SUGAR AND SWEETENERS YEARBOOK May 24, 2000 May 2000, ERS-SSS-228 Approved by the World Agricultural Outlook Board ----------------------------------------------------------------------------- SUGAR AND SWEETENERS YEARBOOK is published by the Economic Research Service, U.S. Department of Agriculture, Washington, DC 20036-5831. This release contains only the text of the SUGAR AND SWEETENERS report -- tables and graphics are not included. Subscriptions to the printed version of the report are available from the ERS-NASS order desk. Call, toll-free, 1-800-999-6779 and ask for stock #SUB-SSS-4033, $25/2 issues. ERS-NASS accepts MasterCard and Visa. ----------------------------------------------------------------------------- Contents Summary U.S. Sugar Current Year, FY 2000 Prices and Policy Developments Forecast for FY 2001 Mexican Sugar Special Article Returns from Mexican Sugar Processing List of Tables Autofax Access to Sugar-Related Data Internet Access to Sugar-Related Data Report Coordinator, Stephen Haley (202) 694-5247 FAX (202) 694-5884 E-mail: SHALEY@ers.usda.gov Principal Contributors, Stephen Haley and Nydia Suarez Database Coordinator/Graphics & Table Design, Fannye Lockley-Jolly Layout & Text Design, Wynnice Napper Approved by the World Agricultural Outlook Board. Summary released May 18, 2000. The next Sugar and Sweetener Situation and Outlook is scheduled for release on September 21, 2000. Summaries and full text of Situation and Outlook reports may be accessed electronically via the ERS web site at www.ers.usda.gov. The Sugar and Sweetener Situation and Outlook is published two times a year and supplemented by a yearbook. To order, call 1-800-999-6779 in the United States or Canada. Other areas please call (703) 605-6220. Or write ERS-NASS, 5285 Port Royal Road, Springfield, VA 22161. Summary On May 11, 2000, Agriculture Secretary Glickman announced that the U.S. Department of Agriculture (USDA) will seek to buy U.S. sugar to reduce the cost of expected sugar program loan forfeitures, support sugar growers, and help stabilize low market sugar prices (press release #0159.00). According to the press release, at least 75 percent of the initial approximately 150,000-ton purchase will be refined sugar and may be followed by additional purchases depending on price and market conditions. USDA expects to save as much as $6 million in administrative costs that the government might otherwise incur from expected loan forfeitures later this summer. The Commodity Credit Corporation (CCC) will purchase the sugar using its authority under the cost reduction options of the Food Security Act of 1985. For now, the CCC plans to store the purchased sugar and will not sell it back into a depressed U.S. sugar market. Beet sugar production for fiscal year (FY) 2000 is estimated at 4.950 million short tons, raw value (STRV). The National Agricultural Statistics Service (NASS) estimated the sugarbeet crop at 33.319 million tons. Area harvested is estimated at 1.527 million acres, and yield is estimated at 21.8 tons per acre. Both sugarbeet acreage harvested and production represent records. Although winter weather conditions were less than ideal for storage, the beets entered storage in good condition and remained in good shape through the winter months. Cane sugar production for FY 2000 is estimated at a record 4.130 million STRV. NASS estimated the U.S. sugarcane for sugar crop at 33.736 million tons. Area harvested is estimated at 940,400 acres, and yield is estimated at 35.9 tons per acre. Both sugarcane acreage harvested and production, if realized, would represent records. Production was particularly strong in Louisiana, where increased acreage harvested and yields combined to produce a record crop. The expected level of sugar tariff-rate quota (TRQ) imports for FY 2000 is now estimated at 1.204 million STRV. As of May 1, 2000, total raw TRQ sugar amounting to 430,316 STRV had entered the U.S. customs territory, or about 34 percent of the amount projected to enter for the fiscal year. As of May 1, 2000, 34,190 STRV or about 89 percent of the remaining refined sugar TRQ, apart from the North American Free Trade Agreement (NAFTA) allocation to Mexico of 27,558 STRV, had entered. Other imports for FY 2000 are estimated at 523,000 STRV. Most of these imports enter under special programs: the Refined Sugar Re-export Program (250,000 STRV), the Sugar-Containing Products Program (125,000 STRV), and the Polyhydric Alcohol Program (15,000 STRV). Estimated sugar derived from sugar syrups entering under HTS 17029040 are at 125,000 STRV. Monthly sugar syrup imports have been steady through February and are expected to maintain the established pace. High-tier tariff imports are estimated at 8,000 STRV. Total deliveries for FY 2000 are estimated at 10.25 million STRV. Deliveries through the end of March amounted to 4.957 million STRV, about 3.7 percent higher than last year through March. Deliveries by beet processors have been particularly strong, 2.2303 million STRV through March, or about 7.2 percent higher than the same period last year. Deliveries by cane refiners, on the other hand, are near the same level as last year, 2.710 million STRV. NASS forecasts sugarbeet area planted for FY 2001 at 1.578 million acres, about 1 percent higher than FY 2000. Beet sugar production for FY 2001 is projected at 4.700 million STRV, with an aggregate U.S. sugar yield expected to be about 3.10 tons per acre. Projection updates throughout the year will continue to take into account the possible plant closings in California beet producing areas. Cane sugar production for FY 2001 is projected at 4.323 million STRV. This sum represents the aggregation of projected production in Florida (2.090 million STRV), Louisiana (1.720 million STRV), Texas (160,000 STRV), Hawaii (330,000 STRV), and Puerto Rico (22,500 STRV). The raw and refined sugar TRQ for FY 2001 has not been established. Other imports for FY 2001 are projected to total 448,000 STRV, including 175,000 STRV under the Refined Sugar Re-export Program, 125,000 under the Sugar-Containing Products Program, and 15,000 under the Polyhydric Program. Sugar syrups under HTS 17029040, and high-tier tariff sugar imports are expected to be at the same levels as in FY 2000. Deliveries for FY 2001 are projected at 10.385 million STRV. This level represents growth over the previous year of 1.3 percent, which is less than the growth of 1.8 percent for FY 2000. Deliveries to industrial users are projected up 2.0 percent, while deliveries to non-industrial users are up 0.3 percent. The USDA currently estimates Mexican sugar production for 1999/2000 at 5.070 million metric tons, raw value (MTRV). Area harvested for sugarcane are estimated at 627,000 hectares, and sugarcane production is estimated at 43.365 million metric tons. The implied sugar recovery rate from the sugarcane is 11.7 percent, which, if realized, would be a record. Sugar consumption for 1999/2000 is estimated at 4.482 million MTRV and exports are estimated at 630,000 MTRV. The U.S. Embassy's Office of Foreign Agricultural Affairs in Mexico City currently forecasts Mexican sugar production for 2000/01 at 5.090 million MTRV. Area harvested for sugarcane is projected at 630,000 hectares, and sugarcane production is projected at 44 million metric tons. Sugar consumption for 2000/01 is projected at 4.482 million MTRV, the same as for 1999/2000, and exports are projected at 700,000 MTRV. On October 1, 2000, Mexico's duty-free access to the U.S. market, as set out in the side-letter agreement to the original NAFTA, increases to the smaller of 250,000 MTRV or Mexico's "net surplus production," which accounts for Mexican consumption of high fructose corn syrup. U.S. Sugar On May 12, 2000, the U.S. Department of Agriculture (USDA) released its