By using a beginning-of-period timing convention for consumption, and by including the Great Depression years in the analysis, we show that on annual data from 1926 to 2009 a standard contemporaneous consumption risk model goes a long way in explaining the size and value premiums in cross-sectional data that include both the Fama-French portfolios and industry portfolios. A long run consumption risk variant of the model also produces a high cross-sectional …t. In addition, the equity premium puzzle is signi…cantly reduced in the models. We argue that in evaluating consumption based models, it is important to include both boom and crises periods, i.e. periods with severe consumption declines as well as periods with strong growth, and that the standard post-war data sample may not be well suited in this respect.