\(E(F) = \exp(b_0 + b_1 \log M_o + b_2 \log M_d + b_3 \log d)\) by the
Poisson score equations (IRLS, as glm(family = poisson));
consistent under heteroskedasticity and keeps zero flows (Santos Silva
and Tenreyro 2006). Returns HC0 sandwich and model-based standard errors,
log-likelihood, AIC and BIC.
Usage
GravityPPML(flows, mass_o, mass_d, dist, tol = 1e-12, max_iter = 100L)
Arguments
- flows
Non-negative flows.
- mass_o, mass_d
Positive origin and destination masses.
- dist
Positive distances.
- tol
Relative deviance tolerance.
- max_iter
Maximum IRLS iterations.
Value
List with coefficients, se_robust, se_model,
fitted, loglik, aic, bic, iterations.
References
Santos Silva, J. M. C. and Tenreyro, S. (2006). The log of
gravity. Review of Economics and Statistics 88, 641-658.
Examples
GravityPPML(c(12, 0, 30, 7, 55, 3), c(5, 5, 9, 9, 20, 20), c(9, 20, 5, 20, 5, 9),
c(1, 3, 1, 2, 3, 2))$coefficients
#> [1] 8.8361998 -0.6109571 -2.5098502 0.8842146
