基于多源数据的全国三大主粮生产风险评估研究  

Study on Production Risk Assessment of Three Major Grain Crops in China Based on Multi-Source Data

作　　者：赵思健[1] 聂谦[1] 张峭[1] 陈爱莲 李越[1] ZHAO SiJian;NIE Qian;ZHANG Qiao;CHEN AiLian;LI Yue(Agricultural Information Institute,Chinese Academy of Agricultural Sciences,Beijing 100081)

机构地区：[1]中国农业科学院农业信息研究所,北京100081

出　　处：《中国农业科学》2024年第21期4276-4289,共14页Scientia Agricultura Sinica

基　　金：教育部人文社会科学重点研究基地重大项目(17JJD910002);国家自然科学基金面上项目(41471426,72073132);中国农业再保险股份有限公司课题项目(ZGNZ2022-JS12)。

摘　　要：【目的】“一省、一作物、一费率”的粗放式定价模式带来逆选择、道德风险、经营乱象等问题,严重制约我国农业保险的健康可持续发展。费率定价离不开风险评估,开展全国三大主粮生产风险评估、制作风险地图,为实现我国主粮保险费率精准化、推进主粮保险高质量发展提供依据。【方法】围绕我国三大主粮作物(水稻、小麦、玉米),收集整理长时间序列的三大风险数据源(单产、灾情和保险数据),以单产数据为核心,结合灾情数据和保险数据,开展基于多源数据的风险评估建模,通过县级风险的低估调整和省级风险的秩相关调整,计算全国三大主粮省级和县级的纯风险损失率,采用分位数法开展全国三大主粮风险分区,制作风险地图。【结果】省级风险的秩相关调整以灾情风险结果为主,保险风险结果为次。经调整,水稻秩相关系数由0.610提升到0.766,小麦由0.547提升到0.748,玉米由0.576提升到0.760。调整后,全国三大主粮平均的风险低估系数为20%—40%,说明全国范围内利用县域单产开展风险评估的平均低估程度在2—4成,玉米低估系数要高于水稻和小麦。从省级尺度上看,黑龙江三大主粮生产风险均处于极高水平,内蒙古水稻、小麦,吉林和辽宁水稻、玉米生产风险处于极高水平,山西小麦生产风险处于极高水平;从县级尺度上看,水稻生产的极高风险(纯风险损失率>4.4%)主要集中在东北三省绝大部分种植县域、内蒙古东北部与东北三省接壤的种植县域。小麦生产的极高风险(纯风险损失率>6.3%)主要集中在内蒙古绝大部分种植县域。玉米生产的极高风险(纯风险损失率>6.9%)主要集中在内蒙古与东北三省、山西和陕西接壤的种植县域,辽宁、安徽、江西的大部分种植县域。从全国833个产粮大县上看,玉米极高和高风险县域数量占比最高(28.1%),水稻次之(25.1%),小麦最�【Objective】The extensive pricing model of“one province,one crop,one premium rate”has brought about problems,such as adverse selection,moral risk and disorderly operation,which seriously restricts the healthy and sustainable development of agricultural insurance in China.Accurate rate pricing cannot be achieved without agricultural risk assessment.Insurance rate pricing cannot be separated from risk assessment.Launching agricultural production risk assessment is an important task to achieve accurate rate pricing for grain insurance and to accelerate the high-quality development of agricultural insurance.【Method】Aiming at the three major grain crops(rice,wheat,and maize)in China,three kinds of risk data sources(yield data,disaster loss data,and insurance data)were collected and organized for a long time series.With yield data as the core,combined with disaster and insurance data,the risk assessment modeling was carried out,throughout the adjustment for underestimation of county-level risks and rank correlation of provincial-level risks,to calculate the pure risk loss rate of the three crops at the county level,and then to use the quantile method in risk zoning for the three crops and produce risk maps.【Result】The rank correlation adjustment of provincial-level risks was mainly based on disaster risk results,followed by insurance risk results.After adjustment,the rank correlation coefficient for rice was increased from 0.610 to 0.766,wheat was increased from 0.547 to 0.748,and maize was increased from 0.576 to 0.760.After adjustment,the average underestimation coefficient for the three major grain crops nationwide was between 20%and 40%,indicating that the average degree of risk underestimation using county-level yield nationwide is between 20%and 40%,with maize having a higher underestimation coefficient than rice and wheat.At the provincial level,the production risks of the three crops in Heilongjiang were all at an extremely high level.The production risks of rice and wheat in Inner Mongolia,rice

关 键 词：农业保险 生产风险 风险评估 多源数据 三大主粮 纯风险损失率 风险地图 

分 类 号：F326.11[经济管理—产业经济]