Examining News Shocks

Diploma Thesis by Henning Weber
December 29, 2003

Abstract:

A fundamental discrepancy surrounds news shocks, that is, a-priori information agents receive about developments in the economy. Whereas news shocks appear relevant empirically, in theoretical real-business-cycle models news about future total factor productivity generates counterfactual predictions. In standard real-business-cycle models these predictions emerge on top of the failures in this class of models, already known in the literature.

This coincidence of empirical evidence and theoretical failure motivates the search for a structure of real-business-cycle models that makes news shocks work. In these models, when good news is indicated, agents anticipate today the positive income effect of tomorrow. When news prove correct and total factor productivity eventually increases, a strong substitution effect occurs. The sequence of these two effects is what causes the adverse consequences of news shocks. The analysis indicates that the main feature responsible for the failure of standard models is the high substitutability of consumption and investment.

However, evaluating a non-standard three-sector real-business-cycle model shows that a low substitutability of consumption and investment is not the only element needed to fix the adverse consequences of news shocks. The results are robust with respect to assumptions that vary the exact means of how new information enters the model. In light of the former results, this robustness suggests that the real structure of the models, as opposed to the informational structure, is the crucial part determining success and failure of news shocks.

This thesis shows that there is no easy way of formalizing news shocks. Nevertheless, it provides a discussion of elements and their combination that appear relevant for guiding future research. Appropriate frictions in the factor markets in combination with a multiple sector model appear promising. Furthermore, one suggestion originally proposed in Cochrane (1994) and refreshed here is to construct models that disseminate news about candidates of shocks inducing wealth effects, but no intertemporal substitution.