Aggregation Bias in Trade Elasticities: The Case of Macedonia

This paper evaluates the bias which may occur when trade elasticities are estimated using data on aggregate trade, instead of using data on bilateral trade. The exercise is done on the case of Macedonia. Elasticities obtained from aggregate-trade data, using the Autoregressive Distributed Lag approach, are compared with the elasticities obtained from bilateral-trade data, using dynamic heterogenous panels techniques. Results point out that the aggregation bias is sizeable and that relying on aggregate data can lead to wrong conclusions about the trade elasticities.