"Using data on workers' flows into and out of employment, unemployment, and not-in-the-labor-force, I construct transition probabilities between "employment" and "unemployment" that can be used in the calibration of economies such as Krusell and Smith's (1998). I show that calibration in Krusell and Smith has some counterfactual features. Yet the gains from adopting alternative calibrations in terms of matching the data are not very large, unless one assumes that the duration of unemployment spells is well above what is usually assumed in the literature"--Federal Reserve Bank of Atlanta web site.