latest supply and use estimates for fiscal year (FY) 2000 and projections for FY 2001. Current Year, FY 2000 Production Beet sugar production for FY 2000 is estimated at 4.950 million short tons, raw value (STRV). The National Agricultural Statistics Service (NASS) estimates the sugarbeet crop at 33.319 million tons. Area harvested is estimated at 1.527 million acres, and yield is estimated at 21.8 tons per acre. Both sugarbeet acreage harvested and production represent records. Although yield is estimated at less than last year's record, generally good growing and harvesting conditions permitted a clean crop, with beets having higher sugar content than last year. Although winter weather conditions were less than ideal for storage, the beets entered storage in good condition and remained in good shape through the winter months. Beets sliced through March equaled 24,880,867 tons, a record that exceeded last year's to-date total by over 550,000 tons. The September 1999 through March 2000 cumulative extraction of sugar from sliced beets is estimated at 279 pounds per ton. This rate of sugar extraction is about 26 pounds per ton greater than last year's rate through March and about 9 pounds per ton greater than the average through March for FY 1995-1998. It is projected that the September - August extraction rate will be about 290 pounds per ton. Adding estimated sugar production from the desugaring of molasses, the total is estimated at 4.950 million tons, adjusted to match the fiscal year. Cane sugar production for FY 2000 is estimated at a record 4.130 million STRV. NASS estimates the U.S. sugarcane for sugar crop at 33.736 million tons. Area harvested is estimated at 940,400 acres, and yield is estimated at 35.9 tons per acre. Both sugarcane acreage harvested and production, if realized, would represent records. Cane sugar in Florida is estimated at 1.980 million STRV. NASS estimates sugarcane for sugar at 15.540 million tons and acreage harvested at 444,000 acres. Although area harvested is estimated at 18,000 acres more than last year, yield is estimated significantly lower at 35.0 tons per acre, compared with 40.1 tons per acre last year. Sugar yield is projected at 4.46 tons per acre, which is lower than last year (5.00 tons per acre) and the previous year (4.57 tons per acre). Sugar recovery (raw sugar production as a percentage of sugarcane for sugar), however, is calculated at 12.74 percent, a record. The harvest was aided by dry, nearly optimal weather conditions through early April. Cane sugar in Louisiana is estimated at a record 1.680 million STRV. NASS estimates sugarcane for sugar at a record 14.355 million tons and area harvested for sugar at a record 435,000 acres. Because of the large sugarcane crop, harvesting began in September, with raw sugar production estimated at 69,300 STRV, which counts towards last year's production total because the fiscal year does not begin until October. Fall harvest conditions were excellent, and the harvest season ended in early January with no exposure to freezing conditions. With another large sugarcane crop projected for next fiscal year (see below), production in September 2000 is again expected to be high. Cane sugar in Texas is estimated at 104,680 STRV. NASS estimates sugarcane for sugar at 976,000 tons and acreage harvested at 28,700 acres. The harvest ended in late February. Harvest conditions were considered good because of dry weather and the absence of freezes. NASS estimates a cane yield of 34.0 tons per acre, the highest since FY 1996. Sugar yield is estimated at 3.65 tons per acre, a record; and the recovery rate is estimated at 10.73 percent (or 215 pounds per ton), the highest since 1979. Cane sugar in Hawaii is currently estimated at 360,000 STRV. Because NASS does not estimate Hawaiian sugarcane acreage until June, the USDA estimate for FY 2000 sugar production was based on a close examination of factors that could change this year's production level from the previous year's level of 384,000 STRV. The most important factor was the closing of the Pioneer Mill on Maui. It was expected that the closure would reduce sugar production between 25,000 and 30,000 STRV. However, technical improvements occurring in other mills, especially by the Hawaiian Commercial & Sugar Company on Maui, were expected to contribute an offset to the decline in sugar production due to the Pioneer closing. While USDA is still estimating 360,000 STRV, the Hawaiian situation is being closely monitored. In November 1999, AMFAC Sugar on Kauai announced that it was closing the Kekaha Sugar Mill on the western side of the island. At the same time, the Company announced that sugarcane harvested from the western operations would be trucked to the eastern operations of its Lihue Plantation. How much of a strain this arrangement will place on production is unclear, but it may affect future production estimates. Cane sugar in Puerto Rico is currently projected at 5,000 STRV. Initial reports indicate that the harvest season ended in early May. Only one plant is operating this year because of hurricane damage to the other. Imports As explained in the January 2000 edition of the Sugar and Sweetener Situation and Outlook, the USDA on October 26, 1999, extended a waiver to C&H Sugar Company, Inc. that allowed it to import an additional 50,000 metric tons, raw value (MTRV) over and above its original re-export license of 50,000 MTRV. The waiver also extended the period over which an equivalent quantity of refined sugar had to be re-exported from the standard 90 days to 5 years. On December 29, 1999, the USDA specified that the re-export period granted in the original waiver be reduced to 180 days. To facilitate the export of the sugar, the USDA allowed the sugar to be exported as either raw or refined sugar or some combination of the two. On February 29, 2000, the USDA announced a 30-day extension of the re-export period up to 210 days. The extension is meant to facilitate the removal, during the fiscal year, of the 100,000 MTRV imported as a result of the October 26, 1999, waiver. Importantly, the USDA also announced that it would accept from the company Certificates for Quota Eligibility (CQEs) issued by any certified authority under the FY 2000 raw sugar tariff-rate quota (TRQ). The CQEs are to be accepted in lieu of exports by the company, provided they are not associated with the importation of raw sugar under the TRQ. On November 2, 1999, the raw sugar TRQ was set at 1.501 million STRV, but only 1.251 million STRV were made available to the office of the U.S. Trade Representative (USTR) for allocation, with the remainder held in reserve, to be made available to the USTR for allocation at the discretion of the USDA. As of May 1, 2000, total raw TRQ sugar amounting to 430,316 STRV had entered, or about 34 percent of the amount projected to enter for the fiscal year. The refined sugar TRQ, announced on October 1, 1999, was set at 66,139 STRV. This amount included 27,558 STRV allocated to Mexico under the side-letter agreement to the North American Free Trade Agreement (NAFTA). This sugar can be shipped as either raw or refined. The specialty sugar allocation, which is a subset of the refined sugar TRQ, was set at 16,155 STRV. As of May 1, 2000, the specialty sugar TRQ has been completely filled. Apart from the NAFTA allocation, 18,035 STRV or about 80 percent of the remaining refined sugar TRQ had entered as of May 1, 2000. The expected level of sugar TRQ imports for FY 2000 is now projected at 1.204 million STRV. This level reflects two adjustments. First, the raw sugar TRQ shortfall is expected to be about 65,000 STRV. Second, CQEs from Jamaica, and St. Kitts and Nevis, totaling 20,770 STRV, were accepted by the USDA in lieu of imports under the raw sugar TRQ, as explained above. Other