Testing for Aggregation Bias in Linear Models.

This paper discusses alternative methods of testing for aggregation bias and proposes direct tests of the discrepancy of the macroparameters from the average of the corresponding microparameters, and derives tests of aggregation bias in the general case where the parameters of interest may possibly be nonlinear functions of the microparameters. The paper also develops a Durbin-Hausman type misspecification test of the disaggregate model. These tests are then applied to disaggregate and aggregate specifications of employment functions for the U.K. economy disaggregated by forty industries. Copyright 1990 by Royal Economic Society.