Are goods goods? Some empirical evidence

Abstract

Interest has recently been shown in the possibilities of extending demand analysis to take account of the characteristics of goods. This paper reports the results of an experiment with 386 respondents who were asked to evaluate some of a set of 17 durable consumer goods by indicating the strength with which each of 21 common characteristics were held to be associated with each of the goods. Discriminant analysis was used to classify the goods on the basis of the responses and 18 of the 21 variables used were statistically significant in the discriminant functions. It was found that, for many goods, the efficiency of classification was low but consistent patterns of misclassification were observed which allowed the identification of groupings of goods with common strengths of characteristics. Five groups were identified - utilities, luxuries, leisure goods, central heating and car - and eleven characteristics were identified as statistically significant discriminators between the groups of goods. A further discriminant analysis on the identified groups of goods markedly increased the efficiency of the classification and confirmed the characteristics which had previously been identified as significant discriminators between the different groups of goods. On the basis of these results, it is argued that analysis can usefully proceed through the comparison of the strengths with which common characteristics exist in different durable goods. It is not justifiable to assume that goods are simply what are thought of as goods, instead certain groups of durable goods with common patterns of characteristics can be identified and analysis could proceed on the basis of those groups.