Forecast errors and financial developments

As central banks have moved towards a forward-looking implementation of monetary policy, the role of forecasts in the policy process has greatly increased. Against this background, this paper looks at the accuracy of forecasts and, more specifically, addresses the question whether forecasts of growth and inflation can be improved by including information from financial markets. The empirical work presented suggests that average forecast errors are not large enough to seriously undermine the basis for forward-looking monetary policies, except in periods of common shocks and at cyclical turning points. It also appears that unexpected changes in non-financial variables are the primary source of forecast errors. Nonetheless, for several countries, forecasts could also be improved by using the information contents of changes in the yield curve and of movements in exchange rates and other asset prices.