www.econstor.eu TRENDS IN INCOME INEQUALITY, PRO-POOR INCOME GROWTH AND INCOME MOBILITY

We provide an analytical framework within which changes in income inequality over time are related to the pattern of income growth across the income range, and the reshuffling of individuals in the income pecking order. We use it to explain how it was possible both for ‘the poor’ to have fared badly relatively to ‘the rich’ in the USA during the 1980s (when income inequality grew substantially), and also for income growth to have been pro-poor. Income growth was also pro-poor in Western Germany, more so than in the USA, and inequality did not rise as much. NON-TECHNICAL SUMMARY Not only is the inequality of family income higher in the USA than most other western developed nations, but also US inequality grew comparatively faster during the 1980s. Descriptions of the US experience typically emphasize that income growth was greater for the rich than for the poor. For example, it has been shown that the income of a family at the eightieth percentile rose sharply during the 1980s, whereas the income of a family at the twentieth percentile hardly changed at all over the same period, or fell slightly. At the same time, there has also been a growing literature about the longitudinal mobility of incomes in the USA, and several recent studies have found that mobility is lower than in Germany. The three facets of the income distribution – inequality trends, differential income growth, and income mobility – have rarely been studied jointly, however. We do so in this paper. We provide a framework in which changes in income inequality over time are related to the pattern of income growth across the income range and the reshuffling of individuals in the income pecking order, and use it to analyze the US experience and to compare it with Germany’s. We show that when income inequality is measured using any member of the generalized Gini class of indices, the change in inequality between two points in time can be additively decomposed into two components, one summarizing mobility in the form of reranking, and one summarising progressivity in income growth (i.e. whether income growth is pro-poor rather than pro-rich). This decomposition framework is used to reassess US income inequality trends during the 1980s, and to explain a potential paradox. That is, it is possible both for ‘the poor’ to have fared badly relatively to ‘the rich’ – the conventional picture of the USA during the 1980s derived from analysis of surveys like the Current Population Survey – and also for income growth to have been pro-poor. Income growth was pro-poor in Western Germany as well, and to a greater extent than in the USA. This, combined with less reranking than in the USA, underlay the relatively small rise in income inequality in Western Germany during the 1980s and 1990s. A key element of our framework is that we track income changes for individuals, rather than income changes for income groups such as ‘the poor’ or in a reference income such as the bottom quintile or the mean income among the poor. (It is the latter changes that have been tracked in most of the literature on poverty and inequality trends.) The composition of the group who are poor changes over time because some individuals fall into poverty and some escape it. Average income growth between 1980 and 1990, say, among those who were poor in 1980 need not equal average income growth over the decade for those who were poor in 1990. Similarly, the individuals with a 1980 income equal to the poorest quintile in 1980 would have experienced a diversity of income growth rates, and few of these individuals would be likely to have a 1990 income equal to the poorest 1990 quintile. Put another way, analysis of income distribution trends using cross-sectional data sets ignores the reshuffling of individuals in the income distribution over time, whereas this mobility is an integral part of our approach.