Why is there endogeneity in this specification?

  • Level: Graduate 
  • Thread starter operationsres
  • Start date
1 reply · 2K views · last reply
operationsres
Messages
99
Reaction score
0
Hi guys,

I am trying to learn why there is endogeneity in the following OLS estimated regression model:


Regression equation (1): ROA at t+2 minus ROA at t equals alpha plus beta times (Leverage at t minus Leverage at t-1) plus epsilon, followed by a note that ROA at t is Return on Assets at time t.



A big part would be because the differenced Leverage independent variable is correlated with the error term, largely because of omitted variables.

But I'm looking for reasons as to why the way the model is specified is conducive to endogeneity?

Thanks.
 
Hi operationsres

I'd say another reason might be an autoregressive behavior between the leverage and the returns.