Abstract
Recent empirical evidence supports the view that the income process has an individual-specific growth rate component [Baker (1997), Guvenen (2007b), and Huggett, Ventura, and Yaron (2007)]. Moreover, the individual-specific growth component may be stochastic. Motivated by these empirical observations, I study an individual's optimal consumption-saving and portfolio choice problem when he does not observe his income growth. As in standard income fluctuation problems, the individual cannot fully insure himself against income shocks. In addition to the standard income-risk-induced precautionary saving demand, the individual also has learning-induced precautionary saving demand, which is greater when belief is more uncertain. With constant unobserved income growth, changes in belief are not predictable. However, with stationary stochastic income growth, belief is no longer a martingale. Mean reversion of belief reduces hedging demand on average and in turn mitigates the impact of estimation risk on consumption-saving and portfolio decisions.
The final version of this article can be found at http://dx.doi.org/10.1016/j.jmoneco.2009.03.005.