We apply a new methodology for identifying pervasive and discrete changes (``breaks'') in cross-sectional risk premia and find empirical evidence that these are economically important for understanding returns on US stocks. Size and value risk premia have fallen off to the point where they are...