Searching for Correct Specification in Spatial Probit Models

Selecting correct specification in spatial model frameworks is a relevant research topic in spatial econometrics. The purpose of this paper is to compare two well-known model selection strategies, Specific-to-General, Stge, and General-to-Specific, Gets, in the context of spatial probit models. These classical strategies are compared using a powerful machine learning algorithm: Gradient Boosting. The paper includes an extensive Monte Carlo experiment to compare the performance of both strategies with small and medium sample sizes. The results show that under ideal conditions, both strategies obtain similar results for medium-sized samples, but for small samples, Stge performs slightly better than Gets. As expected, both strategies are sensitive to misspecification and endogeneity problems, and the rate of true model selection decreases when they appear. The GBM algorithm obtains slightly higher success rates than the proposed strategies. Moreover, under the hypothesis that the GBM algorithm obtains the highest percentage of correctly classified models, the Stge strategy slightly outperforms the Gets strategy