بازنگری کارایی پیش بینی نسبی کارکنان فدرال و پیش بینی کننده های خصوصی: یک رویکرد پویا  Revisiting the relative forecast performances of Fed staff and private forecasters: A dynamic approach

1. Introduction Today, most central bank watchers invest vast resources in the production of good forecasts of inflation and output. Why is this? It is because it helps them to assess the monetary policy stance better in real time, and to form expectations about the likely future interest rate path. Can private forecasters learn something from central bank forecasts of these key macroeconomic variables? Romer and Romer (2000) demonstrated that the Federal Reserve’s (Fed) Greenbook forecasts outperform private forecasts of output and inflation in the United States, but have relationships changed since then, or has the US economy become more predictable, as was suggested by Tulip (2009)? Moreover, the presence of information cascades (see Bikhchandani, Hirshleifer, & Welch, 2008) appears tohave contributed to a better sharing of information and a reduction in the dispersion of private forecasts. For at least three reasons, it is questionable whether the superiority of staff forecasts still holds today. First, the Fed and other main central banks have achieved a high level of transparency, thereby reducing their relative information advantage, with the aim of enhancing the effectiveness of monetary policy (seeWoodford, 2005). In this context, only the Fed’s Summary of Economic Projections is published in real time, while its Greenbook forecasts are released with a lag of five years. Second, given several large macroeconomic shocks (the Great Moderation, oil price shocks, and financial crises) which contributed to changes in the volatility patterns of macroeconomic time series, it is conceivable that the relative forecast performances of staff and private forecasters may have changed. However, because both groups of forecasters have been subject to profound uncertainty in the presence of these shocks, it could be that their performances are different. Third, the superiority of Greenbook forecasts is at odds with the findings of another paper by Romer and Romer (2008), which suggests that the forecasts made by the Federal Reserve Open Market Committee (FOMC) are less informative than the Greenbook forecasts. These forecasts by FOMC policy-makers are informed by the internal Greenbook forecasts, and should not differ fundamentally from the latter. Ellison and Sargent (2012) have challenged this view in their defence of the FOMC policy-makers’ forecasts. Another strand of the literature has argued that monetary policy should be rule-based, in order to deal with economic uncertainty. This could be viewed as suggesting that forecasts are less relevant for policy-makers. Does the Greenbook’s forecast accuracy matter for US monetary policy? The answer depends strongly on the persistence of the macroeconomic processes. If high levels of inflation or output today signal high levels of those variables in the future, for example, a distinction between the current and forecast values may not matter much (see Bernanke, 2010). The distinction between current and forecast levels of inflation has become increasingly important, as has been shown by several oil price shocks, which have led to increases in the overall inflation. Moreover, as the FOMC transcripts make clear, Fed policymakers use forward-looking Taylor rules as a benchmark for their internal discussions. In fact, the FOMC uses a suite of these rules to check for robustness. In order to be a meaningful benchmark for policy, the Taylor rule needs to have good forecasts as inputs.