基于县域单产数据的种植业保险定价模型关键假设检验

Testing for the Key Assumptions of Crop Insurance Pricing Models Using County-level Yield Data

基于定量风险评估的保险精算是种植业可持续发展的重要前提。针对农户级别历史单产数据严重缺乏的情况,研究中常常采用基于县域统计单产的简化模型用于种植业保险定价。这些模型的简化过程依赖于一些重要的假设,而这些假设能否得到数据的支持有待验证。本文回顾了常见的模型简化假设,并使用在湖南省常德市通过抽样问卷调查获取的农户级别历史产量数据一些关键假设进行了检验。结果表明,抽样数据不支持这些模型进行简化的假设,基于县域单产的简化定价模型不能直接应用到我国当前的种植业保险定价中。通过遥感估产手段对历史产量空间分布进行重建,或建立新的种植业保险损失数据采集标准以规范记录保险损失数据,才是解决我国种植业保险定价的治本之道。

It is the most fundamental issue for the sustainability of crop insurance program to have actuarial pricing based on quantitative risk assessment. Due to the severe lack of farm-level historical yield data, existing research usually employ simplified pricing models based on county-level statistical yield data. The simplification of these models relies upon a few critical assumptions. Nevertheless,whether those assumptions can be supported by actual farmlevel data is not known. In this study,we conducted hypothesis testing against the assumptions based on a sample of farm-level historical yield data collected by field survey in Changde City, Hunan Province. Results showed that our sampled data supported none of the model assumptions. Consequently, those simplified models using county-level yield data cannot be applied to the pricing of farm-level crop insurance in China. Therefore, the solution for crop insurance pricing in China is to overcome the data limitation for pricing. It can be achieved through rebuilding the historical spatial distribution of crop yield with remote sensing estimation, or establishing crop insurance loss data acquisition standard to make sound database for the future.