Estimation of scale economies underlying growth and productivity: The empirical implications of data aggregation

Estimation of scale economies underlying growth and productivity patterns i s typically based on aggregated data, raising questions about the potential for aggregation biases. This paper provides empirical evidence on the exis tence and patterns of such biases. We use a cost-based model to estimate sh ort/long-run and internal/external scale effects for U.S, manufacturing dat a at different aggregation levels. Our results suggest that aggregation bia ses in such a model are not substantive. Also, internal scale economies see m more appropriately represented by the aggregate data, whereas more disagg regated data appears preferable for estimation of external or spillover eff ects that occur between industries or sectors.