Hi there,

For an assignment I have to examine the causal effect of working part-time versus working full time on health. The idea of the assignment is to explain why you choose a certain econometric technique, how you deal with potential biases, etc.

I have a panel data set with variables like gender, age, civil status, number of kids, education, health, disability, to what extent health hinders work, having paid employment, weekly hours of paid employment, and survey year. The sample consists of individuals aged 25-54.


I was wondering how I should deal with reverse causality. Someone's health could influence the amount of hours worked (so not only working hours -> health, but also health -> working hours).

I read multiple times about using a FE with IV. However, I don't think I have an appropriate IV in my dataset. Is there another way to correct for reverse causality with panel data? My professor prefers simple techniques over advanced ones, so the "easier" the better.

I read some articles on using lagged variables, but I don't quite understand how to implement that to correct for this.
Also, maybe I should do something with the variable that indicates to what extent health problems hinder work. I could see how this would say something about health affecting working hours.



Thank you in advance (: