high content of green seeds in Canada’s 1999 crop. The accompanying presence of chlorophyll lessens the oil’s stability and raises the cost of removing it through the refining process. The supply tightness experienced in 1998/99 dramatically changed as Canadian 1999/2000 carryout stocks rose sharply to 2.1 million tons. Canadian crush margins were poor, and domestic seed crushing declined marginally to 3.0 million tons. The boom that made Australia the world’s sixth-largest rapeseed producing country continued in 1999. While canola prices fell, Australian grain prices were still much lower. Canola has also been recognized for its positive effects in rotation with grains in Australia. The country’s harvested area rose to 1.75 million hectares, more than double the amount just 2 years earlier. With normal yields, 1999/2000 canola production reached 2.35 million tons, from 1.8 million in 1998/99. Australia’s share of world rapeseed trade has expanded to about one-fifth, as it has benefitted from its proximity to the major import markets (Japan and China). In addition, Australian exports to the EU have expanded, as imports of genetically modified rapeseed from Canada have halted. Australian producers do not yet grow any genetically engineered rapeseed. Australian canola exports rose from 1.4 million tons to 1.9 million in 1999/2000. Unlike rapeseed, which was mostly planted in late 1998, the growing world glut of oilseeds generally discouraged spring 1999 seeding of sunflowerseed. Area declines in Argentina, South Africa, EU, and the United States curtailed growth in global sunflowerseed output to just 0.8 percent in 1999/2000 to 26.4 million tons. Slender European crush margins and restrictions on Ukrainian and Russian sunflowerseed exports contributed to a 15- percent decline in world trade to 3.7 million tons. A 34-percent cut in the EU oilseed payment to Italian farmers contributed to an 8-percent reduction in 1999 EU sunflowerseed area. In addition, a severe drought in Spain slashed sunflower planting and yields. The EU sunflowerseed harvest dropped to 3.2 million tons, one of the smallest in 15 years. Despite this, weak crush margins cut EU sunflowerseed imports by 19 percent to 2.6 million tons. Eastern Europe was the exception to cutbacks in sunflower planting. The sharp 1998 devaluation of the ruble made Russian sunflowerseed production very attractive in 1999. Despite late frosts in early May 1999 that forced some replanting, Russian sunflowerseed area surged 35 percent from the 1998 peak to 5.5 million hectares. Soil moisture conditions were better than in 1998. Russia’s 1999 sunflowerseed harvest reached a record 4.2 million tons, against 3.0 million in 1998. The accumulation of supplies significantly aided domestic sunflowerseed crushers. The bumper crop also buoyed Russian exports at 0.8 million tons in spite of the imposition of a 10-percent export tax. Russian sunflowerseed oil imports dropped sharply on account of the increased domestic production of oil. Romania and Hungary also experienced sharp increases in 1999 sunflower area and production based on decreased wheat and corn planting. Similarly, Ukraine’s expansion of sunflower area by 16 percent boosted 1999 sunflowerseed output from 2.3 million tons to 2.7 million. Although minimal input applications were seen from a lack of farm credit, somewhat better weather also aided a slight improvement in Ukrainian sunflowerseed yields. In the mid-1990's, the Government of Ukraine liberalized marketing of sunflowerseed. As a result, Ukrainian sunflowerseed exports surged in the following years. However, domestic crushing sagged, as the processors were unable to compete effectively against their more efficient Western European counterparts. To address the greatly under-utilized capacity and a shortage of Government revenues, officials advocated a 30-percent export tax on sunflowerseeds. The International Monetary Fund (IMF), which approved a $2.6-billion loan to Ukraine in 1998 to cover foreign debt payments, opposed the tax as detrimental to earning much- needed foreign exchange from one of the country’s largest export products. Nevertheless, effective October 2, the Government enacted a new 23-percent tax on sunflowerseed exports. The IMF subsequently withheld disbursement of loans until the export tax is abolished. Ukrainian sunflowerseed exports subsequently plunged by two-thirds to 0.3 million tons. Conversely, 1999/2000 domestic crush soared 65 percent to 2.1 million tons. Ukrainian exports of sunflowerseed oil (a large proportion of which go to Russia) expanded 71 percent to 0.35 million tons. Much weaker vegetable oil prices scaled back 1999 Argentine sunflower plantings by 11 percent to 3.5 million hectares. A wet harvest also trimmed yields, reducing sunflowerseed output 1 million tons to 6.1 million. Despite the reduction in EU output and higher Ukrainian export duties, Argentina’s 1999/2000 sunflowerseed exports fell 40 percent because of the smaller crop and weaker foreign import demand. Global cottonseed production rose 2 percent in 1999/2000 to 33.1 million tons. Excellent yields in Pakistan were responsible for much of the gain in world cottonseed output. In India, a lower area and inopportune dryness limited 1999 cottonseed output growth to 5.4 million tons. China’s 1999 cotton area fell to an historic low, cutting cottonseed output 15 percent to 6.9 million tons. Thus, global cottonseed supplies remain relatively tight, which boosted world cottonseed exports 9 percent to 1 million tons. World peanut production dropped 0.7 million tons in 1999/2000 to 29.2 million. Dry weather in the major western and southern producing regions of India sharply reduced the 1999 peanut harvest. The Indian crop fell from 7.5 million tons to 5.5 million and was the smallest since 1985. However, large peanut plantings pushed China’s output to a record 12.6 million tons. Also, Argentina’s 2000 peanut production recovered from the weather conditions that marred the poor 1999 harvest. World peanut trade expanded from 1.3 million tons in 1998/99 to 1.5 million. Most of the export gains accrued to the