Search results

This study evaluates the Federal Reserve (Fed) initial and final forecasts of the unemployment rate for 1983Q1-2018Q4. The Fed initial forecasts in a typical quarter are made in the first month (or immediately after), and the final forecasts are made in the third month of the quarter. The analysis also includes the private forecasts, which are made close to the end of the second month of the quarter.

In evaluating the multi-period forecasts, the study tests for systematic bias, directional accuracy, symmetric loss, equal forecast accuracy, encompassing and orthogonality. For every test equation, it employs the Newey–West procedure in order to obtain the standard errors corrected for both heteroscedasticity and inherent serial correlation.

Both Fed and private forecasts beat the naïve benchmark and predict directional change under symmetric loss. Fed final forecasts are more accurate than initial forecasts, meaning that predictive accuracy improves as more information becomes available. The private and Fed final forecasts contain distinct predictive information, but the latter produces significantly lower mean squared errors. The results are mixed when the study compares the private with the Fed initial forecasts. Additional results indicate that Fed (private) forecast errors are (are not) orthogonal to changes in consumer expectations about future unemployment. As such, consumer expectations can potentially help improve the accuracy of private forecasts.

Unlike many other studies, this study focuses on the unemployment rate, since it is an important indicator of the social cost of business cycles, and thus its forecasts are of special interest to policymakers, politicians and social scientists. Accurate unemployment rate forecasts, in particular, are essential for policymakers to design an optimal macroeconomic policy.

This study aims to both test the asymmetric information hypothesis and explore the factors influencing the one‐ through four‐quarter‐ahead Federal Reserve inflation forecasts for 1983‐2002.

Encompassing tests are used to examine the asymmetric information hypothesis. In modeling the Federal Reserve inflation forecasts, the authors are mindful of alternative theories of inflation which emphasize such determinants as cost‐push, demand‐pull and inertial factors.

First, the Federal Reserve inflation forecasts embody useful predictive information beyond that contained in the private forecasts. Second, with the private forecasts controlled for, the near‐term Federal Reserve inflation forecasts make use of qualitative information, and the longer‐term forecasts are influenced by the forecasts of growth in both unit labor costs and aggregate demand as well as the preceding‐quarter inflation forecasts and monetary policy shifts.

The Federal Reserve forecasts are released to the public with a five‐year lag and are currently available up to the fourth quarter of 2002. This limits the use of the most up‐to‐date forecasts desirable for this study.

The factors influencing the Federal Reserve inflation forecasts are basically those emphasized publicly by monetary authorities. This finding points to the Fed's transparency and should thus help enhance its credibility with the public. Also, our results (which shed light on the predictive information in the Federal Reserve inflation forecasts not included in the private forecasts) are of value, since they can help the Fed better predict how inflation will respond to policy actions, and they can help the public form more informative inflationary expectations.