《保险研究》20190902-《保险发展与经济增长的S曲线成立吗？——来自87个经济体的证据》（段白鸽）

［摘   要］保险发展与经济增长之间的S曲线关系是一个重要假说,但在经验上是否成立一直存在争议。本文利用1999～2017年87个经济体保险市场与宏观经济指标数据,将数据质量评估、保险业务结构融入经典的S曲线中,通过构建“一致性”分层模型,采用稳健的实证方法检验S曲线假说,研究发现,假说成立。具体而言,全球寿险、非寿险和总保险市场发展与经济增长之间的S曲线并非估计偏差,而是具有规律性的事实；保险需求收入弹性与经济发展程度显著负相关,且寿险需求收入弹性明显高于非寿险,这与寿险市场更易受外部环境和经济周期影响一致；不同经济体保险需求收入弹性存在明显下行和收敛趋势；各经济体非寿险市场发展趋势表现出比寿险市场更大的一致性。S曲线的稳健估计有助于改进保险深度和保险密度的长期预测,对确立保险市场发展阶段及其与经济增长的关系具有重要启示。

［关键词］保险发展；S曲线假说；分层模型；数据质量评估；保险需求收入弹性

［基金项目］本文获得国家自然科学基金项目(71490734,71673058)、上海市社科规划课题“大数据背景下基于机器学习的长寿与健康风险精算预测建模研究”、中国保险学会教保人身保险高校课题研究基金“低利率环境下我国寿险公司资产负债管理研究”的资助。

［作者简介］段白鸽,复旦大学经济学院风险管理与保险学系副主任、讲师、博士,中国准精算师,研究方向:保险经济学、统计精算与定量风险管理,E-mail:duanbaige@fudan.edu.cn.

Is the S-curve between Insurance Development and Economic Growth Tenable?—Empirical Evidences from 87 Economies

DUAN Bai-ge

Abstract：The S-curve relationship between insurance development and economic growth is an important hypothesis,but it is debatable empirically.The paper used the data of insurance markets and macroeconomic indicators of 87 economies from 1999 to 2017,incorporated data quality assessment and insurance business structure into the classical S-curve,and constructed a consistent hierarchical model to test the S-curve hypothesis with a robust empirical method.The results prove the robustness of the S-curve hypothesis.Specifically,the S-curves between global life insurance,non-life insurance and total insurance market development and economic growth aren’t an estimation deviation but a regular fact.The income elasticity of insurance demand has a remarkable negative correlation with economic development,and the elasticity of life insurance demand is significantly higher than that of non-life insurance,which is consistent with the fact that the life insurance market is more susceptible to external environment and economic cycles.The income elasticity of insurance demand of different economies has a clear downward and convergence trend.The development trend of non-life insurance market in different economies shows greater consistency than that in the life insurance market.The robust estimation of S-curve is helpful to improve the long-term forecast of both insurance penetration and insurance density,and has important enlightenment for determining the stage of insurance market development and its relationship with economic growth.

Key words：insurance development;S-curve hypothesis;hierarchical models;data quality assessment;income elasticity of insurance demand