Inference and Macroeconomic Dynamics

点击翻译按钮获取中文摘要

This project aims at improving our ability to understand, predict and control macroeconomic developments at the national and international level. Long term trend movements in the economy, at time scales exceeding the span of data we have available, raise special problems for inference that are just beginning to be addressed by econometricians. This project will develop ideas for approaching these problems in a Bayesian framework -- making explicit the separate contributions of the data and of information from other sources (prior beliefs) to knowledge of long term dynamic behavior. A special emphasis will be on extending the recent results in this area for univariate models to more practically relevant multivariate models. There is disagreement among macroeconomists on the importance of monetary policy and other forms of aggregate nominal demand management in generating or offsetting business cycles. Increasing our understanding in this area is important not only for improving demand management policy, but also in guiding the restructuring our banking system is undergoing. The international facts about relations among the main macroeconomics aggregates are only beginning to be documented and understood. Patterns of responses to innovations (surprise changes) in monetary variables appear in preliminary work surprisingly robust across countries, time periods, and lists of variables included in the system. This research will document these regularities. These regularities present difficulties for simple standard theories. For example, Keynesian or monetarist theories view monetary policy as a major independent influence on the business cycle. Real business cycle theories treat it as passive and unimportant because of the smooth dynamic functioning of product and asset markets. This research will explore possible interpretations of the regularities, using several kinds of new and recently developed modeling strategies.