Coverage selection for the annual forage insurance based on robust portfolio optimization methods

Abstract This paper explores coverage selection strategies for the Annual Forage insurance program using robust portfolio optimization approaches, such as shrinkage and ensemble learning methods. The objective of the proposed models is to obtain an expected return with the lowest risk possible. Compared to previous efforts, a wider range of coverage selection parameters (i.e., coverage level, productivity factor, and index intervals) is considered. The proposed methods are used to protect cool-season forage production in Texas, using historical market, production, and actuarial data. This work provides empirical evidence on the effectiveness of the Annual Forage program in managing forage production risks.