November 13, 1998

NOTE TO EDITORS

Attached is the latest edition of the New York Fed's Current Issues in Economics and Finance, entitled Rethinking the Value of Initial Claims as a Forecasting Tool.

Using a simple forecasting model, economist Margaret M. McConnell finds that initial claims for unemployment insurance, a weekly measure of layoffs released by the U.S. Department of Labor, consistently send an accurate signal about employment during recessions but not during expansions.

By comparing the accuracy of a forecasting model that incorporates initial claims with one that does not, Ms. McConnell finds that claims serve as a useful tool for forecasting employment fluctuations during recessions. Very early on in economic expansions, however, the model loses its predictive power, and claims actually worsen forecast accuracy.

According to Ms. McConnell there is a relatively simple explanation for the shift in the forecasting power. Although initial claims are a timely measure of layoffs, employment growth depends on both layoffs and hiring. During a recession, these two variables appear to be related in a significant way: when hiring falls early in a recession, layoffs tend to rise, and toward the end of a recession, when hiring begins to recover, layoffs generally fall.

During an expansion, however, the strong inverse relationship between layoffs and hiring disappears, and hiring becomes the most important factor in determining changes in employment. Accordingly, the power of claims to predict employment changes is likely to be relatively weak during periods of sustained growth, concludes Ms. McConnell.

Contact: Douglas Tillett