Collaborative Research: Piecewise Linear Approximations for DSGE Models With Occasionally-Binding Constraints: Solution, Estimation, Model Evaluation, and Applications

University Of Maryland, College Park, College Park MD

Investigators

Abstract

Central banks regulators use mathematical models to understand fluctuations in economic activity and the effects of the bank's policies. Current methods used to analyze the general effects of policy changes on relationships among large-scale economic variables, such as total output, inflation, and interest rate are either not able to capture the effects of some extreme events, such as decreasing interest to zero, or are too slow and costly to implement. This research project will develop new, faster, cheaper, and better methods to analyze the effects of policy changes on large-scale variables in the economy. The results of this research project will therefore contribute to economic science, and more important, allow for a better understanding of how policy changes or external influences are transmitted throughout the larger economy. This will allow economist to provide more informed advice to police makers and in so doing, enhance economic growth and improve the living standards of Americans. The tools developed in this project can also be applied anywhere in the world, thus establishing the U.S. as the global leader in the development of economic analytical tools. Prior to the Great Recession, models of linear relationship among macro variables used by the Fed and other regulatory agencies were able to capture the most important features of aggregate time series and generate accurate predictions. However, during and after the Great Recession, nonlinearities generated by occasionally-binding constraints such as an effective lower bound constraint on nominal interest rates, have become the norm. This research project will develop new techniques to construct nonlinear solutions to macroeconomic models in which occasionally-binding constraints play an important role. The goal of the proposed research is to develop and apply a piecewise-linear solution method that trades off a little of accuracy against computational speed and is scalable to large models used by central banks. The project will also develop a class of flexible models that can be used to assess whether models that impose strong theoretical restrictions are well specified. The results of this project will provide researchers with new analytical tools that will improve the quality of policy advice as well as regulation. This could increase economic growth in the U.S. and improve the lives of citizens. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.