Consumer sentiment and consumption: an update for Australia

Policy-makers, business analysts and the media have always closely monitored consumer sentiment indices. It is widely assumed that consumer sentiment is a good indicator and predictor of household consumption. In this study, we revisit the empirical question raised by Carroll, Fuhrer and Wilcox (1994) of whether there is any additional evidence of a direct and independent influence running from sentiment to consumption. A recent study, by Easaw, Garratt and Heravi (2005) has found, using a similar framework as articulated in Carroll, Fuhrer and Wilcox (1994), that Consumer Sentiment does, in part, predict household expenditure However, the studies by Carroll, Fuhrer and Wilcox (1994) and Easaw, Garratt and Heravi (2005) are based on relatively ad hoc specifications of the aggregate consumption function. The motivation for this paper is to determine whether consumer sentiment has any predictive ability for aggregate consumption when it is examined over a range of aggregate consumption specifications for Australia for the period I979:Q4 to 2006:Q4. The specifications are chosen to represent a spectrum of theoretical aggregate consumption formulations. We find that consumer sentiment is significant when we augment it with a well-specified consumption function.