I am currently working with MEPS data for 2015/2016, panel 20. I have already combined the two datasets and trimmed them down so that the dataset only contains individuals from panel 20. I would like to calculate the proportion of individuals who lose and regain insurance coverage during the period that they are followed and the mean/median time that they spent without insurance. I’m currently playing with various combinations of INS(month)15X and INS(month)16X but would appreciate any tips or potential paths to go down.
What you’ve described sounds like a reasonable approach. There is a full set of monthly health insurance variables that you could explore. While INSMMX will capture any health insurance, perhaps you could uncover distinct variations/patterns and/or movements through different types of insurance by using more specific variables.