This paper develops a version of the Permanent Income Hypothesis in which permanent and transitory components of consumption and labor income are explicitly accounted for. The model is used to derive a restricted vector autoregressive representation of adjusted measures of consumption and saving, which is used to test the theory and to study the dynamic effects of the two • components of labor income on consumption. We find that the restrictions on the VAR are not easily rejected for quarterly post-war U.S. data. An analysis of the restricted VAR leads us to conclude that consumption can be almost entirely explained in terms of the permanent component of labor income.