Contrary to simple theoretical predictions, existing evidence suggests that federal grants do not crowd out state government spending. A legislative bargaining model with endogenous grants documents a positive correlation between grant receipts and preferences for public goods; this correlation has likely biased existing work against measuring crowd-out. To correct for such endogeneity, the model motivates instruments based on the political power of state congressional delegations. Exploiting this exogenous variation in grants, the instrumental variables estimator reports crowd-out that is statistically and economically significant. This endogeneity may explain the flypaper effect, a nonequivalence between grant receipts and private income.