Determining lags in export supply function in India

Abstract

This paper considers a partial adjustment model of the export supply function in Indian economy at the aggregate level. A minimum of one month and a maximum of 12 months is considered as the plausible range of the length of lags in exports supply in India with respect to both the independent variables domestic production and real effective exchange rate. Instead of the distributed lag model, a specific lag in the two independent variables is simultaneously considered in the paper. Different lags in the two variables would yield alternative models for export supply in India. Since the methodology based on nested model is not likely to work efficiently in such cases, an alternative procedure is suggested here to select the most appropriate model which amounts to determining the lengths of the lags in the two independent variables in the export supply function. The paper uses monthly data on Indian exports and other variables for the period January 1982 to July 1993. It appears from the selected model that in Indian economy export supply response lags by 12 months with respect to domestic production and 5 months with respect to real effective exchange rate changes. The extra-sample forecast accuracy of the elasticity of export supply with respect to exchange rate changes before 5 months are 0.52 in the short run and 1.11 in the long run. It is hoped that the suggested method in the present paper is likely to work more efficiently than the nested model method of selecting the most appropriate regression model whenever a choice has to be made from among a set of highly collinear alternative independent variables.