Have you ever wished that Romer and Romer's (2010) identification method for fiscal multipliers was expanded to multiple countries and incorporated both spending and tax policy? Well now you have it (HT - Bruce Bartlett). I've read the first several pages and skimmed the rest. This provides a great overview of the macroeconometric identification problems I was discussing earlier, and presents results for both the corrected and uncorrected estimates to demonstrate the endogeneity bias. The results? Fiscal contraction is contractionary and fiscal expansion is expansionary. Keep in mind, of course, that while this corrects for endogeneity bias it mushes together periods we would expect to be more and less expansionary. I can't get the picture to copy, but there's a nice illustration on page 13 of the extent of the bias.