Fiscal forecast errors: governments vs independent agencies?

The fact that the literature tends to find optimistic biases in national fiscal projections has led to a growing recognition in the academic and policy arenas of the need for independent forecasts in the fiscal domain, prepared by independent agencies, such as the European Commission in the case of Europe. Against this background the aim of this paper is to test: (i) whether the forecasting performance of governments is indeed worse than that of international organizations, and (ii) whether fiscal projections prepared by international organizations are free from political economy distortions. The answer to these both questions is no: our results, based on real-time data for 15 European countries over the period 1999-2007, point to the rejection of the two hypotheses under scrutiny. We motivate the empirical analysis on the basis of a model in which an independent agency tries to minimize the distance to the government forecast. Starting from the assumption that the government’s information set includes private information not available to outside forecasters, we show how such a framework can help in understanding the observed empirical evidence