Does consistent aggregation really matter?

Abstract

Consistent aggregation ensures that behavioural properties which apply to disaggregate relationships apply also to aggregate relationships. The agricultural economics literature which has tested for consistent aggregation or measured statistical bias and/or inferential errors due to aggregation is reviewed. Tests for aggregation bias and errors of inference are conducted using indices previously tested for consistent aggregation. Failure to reject consistent aggregation in a partition did not entirely mitigate erroneous inference due to aggregation. However, inferential errors due to aggregation were small relative to errors due to incorrect functional form or failure to account for time series properties of data.