ISER Working Paper Series 2006-36
Comparisons of income mobility profiles
01 Jul 2006
Information about social mobility is important for assessing the prospects of individuals. It is not enough to observe the distribution of income at one point in time, or even at several points. These snapshot views do not tell us about how the income of each person evolves over time. To examine this, one needs longitudinal data, and these are increasingly available. One also needs appropriate tools with which to summarize mobility and, in contrast to the analysis of poverty or inequality at a point in time, there is relatively little consensus to date about how to describe and evaluate income mobility.
This paper argues that many of the most frequently used summary indices of mobility can be expressed as population averages of a statistic defined at the individual-level that captures the degree of mobility by that individual, for example the proportionate change in her income over time, the absolute value of the change, the change in her rank in the income distribution, etc. In the averaging process used to compile the overall summary measure, these indices treat each individual in the same way, and simply add up the individual mobilities. However, it is often implicit in the debate about mobility that it is a better thing from a societal point of view if the pattern of mobility is such that relatively poor people climb up the income ladder (and are replaced by others at the bottom of the distribution), rather than mobility being driven by improvements for the relatively rich. In other words, social evaluations of overall mobility depend on how income changes are distributed relative to people's positions in the base period income distribution -- whether income growth is relatively 'propoor' or not.
This paper develops the concept of a 'mobility profile' to address these issues. It is shown, first, that mobility profiles provide an evocative graphical summary of the pattern of mobility according to baseperiod position, thereby also clarifying the relationship between different mobility concepts and indicating why different overall summary indices may lead to different conclusions about the extent of mobility. Second, the paper offers a framework for making comparisons of income mobility based on normative considerations that go beyond the comparison of population averages. It shows how nonintersection of mobility profiles for two societies is equivalent to rankings of their mobility according to a large class of summary mobility indices. A new family of mobility measures is also proposed.
The methods are applied to data for ten EU countries between 1996 and 2001 taken from the European Community Household Panel survey. It is shown that individual income growth in countries such as Ireland and Spain was particularly large among those at the bottom of the income distribution. The overall average level of income growth was much the same in Portugal but, by contrast with the other countries, it was much less 'pro-poor'. Portugal therefore performs worse than the other countries according to social evaluation functions in which concerns about the distribution of mobility across the income distribution are taken into account.