The log of gravity revisited
![Thumbnail](/xmlui/bitstream/handle/10234/39421/49226.pdf.jpg?sequence=4&isAllowed=y)
Visualitza/
Metadades
Mostra el registre complet de l'elementcomunitat-uji-handle:10234/9
comunitat-uji-handle2:10234/8643
comunitat-uji-handle3:10234/8644
comunitat-uji-handle4:
INVESTIGACIONMetadades
Títol
The log of gravity revisitedAutoria
Data de publicació
2011-09-23Editor
Taylor & FrancisISSN
0003-6846; 1466-4283Cita bibliogràfica
Taylor & Francis (2013) vol. 45, no. 3, p. 311-327Tipus de document
info:eu-repo/semantics/articleVersió
info:eu-repo/semantics/submittedVersionParaules clau / Matèries
Resum
This paper evaluates the performance of alternative estimation methods for gravity models with heteroskedasticity and zero trade values. Both problematic issues, recently addressed by Santos Silva and Tenreyro in an ... [+]
This paper evaluates the performance of alternative estimation methods for gravity models with heteroskedasticity and zero trade values. Both problematic issues, recently addressed by Santos Silva and Tenreyro in an influential paper, are re-examined here. We use Monte Carlo simulations to compare the Pseudo Poisson Maximum Likelihood (PPML) estimator recommended by Santos Silva and Tenreyro, a Gamma pseudo-maximum-likelihood (GPML), a Non-Linear Least Squares (NLS) estimator and a Feasible Generalized Least Squares (FGLS) estimator with more traditional techniques. Additionally, estimates of the gravity equation are obtained for three different data sets with the abovementioned methods. The results of the simulation study indicate that, although the PPML estimator is less affected by heteroskedasticity than others are, its performance is similar, in terms of bias and standard errors, to the FGLS estimator performance, in particular for small samples. GPML presents however the lowest bias and standard errors in the simulations without zero values. The results of the empirical estimations, using three different samples containing real data, indicate that the choice of estimator has to be made for each specific dataset. There is not a general “best” estimator and it is highly recommended to follow a model selection approach using a number of tests to select the more appropriate estimator for any application. [-]
Drets d'accés
Apareix a les col.leccions
- ECO_Articles [696]