Creating a Smarter Conditional Consensus Forecast

We examine analyst forecasts of 26 macroeconomic statistics over the August 1998 to March 2007 time period. We pose four research questions: (1) Does forecast accuracy persist?; (2) What are the determinants of such persistence?; (3) Do analysts who exhibit these characteristics make more accurate forecasts than the simple consensus?; and (4) Is a quot;smartquot; consensus based on individuals exhibiting these characteristics more accurate than the simple consensus? We show that analyst forecast accuracy persists and is determined by long-term, past accuracy and the overall ability in forecasting all statistics. That is, analyst long-term track record is more important than short-term track record, and track record in forecasting other statistics (general ability) is more important than track record in forecasting the macroeconomic statistic in question (specific ability). We demonstrate that a smart consensus conditioned on forecasts of analysts with superior long-term, general ability outperforms commonly reported mean and median consensus forecasts