imports for FY 2000 are estimated at 523,000 STRV. Most of these imports enter under special programs: the Refined Sugar Re-export Program (250,000 STRV), the Sugar-Containing Products Program (125,000 STRV), and the Polyhydric Alcohol Program (15,000 STRV). Sugar derived from sugar syrups entering under HTS 17029040 are estimated at 125,000 STRV. Monthly sugar syrup imports have been steady through February and are expected to maintain the established pace. High-tier tariff imports are estimated at 8,000 STRV. Most of the high-tier imports have entered for retail sale in areas close to the Mexican border. Unless U.S. sugar prices should rise unexpectedly, it is not currently forecast that additional high-tier tariffs from Mexico will enter. Deliveries, Exports, and Ending Stocks Total deliveries for FY 2000 are estimated at 10.25 million STRV. Deliveries through the end of March have amounted to 4.957 million STRV, about 3.7 percent higher than last year through March. Deliveries by beet processors have been particularly strong, 2.2303 million STRV through March, or about 7.2 percent higher than the same period last year. Deliveries by cane refiners, on the other hand, are only at about the same level of last year, 2.710 million STRV. Demand through March by industrial users of refined sugar has been very strong, about 4.3 percent higher than the same period last year. In particular, deliveries to ice cream and dairy product manufacturers have been running over 20 percent higher: 241,770 tons, actual weight for FY 2000, compared with only 200,020 tons, actual weight for FY 1999, through March. Deliveries to confectionery manufacturers have been running over 6 percent higher than the previous year, and deliveries to bakers and cereal manufacturers have been about 3.9 percent higher. Deliveries to non-industrial users trail in comparison, but are still 1.9 percent higher than last year through March. U.S. sugar exports are comprised of exports under the Refined Sugar Re-export Program and are currently estimated at 229,000 STRV. While imports under the Program are estimated at 250,000 STRV, exports have been reduced by the amount of the CQEs surrendered in lieu of exports, as set out in the USDA press release dated February 29, 2000 (#PR 0102-00), and as explained in the "Import" section above. Total sugar use for FY 2000 is the sum of total deliveries and exports, and is estimated at 10.479 million STRV. Total supply (beginning stocks, production, and imports) is estimated at 12.447 million STRV. The difference of total supply and total use provides an estimate of ending stocks at 1.967 million STRV, implying an ending stocks-to-use ratio of 18.8 percent. This ratio is the highest estimated since FY 1986. Prices and Policy Developments On May 11, 2000, Agriculture Secretary Glickman announced that USDA will seek to buy U.S. sugar to reduce the cost of expected sugar program loan forfeitures, support sugar growers, and help stabilize low market sugar prices (press release #0159.00). According to the press release, at least 75 percent of the initial approximately 150,000 tons purchased will be refined sugar and may be followed by additional purchases depending on price and market conditions. USDA expects to save as much as $6 million in administrative costs that the government might otherwise incur from expected loan forfeitures later this summer. The Commodity Credit Corporation (CCC) will purchase the sugar using its authority under the cost reduction options of the Food Security Act of 1985. For now the CCC plans to store the purchased sugar and will not sell it back into the depressed U.S. sugar market. The Secretary challenged the sugar industry by saying that the USDA expects the sugar industry to rapidly develop conservation and production options that can form the basis of a sustainable sugar policy. He emphasized that relying on continued government purchases over the long term is neither feasible nor realistic. He urged sugar growers and processors to voluntarily increase their participation in USDA's Conservation Reserve Program (CRP). The Secretary said that enrollment in the CRP would help the industry and also improve environmental quality for all Americans. Good weather, productivity improvements, expanding plantings, and increased imports of sugar from foreign countries outside the TRQ are all factors that underpinned the oversupply of sugar. U.S. raw sugar prices (nearby futures, C.I.F., duty-paid, Contract No. 14, New York) averaged 19.43 cents per pound during April, up 97 points from the March average, but down 314 points from a year earlier. July No. 14 futures prices averaged 18.72 cents per pound on May 15. World raw sugar prices, which have been in steady decline since mid-1998, averaged 6.48 cents per pound, 1.04 cents a pound above a year ago. As of May 15, the world raw sugar price averaged 6.89 cents a pound, with a maximum of 7.36 cents a pound and a minimum of 6.65 cents. Although world prices have shown strength recently, they may dip during the summer and rise in the fall due to production cutbacks in the European Union, Brazil, and Thailand. Forecast for FY 2001 Production NASS forecasts sugarbeet area planted for FY 2001 at 1.578 million acres, about 1 percent higher than FY 2000. Since the March 2000 forecast release date, there has been some question whether this forecast level can be realized. The situation is made uncertain by the possible closing of California beet processing plants in Woodland and Tracy. It now seems certain that acreage intended for harvest in the spring of 2001 will not be planted. Taking this information into account and applying State-level 5-year averages of the acreage harvested to planted ratios, the Interagency Sugar Estimates Committee (ISEC) of USDA projects acreage harvested at about 10,000 acres less than last year's total. Beet sugar production for FY 2001 is projected at 4.700 million STRV. This projection is based on projected acreage harvested and on analysis of sugar yield per acre relationships. The analysis typically takes into account trends and the effect of weather conditions, as well as sugar produced through the process of desugaring molasses. The analysis is updated regularly as new information becomes available. Early in the season, the analysis depends on the assumption of normal weather patterns and other current information. On this basis, the ISEC expects the aggregate U.S. sugar yield to be about 3.10 tons per acre. This rate is not as high as that estimated for FY 2000 (3.24 tons per acre), an exceptional year for beet sugar production. Projection updates throughout the year will continue to take into account the possible plant closings in California. Cane sugar production for FY 2001 is projected at 4.323 million STRV. This sum represents the aggregation of projected production in Florida (2.090 million STRV), Louisiana (1.720 million STRV), Texas (160,000 STRV), Hawaii (330,000 STRV), and Puerto Rico (22,500 STRV). Florida cane sugar production for FY 2001 is projected at 2.090 million STRV. Sugarcane acreage harvested for sugar is expected to be at about the same level as FY 2000. Normal weather patterns would imply sugarcane yields falling midway between this year's level (35.0 tons per acre) and that of the previous year (40.1 tons per acre). Analysis of Florida sugar yields indicates an expected value between 4.70 and 4.72 tons per acre. Like the sugarcane yield projection, this amount falls about halfway between this year's yield (4.46 tons per acre) and the previous year's level (5.00 tons per acre). Louisiana cane sugar production for FY 2001 is projected at 1.720 million STRV. Sugarcane acreage harvested for sugar and seed is expected to expand by about 2.2 percent, with most of the increase going to seed