Econometric identification of crop insurance participation

Abstract

This paper shows how econometric identification can be improved in studies making use of crop insurance participation as either an independent or dependent variable. The paper provides the reader with a succinct overview of how crop insurance contracts are priced and how to use publicly available data to derive a novel composite crop insurance design parameter that emulates existing crop insurance rating parameters using a procedure that is based on current actuarial practices. The derived design parameter performs well at predicting historic crop insurance loss-cost ratios and satisfies the requirements for an instrumental variable for a variety of empirical applications related to crop insurance. Representative empirical examples are presented where it is shown that the proposed instrument has favorable two-staged least squares diagnostic tests and is effective at eliminating endogeneity bias.

The effect of crop insurance on agricultural loan delinquencies

“…Results of the robustness checks using alternative IVs are presented in Table 4. In general, estimation results from the IV‐FE models using the alternative instruments suggested by Yu et al (2018), DeLay (2019), and Tsiboe and Turner (2023) remain consistent with our baseline IV‐FE model using the instrument suggested by Miao (2020).…”

“…First, we consider alternatives to the crop insurance participation variable where, instead of the liability ratio variable, we use three different measures of crop insurance participation—percentage of total insured acres, average coverage level, and total premium amounts. Second, we conduct another set of robustness checks where we run the IV procedures proposed by Yu et al (2018), DeLay (2019), and Tsiboe and Turner (2023) (i.e., these three IV approaches as alternative to the Miao (2020) approach described above. Third, in addition to the traditional IV approach, we also implement a recently developed “external‐IV‐free” approach by Kiviet (2013, 2020), called kinky least‐squares (KLS).…”

Farm‐level evaluation of area‐ and agroclimatic‐based index insurance

“…6 We use the historic values of the E rvt and F rvt to estimate base premium rates for each grid and a given T ig . The rating process adapted from (Tsiboe & Turner, 2023a) and is partly based on PRF-RI (Coble et al, 2020) involves the following steps: We first define stochastic historic empirical and payment calculation factors for each index variable/grid/coverage combination as PCF { }…”

“…In the FCIP, the final base rate for a given grid [county] is determined via a spatial smoothing algorithm that takes a weighted average of the grid's [county's] raw rate and those from contiguous grids [counties] (Tsiboe & Turner, 2023a). In this study, the spatial smoothing algorithm was not employed because of technical constraints.…”

“…financing (Ifft et al, 2015;Tsiboe & Turner, 2023a), and reductions in farm exits and disinvestment (Kim et al, 2019). Active insurance markets may also provide benefits to consumers of agricultural products in the form of lower prices and/or reduced volatility (Lusk, 2017).…”