Hi hope everyone is well
I recently read a paper that has confused me a little about the difference between two stages least squares and the fixed effects within estimator.
I also wanted to ask what people think would be best to use for my data.
I am running a regression
logarithmic differences of sales (sales growth) on rd intensity. I have lagged RD intensity by taking the average of RD intensity from 2 years and 3 years prior. This is because common study says RD takes 2 years to show up. But it also solves some endogeneity issues (reversed causality) does this mean I could use it as an instrument in 2SLS? Or shall I just stick with fixed effects within.
Thank you for your time