United States, which benefitted the most from a shortage of high quality peanuts between the two Argentine harvests. World Vegetable Oil Situation Global vegetable oil supplies for 1999/2000 were up 2 percent to a record 92.2 million tons. The increase in world soybean oil production was small (from 24.7 million tons to 24.8 million), as were gains in sunflowerseed oil and cottonseed oil. By comparision, the output gains for palm oil (from 19.2 million tons to 21.1 million) and rapeseed oil (from 12.1 million to 13.8 million tons) were much larger. However, most of the increase in rapeseed oil output was consumed internally by the producing countries, causing little change in world trade. Large global supplies of competing vegetable oils kept pressure on international soybean oil trade. World soybean oil exports declined 9 percent to 7.4 million tons. Consequently, carryover stocks of soybean oil accumulated, primarily in the United States. Brazil’s soybean oil exports weakened to 1.15 million tons from 1.4 million in 1998/99. However, Brazil’s domestic consumption was up 5 percent from 1998/99 to 3.0 million tons. Similarly, Argentine soybean oil exports slipped 0.1 million tons in 1999/2000 to 3.0 million. Smaller world supplies of soybean oil were offset by the robust productivity growth of palm oil producers. Based on favorable rainfall in the preceding 12 months and greater tree maturity, 1999/2000 world palm oil production rose from 19.2 million tons to 21.1 million. Malaysian production grew from 9.8 million tons to 10.5 million. Expanded area and higher yields also hiked Indonesian palm oil output from 5.8 million to 6.5 million tons in 1999/2000. The abundance of palm oil supplies recaptured markets throughout Asia and the Middle East that were lost during shortfalls in the previous 2 years. Malaysian palm oil exports increased 9 percent to 8.8 million tons. Likewise, ample domestic supplies, liberalized trade, and a weaker foreign exchange rate lifted Indonesian exports from 3.1 million to 3.5 million tons. In 1999/2000, China surpassed the United States as the world’s second-largest vegetable oil producer, behind Malaysia. Considering its huge domestic soybean and rapeseed oils output, China did not need as many imports of vegetable oil. China reduced 1999/2000 imports of soybean oil (to 560,000 tons from 950,000 in 1998/99) and palm oil (to 1.2 million tons from 1.275 million in 1998/99). The partial withdrawal of China from the world vegetable oil market shifted the focus onto India, which has become the world’s largest importer. India’s Government remains under pressure to support product prices for domestic oilseed farmers and processors. In late December, the Indian Government announced an increase in the effective import tariff on refined edible oils, from 16.5 percent to 27.5 percent. Tariffs on crude oil imports were left unchanged at 16.5 percent, although this rate was restricted to processors and excluded traders. Import tariffs on oilseeds (which were already a prohibitive 44 percent) were unchanged. In June, the Government again raised the effective duty on refined oils to 44 percent. However, the higher Indian tariffs did little to stem the tide of vegetable oil imports. The problems of the Indian oilseed sector, respecting adverse weather, poor efficiency, low farm returns, and no oilseed imports make it hard to ensure adequate domestic oil output. Traders anticipated the duty hikes and sold stocks that had been imported prior to the increases. At the same time, the Malaysian 1999/2000 average palm olein price fell 35 percent below the 1998/99 average to $314 per metric ton. Each increase in import tariffs just forced exporters to further cut prices to relieve the glut. Given a strong consumption response (and some stock building) to the low world prices, protecting domestic processing industries became very difficult, and large Indian oil imports were unavoidable. The higher tariff differential for palm olein succeeded in substituting more crude oil imports for refined oils, but total imports were only modestly curtailed. Lower refined oil imports reduced Indian 1999/2000 soybean oil imports from 830,000 to 770,000 tons. Palm oil imports (crude and refined) increased minimally from 2.9 million tons in 1998/99 to 3.1 million. Indian sunflower oil imports modestly increased from 550,000 to 570,000 tons. One successful intention of the changes was to encourage greater use of oil refining capacity in India. India’s higher import duties on refined palm oil particularly affected Malaysian exports. Malaysia sends about one-fourth of its palm oil exports to India, of which nearly all was formerly palm olein. After India imposed higher import duties on palm olein, Indonesia (which exports mostly crude palm oil) gained an advantage in the Indian market over Malaysian olein exporters. For years, Malaysia has had a differential export tax between exports of refined and crude palm oil to help develop its own refining and oleochemical industries. The Malaysian Government responded by extending exemptions of its export tax on crude palm oil to help keep its exports competitive and minimize the accumulation of domestic stocks. Malaysia’s 1999/2000 exports surged 9 percent to 8.8 million tons, but larger gains in output swelled domestic stocks to a record high 1.3 million tons. Indonesia countered by reducing its own export taxes on crude palm oil to 5 percent and to 2 percent for refined oils. Aside from India, many other countries imported record volumes of palm oil and soybean oil. Throughout the Middle East, Africa, and Latin America, bargain prices encouraged vegetable oil consumption in these price-sensitive markets. The tightness of global coconut oil supplies that prevailed in 1998/99 eased somewhat in 1999/2000. Although world coconut oil production increased from 2.7 million to 3.1 million tons in 1999/2000, this was still below yields of the mid-1990’s. Virtually all of the increase was due to improved output in the Philippines, the world’s largest producer. Coconut oil prices in Western Europe declined from an average $748 per ton