Quantitative methods for monetary and fiscal policy

Princeton University, Princeton NJ

Investigators

Abstract

One component of this project will explore the possibility of explaining the nature of delay and inertia in economic behavior by importing into standard dynamic economic theory from communications engineering the notion of the Shannon capacity of an information channel. This has already been shown to imply realistic modifications of the predictions of the permanent income theory, and appears ripe for further extension. Another component will improve and apply computer code for finding Taylor series expansions of solutions to dynamic stochastic equilibrium models. Such models are seeing increasing use in central bank policy analysis, so that wider availability of these methods will find quick application to important problems. There is a dispute, relevant to the evaluation of reform proposals for our monetary institutions, over how much and in what way monetary policy has changed in the US over the course of the twentieth century. A component of this research project will apply Bayesian model comparison methods to look further into the sources of conflicting conclusions in the literature on these issues. Broader Impact Recent developments in econometric methodology and in its connection to decision theory have made it practical to take new approaches to quantitative modeling for monetary policy analysis. The local expansion programming and the Bayesian time-varying parameters and model choice methods in other components of this proposal are likely to find application in the central bank revisions of their methods. One component of this project is to continue existing contacts of the investigator with central bank research departments, both to disseminate the results of this project's research and to learn more about the analytic bottlenecks in the policy process.