On the loss structure of federal reserve forecasts of output growth

Abstract

Utilizing recent techniques with comparable private forecasts as benchmarks, we test the rationality of the Federal Reserve forecasts of growth under flexible loss. Our findings for 1984–2006 indicate that these forecasts are rational (efficient) and directionally accurate under symmetric loss and are thus of value when similar cost is assigned to both incorrect upward and downward predictions. Over-predictions (under-predictions) are costly when the Fed responds by implementing an unnecessary contractionary (expansionary) monetary policy and thus causes lower growth (higher inflation). Symmetric loss thus indicates that the Fed is equally concerned about both low growth and high inflation. Further results reveal that the private sector closely replicates the Federal Reserve forecasts released to the public with a five-year lag, suggesting that one can utilize the readily available private data as proxies for the not-yet-released Federal Reserve forecasts.