My endogenous variable is a dummy variable indicating foreign director presence (Fdummy)
Regressing this with firm performance and relevant covariates has been giving coefficient values positive around 0.5.
This would translate to a 0.5 percentage point improvement to firm performance measured by Return on Assets.
So far this has been robust to comparing between fixed effects and pooled OLS models.
However using instrumental variables in 2SLS regression has been giving strange results.
Not only has the sign switched to negative (Not implausible) but the coefficient is now around -20
Surely this cannot be correct?
How should I interpret this result, surely it does not infer foreign director presence causes a fall in ROA by 20 percentage points
Is there anything that I can do to sort this
Regressing this with firm performance and relevant covariates has been giving coefficient values positive around 0.5.
This would translate to a 0.5 percentage point improvement to firm performance measured by Return on Assets.
So far this has been robust to comparing between fixed effects and pooled OLS models.
However using instrumental variables in 2SLS regression has been giving strange results.
Not only has the sign switched to negative (Not implausible) but the coefficient is now around -20
Surely this cannot be correct?
How should I interpret this result, surely it does not infer foreign director presence causes a fall in ROA by 20 percentage points
Is there anything that I can do to sort this
