Interest rate forecasting and the financial crisis: a turning point in more than just one way

Abstract

Investors frequently rely on forecasts published by professional analysts. During the financial crisis uncertainty has substantially risen. Not surprisingly experts' predictions should be more than welcome for decision makers. After modelling the long-term relationship between the three month EURIBOR and the Consensus Economic forecast by using co-integration analysis this paper tests for changes in the accuracy of forecasts for the three month EURIBOR. We use traditional evaluation methods like sign accuracy tests, turning point analysis and the root mean square error (RMSE). We find evidence for a crisis related structural break. Checking for quality changes it can be stated that the forecasters' accuracy after the first months of the crisis is not much better in general. Furthermore, neither before nor after the structural break the analysts' forecasts did outperform a random prediction. But it can at least be stated that there is a significant improvement in the turning point prediction.

How do firms manage their interest rate exposure?

“…Financial decision makers often rely on interest rate forecasts from professional analysts, where especially in times of rising uncertainty the relevance of accurate predictions increases. With our sample period falling into the final chords of the financial crisis and particularly into the European sovereign debt crisis, forecasts move more into the focus of financial market participants (Kunze et al, 2015). On the other hand, the literature evaluates and challenges the accuracy and usefulness of forecasts for a long time (Friedman, 1980;Belongia, 1987;Greer, 2003;Baghestani et al, 2010)[14].…”

“…As our non-financial French sample firms mainly focus on EUR interest rates, especially the development of the EURIBOR benchmark rate is of particular importance. Kunze et al (2015) and Kunze et al (2017) assess, amongst others, the accuracy of three months EURIBOR rates and indicate that the quality of forecasts declines during financial crisis events due to increasing forecast errors. In this respect, Kunze et al (2017) point to the usefulness of continuously monitoring uncertainty measures such as the dispersion of forecasts as well as forecast errors instead of merely focusing on point predictions: the authors are able to link these uncertainty measures to real economy events and suggest that they could be used as additional warning indicator.…”