The Log of Gravity

Export promotion agencies: Do they work?

“…2 4 digit goods, and the liberal classification includes 533 goods. 12 We also estimate the equation in the first column using a Poisson estimator and a robust regression estimator to correct for Jensen-inequality biases in the coefficients (Santos-Silva and Tenreyro, 2006), outliers or other forms of heteroscedasticity. Results are qualitatively the same and not statistically different from the ones reported here.…”

Trading wastes

“…(12) are not independent of the right-hand side covariates. Indeed, the similarity of the OLS results in column (1), which are shown by Santos Silva and Tenreyro [57] to be inconsistent for this reason, and the two-step polynomial results in column (3) are suggestive. The HMR two-step estimator is intended to control for both country pair selection (via the Heckman correction) and bias due to unobserved firm selection into export markets.…”

“…To account for unobserved multilateral resistance terms due to country importer and exporter heterogeneity Anderson and van Wincoop [3] and Santos Silva and Tenreyro [57] augment Eq. (1) by including importer and exporter fixed effects, d I and d E , implying Eq.…”

“…(2) using linear regression techniques. However, Santos Silva and Tenreyro [57] show that linear regression techniques on a log-linearization of Eq. (2) are inconsistent as the log-linearized error is dependent on the covariates and, further, log-linearization inadequately accounts for values of zero in the dependent variable.…”

“…Using the gravity model and its established foundations in international trade theory, the empirical hypothesis that relative levels of environmental regulation across countries are an important determinant of waste trade is explicitly tested. An empirical strategy similar to the Poisson pseudo-maximum-likelihood estimator proposed by Santos Silva and Tenreyro [57] is employed on a cross sectional dataset on bilateral waste trade between 92 developed and developing countries.…”