Permanent income and transitory variation in investment and output

We use the permanent income hypothesis as the framework to analyze a number of results from recent empirical macroeconomic research. First, we demonstrate that the "productivity" shock isolated in both the three- and six-variable models of King et al. [Am. Econ. Rev. 81 (1991) 819] depend mainly on the time-series behavior of the reduced form consumption residual. A permanent income hypothesis interpretation of this finding is that consumption fully reflects the implications of long-lived shocks for the common stochastic trend in consumption, investment and output. Further, for the three-variable model, shocks to consumption have permanent effects on the levels of the series while shocks to investment and output have only transitory effects, given that consumption is ordered first in a causal ordering. The permanent income interpretation is that shocks to productivity are changes in permanent income, and hence, are fully reflected through consumption decisions. From this perspective, consumption shocks are proxies for the changes in permanent income generated from the "productivity" shocks. The results complement the findings in Hall [J. Polit. Economy 96 (1978) 339], Campbell [Econometrica 55 (1987) 1249] and generalizes Cochrane's [Quart. J. Econ. 109 (1994) 241] analysis to a model that includes investment.