New nonlinear estimators of the gravity equation

The gravity model of international trade is often applied by economists to explain bilateral between countries. Nevertheless, some estimation practices have been subject criticism, namely how zero values and the heteroskedasticity are handled. This paper proposes new nonlinear techniques address these issues. In particular, we propose standard generalized versions Heckman two-step approach that do not require log-linearization equation corrects for non-random selection bias, a least squares estimator can be viewed as an iterative version normal family Quasi-Generalized Pseudo-Maximum-Likelihood estimator. Monte Carlo simulations show our proposed estimators outperform existent linear very efficient in correcting bias reducing deviation estimates. Empirical results previous studies overestimated contribution variables such importer’s income, distance, remoteness, agreements, openness.