Abstract:

Weather index insurance, as an innovative insurance product, serves as a key risk management tool in mitigating agricultural meteorological disasters. It features low operating costs and reduced susceptibility to moral hazard and adverse selection. Nevertheless, it faces the dual challenges of base risk and systemic risk. This paper systematically provided a systematic review of current research on the control of base risk and systemic risk in weather index insurance, with a particular focus on the risk modeling perspective. It examined existing methodologies in key areas such as weather index design, modeling of the relationship between weather indices and crop yields, and the modeling of agricultural meteorological systemic risks. The strengths and limitations of these approaches were analyzed, along with emerging trends. Finally, the study identified areas for further research, aiming to provide valuable insights for the optimizing design of weather index insurance in China, advancing of agricultural meteorological risk management and the safeguarding of food security. The results indicated that multi-source data integration, multiple disaster causes, multiple indicators and customized weather indicators were the main approaches to control base risk at the level of weather index design. At the modeling level of the weather index-yield relationship, high dimensionality, asymmetry, nonlinearity and interactions were key to managing base risk. Spatiotemporal dependence modeling of multiple variables was the primary focus for measuring systemic risk, while reinsurance, purchasing financial derivatives and expanding risk pools were the primary strategies for mitigating systemic risk. The coupling mechanisms of multiple disaster factors remained unclear, and the impact mechanism of crop damage was not well understood. Existing research had not effectively addressed the quantitative assessment of crop yield loss due to the interaction of multiple disaster types. Future research might explore the complex dependencies among various meteorological risks and their compound effects on crop yield reduction through data and model-driven approaches. In terms of product design, most researches adopted a step-by-step approach, which could lead to cumulative errors and result in a mismatch between the final product and the intended objectives. Future research should further explore an integrated framework for the optimal design of weather index insurance. Moreover, base risk and systemic risk exhibit a trade-off and may transform into each other. Therefore, it is essential to examine the boundary of insurance liability from a holistic perspective that considers both types of risk.

Key words: Weather index design, Index-yield relationship, Multivariate spatio-temporal dependence, Data and model