in 1998/99 to $557. Consequently, world coconut oil exports rebounded from 1.4 million tons in 1998/99 to 1.7 million. World ending stocks remain relatively tight, however. Situation for Other U.S. Oil Crops Cottonseed A Southern drought caused acreage abandonment as well as below average cotton yields in 1999, reducing U.S. cottonseed production to 6.4 million short tons. Despite this setback, a rebound in acreage and yields from the large 1998 deficit somewhat eased tight supplies of cottonseed. Imports (mostly from Australia) extended 1998/99 supplies by soaring to 207,000 tons and again rose in 1999/2000 to 308,000 tons. U.S. cottonseed exports rebounded to 198,000 tons in 1999/2000, with about 90 percent shipped to Mexico. Continuing a trend that started in 1998/99, cottonseed feeding (at 3.5 million tons) exceeded crushing demand. Greater cottonseed supplies also permitted 1999/2000 crushing to recover to 3.1 million tons from the previous year’s slide of 2.7 million. Yet, chronically low rates of return forced permanent closures of several U.S. cottonseed processing plants in 2000. Lower cottonseed oil prices benefited domestic oil demand, which increased to 837 million pounds versus 772 million in 1998/99. Also, cottonseed oil exports recovered somewhat, from 111 million pounds to 141 million. However, oil exports in the last 2 years have been well below annual volumes shipped over the last decade because shipments to Central American countries have sharply declined. Cottonseed prices eased considerably from the 1998/99 average of $129 per ton to $89. Oversupplies of competing vegetable oil were largely responsible, which depressed season average cottonseed oil prices from 27.3 cents in 1998/99 to 21.5 cents per pound. On the other hand, higher prices for soybean meal supported 1999/2000 cottonseed meal prices at $127 per ton compared with $110 the preceding year. Peanuts The 1999 national peanut poundage quota was raised 1 percent to 1.18 million short tons (2,360 million pounds). The 1996 farm legislation froze the national average loan rate for quota peanuts at 30.5 cents per pound. The loan rate for additional (over-quota) peanuts was also unchanged at 6.6 cents per pound. While U.S. peanut planting was up 1 percent in 1999 to 1.535 million acres, a hot and dry summer in the Southeast increased abandonment and cut harvested acreage 2 percent from 1998. Hurricanes in September and October also dumped excessive rain on North Carolina and Virginia peanuts, reducing output in those States by 18 percent. In the Southwest, although poor rainfall caused greater abandonment of dryland acreage, the average yield on irrigated acreage was quite good. U.S. peanut production declined 3 percent in 1999 to 3,829 million pounds. Domestic food use of peanuts improved to 2,233 million pounds, up 80 million from 1998/99. Consumption of peanuts for peanut butter (the major use in the United States) significantly accelerated in 1999/2000, rising nearly 4 percent to 772 million pounds (in- shell basis). Consumption for snack peanuts rose 13 percent to 394 million pounds, but candy use fell 7 percent to 355 million pounds. U.S. peanut imports rose from 155 million pounds to 178 million. But a shortage of Argentine production helped U.S. peanut exports expand to 727 million pounds (farmer stock equivalent) in 1999/2000, up from 562 million a year earlier. About 60 percent of U.S. peanut exports go to West Europe, a premium quality market. A larger quantity of excess quota peanuts also swelled domestic crushing from 460 million pounds to 713 million. Even with robust growth in exports and domestic use, ending stocks of peanuts fell only modestly to 1,233 million pounds from the very large 1998/99 carryover of 1,392 million. The 1999/2000 national average farm price for peanuts was 25.4 cents per pound, down from 28.4 cents the previous season. Lower prices and a smaller harvest reduced the value of the 1999 peanut crop to $973 million. With a large expansion in crush, U.S. peanut oil production increased 58 percent to 229 million pounds. The abundance of domestic peanut oil supplies cut back U.S. imports to a more typical level of 12 million pounds. Similarly, peanut meal output jumped to 146,000 short tons. Peanut oil prices slumped from 40.6 cents per pound to 35.4 cents in 1999/2000, although peanut meal prices rose $5 to $108 per ton. Sunflowerseed As with soybeans, a maximum marketing loan for minor oilseeds supported increases in 1999 acreage. Sunflower producers intended to plant 3.96 million acres in 1999, up from 3.55 million in 1998. However, persistent rains during May and June in North and South Dakota and Minnesota held down the acreage that farmers could plant. Instead of an 11-percent intended increase in plantings, U.S. sunflower acreage changed little. Expansion was evident in the newer areas of Colorado, Kansas, Nebraska, and Texas, but plantings in North Dakota (the largest producing State) dropped by 290,000 acres. The geographic shifts in turn decreased oil-type sunflowerseed relative to non-oil type production, which moderated the national average yield. While weather conditions improved in the summer, excess moisture problems returned later in the fall. Wet weather during flowering promoted a head rot disease (sclerotinia). This disease was reported present in 80 percent of North Dakota sunflower fields, which damaged seed yields and quality. Late frosts enabled sunflowers to reach maturity, but seed weights were significantly diminished. The national average yield fell to 1,262 pounds per acre, the poorest since 1995. U.S. sunflowerseed production, at 4,342 million pounds, dropped 18 percent from the bumper 1998 harvest. With tighter seed supplies and very weak oil prices, sunflowerseed crushing in 1999/2000 declined slightly to 2,511 million pounds. A sharp drop in sunflowerseed oil exports to India, Colombia, and Turkey restricted domestic processing. Exports of sunflowerseed oil for 1999/2000 fell to 630 million pounds. Despite solid demand for confection-type sunflowerseed, oil-type seed exports to the Netherlands, France, and Spain