The Sample Selection Model: Application on The Farmers' Decision of Rice Acreage

Woraphon Yamaka, Chaowana Phetcharat, Nattamon Teerakul, Thaya Navanugraha

Abstract

This study explores the estimation and inferences of a sample selection model with ordered-multinomial response outcomes. The model consists of two equations, namely, ordered response and regression equations. The estimation of parametric model is inconsistent when the distribution is unknown. To overcome this problem, we propose the primal generalized maximum entropy estimator to the model based on a constrained sample selection model with multinomial response outcomes. This method is also applied to examine farmers' decision on the rice acreage and perception of weather risk. The result of this study shows that the gender and age affect the farmers' subjective probability of the weather risk.