Generating detailed commodity forecasts from a computable general equilibrium model

The largest computable general equilibrium (CGE) models currently in operation produce forecasts for about 100 commodities (goods and services). This level of detail may seem overwhelming to macroeconomists but is often inadequate for micro planning. For example, a forecast for business services (a typical commodity at the 100-level) is of marginal interest in planning educational programs for sub-categories of business services such as accountancy, advertising and architecture. As a step towards generating information for micro planning, this paper describes a tops–down method for disaggregating CGE forecasts. The method relies on detailed sales data often collected by input–output sections of statistical agencies. An application is reported in which forecasts from a 114-commodity CGE model are disaggregated into forecasts for 780 commodities. Within each of the 114 core commodities, differences in prospects are forecast for sub-commodities reflecting differences in their sales patterns and in the degree to which they face import competition.