fell sharply. Total sunflowerseed exports declined to 438 million pounds in 1999/2000. U.S. year ending seed stocks declined only slightly, to 491 million pounds, from the very large 1998/99 carryover of 508 million. U.S. farm prices for sunflowerseed averaged $7.53 per cwt, well below the $9.30 marketing loan rate and the lowest since 1986/87. Farmers received $120 million of loan deficiency payments on 80 percent of 1999 sunflowerseed production. The farm value of the harvest was $350 million. Other Minor Oilseeds Despite farmers’ intentions to expand U.S. canola acreage, a wet spring caused plantings to drop 39,000 acres in 1999 to 1.076 million. In addition, below-average yields further limited production to 1,364 million pounds, which was down 12 percent from 1998. Weak processing margins limited the increase in domestic canola seed crushing from the 1,533 million pounds in 1998/99 to 1,588 million. Consequently, 1999/2000 canola seed imports declined 150 million pounds to 534 million. A decline in U.S. canola exports to 299 million pounds helped offset the reduction in domestic supplies. Low domestic vegetable oil prices slowed the increase of U.S. canola oil and canola meal imports to 1.1 billion pounds and 1.2 million tons. U.S. flaxseed farmers intended to increase 1999 planted area by 55 percent. However, because of excessive spring wetness in the Northern Plains, farmers only managed to seed a more modest 15- percent increase to 387,000 acres. Ample moisture helped yields to top the 1998 record. The 1999 U.S. flaxseed harvest of 7.9 million bushels was the largest since 1986. The abundance of domestic production moderated imports from Canada to 6.6 million bushels. Crushing demand expanded 8 percent to 11.5 million bushels. Relatively liberal flaxseed supplies cut the U.S. season-average farm price from $5.05 to $3.79 per bushel. Thus, with a loan rate of $5.21 per bushel, 1999/2000 flaxseed loan deficiency payments totaled $9.3 million. U.S. safflower acreage also fell in 1999 to 275,000 acres. But slightly better yields than the poor 1998 conditions limited the reduction in 1999 safflowerseed production (of 405 million pounds) to 2 percent. Weaker 1999/2000 export shipments of safflowerseed oil to Japan curbed domestic crushing and expanded U.S. oil consumption. Other Fats and Oils Highlights Corn Oil As world vegetable oil supplies multiplied in 1999/2000, price incentives for most oil producers weakened. Domestic soybean processors cut back production early in 2000, which helped stabilize soybean oil prices. However, corn oil production steadily expanded throughout the year. Corn oil output is less sensitive to a drop in vegetable oil prices because of its secondary importance in the profitability of corn wet milling. A sharp increase in gasoline prices in 2000 spurred the demand for ethanol, which is a fuel additive. Consequently, supplies of corn germ (which is a byproduct of ethanol production and the raw material for extraction of corn oil) expanded. U.S. corn oil production rose from 2,374 million pounds in 1998/99 to 2,554 million in 1999/2000. Corn oil prices continued to slide, from 22 cents per pound in October 1999 to 10.5 cents in October 2000. Thus, the typical premium for corn oil against soybean oil declined in 1999/2000, and by summer a 1-cent-per-pound discount had emerged. Corn oil prices averaged 17.8 cents per pound in 1999/2000, the lowest since 1969. Part of the growth in U.S. corn oil output was consumed domestically, with domestic disappearance rising from 1,394 million pounds to 1,473 million. Yet, ending stocks increased even more, from 135 million pounds in 1998/99 to 264 million. U.S. corn oil exports slipped to 970 million pounds, from 989 million in 1998/99. Lower exports to South Korea, Venezuela, Spain, and Jordan outweighed an expansion in trade with Italy. Imported Oils As world palm oil prices plunged in 1999 and 2000, U.S. palm oil imports surged from 284 million pounds in 1998/99 to 345 million in 1999/2000. The United States imports very little crude palm oil; primarily obtaining refined palm oil from Malaysia and Indonesia. The improvement in world coconut oil supplies raised American imports in 1999/2000, which had fallen to the lowest volume in a decade in 1998/99. Foreign shipments to the United States totaled 926 million pounds, up from 791 million in 1998/99. However, the pace of coconut oil imports did not accelerate until May 2000. Stocks tightened in the first half of the marketing year, which curbed 1999/2000 domestic disappearance of coconut oil 9 percent from the previous year’s 1,021 million pounds. The recovery in coconut oil imports slowed U.S. palm kernel oil imports (an alternative lauric oil) from 401 million to 385 million pounds. Global olive oil production slumped to 2.2 million tons in 1999/2000 from 2.5 million in 1998/99. Ample carryover stocks cushioned the impact of lower output, so that global consumption and trade continued to creep higher. The United States, which is one of the world’s largest olive oil importers, raised imports from 375 million pounds to 417 million in 1999/2000. Italy shipped about 75 percent of U.S. olive oil imports, followed by Spain with about 15 percent of the total. Animal Fats Poor returns for hog production in 1998 and 1999 caused a contraction in size of the U.S. herd. Hog slaughtering slowed by 4 percent in 1999/2000, which was offset to a degree by higher animal weights. Consequently, domestic lard output slipped to 1,069 million pounds in 1999/2000 from 1,106 million the previous marketing year. The season average price for lard fell to 13.6 cents per pound, the lowest since 1992/93. By summer of 2000, lard prices had dropped toward 10 cents per pound, a level not seen since 1972. Cheap lard prices helped U.S. exports expand from 140 million pounds in 1998/99 to 189 million in 1999/2000. The principal foreign buyers were Mexico, Hong Kong, and Canada, which accounted for over 80 percent of U.S. lard exports. However, pressure from ample supplies of tallow and other vegetable oils pushed domestic lard disappearance lower, to 886 million pounds