Replacing Judgment by Statistics: Constructing Consumer Confidence Indicators on the Basis of Data-driven Techniques

This article compares the properties of the European Commission's Consumer Confidence Indicator (CCI) for the euro area with three alternative indices which differ from the former in that they (i) consider a richer set of survey questions and (ii) are the result of data-driven statistical techniques, rather than the simple arithmetic mean of the input series. The alternative indicators are shown to perform only slightly better than the CCI in tracking real private consumption growth and to fail to produce significantly better forecasts of expansions and contractions in private consumption, once information from relevant, timely available hard data is controlled for. The conclusions change, however, if the analysis is re-conducted on well-defined subsets of survey questions. Concretely, the application of the alternative construction techniques to a data set which is limited to questions about consumers' personal finances produces an indicator which, combined with relevant macro-economic time series, yields significant improvements in forecasting expansions and contractions in private consumption.