In recent y ears, simple policy rules have received attention as a means to a more transparent and eective monetary policy. Often, however, the analysis is based on unrealistic assumptions about the timeliness of data availability. This permits rule specications that are not operational and ignore diculties associated with data revisions. This paper examines the magnitude of these informational problems using Taylor's rule as an example. First, I construct a database of current quarter estimates/forecasts of the quantities required by the rule based only on information available in real time. Using this data I reconstruct the policy recommendations which w ould have been obtained in real time. I demonstrate that the real-time policy recommendations dier considerably from those obtained with the ex post revised data. Within-year revisions in the policy recommendations are also quite large with a standard deviation exceeding that of the quarterly change of the federal funds rate. Further, I show that estimated policy reaction functions obtained using the ex post revised data can yield misleading descriptions of historical policy. Using Federal Reserve sta forecasts I show that in the 1987-1992 period simple forward-looking specications describe policy better than comparable Taylor-type specications, a fact that is largely obscured when the analysis is based on the ex post revised data.

We examine the reliability of alternative output detrending methods, with special attention to the accuracy of real-time estimates of the output gap. We show that ex post revisions of the estimated gap are of the same order of magnitude as the estimated gap itself and that these revisions are highly persistent. Although important, the revision of published data is not the primary source of revisions in measured output gaps; the bulk of the problem is due to the pervasive unreliability of end-of-sample estimates of the trend in output. Multivariate methods that incorporate information from inflation to estimate the output gap are not more reliable than their univariate counterparts. © 2002 President and Fellows of Harvard College and the Massachusetts Institute of Technology.

No abstract