Academic journal article Economics & Sociology

The Level of Wealth and the Financial Condition of Households in Relation to the Results of Consumer Surveys

Academic journal article Economics & Sociology

The Level of Wealth and the Financial Condition of Households in Relation to the Results of Consumer Surveys

Article excerpt


The values of such indicators as the Consumer Confidence Index or the Index of Consumer Sentiment are generally considered to be indications regarding the consumer activity of households. The predictive usefulness of such data is one of the most frequently discussed topics (Ludvigson, 2004, p. 29). The focus of interest is usually on the strength and type of the relationships between the qualitative series and fluctuations in consumption (Souleles, 2004, p. 40; Dees, Brinca, 2013, pp. 6-9). The main goal of many previous studies into synthetic indicators of consumer confidence was to test their predictive usefulness in the context of various categories of expenditure, including expenditure on durable goods (Howrey, 2001, p. 175; Souleles, 2004, p. 40). However, mathematical and statistical analyses of the relationship between quantified consumers' answers and the estimates of effective demand do not always render the expected results. This applies to the strength of the interdependencies, their direction, as well as time lags. For example, the results of a study of a well-known synthetic index - the Michigan Consumer Sentiment Index (CSI) - have shown that it is characterized by the surprisingly low strength of its relationship to the volume of durable goods purchases (Kwan, Cotsomitis, 2004, p. 139).

One of the key questions that appear in scientific publications relates to the information content of confidence indicators - whether they contain any additional information beyond that which is already contained in the quantitative variables (Fuhrer, 1988). Some analyses show that this is indeed the case (Carroll, Fuhrer, Wilcox, 1994, p. 1397; Ludvigson, 2004, p. 29). Various tests show, however, that a lot depends on the wording of the questions asked, the size of the sample, the method used, as well as the quantification of the responses and the construction of the indices (Bram, Ludvigson, 1998, pp. 60-61; Das, van Soest, 1997, p. 138).

In the context of the information content it is worth considering the possible causes of the prognostic usefulness of sentiment indicators. In other words, is consumer confidence an independent factor that not only makes it possible to predict consumer spending but is also its cause? Or does perhaps the prognostic value of qualitative indicators in relation to consumption stem from the fact that they are a reflection of certain economic phenomena, which can cause the volume of purchased goods to increase or decline (Carroll, Fuhrer, Wilcox, 1994, p. 1398; Fuhrer, 1993, pp. 34-35)?

This article has adopted the latter perspective, by assuming that consumers' spending depends on their income, which in turn is related to the situation on the labour market and overall economic activity (Lovell, Tien, 2000). Thus it is assumed that the prognostic value of qualitative indicators of consumer sentiment depends to a large extent on changes to consumers' wealth, which is related to, among other things, their earnings (Bram, Ludvigson, 1998, p. 69). From this point of view, the indicators of consumer confidence cannot be predictors of consumption as well as its direct cause (Howrey, 2001, pp. 198, 204-205). Rather, they are a set of opinions on the overall economic situation, and in particular, assessments based on the current and predicted income of households. In this sense, the indicators can predict changes in consumer spending only indirectly, through the information which they contain relating to income. It is worth noting that adopting this point of view assumes non-compliance with the basic version of the permanent income hypothesis (PIH). If all consumers behaved in accordance with the assumption of the basic PIH model, consumption would follow the random-walk process so that it would be impossible to predict it on the basis of any information previously known to consumers (Mehra, Martin, 2003, p. 52; Dees, Brinca, 2013, p. 2; Acemoglu, Scott, 1994, p. 5).

The adopted approach is similar to the interpretation of prognostic usefulness whereby sentiment might predict spending without being an independent causal force (Mehra, Martin, 2003, p. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed


An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.