On Factor Substitution and Effective Tariff Rates
The purpose of a recent paper by Herbert Grubel and Peter Lloyd [5] was to examine the nature and magnitude of the bias introduced by the neglect of factor substitution between produced and primary inputs when calculating effective rates of protection (ERP). They concluded that the bias from neglecting substitution possibilities between produced and primary inputs is usually small. The purpose of this paper is to make three comments on their conclusions. Firstly, there is an error, or possibly a misprint, which can be misleading. Secondly, because Grubel and Lloyd restrict their calculations to a narrower range of parameter values than is empirically reasonable their results understate the magnitude of the bias which one might reasonably expect to encounter in empirically grounded calculations. Thirdly, Grubel and Lloyd conclude that the bias from neglecting substitution possibilities is usually small 3 without discussing criteria for determining when a bias becomes significant. It is shown here that the magnitude of the bias encountered when using empirically relevant data is very significant because the ranking of production processes by their calculated ERP can vary greatly with biased estimates. These comments have significant implications for the notion that the elasticity of substitution is not an important parameter in calculating the ERP.