For those working with a panel data set noted for its "large N, small T," a recent discussion here (from the StataList) might help save some time when it comes to working through challenges posed by serialcorrelation and heteroskedasticity. While these issues are especially common in corp finance and securities papers, as the latter comments in the discussion imply the magnitude of the problem (for 'short panel' data sets) remains somewhat of a contested point.
Comments