Context

Given the lack of availability of the randomised control trial (RCT) data on the impact of the JKN program, we are exploiting the rich nation-wide socio-economic household survey to establish a quasi-experiment study.

One key challenge we face is how to identify the multi-valued treatments owing to the multiple types of insurance available and in use by individual household members.

We have employed a rigorous strategy in selecting the insurance beneficiaries to be evaluated, as well as improved the pseudo-panel approach to ensure we have an efficient estimator when analysing the independent cross-section survey data.

We are also working to define a favour-causal relationship methodology owing to the data limitation.