Hello everyone, I am currently facing some difficulties with an approach. I’m using the technique proposed by [Balli, H. O., & Sørensen, B. E. (2013). Interaction effects in econometrics. Empirical Economics, 45(1), 583–603] to estimate interaction effects with an instrumental variable (IV). However, I’m applying Lewbel’s approach.
First, I run: reg performance $covlist distress
hen I create an interaction between telework and distress which gives me: Dtel = telework * distress
Then I re-estimate the regression using: ivreg2h performance $covlist (PD PDtel = I), ro
However, I find that the coefficients of PD in both cases are nearly identical. What solutions would you suggest, please?
First, I run: reg performance $covlist distress
hen I create an interaction between telework and distress which gives me: Dtel = telework * distress
Then I re-estimate the regression using: ivreg2h performance $covlist (PD PDtel = I), ro
However, I find that the coefficients of PD in both cases are nearly identical. What solutions would you suggest, please?
