We measure the length of time it takes for naturalized citizens to participate in U.S. elections at the same rate that native-born voters participate.  Using voter turnout data for U.S. presidential and midterm elections from 1996 through 2006, we show that additional years since immigration do not increase naturalized citizen turnout relative to native turnout, all else being equal.  These results survive controlling for demographic and election characteristics, using different estimators, and correcting for bias from immigrant self-selection into citizenship.  Our findings suggest that unobserved heterogeneity correlated with an immigrant's year of arrival explains all of the variation in turnout among naturalized citizens living in the United States for different amounts of time.