河南省玉米区域产量保险费率厘定实证研究

An Empirical Research on Ratemaking of Regional Corn Yield Insurance in Henan Province

自2008年河南成为种植业保险试点以来一直实行全省统一费率,极易诱发逆选择和道德风险问题,从而制约种植业保险的可持续健康发展。根据国外经验区域产量保险可有效削弱逆选择和道德风险。开展区域产量保险的前提是风险区划并厘定各分区的费率。本文从理论层面运用简单的动态博弈模型阐述区域产量保险的优点,并以河南18个地市1996~2015年的玉米单产、玉米种植面积和农作物种植面积为依据对河南省玉米区域产量保险展开实证研究。首先运用ARIMA模型估计18个地市1996~2015年玉米的趋势单产,进而求得各地市历年的玉米单产损失率;在此基础上通过非参数信息扩散模型求出各地市单产损失率的概率分布,并得出损失率超过某一阀值的概率;考虑到生产风险的复杂性又分别计算出各地市单产变异系数、专业化指数和生产效率指数三类风险指标,并结合单产损失率概率分布共四种风险指标综合考量各地市玉米生产风险水平;通过系统聚类分析法结合河南玉米生产现实状况和经营成本把河南省玉米生产划分为高、中、低三等级风险区域;最后分别计算出各分区95%和100%保障水平下的期望损失率,即为各风险区域玉米产量保险分别在95%和100%保障水平下的纯费率。

Since Henan became a pilot region for crop insurance in 2008, it has been using the same premium rates in the whole province, which can easily induce adverse selection and moral hazard, thus restricting the sustainable and healthy development of crop insurance. According to foreign experience, regional yield insurance can effectively reduce adverse selection and moral hazard, the premise of which is risk zoning and ratemaking for each zone. On the theoretical level, this article used simple dynamic game model to describe the advantages of regional yield insur- ance. On the empirical level, it carried out an empirical research based on the corn unit yield, corn planting area and crop planting area data of 18 cities in Henan province from 1996 to 2015. First,the ARIMA model was used to esti- mate the trend yield of com in these 18 cities from 1996 to 2015 ,and then the unit yield loss rate of corn in each city was obtained. Based on the non-parametric model of information diffusion, the probability distribution of the loss rate was obtained, as well as the probability of the loss rate exceeding a certain threshold. Considering the complexi- ty of the production risk, this article respectively calculated three indicators ： the unit yield coefficient of variation in each city, the specialization index and the productivity index, and combining the probability distribution of unit yield loss rate, a total of four indicators of risk, the corn production risk levels of each city was examined. Based on the hierarchical clustering analysis and actual corn production situation in Henan, Henan corn production regions were divided into high,medium and low risk zones. Finally, the expected loss rate of 95% and 100% assurance level of each zone was calculated respectively, which was the pure premium rate of the corn yield insurance of each risk zones at 95% and 100% protection levels.