The permanent income hypothesis under permanent-transitory confusion

This paper empirically examines the Permanent Income Hypothesis, relaxing the typical assumption that households immediately determine the duration of any income shock. Instead, permanent-transitory confusion is imposed, and a multi-state Kalman filter is used to estimate the permanent and transitory components of income shocks. It is shown that tests of the PIH are sensitive to assumptions made concerning households information and that, in general, imposing permanent-transitory confusion results in a superior specification.