医疗保险、健康异质性与精准脱贫——基于贫困脆弱性的分析

医疗保险能否有效减少贫困,尚未达成共识。本文利用CHARLS全国调查数据,基于资产度量个体的贫困脆弱性(即陷贫概率),实证分析医疗保险参保行为及其保障水平对贫困脆弱性的影响。研究表明,参与医疗保险和提高医疗保险保障水平的减贫效果都存在健康异质性,对健康状况差的群体有显著的减贫效果,而对健康状况好的群体减贫效果不显著。此外,本文证实了改善劳动供给是医疗保险减贫的重要中介渠道之一。

Health Insurance,Health Heterogeneity,and Targeted Poverty Reduction: A Vulnerability to Poverty Approach

Disease is the primary cause of poverty in China. Following the implementation of the Chinese poverty reduction program, the fraction of diseases caused by poverty increased from 42.2% of the impoverished population in 2013 to 44.1% in 2015, which corresponds to nearly 20 million people. Worldwide, approximately 100 million people fall into poverty as a result of medical expenditure every year, and 150 million suffer from the economic shock of medical expenses. Thus, disease related poverty is not only a primary concern of China but also a global challenge.
Health insurance is an important tool for managing health risks and mitigating the economic shock of disease. It is a widely accepted hypothesis that health insurance alleviates the poverty problem. However, the empirical evidence on the effectiveness of health insurance in alleviating poverty remains inconsistent. Some research suggests that health insurance has a significant effect in reducing poverty, and that significance increases over time. Other research suggests that this effect is insignificant or minimal because of a low degree of coverage, inaccurate targeting of groups, and the long and complex process of making claims. In this paper, we examine this issue from a new perspective, i.e., the heterogeneity of health status, to explain the inconsistent poverty reducing effects of health insurance.
We address the following two research questions. First, we examine whether the impact of health insurance on poverty reduction depends on health status by focusing on the change from having no health insurance to having health insurance. Second, we examine whether the impact of increased health insurance coverage on poverty reduction depends on health status by focusing on the change from low health insurance coverage to high coverage.
Using data from the China Health and Retirement Longitudinal Study (CHARLS) in 2011, 2013, and 2015, we construct an asset-based “vulnerability to poverty” index (i.e., the probability of poverty) at the individual level. Next, we use propensity score matching with difference-in-difference (PSMDD) to estimate the effect of health insurance participation on the poverty reduction in different health groups. We also use a multivariate regression to estimate the impact of health insurance coverage on poverty reduction, and find that an individual's health status moderates the relationship between health insurance and poverty reduction. Participation in health insurance and increasing the health insurance coverage reduce the levels of poverty among poor-health individuals. However, this poverty reduction effect is insignificant for good-health individuals. In addition, we further analyze the mediating effects of the labor supply and medical expenses on poverty reduction, and find that health insurance reduces poverty through improving the labor supply.
This paper makes the following contributions to the literature. First, we theoretically and empirically fill the research gap on whether health insurance has different poverty reduction effects on populations with different health status. Second, the literature focuses on analyzing the impact of health insurance participation, i.e., the treatment effect of having health insurance compared with no health insurance. We extend this analysis by examining the impact of the degree of health insurance coverage, considering the over 95% health insurance coverage rate in China. Third, we examine the mediating effects of the labor supply and medical expenditure on the relationship between health insurance and poverty reduction. We provide the first evidence showing how health insurance reduces poverty through improving the labor supply.
Our empirical analyses show that participation in health insurance and improving the degree of coverage have significant effects in reducing poverty under certain conditions, namely, a relatively long period of insurance coverage and targeting people in need (i.e., those with poor health). Considering the limited resources available for poverty reduction programs, we suggest that more resources should be allocated to the unhealthy poor. Policy makers should consider the heterogeneity in the health status of the population and consider offering different health status groups different reimbursement rates.
Our paper provides several directions for future research. First, more empirical tests are needed to determine whether different insurance products have different effects in reducing poverty. Second, because progressive coverage may create a moral hazard, researchers should develop more effective approaches to control for the moral hazard in poverty reducing insurance products.