Questions & Answers
- What are long-term baseline projections?
- When are the projections released?
- What is the difference between a baseline projection and a forecast?
- What are some applications of the baseline?
- What methods are used by USDA to prepare its 10-year baseline projections?
Each year, USDA makes 10-year projections of the food and agriculture sector. The commodity projections are used to forecast farm program costs and to prepare the President's budget. The projections reflect a number of assumptions that are spelled out in a baseline scenario and cover agricultural commodities, agricultural trade, and aggregate indicators of the U.S. farm sector such as farm income.
The Departmental baseline report is released in February each year.
Baseline projections focus on longer term underlying trends based on a set of assumptions, while forecasts focus more on predicting actual outcome within a shorter time frame (1 or 2 years). A USDA "baseline" projection represents one plausible scenario for the next 10 years. These projections assume no shocks, but instead are based on specific assumptions for the macroeconomy, policy, weather, and international developments. Such conditioning assumptions are usually designed to provide a neutral backdrop for the projections to allow the analyses to focus on key long-term underlying factors. For example, macroeconomic assumptions for baseline projections are usually "smoothed," without recessions or economic booms, and agricultural policies are typically assumed to remain unchanged from current law. In contrast, forecasts incorporate additional information that departs from the neutral assumptions of baseline projections and are designed to lead to predictions of actual outcomes.
The commodity projections in the baseline are used to forecast farm program costs and to prepare the President's budget. As a neutral policy scenario, the baseline provides a useful basis of comparison for analysis of alternative polices and market developments. Examples of baseline applications include the following:
- evaluating the effects of changes in the renewal fuel standard and ethanol production on U.S. agricultural commodity markets and farm income; and
- analyzing the relationship of U.S. agricultural trade to the economies of developing countries, and comparing these countries' income changes and exchange rate movements with the baseline scenario.
The U.S. projections are developed by commodity-specific analytical tools with input from members in the Interagency Agricultural Projection Committee (IAPC). The final U.S. projections are incorporated into a U.S. model that is part of the international baseline modeling system.
The international models use economic behavioral relationships to project production, use, and trade quantities based on the world and domestic commodity market prices and assumed macroeconomic conditions. Domestic market prices are usually linked to world prices, but the models filter the world prices through import and export tariffs and domestic agricultural policy variables, such as support prices and producer and consumer subsidies.
The Excel-based country and regional model equations are extracted, converted to Fortran, and run together with a linking procedure to allow a global solution for equilibrium prices and trade. The linking method iteratively adjusts world prices until imports and exports balance across all countries for each major agricultural commodity market. The solution algorithm uses a numerical version of Newton’s method, with each slope (trade change versus price change) observed from the previous price-adjustment iteration. The price adjustments are repeated within Gauss-Seidel loops until all commodity markets are simultaneously cleared (meaning that world imports and exports are balanced). During the solution, each country model receives the adjusted world prices and responds, as domestic prices change in response to world price changes, generating changes in its production, consumption, imports, exports, and other variables.
The primary source of historical supply and demand data is the USDA, Foreign Agricultural Service (FAS) Production, Supply and Distribution (PSD) database. The database includes most grains, oilseeds and products, cotton, sugar, livestock, and livestock products. Supply-side variables include beginning stocks or animal inventory, area harvested or slaughter, yield, production, and imports. Demand-side variables include food, feed, industrial use, other use, ending stocks, and exports.
Some livestock product data are unavailable or no longer are updated for several country/commodity pairs in the PSD dataset. To include more market participants, the missing data may be pulled from the Food and Agriculture Organization Corporate Statistical Database (FAOSTAT) for food balances, livestock production, and livestock trade. The commodities considered are beef and veal, pork, poultry (chicken and turkey) meat, eggs, and lamb and mutton. Any selection of FAOSTAT data is made across the supply and use attributes and years for an entire country/commodity pair. FAOSTAT data series may be used if they are at least as recent as the corresponding PSD data series. PSD is selected as the source where PSD contains current meat data, since the final data years in FAOSTAT end earlier than the final data years for currently updated PSD series.
*For more information on the methods used in the international baseline projection process, please see USDA, Economic Research Service (ERS) Technical Bulletin Number 1951: The ERS Country-Commodity Linked System: Documenting Its International Country and Regional Agricultural Baseline Models
And in Technical Bulletin Number 1956: Structure of the USDA Livestock and Poultry Baseline Model