Mordecai Kurz --
Diverse Beliefs, Forecast Errors and Central Bank Policy
Click to view a PDF (Adobe Acrobat
PDF) file.
Abstract
January 2005
The Fed and private forecasters have not been able to forecast inflation and GDP growth very well and
have assessed incorrectly the effects of monetary policy actions on these variables. The mistakes are quantitatively
different across three monetary regimes studied here, but we also find significant qualitative similarity in the errors
of assessing the stabilization effects of monetary policy. Over the period of study (1965-1995) we find that
-
(i) for long horizons the Fed and private inflation forecasts (i.e. over 3 quarters) either underestimated or
overestimated future developments: during low inflation the forecasts were too low and during high inflation
they were too high. The same pattern is true for GDP growth forecasting but over all horizons;
-
(ii) the Fed inflation forecasts over-estimated the stabilization effects of monetary policy over short horizons
and under-estimated these effects over long horizons;
-
(iii) in forecasting GDP growth all forecasters for almost all horizons underestimated the stabilization effect
of interest rates on GDP growth in all three monetary regimes.
Forecasters underestimated both the
stimulating effects of low interest rates as well as the inhibiting effect of high interest rates on GDP growth.
The quality of the Fed�s forecasts of inflation and GDP growth is sufficiently poor to conclude that monetary policy
decisions which are unexpected by the private sector should be confined to special circumstances and to emergencies
such as wars or financial crises. We present extensive evidence in support of the perspective that private forecasters
and the Fed use different models to forecast inflation and GDP growth: heterogeneity of forecasting models is the
norm in the marketplace. We study in detail the similarity and differences of such models.
Working Papers Index