Journal Article

Modeling trends in macroeconomic time series

Abstract: How predictable are real GNP, prices, and other macroeconomic data over long time horizons? The answer depends on the nature of their trends. In this article, Nathan S. Balke describes alternative models of trend for economic data, discusses the implications of these models for forecasting and business-cycle analysis, and reviews some of the existing evidence for and against various models of trend. ; In addition, Balke conducts a case study of real GNP and the price level. He finds that a simple linear time trend may adequately reflect the long- run behavior of real GNP. The price level, on the other hand, appears to be affected by infrequent but dramatic events that have long-lasting effects. Consequently, the price level is much more difficult to forecast.

Keywords: Macroeconomics; Econometric models; time series analysis;

Bibliographic Information

Provider: Federal Reserve Bank of Dallas

Part of Series: Economic and Financial Policy Review

Publication Date: 1991

Issue: May

Pages: 19-33