production. For a second year in a row, Louisiana sugarcane acreage will exceed that of Florida. It is expected that the portion of the crop from the high-yielding variety LCP85-384 should continue to increase and result in continued sugarcane yield gains. With normal weather, sugar yields should be in the neighborhood of 3.92 to 3.93 tons per acre. These levels, if realized, would represent records. Texas cane sugar production for FY 2001 is projected at 160,000 STRV. Although preliminary, it is expected that Texas sugarcane area for sugar and seed will be about 47,000 acres, which, if realized, would represent growth of about 15,800 acres over FY 2000. The acreage growth is consistent with the high percentage (8.01 percent) of the FY 2000 sugarcane acreage reserved for seed. The Texas projection, like the others, assumes normal weather. There seems to be some concern, however, about reservoir water levels for irrigation if current dry conditions persist through the year. Hawaii cane sugar production for FY 2001 is projected at 330,000 STRV. As explained above, there is some concern about production on Kauai with the closing of the Kekaha Sugar Mill and the planned transport of sugarcane across the island to the Lihue Plantation. Although sugar produced to date for FY 2000 is still on track to meet the FY 2000 projection level of 360,000 STRV, this may be an artifact of strong calendar 1999 production from the last 3 months of the year counting toward the current fiscal year's total level. Puerto Rico sugar production for FY 2001 is projected at 22,500 STRV. It is expected that the two sugar mills on the island will be back in production next year. Sugarcane area is expected to expand to about 11,900 acres, with plans for as much as 300,000 tons of sugarcane. Non-TRQ Imports and Exports The raw and refined sugar TRQ for FY 2001 has not been established. Other imports for FY 2001 are projected to total 448,000 STRV, including 175,000 STRV under the Refined Sugar Re-export Program, 125,000 under the Sugar-Containing Products Program, and 15,000 under the Polyhydric Program. Sugar syrups under HTS 17029040, and high-tier tariff sugar imports are expected to be at the same levels as in FY 2000. Deliveries Deliveries for FY 2001 are projected at 10.385 million STRV. This level represents growth over the previous year of 1.3 percent, which is less than the growth of 1.8 percent estimated between FY 1999 and FY 2000. Growth in deliveries to industrial users is projected at 2.0 percent, while growth in deliveries to non-industrial users is projected at only 0.3 percent. At this point, it is difficult to factor in the effect of possible continued low prices for wholesale sugar. Whether low prices will be passed through to retail or affect industrial demand for sugar will be closely watched. Mexican Sugar The U.S. Department of Agriculture (USDA) currently estimates Mexican sugar production for 1999/2000 at 5.070 million metric tons, raw value (MTRV). Area harvested for sugarcane is estimated at 627,000 hectares, and sugarcane production is estimated at 43.365 million metric tons, for a yield of 69.2 tons per hectare. The implied sugar recovery rate from the sugarcane is 11.7 percent, which, if realized, would be a record. The harvest has been aided by good weather and dry conditions. According to Mexican sources, about 37.57 million tons of sugarcane have been harvested through May 7, 2000, and about 4.2 million MTRV, tel quel, of sugar has been produced. Sugar produced to date has only been surpassed in the 1997/98 season. Figure x shows the cumulative sugar recovery rates (tel quel basis) for the last few years. This year's rate has been consistently above comparable period rates, reflecting good weather and improved milling performance. Sugar consumption for 1999/2000 is estimated at 4.482 million MTRV. Despite improved consumer purchasing power, consumption growth has been slow because of the increased usage of alternative sweeteners. Industry sources indicate, however, that high compensatory duties on high fructose corn syrup (HFCS) imported from the United States has aided the demand for sugar in the current marketing year. Sugar exports for 1999/2000 are estimated at 630,000 MTRV. Mexico's access to the U.S. market for FY 2000 is limited to 25,000 MTRV under the North American Free Trade Agreement (NAFTA), plus an additional 2,954 MTRV covered under the FY 2000 refined sugar tariff-rate quota. Ending stocks are estimated at 623,000 MTRV. The Mexican Government currently provides final assistance to cover storage costs for up to 600,000 MTRV. The USDA currently forecasts Mexican sugar production for 2000/01 at 5.090 million MTRV. Area harvested for sugarcane is projected at 630,000 hectares, and sugarcane production is projected at 44.000 million metric tons. These projections indicate insignificant growth. The Mexican sugar industry is harmed by current low world sugar prices and by longstanding financial problems. The Mexican Sugar Financing Institution (or FINA) initiated plans in 1999 to restructure about US$2.0 billion of the sugar mills' debt. FINA is reportedly no longer a provider of credit to the industry but rather is functioning solely as a debt collector. The largest Mexican sugar group, the Consorcio Industrial Escorpion, whose debt is estimated at $800 million, has presumably been given to October 2000 by FINA to locate buyers for its mills. (Recent reports, however, indicate that recently-secured short term financing may have forestalled mill sales for the time being.) On May 4, 2000, the second largest Mexican sugar producer, the Grupo Azucarero Mexicano (GAM), asked for protection under existing bankruptcy laws, which are due to become much more stringent later in the month. GAM is reportedly indebted in the neighborhood of US$126 million. Sugar consumption for 2000/01 is projected at 4.482 million MTRV, the same as for 1999/2000. Refined sugar purchases by the Mexican soft drink industry are expected to be between 1.0 and 1.2 million metric tons. Lower-priced sugar this year has reportedly displaced sales of the HFCS in the soft drink industry. Mexican sugar exports are projected at 700,000 MTRV, and ending stocks are currently projected at 531,000 MTRV. The Mexican Government's financial assistance for storage up to 600,000 MTRV will not continue in 2000/01. According to the NAFTA side-letter agreement, Mexico's duty-free access to the U.S. market will increase to the smaller of 250,000 MTRV or Mexico's "net surplus production". The net surplus production formula is the difference between Mexico's projected production in metric tons, raw value, less the sum of projected consumption of sugar in metric tons, raw value, and projected consumption of HFCS in metric tons. Economic Research Service (ERS) analysis projects HFCS consumption in 1999/2000 in the neighborhood of 475,000-500,000 metric tons. The NAFTA specifies that U.S. and Mexican officials consult by July 1 of each of the first 14 years beginning in 1994 to determine by what quantity either nation is projected to be a net surplus producer in the next marketing year. Returns from Mexican Sugar Processing Stephen Haley 1/ ----------------- 1/ Agricultural Economist, Specialty Crops Branch, Market and Trade Economics Division, ERS. ----------------- Abstract Mexico is soon expected to be a larger supplier of sugar to the U.S. market. The magnitude of Mexican sugar sourcing depends on the profitability of the Mexican sugar milling sector. This article analyzes trends in gross returns from sugar sales and production costs for the Mexican milling sector, and analyzes the contribution of important determinants to earnings per hectare. Technological adaptions have been fundamental in making the Mexican processing sugar sector profitable since