Despite some obvious comparative advantages, panel data sets invite traditional complications as well. These complications include (but are not limited to) heteroskedasticity, autocorrelation, and cross-sectional dependence issues. How to manage such complications can be more “art” rather than “science,” and