Hi everyone, I am a MSc's student and I am trying to figure out wether Proxy advisors presence (advice provided) in a company could have a positive impact in corporate performance. Dependent variables are example: Number of advice received in a year by a company, number of advice for specific item on agenda ( I have 5 categories of vote for instance M&A, CSR etc.). I have several control variables and I tried running OLS model having in turn as dependent variable (ROA, Tobin's Q , IROA (roa company - avg industry roa), ITobin'sQ) and in turn the independent variable selected. Furthermore I have severa control variables such as Firm size, %institutional shareholders etc.
I found an issue running OLS, Heteroskedasticity, solved using robust std errors. Then The other issue is Endogeneity:
I tried running IV 2sls models having as instrument variable (or lagged values of ROA/Tobin'sQ or average industry values) But since I have several independent variables to be tested once per time in a dedicated model, in several models instruments are weak and literature not provide instruments since there is almost no articles studying impact on performance of Proxy advisors. How could I treat endogeneity issue? What do you suggest to provide about database/models for your understanding?
I found an issue running OLS, Heteroskedasticity, solved using robust std errors. Then The other issue is Endogeneity:
I tried running IV 2sls models having as instrument variable (or lagged values of ROA/Tobin'sQ or average industry values) But since I have several independent variables to be tested once per time in a dedicated model, in several models instruments are weak and literature not provide instruments since there is almost no articles studying impact on performance of Proxy advisors. How could I treat endogeneity issue? What do you suggest to provide about database/models for your understanding?