the early 1990's. Although high domestic prices that are in part traceable to supporting Mexican Government policies have played their role, the industry has done much to advance its production potential. The data show that processing mills are diverse but almost all show consistently positive rates of return. While some mills are more vulnerable than others with respect to price and supply shocks, most have potential for continuing gains in production. Analysis shows that mills that produce more relative to their rated capacities and have longer campaign lengths receive higher per-unit returns. Results also indicate that mills with lower earnings have relatively more to lose by not maintaining the pace of adapting technological improvements. An associated implication is that the marginal return to technological adaptation is higher for lower-earning mills. Keywords: costs of production, Mexico, sugar, sugarcane. Returns from Mexican Sugar Processing The U.S. sugar sector is increasingly dependent on developments within the Mexican sugar industry. With the North American Free Trade Agreement (NAFTA) in place, Mexico is soon expected to be a larger supplier of sugar to the U.S. market. Under the terms of the side-letter agreement to the NAFTA, Mexico's duty-free access to the U.S. sugar market will increase up to a maximum of 250,000 metric tons, raw value (MTRV) as of October 1, 2000. By fiscal year (FY) 2008, there will be no quantitative limit on Mexican duty-free imports. The NAFTA high-tier tariff schedule, currently at 12.09 cents a pound for raw sugar and 12.81 cents a pound for refined sugar, decreases at a rate of 1.51 and 1.60 cents a pound for raw and refined sugar for each fiscal year through 2007. The U.S. Department of Agriculture's (USDA) sugar baseline shows high-tier sugar imports from Mexico at 121,000 short tons, raw value (STRV) in FY 2003 and above 600,000 STRV for FY 2004 through 2007 (USDA, 2000). By FY 2004, 8.1 percent of U.S. sugar consumption is projected as coming from Mexico. This rises to above 9 percent in FY 2005 and grows steadily thereafter. 2/ Downward price adjustments emanating from the expansion of U.S. sugar supply from duty-free Mexican imports cannot be directly offset by the USDA. ------------ 2/ For a full exposition, see special article entitled "Conceptual Overview of the U.S. Sugar Baseline: Incorporating the Effects of the North American Free Trade Agreement," in the Sugar and Sweetener Situation and Outlook. SSS-227, January 2000, pp. 11-19. -------------- Under the terms of the Uruguay Round Agreement (URA), the United States agreed to bind sugar imports under the sugar tariff-rate quota (TRQ) at a minimum access level of 1.256 million STRV. With record increases in U.S. sugar production, the USDA does not project sugar TRQ imports above minimum access levels though 2010. Also, the declining NAFTA high-tier tariff implies an eventual upward limit on U.S. sugar prices. The incentive for Mexico to change its level of high-tier sugar exports is signaled by a pricing threshold. This threshold is the sum of the world price for sugar, marketing costs, applicable premiums and discounts, and the NAFTA high-tier tariff. As the tariff is reduced each year, the threshold is reduced and the likelihood of increased imports is enhanced. As long as Mexico has enough sugar to ship at the threshold level, this level sets the price of sugar in the U.S. market. Without specific policy interventions, a U.S. price higher than the threshold would result only if Mexican export potential were limited. In this case, pricing would be higher than the implied threshold level but less than the level implied if there were no high-tier tariff imports from Mexico. Mexican export potential is influenced by a variety of factors. On the demand side, the chief influence is how much high fructose corn syrup (HFCS) will substitute for sugar in beverage and food processing industries. According to the original NAFTA, the tariff on imports of U.S.-sourced HFCS was to reach zero in 2004. Although the Mexican Government currently levies an anti-dumping duty on HFCS imports from the United States, the incentive to eventually switch from sugar to HFCS is huge and could therefore make available for export over an additional million tons of sugar beyond that already projected. On the supply side, an important question is how much Mexican sugar production can be expected to expand in the future. Garcia and others are pessimistic regarding Mexican production potential (Garcia and others, 1999). They cite the preponderance of unsustainable high-cost sugar mills and the heavy indebtedness (about $2.5 billion) of the largest of the sugar groups that own many of the mills. Although the USDA does not do an official baseline for Mexican sugar, it implicitly assumes that production will continue to expand through 2010 and that Mexico will have over 1.0 million STRV to export. The purpose of this article is to examine more closely factors underlying Mexico's capacity to produce sugar. In particular, it analyzes trends in gross returns from sugar sales and production costs for the Mexican milling sector. These data are made available by the Fideicomiso Para El Mercado Azucar ( FORMA). Besides sales and cost data, FORMA makes available data on area harvested in hectares (ha), sugar and sugarcane production, sucrose recovery rates, campaign length, petroleum use in milling, and milling capacity levels. These data are reported on factory, State, and national levels for the 1988-98 marketing years. They provide a rich data set that is amenable to analysis. In order to account for inflationary trends in the data reported in pesos, these data have been converted into U.S. dollars at the prevailing exchange rates. The hypothesis of this paper is that technological adaptions have been fundamental in making the Mexican processing sugar sector profitable since the early 1990's. Although high domestic prices that are in part traceable to supporting Mexican Government policies have played their role, the industry has done much to advance its production potential. The data show that processing mills are diverse, but almost all show consistently positive rates of return. While some mills are more vulnerable than others with respect to price and supply shocks, most have potential for continuing gains in production. The analysis is carried out in three ways. The first way examines earnings, gross returns, and costs across firms for the period 1993-98. The goal is to determine which variables within the available data set help explain earning differences between factories. Although this analysis reveals certain illuminating relationships based on capacity usage measures, its basic conclusion is that differing endowment factors differentiate firms according to earnings potentials. This conclusion suggests a second way of analyzing the data based on the examination of subsets of the data across time. This analysis emphasizes the effects of exogenous price movements and adaption of technology to improve earnings for groupings of mills along endowment, ownership, and location characteristics. The third analytical method modifies the second by using its results to project earnings out to 2011. This sensitivity analysis draws out the implications of estimation results for future developments within the Mexican sugar sector, about which there is much interest and speculation. Cross-Factory Analysis Figure A-1 shows net earnings per hectare (revenues less costs divided by total hectares harvested) and sugar production at the national level for the period 1988 through 1998. The data reveal contrasts in the two periods. The first covers the years 1988-92 where the average earnings per hectare amounted to $529.85. The average for 1993-98 is