Unstructured data file and HAC robust standard errors
Posted: Thu Nov 02, 2017 8:26 am
Hello
I have created an Unstructured data file to estimate the regressions where 250 firms have several variables' observations across several years, however each firm has a different number of years for which data are available, hence, unstructured file. Does it make much sense to estimate an OLS regression using HAC robust standard errors in the case of the unstructured, rather than balanced, file? Autocorrelation, for which HAC corrects, assumes no residuals correlation over time. But with an unstructured data file, I am not specifying any time periods, so how does Eviews then performs such a correction? Thank you all!!
I have created an Unstructured data file to estimate the regressions where 250 firms have several variables' observations across several years, however each firm has a different number of years for which data are available, hence, unstructured file. Does it make much sense to estimate an OLS regression using HAC robust standard errors in the case of the unstructured, rather than balanced, file? Autocorrelation, for which HAC corrects, assumes no residuals correlation over time. But with an unstructured data file, I am not specifying any time periods, so how does Eviews then performs such a correction? Thank you all!!