Testing for excess sensitivity in consumption: A state-space unobserved components approach

This paper tests the joint rational expectations/permanent income hypothesis using a state-space/unobserved components model to specify anticipated and unanticipated elements of permanent, transitory, and seasonal components of the income path and measure their relationships with consumption. The results do not find significant excess sensitivity. However, a similar test that accounts for seasonality using X11-adjusted data does, thus indicating that seasonal filtering may be responsible for previous empirical rejections of the theory. Since the model includes consumption information in estimating the different components of income, it also improves upon univariate investigations of the relative magnitudes of permanent and transitory disturbances to the income path.