almost 44 percent higher at $761.16. The first period corresponds to that period when the mills were undergoing privatization after having been run by the Mexican Government. Most of the mills had been suffering from a lack of investment. Many of the mills had been sold at inflated values that would later contribute to the debt repayment problems experienced by some sugar companies. In this pre-NAFTA period, many analysts doubted the potential of the industry to produce at levels high enough to justify exports. Sugar production, along with earnings, grew in the second period. It increased an average of 940,000 metric tons, tel quel (MTTQ) over the first period average, up to 4.332 million MTTQ. Sugar production in 1998 grew to a record 5.174 million MTTQ. Table A-1 presents some of the earnings data for the individual sugar mills for 1993-98. There are 64 mills but only 61 have been operating since 1992. The mills are organized into six categories based on the average earnings per hectare. Group no. 1 is comprised of the 11 mills (not counting the inoperative mills) with the lowest earnings. Groups numbered 2 through 6 represent grouped mills of 10 each and are ordered on the basis of ascending earnings per acre. Columns 2 and 3 show the companies that own each of the mills and the Mexican State where each mill is located. Data include gross returns per ton, costs per ton, daily cane milling capacity, average sugarcane harvested, and average campaign length. Analysis of Data through Scattergrams Relationships between the data are not immediately apparent. Figures A-2 and A-3 show scattergrams between average earnings per hectare and milling capacities and sugarcane milled. The regression lines shown in the figures represent the relationship between the variables: a positive or negative slope would indicate that as one variable changes, so does the other, either in the same (positive slope) or opposite (negative slope) direction. The flatness of the lines in both figures indicate no relationship between earnings per hectare and capacity or volume of sugarcane milled for the mills. If there were economies of scale in sugar milling, positive relationships would have been expected. Further pursuit yields more insight. Figure A-4 shows a scattergram between earnings per hectare and capacity usage (that is, the ratio of average daily sugarcane milled to rated daily capacity). The regression line slopes upward. The coefficient on the slope parameter indicates that as capacity usage increases by 1 percent, average earnings per hectare increases by $10.27. This result, while statistically significant, explains less than 9 percent (R2 = 0.0881) of the variance in earnings per hectare. Another constructed variable from the data set that measures capacity usage is total sugarcane milled relative to daily capacity. This measure essentially takes the previously defined capacity usage measure and multiplies it by the campaign length. Figure A-5 shows the scattergram and regression line. The line slopes upward and an increased amount of earnings variance is explained, about 15 percent, or 6 percent more than the daily capacity usage measure. Regression Analysis These scattergram results suggest that more sophisticated regression analysis involving gross unit returns and unit costs would likely benefit from the inclusion of capacity usage measures as explanatory variables. Also, the low level of the R2 measures from the scattergram analysis suggest that other factors may help explain earnings differences between mills. One possibility is that firm ownership or geographical location may influence earnings differences between mills. A regression model encompassing these factors is specified as follows: The variable Z represents either gross returns per ton or costs per ton on a factory-level basis. The general explanatory variables X are capacity usage variables. In one case, average daily capacity usage and number of campaign days can be used in a single regression. In an alternative case, the two measures are combined in the single measure of total sugarcane milled relative to a mill's daily grinding capacity. The variables FIRMi and STATEj are indicator variables whose respective values are equal to 1 if a factory is owned by the ith firm or located in the jth State; the value is zero, otherwise. In results discussed below, firm-specific and State-specific effects are not both included in a single equation. This means that if firm-specific (location-specific) effects are analyzed in an equation, the Bj (Bi) coefficients corresponding to the location-specific (firm-specific) are restricted to zero. In other words, multiple regressions are run including only one set of specific factors at a time, and then the specifications are compared for their relative advantage in making out-of-sample predictions. 3/ ------ 3/------------ The Schwarz Information Criterion (SIC) is used. It is a measure of the 1-step ahead out-of-sample prediction error variance. A smaller value associated with an equation indicates that the equation has better forecasting ability relative to those equations with higher values. -------------- This approach potentially allows a determination of which specific factors are better in explaining earnings differences among mills. Table A-2 shows the estimation results for eight cases. Cases numbered 1-4 examine gross returns per ton, and cases 5-8 examine costs per ton. Cases 1, 2, 5, and 6 include firm-specific factors, while cases 3, 4, 7, and 8 consider State-specific factors. Cases 1, 3, 5, and 7 include two measures of capacity usage (daily capacity usage and campaign length), while cases 2, 4, 6, and 8 consider the composite measure (total sugarcane milled relative to daily capacity). Originally, all firm- or State-specific variables were included in their respective equations. Most of the estimated coefficients did not test differently from zero--only those whose t-statistics indicate statistical significance are presented in the table. For the analysis of gross returns per ton, results indicate that all the capacity usage measures make a positive and significant contribution to accounting for differences among mills. Judging from the lower Schwarz criterion values (cases 2 and 4 relative to 1 and 3, respectively), the single measure of total sugarcane milled relative to capacity is more successful in explaining differences. This result is not unexpected considering that daily capacity usage and campaign length have a positive correlation coefficient of over 0.55. Each affect is likely partially picking up the effect of the other. Consideration of firm-specific effects (cases 1 and 2) indicate that mills owned by the Grupo Porres and the Consorcio Industrial Escorpion (CAZE) have higher returns than accounted for by capacity usage alone. Likewise, consideration of State-specific effects (cases 3 and 4) indicate that mills in the State of San Luis Potosi have higher returns and that mills in the State of Sinoloa have lower returns than accounted for by capacity usage alone. Lower Schwarz criterion values associated with cases 3 and 4 imply that State-level characteristics have more predictive power in forecasting than firm-specific effects. Cases 5-8 examine costs per ton for the mills. The capacity effects are much less certain compared with their effects on gross returns per ton. In cases 5 and 6 (the firm-specific cases), the coefficients on campaign length and total sugarcane milled relative to capacity are statistically significant and negatively related to unit costs, as would be expected. In cases 7 and 8, these same coefficients cannot be distinguished from zero. The State-specific effects are sufficient to produce a relatively high adjusted R2s above 0.60 and much lower values for the Schwarz