Over the past decade, the amount of poverty reduction resulting from a given rate of economic growth has varied in close step with income distribution. Work by Hanmer and others reveals that, on average, a growth rate of 10 percent reduced the poverty headcount (the percentage of people living on less than $1 a day) by 9 percent in countries where income was fairly equally distributed. However in countries where income was unequally distributed, a growth rate of 10 percent reduced the poverty headcount by only 3 percent.

The World Bank estimates that developing countries will grow at 4 percent per capita per annum until 2015. What are the implications of this growth rate for meeting the DAC (OECD's Development Assisstance Commitee) income-poverty target? As the chart (see box) shows, if the 4 percent growth rate is accompanied by low income inequality then the DAC target is easily met, and poverty is halved by 2006. If, on the other hand, high income inequality accompanies growth then the DAC target is not achieved. The 4 percent growth rate forecast is optimistic to begin with.

Higher rates of growth help, but only to a limited extent. The same chart shows that shifting from a high inequality to a low inequality growth path has a greater effect on poverty reduction than adding an extra 1 percent to the growth rate. In fact, in the high income inequality scenario it would require an extra 5 percent growth (to make a total rate of 9 percent) per capita per annum to meet the target. This figure would be without historical precedent. Even the Southeast Asian economies only grew 5.5 percent per annum between 1965 and 1997.

High and low income inequality are defined in the calculation respectively as Gini coefficients above and below 0.43 (see Unbottling the Gini, below). The table shows Gini coefficients for some large developing countries and some developed market economies. Income distribution varies from country to country but there are more high-inequality countries in Latin America and the Caribbean and in sub-Saharan Africa than in Asia, Europe and North America. If income inequality remains unchanged in Latin American and sub-Saharan African the desired poverty target will not be achieved.

Unbottling the Gini The established way to measure the degree of inequality in the distribution of household or individual incomes is the Gini coefficient or 'Gini'. There are a number of different ways to define income and it can be useful to look at the sources of income individually or by household. But the outcome is the same - the higher the Gini, the more extreme the observed inequality. A graphical plot of the cumulative income distribution by population percentiles yields the Lorenz curve (see charts with the Zimbabwe article, overleaf). The further this curve deviates from the diagonal (marking total equality in distribution of incomes across the population), the more unequal the observed distribution. Formally, the value of the Gini coefficient lies between zero and one and is the ratio of the area between the line of total equality and the Lorenz curve, to the entire triangular area beneath the diagonal.

So the good news is that the DAC income-poverty target is attainable - provided that significant improvements take place in income distribution. These can be achieved ex-ante, by designing growth strategies that increase disproportionately the incomes of the poorest, or ex-post, by redistributing income through taxation. Many questions arise. What is the recipe for income-redistributing growth? Is there a trade-off between growth and distribution? Are ex-post strategies of redistribution feasible? These questions are far from new. Indeed, to a large degree, they are the very questions on which the development studies profession is founded. Nevertheless, they have been neglected in recent years. Does current research offer new perspectives? Articles in this issue of Insights offer six main conclusions. They are that: