Question

"The Impact of Health Insurance on Health," Levy and Meltzer pose a question – "Why is it so difficult to estimate the effects of health insurance on health?" The article goes on to state, "…the problem is that health insurance is an endogenous variable." Briefly describe what is meant by the statement, "health insurance is an endogenous variable." Why is the causal impact of health insurance on health an important/relevant question for policymakers?

Levy and Meltzer summarize the literature on the impact of health insurance on health, and in doing so, they divide studies into three “groups,” observational studies, natural experiments, and experimental studies. What are natural experiments? How is this different from an observational study or an experimental study? What is a limitation associated with natural experiments?

Answer 1

Endogenous variables are similar to dependent variables where its value is determined by other variables which are exogenous (independent). Now as per the article health insurance is considered to be endogenous variable, health insurance is dependent on various other factors which are beyond control and depends on several factors such as the level of income, the availability of insurance services, the level of savings, the premiums charged, etc. It might be also the case that only the rich opt for health insurance, or only those who want to be healthy opt for insurance, which could drastically change the results. An individual opts for health insurance because of a number of myraid reasons which ultimately determines the health of an individual.

The causal impact of health insurance on health is an important question for policymakers because they want to determine the efficacy of health insurance, of whether they truly impact the health status of the society, in order to broaden the scope of use and implement effective measures to increase health in a society. Policymakers are ultimately focused on improving health conditions and they want to see whether health insurance could remedy that.

Natural experiments are experiments which are undertaken where naturally a control group arises, for example when a policy of health insurance is changed and one can determine the effects of this change. Observational study on the other hand is when a researcher does not control any variable and merely observes the data, such as number of females and males opting for health insurance. Experimental study on the other hand is when variables are controlled by the researcher, for example health insurance cover is randomly given to individuals while the other group is not.

Limitation associated with natural experiment is that it might rarely occur which leaves less room for experimentation. All variables are not able to be controlled by the researcher. Correlations might occur based on assumptions.