We study Federal Open Market Committee members' individual forecasts of inflation and unemployment in the period 1992-2004. Our results imply that Governors and Bank presidents forecast differently, with Governors submitting lower inflation and higher unemployment rate forecasts than bank presidents. For Bank presidents we find a regional bias, with higher district unemployment rates being associated with lower inflation and higher unemployment rate forecasts. Bank presidents' regional bias is more pronounced during the year prior to their elections or for nonvoting bank presidents. Career backgrounds or political affiliations also affect individual forecast behavior.
KEYWORDS
FOMC, individual characteristics, individual forecasts, regional bias
| INTRODUCTIONWe analyze the forecast determinants of Federal Open Market Committee (FOMC) members. Individual forecasts of key macroeconomic factors (such as unemployment and inflation rates) are crucial indicators for determining optimal monetary policy when a forward-looking policy rule is considered. 1 Since the FOMC is a committee consisting of 12 voting members (seven members of the Board of Governors (BoG) 2 and five voting regional Federal Reserve Bank presidents), disagreement about the optimal monetary policy stance (as shown in many studies) not only leads to dissenting votes in the FOMC but may also lead to dispersion of forecasts among FOMC members. Thus analyzing the determinants of real-time inflation and unemployment rate forecasts may improve our understanding of differences in the monetary policy preferences among FOMC members. This paper aims to analyze the determinants of FOMC members' individual inflation and unemployment rate forecasts in the period 1992-2004. We add to the existing literature not only by considering a broadest set of potential regional, individual, and political characteristics, but also by testing determinants (such as career backgrounds, electoral cycles, political affiliation) not previously considered in the FOMC forecasting literature. Our results indicate considerable differences in the economic forecasts between Governors and Bank presidents. Moreover, forecasts of Bank presidents are influenced by the unemployment rate in their 1 Haldane and Batini (1999), as well as Rudebusch and Svensson (1999), compare monetary policy rules and conclude that forward-looking rules including forecasts of inflation instead of actual values can improve monetary policy performance relative to a simple benchmark rule (such as proposed in Taylor, 1993). Clarida, Gali, and Gertler (1998) find that so-called G3 central banks (Germany, Japan, USA) implicitly have followed a type of inflation-targeting regime since 1979, and thus all three central banks indirectly applied a forward-looking framework. Orphanides (2003) find periods -such as during the 1970s or mid-1990s-when a forecast-based rule did an even better job in describing FOMC's monetary policy. Orphanides and Wieland (2008) use a forecast-based Taylor rule framework and conclude...