According to the “welfare magnet” hypothesis, migrants with a high likelihood of claiming benefits cluster in the most generous welfare systems. After the introduction of the freedom of movement for Eastern European workers, EU-15 countries can thus be expected to reduce public benefits in order to avoid becoming “welfare magnets”. However, OECD data on benefits do not support the prediction of a race to the bottom in protection levels. Using data from the EU-LFS 2004 to 2011, I analyze the determinants of migration flows and find that, in contrast to theory, welfare state variables do not significantly affect migration flows when controlling for temporary political restrictions of the freedom of movement (2+3+2 rule). This explains why the pressure to modify national welfare spending is small. Furthermore, evidence is found that the restrictions completely offset the incentive effects of work-related pull factors and thereby hamper an efficient allocation of labor across national borders.