中国农业气象 ›› 2025, Vol. 46 ›› Issue (01): 62-70.doi: 10.3969/j.issn.1000-6362.2025.01.006

• 农业气象保险专刊 • 上一篇    下一篇

引入标准化降水蒸散指数(SPEI)的区域产量保险定价研究

夏梦,刘莉,熊涛,刘笑天   

  1. 1.中国气象科学研究院&中再巨灾风险管理股份有限公司·气象风险与保险联合开放实验室,北京 100081;2.华中农业大学经济管理学院,武汉 430070;3.中国气象科学研究院,北京 100081
  • 收稿日期:2024-04-03 出版日期:2025-01-20 发布日期:2025-01-17
  • 作者简介:夏梦,E-mail:meng.xia@webmail.hzau.edu.cn
  • 基金资助:
    气象风险与保险联合开放实验室开放基金项目(202300F9);国家自然科学基金青年项目(72203069);国家社会科学基金重大项目(22&ZD079)

Research on Area Yield Insurance Rating with Incorporating Standardized Precipitation Evapotranspiration Index (SPEI)

XIA Meng, LIU Li, XIONG Tao, LIU Xiao-tian   

  1. 1. Chinese Academy of Meteorological Sciences & China Re-catastrophe Risk Management Company Ltd·Joint Open Lab on Meteorological Risk and Insurance, Beijing 100081, China; 2. College of Economics and Management, Huazhong Agricultural University, Wuhan 430070; 3. Chinese Academy of Meteorological Sciences, Beijing 100081
  • Received:2024-04-03 Online:2025-01-20 Published:2025-01-17

摘要: 科学合理地厘定保险纯费率是区域产量保险持续、高效发展的基础。区域产量保险定价中纳入气候预测信息是提高保费厘定准确性的重要手段之一。本研究基于美国农业部风险管理局(Risk Management Agency,RMA)厘定区域产量保险纯费率法,引入标准化降水蒸散指数(Standardized precipitation evapotranspiration index,SPEI)构建SPEI−RMA模型,计算东北三省黑龙江、吉林和辽宁的玉米区域产量保险纯费率;引入样本外博弈框架,对比分析SPEI−RMA方法与RMA方法在区域产量保险定价中的表现,以期为设计更精准的区域产量保险产品提供科学依据。结果表明:相较于RMA方法,SPEI−RMA方法拟合作物单产趋势时,黑龙江省、吉林省和辽宁省的决定系数(R2)分别提高了0.044、0.088和0.153,均方误差(MSE)分别降低了0.068、0.067和0.213,有效提高了模型估计精度。利用两种方法厘定三省的纯保费,纯保费的绝对差异值中位数分别为0.105、0.114和0.087。通过样本外博弈发现,在70%、80%和90%的保障水平下,SPEI−RMA方法能更精准厘定保险纯费率,可在经济和统计意义上获得显著收益。

关键词: 农业保险, 区域产量保险, 费率厘定, RMA, SPEI?RMA

Abstract:

Developing actuarially fair premium rates is essential for the sustainable and efficient advancement of area yield insurance. Incorporating climate forecast information is one of the important means to enhance the accuracy of premium calculation in area yield insurance pricing. This study introduced the Standardized Precipitation Evapotranspiration Index (SPEI) into the pure premium rate making method of the U.S. Department of Agriculture's Risk Management Agency (RMA) to construct an SPEIRMA model. The model was applied to calculate pure premium rates for corn area yield insurance in Heilongjiang, Jilin, and Liaoning provinces of Northeast China. An out-of-sample game framework was employed to compare the performance of the SPEI−RMA method with the traditional RMA method in insurance pricing, aiming to provide a scientific basis for designing more precise area yield insurance products. Results indicated that when fitting the trend of crop yield, compared to the RMA method, the SPEI−RMA method increased the coefficient of determination (R²) by 0.044, 0.088, and 0.153 and reduced the mean squared error (MSE) by 0.068, 0.067, and 0.213 for Heilongjiang, Jilin, and Liaoning, respectively, thereby enhancing model estimation accuracy. The median absolute differences in pure premiums between the two methods are 0.105, 0.114, and 0.087 for the three provinces. Out-of-sample game result revealed that at coverage levels of 70%, 80%, and 90%, the SPEIRMA method more accurately determines pure premium rates, yielding significant economic and statistical benefits. 

Key words: Agricultural insurance, Area yield insurance, Rate making, RMA, SPEI-RMA