My name is Valentina and I am a Master student, currently working on my MSc thesis. I am examining the effect of Politicians on the board of directors (IV) on corporate risk-taking behavior (DV). My time frame is 2005-2010, and I have 51 companies in the sample. I am examining the differences in the risk-taking level of companies that are politically connected compared to the ones that are not.

My data set includes the following variables: 1) IV: Politicians on the board (dummy: Yes/No), 2) DV: Risk-taking, 3) Control variables: firm size (log total assets), regulation of industry and state-ownership (both dummies). The problem is that my IV and all control variables are dynamic and change with both Time and Company. However, Risk-taking is the deviation of the firmís EBITDA/Total Assets from the country average for the corresponding year. Then, the standard deviation of this measure is calculated for each firm. Hence, each company has only ONE risk-taking variable.

I was initially aiming for a fixed-effects panel data regression model but this is not working anymore. Can you suggest something else to cope with this issue?

Any tips and help are greatly appreciated!