criterion. These results imply that unit costs are similar in most producing areas except for these States, arranged in order of increased unit costs: Jalisco (6 mills), Tamaulipas (2 mills), Michoacan (4 mills), and Sinaloa (4 mills). These States account for about 23 percent of total hectares harvested in Mexico. Mills in Sinaloa seem the worst off, considering their lower relative gross returns (cases 3 and 4) and their higher unit costs. Summarizing, mills that produce more relative to their rated capacities and have longer campaign lengths receive higher per-unit returns. There is less variation on unit costs except for mills in 4 of the 15 States where there is sugar production. Better statistical results from inclusion of State-specific factors suggest that agronomic or locational conditions rather than ownership patterns are likely more important for differentiating mills on the basis of earnings. Times-Series Analysis The next step in the analysis is to extend it dynamically by considering what factors are important in causing earnings to change over time. The static results presented in the previous section suggest diversity across mills that probably hinders accurate analysis of factors accounting for the changes. For this reason, the data are divided into various subsets according to the level of earnings (the six groupings previously described), ownership by the 14 major sugar-processing companies and the independents considered together, and the 15 States. Data Analysis Table A-3 presents some basic production and earnings data for the groupings. Within each grouping, the members are organized on the basis of ascending earnings per hectare. In order to emphasize changes through time, averages for the early (1988-92) and more recent (1993-98) periods are presented side-by-side for comparison. The changes have been noteworthy. Total hectares harvested have increased 5.2 percent; sugarcane yields have increased 6.6 percent; and most remarkably, sugar yields per hectare have increased by nearly 20 percent. Gross unit returns have increased 26 percent, but unit costs have increased only 21 percent. On a sugarcane-metric-ton basis, earnings have increased almost 35 percent over the early period. Within each of the groupings, both sugarcane yield and sugar yield data increase significantly, going from the lower earning group member to the highest. For the grouping based on earnings, the average increase is 6.3 mt/ha for sugarcane and 0.98 mttq/ha for sugar. Corresponding increases are 1.3 mt/ha and 0.20 mttq/ha for the company classification, and 3.4 mt/ha and 0.46 mttq/ha for the State classification. Although these increases have little value in themselves, they indicate that earnings are correlated with productivity in both production and processing. Annual computed increases in sugarcane and sugar yields, on the other hand, do not correlate with the within-group ordering. It is more relevant to note that across all earning group members, and across most company and State group members, annual yield growth rates have been significantly positive. This suggests that productivity gains have been distributed fairly evenly across almost all mills. Table A-4 shows how mills are distributed across the company and State groupings by earnings per hectare. Individual companies own mills that typically span across the earnings spectrum. The same is true for mills in individual States except possibly for Sinoloa, Campeche, Tamaulipas, and Quintana Roo at the low end, and for Puebla and Morelos at the high-earnings end. The ninth and tenth columns show average earnings per hectare, and the eleventh column shows the ratio of the later period to the earlier. Earnings growth has been evident in all companies and in almost all States. The aggregate increase has been almost 44 percent. Regression Analysis The major hypothesis of this article, supported by the analysis presented thus far, is that productivity growth and technology adaption have been the primary factors in increasing earnings of Mexican sugar mills. The next step is to explicitly test the hypothesis. The model employed for the testing is as follows: Earnings per hectare are modeled as a function of the unit value of sugar sales (average sugar prices faced by mills), sugar yield per hectare, and a time trend. Sugar yield is interpreted as an output measure for technological adaptation. Sugar yield has both trend and weather-related random aspects. As seen in appendix table 2, the weather-related portion is correlated highly with sugarcane yields, leaving a trend portion that is interpreted as that portion of output gain from a fixed land input measure. As seen both in the appendix table and in table A-3, the sugar yield growth rate has averaged about 1.90 percent a year over 1988-98 and has been statistically significant across almost all elements of the three data groupings (earnings level, mill ownership, and State location). The value of unit sugar sales is interpreted as the average price at which mills have sold their sugar. It, along with sugar yield, is expected to be highly correlated with earnings per hectare. The time trend variable is included to account for miscellaneous effects not specifically captured by the other two variables. Regression results for the three groupings and the Mexican industry as a whole are displayed in appendix table 3. For the earnings grouping, the adjusted R2 are in the range 0.628-0.919, with an average of 0.763. Serial correlation is not a problem. Only for group members 1 and 2 (representing the 21 poorest performing mills) are the time trend coefficients significantly different from zero, and they are negative. Results for the company and State groupings are roughly similar. With the exception of two cases, all adjusted R2 are greater than 0.500. The average R2 for the company regressions is 0.768, and is 0.838 for the State regressions. The whole Mexico adjusted R2 is 0.818. Coefficient values on the sugar price and sugar yield variables that are significantly different from zero are shown in table A-5. The values of the sugar price coefficients tend to be slightly under the value of 1.0. This is true for 6 of the 6 cases for the earnings' grouping; 10 of the 15 cases for the company grouping, and 7 of the 15 cases for the State grouping. The coefficient value for whole Mexico cases is 0.754. The coefficient value on the values of the sugar yield coefficients, on the other hand, are mostly greater than one, indicating that earnings per hectare are elastic with respect to changes in yield values. This is true for 6 of the 6 cases for the earnings' grouping; 11 of the 15 cases for the company grouping, and 9 of the 15 cases for the State grouping. The coefficient value for whole Mexico case is 2.026. This indicates that the effect of sugar price changes are likely very uniform on earnings per hectare across the grouping.4/ ----------------- 4/ The formal test run is the Wald Coefficient Test. It measures how close an unrestricted regression comes to a regression with restrictions placed on particular coefficient values. The resulting statistic follows a chi-square distribution if the unrestricted estimates are close to those in the restricted case. The Wald statistic for the sugar price coefficients restriction is 3.37, which is consistent with the chi-square distribution. ---------------- Applying the test to the sugar yield coefficients yields the opposite result: the coefficients differ according to the grouping of the data for analysis. The coefficient value for the group no. 1 is 3.516, which is about 2.5 times as much as the average of group nos. 2-6. Applying the test while excluding group no. 1 still indicates that the yield coefficients differ significantly from each other for group nos. 2-6. Beta Coefficients