I am running a regression in E-views with cross-section data for a sample of 314 companies to check the factors that affect the abnormal return in the M&A deals.
The regression is
CAR ROE_ACQUIRER ROE_TARGET TARGET_DISTRESS PAYMENT_METHOD
where CAR is the Cumulative Abnormal Return calculated through event study methodology. And Target distress, Payment method are dummy variables
The problem is that I have both autocorrelation and heteroscedasticity problems and negative data that I cannot use the log and all my variables are not significant except the target distress
My question is, how can I deal with the autocorrelation and heteroscedasticity with other methods, other than the log and the first difference.
Many thanks, Budor
