How Well Do Diffusion Indexes Capture Business Cycles? A Spectral Analysis
The potential benefits of small surveys of the economy are well known--surveys can supply measures of economic activity quickly and at a relatively low cost. But potential drawbacks are also recognized, namely that small samples may produce predominantly noise and, because of the types of data collected, analytical options are often limited. From a monetary policymaker's perspective, another factor is also very important--though often overlooked. Even when accurate, survey data may reflect deviations from trend that occur at cycles outside of those relevant for monetary policy, i.e., very short or very long cycles. We construct power spectra for manufacturing surveys and find that results generally follow business-cycle-length cycles. These findings provide one method by which to gauge the value of diffusion indexes and suggest that most, though not all, indexes generally reflect business cycle conditions.
Receive an email notification when Economic Quarterly is posted online: