Dear Stata Pros ,

Being interested in the effect of the share of women in board position on return on assets of a company.
Running a regression with adding the following control variables:
Dummies for industry, region, and year, log total assets, the size of the board, demographics of the workforce (like age) and of the board members.

My question:

Why (and in which direction) could the coeffcient on the share of female board members may be biased (Distinguishing between time-varying and time-invariant unobservables)
What do you think?


Thank you very much for your help!

Kind regards
Mimi