The purpose of this paper to develop an empirical methodology for managing spatial basis risk in weather index insurance by studying the fundamental causes for differences in weather risk between distributed locations.
The paper systematically compares insurance payouts at nearby locations based on differences in geographical characteristics. The geographic characteristics include distance between stations and differences in altitude, latitude, and longitude.
Geographic differences are poor predictors of payouts. The strongest predictor of payout at a given location is payout at nearby location. However, altitude has a persistent effect on heat risk and distance between stations increases payout discrepancies for precipitation risk.
Given that payouts in a given area are highly correlated, it may be possible to insure multiple weather stations in a single contract as a “risk portfolio” for any one location.
Spatial basis risk is a fundamental problem of index insurance and yet is still largely unexplored in the literature.
