FOR MANY YEARS, poverty in urban areas of Africa, Asia, Latin America and the Caribbean has been identified by spatial area (e.g. settlement, neighbourhood, inner city) as often as it has been referred to by social group (e.g. low-income women, the homeless). This has perhaps given inadequate recognition to poverty differentiation within urban poor groups or, to put it another way, it has contributed to an assumption that the poor share similar characteristics and face the same difficulties. The focus on chronic poverty that has been of growing interest in recent years is particularly relevant to our understanding of urban poverty because it recognizes differences within “the poor”. Chronic poverty is a concept that takes into account the length of time that an individual or household experiences poverty, as well as the depth of poverty. Conceptually, it challenges us to differentiate between those who are poor at any specific point in time and those who are poor over a long period of time, and between the always (chronically) poor and those for whom poverty is transitory. While in the work of the Chronic Poverty Research Centre, the suggested period for “chronic poverty” is five years, it should be recognized that such specificity is somewhat arbitrary, and other definitions have also been used. As illustrated in recent papers, this often reflects the periods for which data are available. For example, Kedir and McKay use three surveys (1994, 1995, 1997) in their analysis of urban Ethiopia, and suggest that the chronically poor are those who are poor in all three surveys over a four-year period. While much of the analysis has been on incomes and the period of time without adequate income, the concept can also be applied to inadequate access to basic services such as water and education. While, conceptually, the addition of a time dimension is easy to understand, methodological approaches vary. Kedir (in this volume) explores a number of different methodologies for the study of chronic urban poverty in Ethiopia. He argues that in analyzing quantitative income and expenditure data, the use of precise differentiated price indices is significant, with different price indices producing different movements in poverty trends. (For example, in one case, an apparent reduction in poverty became an increase through the use of a different aggregate price measurement.) Hence, the use of a single set of prices, or even the use of different prices for urban and rural areas, may produce misleading indications of relative levels of poverty and the permanence (or not) of poverty. However, while such details matter with respect to specific estimates, the significance of the concept lies in the questions it raises for the analysis of poverty and poverty reduction. A consideration of the length of time that individuals and households have incomes that fall below the poverty line (by poverty line, we mean an income level that is established as necessary to