Clusters of attributes and well-being in the USA

Using ARIMA models and entropy, the dynamic evolution of several functions of aggregate income and other attributes of well-being is analysed for statistical 'similarity' in order to determine potentially distinct dimensions in multidimensional analysis of welfare and quality of life in the USA. The entropy metric compares entire distributions and is more general than principal components and other correlation-based techniques for clustering. To help macroeconomic policy makers, we compare the distribution of several composite measures of well-being which include income, with the distribution of some common measures of aggregate income over the period 1915-1995. Per capita disposable income and growth in GNP are statistically distinct dimensions of well-being. Copyright © 2001 John Wiley & Sons, Ltd.