A general equilibrium model of international trade with heterogeneous firms, under the assumption that the distribution of productivity across firms is Pareto, delivers systematically diff erent power law exponents for exporting and non-exporting firms. In this setup, the presence of international trade systematically a ffects the firm size distribution to make it more heavy-tailed. This model predicts that the power law exponent for exporters should be strictly lower in absolute value than the power law exponent for non-exporters. Following the propositions made in the literature, we estimate the power law exponent for a large sample of Turkish firms. We also question the applicability of the OLS regression in the context of power law estimation and provide maximum likelihood estimates, which have been proven to be consistent and effi cient in this context. Along with the maximum likelihood estimates, we also provide the CDF and ln(Rank-1/2)-ln(size) estimates. Our results provide supporting evidence for the theoretical predictions, according to which the distribution of firm size has heavier tails due to exporting behaviour.
展开▼