in 1999/2000 from 987 million in 1998/99. Rising cattle slaughter rates and heavier animal weights edged 1999/2000 U.S. tallow production up to 1,678 million pounds from 1,677 million the year before. The brisk pace of domestic disappearance in recent years continued into 1999/2000, rising 107 million pounds to 1,467 million. Alternatively, U.S. exports of edible tallow weakened from 322 million to 224 million pounds in 1999/2000, of which Mexico, Canada, and South Korea comprised 82 percent of the total. Despite strong livestock and pet food demand, the domestic surplus weighed down the 1999/2000 average tallow price to 13.2 cents per pound, from 15.1 cents the preceding year. End Uses of Fats and Oils U.S. output of salad and cooking oils rose a modest 3 percent in 1999 to 7,701 million pounds. Exports fell back from the 1998 peak of 834 million pounds to 649 million. Thus, nearly all of the additional salad and cooking oil supplies in 1999 were consumed domestically, which surged 7 percent to 8,030 million pounds. On a per capita basis, U.S. consumption for these uses increased from 27.8 pounds in 1998 to an all-time high of 29.4 pounds. Production of baking and frying fats increased 4 percent in 1999 to 5,945 million pounds. U.S. per capita consumption of shortenings increased to 21.6 pounds. According to Bureau of Census data, U.S. consumption of fats and oils for margarine declined again in 1999. Margarine production declined 2 percent to 2,274 million pounds, while consumption fell an equivalent amount. Consumption of soybean oil, which accounts for nearly 95 percent of the fats and oils used for margarine, declined accordingly. Consumption of oils for inedible uses increased 2 percent in 1999 to 6,733 million pounds. Animal feeds accounted for virtually the entire rise in 1999 nonfood consumption of fats and oils, offsetting a 7-percent decline in fatty acids consumption. The consumption of soybean oil in inedible products expanded 9 percent in 1999 to 588 million pounds. Sharply higher import prices in 1999 curtailed U.S. inedible consumption of coconut oil (mainly for fatty acids and soap) 18 percent to 368 million pounds. Consumption of linseed oil increased 3 percent to 82 million pounds. AN ASSESSMENT OF A ‘FUTURES METHOD’ MODEL FOR FORECASTING THE SEASON-AVERAGE FARM PRICE FOR SOYBEANS Erik Dohlman, Linwood Hoffman, Randall Schnepf, and Mark Ash U.S. Department of Agriculture, Economic Research Service 1/---- ------ 1/ Agricultural economists, Market and Trade Economics Division, Economic Research Service ------- Abstract: Futures prices are used to develop historical forecasts of U.S. soybean season-average farm prices (SAFP) during crop-years 1981/82 to 1999/2000. The method for forecasting soybean SAFP is outlined and the accuracy of these forecasts is assessed by comparing them with actual season- average farm prices during those years. The accuracy of ‘ futures method’ forecasts is also compared with those published monthly by USDA in the World Agricultural Supply and Demand Estimates (WASDE) reports. Findings suggest that both the futures method and WASDE forecasts are generally accurate and comparable, but the futures method provides more accurate forecasts by some criteria. Keywords: Soybeans, price forecasts, futures prices, futures- method forecast, season-average farm price Introduction The U.S. Department of Agriculture (USDA), in its efforts to provide reliable market information on agricultural products, develops short-run forecasts of production, use, and trade for numerous agricultural commodities, including soybeans. Based on expected supply and demand conditions, USDA also issues forecasts of annual commodity prices on a monthly basis, and these projections are used as an important planning tool by both the private and public sectors. For producers, forecasts of the season-average farm price (SAFP) can affect marketing decisions. Furthermore, producers and users of agricultural commodities rely on forecasts to manage income and price risk. For policymakers, accurate forecasts can be important for budgetary purposes related to farm programs. Given the importance of price forecasts to market participants, the objectives of this study are twofold. First, we construct an alternative set of monthly soybean season-average farm price forecasts using the ‘futures method’ model previously developed by Hoffman and Davison (1992), and assess the accuracy of these forecasts by comparing them with actual season-average farm prices during crop years 1981/82 to 1998/99. Second, we compare the accuracy of futures method forecasts to those published monthly by USDA in the World Agricultural Supply and Demand Estimates (WASDE) report. Our aim is to determine whether the futures method represents a generally reliable approach to forecasting commodity prices, as well as to provide an overall assessment of WASDE and futures method forecast accuracy. In addition to our main objectives, we also explore whether the accuracy of futures forecasts improves when futures markets gain access to new information from the most recent WASDE report. That is, are forecasts based on futures prices immediately following the release of WASDE more accurate than those made just prior to the WASDE release. Intuitively, this makes sense. WASDE SAFP projections represent the sum of all publicly available market- related information, but some of this information, such as USDA’s National Agricultural Statistics Service (NASS) survey-based data on crop yields, are not made available to the public until the WASDE’s release. Although market participants may anticipate this information, futures forecasts following the release of the WASDE should represent the most up-to-date composite of public and privately held information. To test this conjecture, we develop two separate forecasts of SAFP using the futures method – one based on futures price data available prior to the release of WASDE, and the other based on futures price data immediately following the release of WASDE. The following section describes the method used to develop monthly forecasts of annual season-average