Comparing Federal Reserve, Blue Chip, and time series forecasts of US output growth

摘要

We evaluate the predictive content of Federal Reserve and Blue Chip forecasts of output growth by utilizing two comparable forecasts as benchmarks: a univariate autoregres-sive (AR) model, and a vector autoregressive (VAR) model which includes output growth, growth in residential investment, and consumers' assessments of business conditions. We first show the forecasts are all directionally accurate, free of systematic bias, and efficient. Second, the asymmetric information hypothesis cannot be supported. Third, the Federal Reserve and private forecasts are generally less informative than the VAR forecasts and thus lack past information on residential investment growth and consumers' assessments of business conditions.