Study on the Accuracy of Leading Indicators - ZEW Indicator Provides Highest Forecasting Quality on a Six-month Horizon

Research

From time to time, the scientific sphere and the public call into question the forecasting quality of leading indicators of economic development. A recent study by the Centre for European Economic Research (ZEW), Mannheim, based on econometric analyses now shows that the ZEW Indicator of Economic Sentiment as part of the ZEW Financial Market Survey as well as the ifo business expectations by the ifo Institute in Munich have a reliable signal effect for the future development of the annual growth rate in German industrial production.

Forecasts based on one of these two indicators have proven far better than non-indicator-based forecasts. The significant one-month lead of the ZEW Indicator of Economic Sentiment over ifo business expectations was once again reaffirmed.

The central findings of the current study have been published as part of the ZEW Growth and Business Cycle Analyses in June 2005 (ftp://ftp.zew.de/pub/zew-docs/kr/kr0502.pdf). The study further suggests that the ZEW Indicator of Economic Sentiment is most suitable for medium-term forecasts, whereas ifo business expectations provide higher short-term accuracy. Within the research period from January 1997 to September 2004, the ifo indicator outperformed the ZEW indicator in terms of short-term forecasts (one to three months). The latter, however, leads in providing six-to-twelve month forecasts. Its highest forecasting quality is reached on a six-month horizon, which corresponds to the exact period of time used as basis for the survey among financial market experts.

A possible explanation for the different leading characteristics of the two indicators may be that the experts surveyed by ZEW focus on other information than the companies interviewed by the Ifo Institute. All in all, it seems that both indicators complement each other not only in terms of their forecasting horizon, but also by reflecting the expectations of different groups of participants.

The current study further analysed the forecasting quality over time. The investigation period was therefore separated into two periods of time. It appears that accuracy of the forecasts based on both indicators has increased over time. Furthermore, so-called trend forecasts can be used to confirm that the projected direction of development very often matches the actual trend. The results of these qualitative forecasts thus testify the accuracy of indicator-based projections.

The current ZEW study builds on a ZEW examination conducted by Felix Hüfner and Michael Schröder in 2002 and adds about two more years of research to it. In the study, the researchers applied Granger causality tests to determine the past leading indicator characteristics of both leading indicators. In this context, it can be demonstrated that the ZEW Indicator of Economic Sentiment has a one-month lead over the ifo indicator in terms of Granger causality. The study also creates forecasting models, which show how accurately both leading indicators can project the future annual growth rate of industrial production. To determine and compare the quality of the forecasts, researchers conduct a naive prognosis, which is not based on indicators but exclusively on past values of the growth rates that were to be forecasted.

Contact

Dr. Sandra Schmidt, E-mail: http://s.schmidt@zew.de