Health insurance is a major component of human welfare and a central pillar of all social insurance and welfare systems. Health is considered a central element of the Human Development Index and of other indices that measure the development state of a nation. Besides the direct impact on wellbeing, ill health also reduces labour productivity and income and can lead to poverty or vulnerability. Many countries have therefore developed insurance systems to ameliorate the negative consequences of health shocks and disability. In this project we estimate the impacts of health insurance reforms and interventions in countries at different development stages: high, middle and low-income countries. The evaluation results will provide evidence that can help to improve the design of insurance systems. Examining the effects across different countries will improve our understanding about heterogeneity in human behaviour across countries and cultures.
In the high-income countries we study well-established health insurance systems and evaluate policy reforms that aim at improving efficiency or at adjustments to demographic and technological changes. The link to middle and low-income countries comes from the fact that developing countries try to learn from experiences in high-income countries and implement health insurance systems that have proven successful elsewhere. We are going to focus on countries with underdeveloped health insurance systems, which effectively only cover a small fraction of the population. Different health insurance policy settings will be experimentally evaluated in these countries at a small scale. From a comparison of evaluation results across different countries we can learn which types of health insurance policies are transferable from high to lower-income countries, how cultural backgrounds interact with incentives generated by insurance coverage schemes, and what problems might arise during the implementation and adoption of healthcare policies in developing countries.
Our project is mostly empirical and based on the analysis of micro-data of individuals or households with state-of-the-art econometric methods. Since reforms of healthcare systems do not happen at random, we need to rely on strong identification designs in order to provide credible estimates of causal effects. We will focus on two identification strategies. In countries where reliable and informative high-quality administrative data systems exist, we will use non-experimental approaches mostly relying on regression discontinuities to estimate impacts of reforms. For most lower income countries, reliable administrative data sources do not exist, such that we will use experimental approaches, where we implement randomised controlled trials (RCT) together with local partner organisations in middle and low-income countries.
As examples of high-income countries, we will focus particularly on Austria and Germany. This part of the project starts with examining the labour market responses to incentives in the health insurance system, exploiting several reforms of health and disability insurance. Both countries enjoy contributions-financed health insurance systems and have embarked on several comprehensive reforms in recent years. Discontinuities in the institutional rules and changes due to policy reforms give rise to credible identification designs that allow us to exploit quasi-experimental settings. We are planning to base our analysis on high-quality matched employer-employee records covering the universe of private sector workers. A key component part is the analysis of heterogeneity in responses and impacts across the population and across causal channels. In the longer run, we also plan to study the health-related effects of reforms in the health insurance and disability programmes. But the access to confidential data on health expenditures has to be secured first. To complement the evidence on behavioural responses to the incentives in health insurance systems based on administrative register data, we will study reform-related attitudes and perceptions in the population on the basis of survey data. We will draw on information in the German Internet Panel (GIP) about subjective impressions about the general designs of health insurance, preferences for fairness, or perceptions of health risks.
In low and middle-income countries, health insurance systems had often been introduced on the basis of the Bismarckian model, where social security contributions are deducted from wages and salaries. This contributions-financed health insurance, however, covers only the formal sector such that in countries with large informal sectors a large fraction of the population is uninsured or has access only to poor services. Providing better insurance against health shocks is important. We will study alternative insurance products or insurance systems and estimate behavioural responses and impacts on labour, health and poverty outcomes. Insurance cover will be extended to the informal sector and combined with modifications of insured risks such as the incorporation of sick pay, preventive healthcare or restrictions to catastrophic health risks only. In low-income countries, administrative health insurance data is often of very poor quality such that we will make use of randomised controlled trials (RCT) to estimate the responses to and impacts of insurance. In order to analyse the sustainability of the effects, long-term follow-up data collections will be done.