Using survey data to correct the bias in policy expectations extracted from fed funds futures

Many studies estimate risk premiums on federal funds futures to extract monetary policy expectations by assuming that average forecast errors of the expectations are zero or that survey forecasts are good proxies for the expectations. These assumptions, however, may fail due to an unanticipated declining trend in the federal funds rate and to survey respondents' strategic behavior. Consequently, the premiums estimated under these assumptions may be biased. We propose a new method to estimate the premiums and find that the premiums have been often negative since 2000, which is generally consistent with the negative betas observed in the 2000s.