Stochastic Fiscal Projections with Cyclical and Long-Term Uncertainty

Several recent efforts to improve budget forecasting techniques have emphasized the uncertainties inherent in projections of economic behavior and demography. Previous attempts to present stochastic budget projections for the U.S. have successfully incorporated sources of long-run uncertainty but have not attempted to model business cycle fluctuations realistically. While perhaps reasonably accurate in the long run, stochastic models that ignore cyclical uncertainty offer relatively few insights into the short-run uncertainties that may be of greater concern to policymakers. This paper develops a new and parsimonious framework for incorporating both cyclical and long-run uncertainty into a single, consistent set of stochastic budget projections, and it presents new long-term budget forecasts for the U.S. federal government and combined state and local governments that highlight the role of short-term uncertainty. ∗Edwards: Postdoctoral Scholar, Morrison Institute of Population and Resource Studies and Department of Biological Sciences, Stanford University. Lee: Professor of Demography and Economics, University of California, Berkeley. Tuljapurkar: Morrison Professor of Population Studies, Department of Biological Sciences, Stanford University. Corresponding author: [email protected]. A PDF version of this paper is available at http://www.stanford.edu/~ryanedw. VERY PRELIMINARY. Please do not quote.