Aggregate versus disaggregate analysis: corn and soybean acreage response in Illinois

Tradeoffs between aggregation and specification errors and their implications for forecasting and policy analysis makeit difficult to determine the most appropriate level of aggregation for econometric analysis. Aggregation introduces bias in parameter estimates, but may reduce the effect of specification errors in the micro-level relationships.Here, tests and procedures for assessing the tradeoffs are applied to com and soybean acreage response equations in Illinois. The results indicate that modeling acreage response in Illinoisat the state level, rather than at the crop reportingdistrictleve~ leads to little loss in predictiveaccuracyand provides parameterestimatesmore consistent with expectations