We decompose permanent earnings risk into contributions from hours and wage shocks. In order to distinguish between hours shocks and labor supply reactions to wage shocks we use a life-cycle model of consumption and labor supply. Estimating our model with the Panel Study of Income Dynamics (PSID) shows that both permanent wage and hours shocks play an important role in explaining the cross-sectional variance in earnings growth, but wage risk has greater relevance. Allowing for hours shocks improves the model fit considerably. The empirical strategy allows for the estimation of the Marshallian labor supply elasticity without the use of consumption or asset data. We find this elasticity on average to be negative, but small. The degree of consumption insurance implied by our results is in line with recent estimates in the literature.