湖南省种植业保险分类定价研究
发布时间：2018-08-14 10:27
【摘要】：农业是国民经济的基础，，因此做好农业相关工作在我国具有非常重要的社会意义。农业保险有助于稳定农户收入，迅速恢复农业生产，实现国家对农业产业的宏观调控，是我国发展农业的重要工具之一。然而由于我国政策性农业保险开展工作的时间较短，有效的承保数据和损失数据非常有限，农业保险费率的确定缺乏精准有效的技术支持。试点初期出于谨慎考虑，只能在全国采取统一费率。虽然易于操作，却带来了道德风险和逆向选择问题。根据精算理论，如果承保对象足够多，就可以根据风险特征的不同对个体进行风险分类，在此基础上厘定各个风险等级的费率，消除道德风险和逆向选择。 本文通过阐述并对比了目前比较流行的几种风险分类的费率厘定方法，例如最小偏差法、非参数核密度法、分层贝叶斯模型以及广义线性模型。通过结合湖南省政策性农业保险关于种植业保险的经营数据和气象数据特征选择了广义线性混合模型对数据建模，认为广义线性混合模型基于指数组分布，可以对非正态分布的数据进行回归分析，但它又保留了很多正态线性回归模型的常规思想，将复杂且相关的数据通过随机效应进行拟合。同时结合广义线性模型做实证分析，对湖南省种植业保险中棉花、油菜和水稻品种进行费率分类定价研究，通过实证分析得出广义线性混合模型中湖南省种植业保险费率厘定的结果。最后对在湖南省具体实施这种定价方法过程中，要求农业保险市场需要进行完善和改进的地方，提出了一些政策性建议。
[Abstract]:Agriculture is the foundation of the national economy, so it is of great social significance to do the agricultural work well in our country. Agricultural insurance is one of the important tools for the development of agriculture in China, which is helpful to stabilize the income of farmers, restore agricultural production quickly, and realize the macro-control of agricultural industry by the state. However, due to the short time of policy agricultural insurance in our country, the effective underwriting data and loss data are very limited, and the determination of agricultural insurance rate lacks accurate and effective technical support. Pilot early out of careful consideration, can only be adopted in the national flat rate. Although easy to operate, it brings moral hazard and adverse selection problems. According to actuarial theory, if there are enough insured objects, individuals can be classified according to different risk characteristics, and on this basis, the rate of each risk grade can be determined, and moral hazard and adverse selection can be eliminated. In this paper, several popular methods of rate determination for risk classification, such as minimum deviation method, nonparametric kernel density method, hierarchical Bayesian model and generalized linear model, are described and compared. Based on the management data and meteorological data characteristics of policy agricultural insurance in Hunan Province, the generalized linear mixed model is selected to model the data. It is considered that the generalized linear mixed model is based on exponential group distribution. Regression analysis of non-normal distribution data can be carried out, but it retains the conventional idea of many normal linear regression models, and the complex and relevant data are fitted by random effect. At the same time, with the generalized linear model to do empirical analysis, Hunan Province planting insurance in cotton, rape and rice varieties of the rate classification pricing research, Based on the empirical analysis, the results of determining the insurance rate of Hunan planting industry in the generalized linear mixed model are obtained. Finally, some policy suggestions are put forward in the course of implementing this pricing method in Hunan Province, which require the agricultural insurance market to be perfected and improved.
【学位授予单位】：湖南大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：F842.66
本文编号：2182565
[Abstract]:Agriculture is the foundation of the national economy, so it is of great social significance to do the agricultural work well in our country. Agricultural insurance is one of the important tools for the development of agriculture in China, which is helpful to stabilize the income of farmers, restore agricultural production quickly, and realize the macro-control of agricultural industry by the state. However, due to the short time of policy agricultural insurance in our country, the effective underwriting data and loss data are very limited, and the determination of agricultural insurance rate lacks accurate and effective technical support. Pilot early out of careful consideration, can only be adopted in the national flat rate. Although easy to operate, it brings moral hazard and adverse selection problems. According to actuarial theory, if there are enough insured objects, individuals can be classified according to different risk characteristics, and on this basis, the rate of each risk grade can be determined, and moral hazard and adverse selection can be eliminated. In this paper, several popular methods of rate determination for risk classification, such as minimum deviation method, nonparametric kernel density method, hierarchical Bayesian model and generalized linear model, are described and compared. Based on the management data and meteorological data characteristics of policy agricultural insurance in Hunan Province, the generalized linear mixed model is selected to model the data. It is considered that the generalized linear mixed model is based on exponential group distribution. Regression analysis of non-normal distribution data can be carried out, but it retains the conventional idea of many normal linear regression models, and the complex and relevant data are fitted by random effect. At the same time, with the generalized linear model to do empirical analysis, Hunan Province planting insurance in cotton, rape and rice varieties of the rate classification pricing research, Based on the empirical analysis, the results of determining the insurance rate of Hunan planting industry in the generalized linear mixed model are obtained. Finally, some policy suggestions are put forward in the course of implementing this pricing method in Hunan Province, which require the agricultural insurance market to be perfected and improved.
【学位授予单位】：湖南大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：F842.66
【相似文献】
相关期刊论文 前1条
1 徐松;陈丹丹;;论我国财产保险业的分类监管[J];中国新技术新产品;2009年05期
相关重要报纸文章 前3条
1 罗三秀;保险分类监管新制度将于明年开始运行[N];证券日报;2008年
2 记者 关媛媛;科技保险分类补贴标准出台[N];重庆日报;2008年
3 ;保监会：明年将全面推进保险分类监管[N];上海证券报;2008年
相关硕士学位论文 前1条
1 郭美君;湖南省种植业保险分类定价研究[D];湖南大学;2014年
本文编号：2182565
本文链接：https://www.wllwen.com/jingjilunwen/bxjjlw/2182565.html