soybean prices with futures, and illustrates the method with a November 1999 forecast for the 1999/2000 crop year. We then compare the historical accuracy of the futures forecasts with WASDE forecasts by calculating the mean absolute percentage error (MAPE) of the forecasts during crop years 1981/82 to 1998/99. Next, the average (1981/82 to 1998/99) absolute percentage error for each forecast month is examined separately to see if there is any pattern to differences between the alternative forecasts over the course of the crop year. We conclude with a brief summary. Overview of Futures Forecasting Method Using the futures method, forecasts of monthly average prices received by U.S. farmers are made for each month of the crop year starting with September. Price forecasts for each month of the crop year are initially based on the current month’s futures price for the nearest contract maturing after the month being forecast (referred to as the ‘nearby futures contract’ ). Most market participants understand that the futures market is a composite indicator of anticipated supplies and demands and that current futures prices therefore provide important information about cash prices on future dates. However, participants also need to be able to forecast a price at the location and time when they plan to buy or sell. Thus, they need to predict the ‘basis,’ the difference between the futures price and the local price. The futures method employed here uses an historical monthly average basis (historical monthly farm price received minus historical monthly average futures price for the nearby contract) that is subtracted from the current nearby futures prices to yield a monthly U.S. average farm price forecast for each month of the crop year. The 12 monthly price forecasts are then multiplied by their 5-year historic share of annual marketings and summed to produce a weighted season-average farm price forecast. As estimated monthly farm prices become available, the predicted season-average farm price becomes a composite of actual and forecasted prices. Basis The difference between a farm (henceforth ‘cash’ ) price received at a specific location and the price of a particular futures contract is known as the basis. The basis tends to be more stable or predictable than either the farm price or futures price. Factors that can affect the basis include local supply and demand conditions for the commodity and its substitutes, handling costs, transportation and storage costs, and market expectations. The basis used in this analysis is a composite of these factors and represents an average of U.S. conditions. The basis in this study is defined as the difference between the monthly U.S. average cash price received by producers and the monthly average settlement price for the nearby futures contract. For example, the September basis is the difference between the September average cash price received by producers less September’s average settlement price of the November futures contract. A 5-year moving average of these bases, used to eliminate distortions that may occur in any given year, is updated at the end of each crop year. Thus, data for the 1976 through 1980 crop years establish the historical basis used to develop the 1981 crop year futures forecast. Data Historical daily soybean futures settlement prices for crop years 1976 to 1999 are obtained from TechTools data service. Historical cash prices were acquired from USDA’s (NASS) Agricultural Prices, and weights for monthly marketings were obtained from USDA’s (NASS) December issues of Crop Production (prior to 1998) and November issues of Agricultural Prices (1998 to present). Procedure and Illustration of Futures Method Table 1 illustrates the method used to forecast the 1999/2000 crop year season-average soybean price in November 1999. Although the futures method forecast for 1999/2000 has been updated through August 2000, we present the November 1999 forecast to more clearly illustrate that SAFP forecasts are, in general, a composite of actual and forecasted monthly prices. It should be noted that our assessment of the accuracy of the futures method for crop years 1981/82 to 1998/99 is based on all 12 monthly forecasts for each year. Recall that we use the futures method to produce two alternative forecasts of the SAFP – one using a 2-day average futures settlement price available just prior to the release of that month’s WASDE, and one using a 2-day average settlement price following the WASDE release. For simplicity of presentation, only the first (pre-WASDE) forecasts are shown in table 1. Seven steps are involved in the forecast process, illustrated here with the November 1999 forecast of the 1999/2000 crop year SAFP: Futures settlement prices are gathered for the contracts that will mature during the forthcoming year (line 1). When pre-WASDE settlement prices are used, the 2-day average futures price for the January, March, May, July, and September (2000) contracts available on November 8th and 9th were selected (WASDE was released on November 10). Estimates of actual monthly prices received are available from NASS and used for September and October 1999. The October 1999 price represents a mid-month estimate published in that month’s issue of Agricultural Prices (the price is updated the following month). The November 1999 contract is not used for reasons discussed below. The monthly futures prices are based on the settlement prices of the nearby contracts. For example, the futures prices for November and December represent the November (8th and 9th) average settlement price of the nearby January contract. The futures prices for January and February are based on the November settlement prices for the nearby contract for those months (March). During months in which a futures contract matures, the next contract month is used because futures contracts are affected by a decline in liquidity during the month of maturity. Although the September 2000 futures contract falls outside of the current crop year, this contract is used to establish the monthly futures price for August 2000. A 5-year moving average