Are OECD forecasts rational and useful? A directional analysis

Tests of direction are employed to evaluate the rationality and usefulness of a large data set of semi-annual forecasts for the G7 economies made by the OECD. Changes in the main components of aggregate demand and output, inflation and the balance of payments are predicted up to three half-years ahead. In total, we inspect the signs of 14 184 pairs of forecasts and outcomes. The results indicate that all these forecasts are rational and, looking ahead 6 months, generally useful. However, there is no evidence that longer term forecasts – with a 1 year or 18-month horizon – are valuable. With very few exceptions, they are no better than a naive model that always predicted the same direction of change.