The problem of aggregation over micro units has had along tradition in
the econometrics Literature, streching backto the pioneering work of Theil (1954). In this Literature two issues in particular have attracted attention.
The first concentrates on the przdiction problem of choosing whether to use macro or micro equations to predict aggregate variables. This issue was raised by Grunfeld and Griliches (1960) and is further addressed in a recent paper by pesaran, pierse and kumar (1989) (P.P.K). In PPK a
generalised prediction criterion and a formal statistical test of hypothesis of perfect aggregation are developed. The present thesis considers the second accelleration mechanism, then by useing the above methods; restricted disaggregated equations are estimated. Imposition ofrestrictions, in the calculation and testing ofaggregation bias, provide intersting results. In additions the null hypothesis ofaggregation bias has been rejected in %1 significance level.