of the basis (cash prices minus the monthly average settlement price for the nearby futures contract) for each month is entered (on line 3). A forecast of the monthly average farm price (line 4) is computed by adding the basis (line 3) to the monthly futures prices (line 2), except when NASS monthly or mid-month price estimates are known. The NASS monthly average farm price is entered on line 5 as it becomes available. In this example, the September price is for the entire month and the October price is a mid-month estimate. In December, the estimate for October would be updated and a mid- month estimate for November would be included. The NASS price estimates and forecast farm prices are spliced together in line 6. The November 1999 forecast of SAFP for crop year 1999/2000 will be based on actual price data for September and October, and forecasts for the remaining 10 months. A 5-year average of monthly marketing shares (in percents) by soybean producers (line 7) is used to weight the monthly farm prices (forecast or actual), yielding the final November 1999 forecast of the 1999/2000 soybean SAFP (line 8). The November 1999 forecast of the 1999/2000 SAFP based on pre- WASDE futures information was $4.64/bushel. Although the actual 1999/2000 SAFP for soybeans is not yet available, this figure compares very favorably with the most recent (August 2000) WASDE point estimate of $4.65/bushel for the current crop year. In the months following the November forecast, the (pre-WASDE) futures forecast fell to about $4.55/bushel before climbing to a peak of just over $4.80/bushel in May 2000. The futures forecast then began to converge towards the WASDE estimate in June, July, and August (fig. 1). The futures forecasts based on post-WASDE release futures data were all within about 10 cents per bushel of the pre-WASDE forecasts and the difference averaged about 4 cents/bushel. In November, the post-WASDE forecast was about 10 cents per bushel lower (at $4.54/bushel) than the pre-WASDE forecast. The difference is probably due to new information conveyed by the November WASDE report. USDA lowered its mid-point forecast of soybean SAFP by 15 cents per bushel due in part to diminished export prospects. The result was a less accurate forecast of the probable 1999/2000 soybean SAFP, but one still more accurate than the November WASDE mid-point projection of $4.85/bushel. Compared with the WASDE price estimates, the futures price forecasts ranged from as much as 20 cents a bushel above the WASDE mid-point forecast in September 1999 to 31 cents a bushel below the WASDE projection in November 1999. Since the actual season-average farm price for soybeans has not yet been established and just one year’s worth of projections are represented here, these comparisons are somewhat less meaningful than the historical analysis of forecast accuracy for the crop years 1981/82 to 1998/99 presented in the next section. Forecast accuracy of the futures method and WASDE (1981/82 to 1998/99) In this section, we examine the historical (1981/82 to 1998/99) accuracy of soybean SAFP forecasts published in USDA’s WASDE reports as well as the accuracy of the two alternative forecasts developed using the futures method. This analysis is designed to help us gauge the general accuracy of the WASDE projections, and to judge whether the futures method represents a reasonable alternative approach for developing such forecasts. Initially, forecast accuracy is assessed by calculating the mean absolute percentage error (MAPE) for each forecast (WASDE or futures) over the entire crop year. That is, for a given crop year, the MAPE gives the average percentage difference between each month’s (September through August) forecast of SAFP and the actual SAFP. We then examine the average absolute percentage error of the monthly forecasts. For instance, the average absolute percentage error for the September WASDE report is the average of the September forecast errors over the 18 years examined. It should be remembered that the WASDE and futures forecasts of SAFP are composites of projected and actual (NASS estimates of) monthly cash prices as they become known. Yearly Forecast Errors (1981/82 – 1998/99) Figure 2 and the accompanying table present the mean absolute percentage errors for the WASDE and the futures method for crop years 1981/82 to 1998/99. The MAPE is a summary of monthly errors during each crop year and therefore masks fluctuations of the errors over the course of the crop year. Nevertheless, it provides a general sense of the overall accuracy of the alternative forecasts as well as a basis for comparison between the forecast methods. Since the results for the pre-WASDE and post-WASDE futures method were similar, figure 2 compares only the pre-WASDE futures forecasts with the WASDE. The accompanying table provides the results for all three methods. The MAPE for each of the three forecasts ranged from a low of 0.56 percent for the 1985/86 post-WASDE release futures method to a high of over 7 percent for the 1987/88 WASDE projections. By the MAPE criteria, it appears that the futures method holds a slight advantage over the WASDE in forecasting soybean SAFP. The average MAPE over the eighteen observations was 2.96 percent for the WASDE, 2.45 percent for the pre-WASDE release futures method, and 2.38 percent for the post-WASDE release futures method. The WASDE projection out-performed one or both futures forecasts in 8 out of 18 years, but in the other years, the WASDE errors tended to exceed those of the futures method by a fairly large margin – particularly in 1983, 1984, and 1987. As indicated in figure 2, the SAFP forecast errors for the WASDE and futures method tend to be highly correlated, generally falling or rising from previous year’s errors in tandem. In addition, the tendency of all three forecasts was to somewhat over-estimate the soybean season-average farm price. For each method, about 55 percent of the 216 monthly forecasts over- estimated the final SAFP, but the simple mean error of all monthly forecasts was lowest for the WASDE (0.17 percent versus 0.36 percent for the pre-WASDE