It is difficult to gauge the importance of how explanatory variables account for changes in a dependent variable by focusing on elasticities alone. An elasticity reports how much the dependent variable changes when the independent variable changes by 1 percent. If there is more than a single independent variable in an equation, it may be that the likelihood of one variable changing by a single percentage point is vastly different from the likelihood of the other independent variable changing by the same amount. Essentially, one of the independent variables may be much more variable than the others. In order to adjust for differing relative variances, a regression coefficient can be weighted by the ratio of its standard deviation to the standard deviation of the dependent variable. The interpretation of the resulting coefficient, called a beta coefficient, is that a one-standard deviation change in the independent variable leads to standard deviation change in the dependent variable equal to the value of the beta coefficient. The normalization process that transforms regression coefficients into beta coefficients allows the effects of differing independent variables to be directly comparable. Table A-5 reports beta coefficients for the sugar price and sugar yield. The underlying model from which the beta coefficients are calculated is the non-logarithmic version of the model defined in the equation above. The sugar price beta coefficient values average about 0.70; and except for group no. 3, are pretty close in value to each other. The sugar yield beta values show more variability. They decrease in size, going from the low earnings group to the higher groups. The last column shows the ratio of the sugar yield betas to the price betas. The descending pattern going from group 1 to 6 is very evident, indicating that sugar yield changes have a relatively less effect on earnings for higher earning firms. An implication is that the mills with lower earnings have relatively more to lose by not maintaining the pace of adapting technological improvements. An associated implication is that the marginal return to technological adaptation (which of course must be weighed against marginal costs) is higher for lower-earning mills. Simulation Analysis The estimation results presented thus far can be used to project earnings per hectare into the future. The goal, however, is not to forecast but rather to illustrate the importance of technological adaption to the Mexican sugar industry. Certain exogenous assumptions must be made, including that estimated relationships will hold for the future. Table A-6 shows the equations of a projections model for Mexican sugar earnings per hectare. The equations track each of the elements of the earnings grouping. Model coefficient values for the first three equations are taken from the results in the appendix tables 1-3. Cane yield is a function of trend and hectares harvested. Sugar yield is a function of trend and cane yield. Earnings per hectare are a function of trend, sugar price, and sugar yield. Sugar prices faced by mills are a function of a national price, which is assumed constant to its 1998 value throughout the baseline simulation. Hectares harvested is assumed to grow at a 1-percent rate throughout the projections period. Figure A-6 shows baseline results for earnings per hectare for each of the earnings group members. The two lowest groups (1 and 2) are projected to have flat earnings throughout the 10-year projections horizon. Earnings per hectare are 3 percent less in 2011 than in 2001 for group no. 1, and are only projected 4.7 percent higher for group no. 2. Yearly sugar yield growth rates of 1.4 percent for group no. 1 and 2.0 percent for group no. 2 are insufficiently high to counter the observed downward trend in earnings for these two groupings. The other four earnings groups are projected to have high earnings growth. Relative to 2001, earnings per hectare in 2011 are projected 53 percent higher for group no. 3, 30 percent for group no. 4, 40 percent for group no. 5, and 23 percent for group no. 6. Each of these groups are projected to have good yearly sugar yield growth: 2.7 percent - group no. 3; 1.6 percent - group no. 4; 1.5 percent - group no. 5 ; and 1.4 percent - group no. 6. Unlike groups 1 and 2, there is no negative trend growth to offset to achieve higher earnings per hectare. To highlight the importance of technological adaption, three scenarios are run that remove the effects of the following: (1) sugarcane yield growth; (2) sugar yield growth; and (3) combined sugarcane and sugar growth. Table A-7 shows projected earnings for 2006, the middle of the projection period, for the groups for the base and for the scenarios. Percentage changes from the base are shown in the three right-most columns. The percentage changes attributable to the lack of technological adaptation are inversely proportional to the earnings ranking of the mills; that is, the largest percentage changes are concentrated in the lower earnings groups 1 and 2 and become smaller with the higher earning groupings. As would be expected, the percentage changes from scenarios 2 and 3 are correlated very highly with the sugar yield beta coefficients from table A-5: scenario no. 2, 0.978; and scenario no. 3, 0.963. The results imply that technology adaption is particularly crucial to sugar mills with lower returns. Another way to see the importance of technological adaptation to lower earning mills is illustrated in figure A-7. Two scenarios are run: the first specifies a 25-percent decrease in the national sugar price, and the second imposes a 10-percent increase in sugar yields on top of the price decrease scenario. The effect of the price decrease hits the lower earning groups 1 and 2 especially hard: 40 percent and 23 percent earnings drop per hectare. Earnings per hectare reductions for group nos. 3-6 are between 12 and 18 percent. The sugar yield effect serves to offset the price reduction effect. The offset effect is the strongest for groups 1 and 2: earnings per hectare are higher than in the baseline. The offset effect becomes progressively smaller going from group no. 3 up through no. 6. Extensions There are many uncertainties for the Mexican sugar producing sector. Many of the firms are heavily indebted, and returns to selling on the world market have been low. The terms under which they will be able to ship into the U.S. market are in dispute. The industry faces demand-side competition from high fructose corn syrup that is imported from the United States and also produced in Mexico. Relationships between the mills and sugarcane growers have not been cordial. Nonetheless, the sector has realized gains in production that were not predicted by most observers only a few years ago. Although analysis presented in this article points to further potential gains, the future remains hard to predict with a high degree of reliability. Additional analysis is needed for enhancing forecast ability. This article has focused on the demand by processors for hectares planted to sugarcane. More work needs to be done on the suppliers of area, or actual producers. Data on alternative crops in each of the producing areas, especially cost and return data, are needed. Emergent technologies that enhance yields and sucrose recovery need to be analyzed on their potential to expand production. This work, along with that on the demand-side, will lead to better forecasting ability. References Garcia, Luis R., Thomas H. Spreen, Gretchen Greene. "Structural Reform and Implication for Mexico's Sweetener Market," paper presented at conference entitled "Sweetener Markets in the 21st Century," sponsored by The University of Florida, Department of Food and Resource Economics, Miami, FL., November 1999. END_OF_FILE