Title: A forecast-based consumer sentiment index
Authors: Claudio Lupi - University of Molise (Italy) [presenting]
Giancarlo Bruno - ISTAT - Italian National Institute of Statistics (Italy)
Marco Centoni - LUMSA (Italy)
Abstract: In recent years there has been a growing interest in using consumer surveys as leading indicators of real economic activity, especially to assess the future path of private consumption. Consumer sentiment indices are widely used and analysed with this aim, but are generally produced from consumer surveys using rather ad hoc criteria and weights. However, consumer surveys amount to very large data sets whose value is probably still underrated. We offer a forecast-based alternative measure of the consumer sentiment with a well defined economic meaning. The indicator is based on the canonical correlations between consumption and finely disaggregated consumer survey data. The underlying idea is to exploit the property of canonical correlation in order to find the linear combination of the consumer survey items that maximises correlation with final consumption. The new indicator is computed on Italian consumer survey data and is derived for total final consumption expenditure of household and for consumption expenditure by durability. Some comparisons with the existing consumer sentiment indicator are derived and discussed.