futures forecasts and 0.30 percent for the post-WASDE futures forecasts). Monthly Forecast Errors (September–August) Not surprisingly, the accuracy of SAFP forecasts for each method tends to improve over the course of the crop year, as actual monthly prices are incorporated into the forecasts. Interestingly, as shown in figure 3 and the accompanying table, the WASDE and futures method forecasts perform similarly during the first monthly projection (September) of the crop year SAFP. The 18-year average (of absolute) September forecast errors ranged from a low of 6.35 percent for the WASDE projection to a high of 6.85 percent for the pre-WASDE futures forecast. In the following months, particularly November through March, however, the WASDE projection errors consistently exceeded the futures forecast errors. Between November and February, the difference averaged more than 1 percentage point per month. Why the WASDE forecast errors exceed the futures forecasts during these months is difficult to determine. One suggestion is that over the time period examined (1981/82–1998/99), WASDE projections of (U.S.) domestic use tended to be under-estimated while ending stocks were over-estimated. A look at statistics on the reliability of monthly WASDE projections between November and March (1981/82 to 1998/99) confirm this impression. The expected impact would be a consistent under-estimation of the SAFP, but a closer look at monthly WASDE forecast errors does not support this conclusion. The simple average of errors for November, December, and January were positive, meaning price forecasts were slightly over-estimated during these months. In any event, this suggestion does not explain differing magnitudes of WASDE and futures method forecast errors, only a potential pattern to WASDE forecast errors (which is not apparent). Another suggestion is that the difference between WASDE and futures method forecast errors from November to February may be related to uncertainties about South American soybean production. Soybean planting in South America typically occurs in October, with harvest beginning in March. Less accurate or timely information on these crops could contribute to forecasting errors, but again, it is unclear that this would have a greater impact on WASDE forecasts than those based on the futures method. It should be pointed out that, regardless of the source of the WASDE forecast errors, the accuracy of WASDE forecasts made during November through March have improved significantly during the 1990’s, while those of the futures method have actually worsened slightly. Compared with the 1980’s (1981/82-1989/90) period, the average November-March WASDE forecast error decreased by more than 1 percentage point in the 1990’s (1990/91-1998/99), whereas futures forecast errors increased by a little more than 0.1 percentage point during the same interval. This may reflect improved information, analysis, or modeling efforts by the USDA. Summary and Conclusion The goals of this analysis were twofold: to develop and illustrate the use of the futures method model for forecasting the season-average farm price for soybeans, and to assess and compare the historical accuracy of this method with USDA’s farm price forecasts published monthly in WASDE. Our findings suggest that both the WASDE and futures method provide reasonable and generally accurate price forecasts. By the mean absolute percentage error (MAPE) criteria, the futures method slightly outperformed the WASDE projections, but a simple average of all (216) monthly forecast errors indicates that the WASDE does not overestimate the SAFP as much as the futures method forecasts. In addition, there is little to distinguish the WASDE from the futures method in terms of beginning-of-the-crop-year accuracy. The futures method is typically more accurate between November and March of the crop year, but the differences are narrowing. Finally, the MAPE of futures forecasts based on post-WASDE release futures prices are on average lower than pre-WASDE futures forecasts--indicating that information conveyed by WASDE reports improve futures method forecasts--but the difference is minor. In conclusion, the futures method of forecasting the season- average-farm-price of agricultural commodities represents a useful tool for analysts and market participants seeking a cross- check to USDA projections. Future research on the method could examine alternative methods of estimating the basis and marketing weights, such as using a 5-year moving olympic average (omitting the high and low figures) rather than a simple moving average. Improved estimates of these variables should enhance the overall accuracy of price forecasts. Another avenue would be to examine the historical accuracy of other forecasting tools that have been used to project commodity prices, such as time series autoregressive-integrated-moving-average (ARIMA) models. Using the ARIMA method, Vroomen and Douvelis (1993) developed forecasts of soybean SAFP for crop years 1989/90 to 1991/92 with results similar to WASDE and futures method forecasts, but it is unclear whether the accuracy of this method would be sustained over the longer run. References Hoffman, L, and C. Davison. ‘Forecasting U.S. Soybean Prices with Futures Prices.’ Oil Crops: Situation and Outlook Report, ERS, USDA. January, 1992. Hoffman, L. and J. Balagtas. ‘Providing Timely Farm Price Forecasts: Using Wheat Futures Prices To Forecast U.S. Wheat Prices at the Farm Level.’ The 10th Federal Forecasters Conference -- 1999: Papers and Proceedings. Bureau of Labor Statistics, Washington, DC, June 24, 1999. U.S. Department of Agriculture, National Agricultural Statistics Service. Agricultural Prices. Various issues. U.S. Department of Agriculture, National Agricultural Statistics Service. Crop Production. Various issues. U.S. Department of Agriculture, World Agricultural Outlook Board. World Agricultural Supply and Demand Estimates. Monthly issues, 1976-2000. Vroomen, H. and G. Douvelis. ‘A Short-Run Forecasting Model for Soybean Prices’ Oil Crops: Situation and Outlook Report, ERS, USDA. January 1993. END_OF_FILE