Out of Sample Forecast of Swedish GDP Growth by the Economic Sentiment Indicator in the Euro Area : A Bayesian Approach

Abstract: In this paper, the predictive capabilities of the Economic Sentiment Indicator (ESI), based on business and consumer surveys in the Euro area, are evaluated by out-of-sample forecasts of Swedish GDP growth. A steady state Bayesian VAR-model is applied to quarterly data from 1996 to 2014. The results show that the inclusion of the ESI improves the forecasting performance, both in the point predictive measurement Root Mean Square Errors and in the forecast density sharpness measurement Log Predictive Density Scores. These findings suggest that international confidence indicators may prove useful in forecasting macroeconomic trends for small open economies.