A Weighted Spatio-Temporal Model for County Yields

This paper seeks to tackle the area yield distribution modelling problem by considering not only a spatial and temporal modelling framework, but also by treating both county and farm level information. We propose a model which combines on the one hand the cross-sectional spatially distributed data about temperature and precipitation over time and on the other hand the information relative to the acreage allocation at the farmers level. We demonstrate that beyond its forecasting power our model more adequately captures the dynamic covariation between the county level yields and the estimated local weather conditions compared with alternative models.

Keywords: Crop yields, crop insurance pricing, SARIMA, Gaussian process

JEL Classification: D24, O13, Q16, Q54

Suggested Citation: Suggested Citation

Ames, Matthew and Ames, Matthew and Bagnarosa, Guillaume and Gao, Suikai and Matsui, Tomoko and Peters, Gareth, A Weighted Spatio-Temporal Model for County Yields (July 3, 2019). Available at SSRN: https://ssrn.com/abstract=3447047 or http://dx.doi.org/10.2139/ssrn.3447047