IMPROVING THE INFLATION RATE FORECASTS OF ROMANIAN EXPERTS USING A FIXED-EFFECTS MODELS APPROACH

This article proposes an empirical econometric approach to improve the degree of accuracy for predictions made by Romanian experts in forecasting. Several fixed-effects models are constructed using the inflation and unemployment rate actual values and the forecasts provided by the European Commission, the National Commission for Prognosis and Dobrescu’s model over 2001-2014. The predictions based on these fixed-effects models did not improve the forecasters’ accuracy, but combined predictions of these models and naive forecasts brought a statistically significant improvement for projections made by Romanian forecasters on the horizon 2011-2013. This assumption was proved by common accuracy measures and Diebold-Mariano test.