Speakers:

Lucia Hanmer and Felix Naschold, ODI.

Stephen Ackroyd, Oxford Policy Management.

Susan Joekes, IDS.

Rosemary McGee, Christian Aid .

1. Jon Wilmshurst, ODI Associate, welcomed participants to this first meeting in the summer series Halving Poverty by 2015: Have We Got What it Takes? The target of halving extreme income-poverty by 2015 was agreed at the World Summit on Social Development in 1995 and was subsequently adopted by the Development Assistance Committee of the OECD. The DAC also adopted a number of human development targets covering important dimensions of poverty such as infant mortality, reproductive health care, gender disparities and education. This series of talks and panel discussions would address the substantial challenges that meeting these targets will pose for developing countries and the world community.

2. Lucia Hanmer then introduced an interim report on the findings of a study commissioned by DFID as a contribution to the preparation of the World Development Report 2000/2001, Meeting the International Development Targets: What Are the Prospects and Key Challenges? This consisted of two parts: a global study, making econometric projections based on pooled data and economic growth forecasts for a large number of countries, and three country studies for Tanzania, Pakistan and Uganda. The global study is more up-to-date and more comprehensive than previous exercises of the same sort and involves methodological improvements.

3. A key finding of the global study concerns the way the effectiveness of economic growth in reducing poverty is affected by the level of income inequality in a country. The "poverty elasticity" is the percentage change in the proportion of a population in poverty associated with a one per cent increase in real GDP per capita. The study confirmed that this varies systematically with inequality, measured by the Gini coefficient. Countries with a Gini coefficient below 0.43 have a poverty elasticity that is nearly three times as large as for those where there is more inequality. Growth was also found to be more effective in reducing poverty where: government policies are "good", in terms of the Sachs-Warner index of interventionism; capital is not severely constrained and is efficently used; and, in sub Saharan Africa, where the pattern of growth favours agriculture.

4. These findings are based on the data on past performance of countries. When they are combined with the forecasts of future growth for countries and regions, in a model which projects future levels of poverty, the results are equally striking. At the global level, the prospects for reducing poverty are affected much more by the assumptions that are made about policy scenarios, including the level of inequality, than by the more or less optimistic assumptions made about likely growth performance. According to the projections, the target of halving extreme poverty by 2015 will be met if and only if low levels of income inequality accompany growth.

5. The regional breakdowns reported by Lucia Hanmer are, as expected, diverse. At one extreme, East Asia and the Pacific make the income-poverty target on both best-case and worst-case scenarios, the financial crisis having set the region back by about four years in this respect. At the other extreme, three regions fail to meet the target even on best-case assumptions. These are sub-Saharan Africa, the Middle East and North Africa, and Eastern Europe and Central Asia (where poverty incidence has risen precipitously in the 1990s).

6. The investigation of human-development targets considered two alternative health scenarios. The Better Health Scenario assumes no increase in HIV prevalence, increased availability of health services and continued technological progress. The other scenario has HIV prevalence rising to 22.5 per cent in all regions by 2015, with this swamping health services and technological progress.

7. Globally, the infant mortality target (one third of the 1990 level by 2015) is exactly met under the more optimistic scenario, but is nowhere near achieved on the AIDS Pandemic assumption. Universal primary school enrolment emerges as easily attainable in all regions except sub-Saharan Africa, but quality concerns raise questions about this target. Gender equity in school enrolment, finally, is only very weakly correlated with per capita income and income inequality. Policies to promote women's empowerment are the crucial factor here.

8. Lucia Hanmer concluded with various caveats. They included a warning that growth rates often fluctuate widely around forecast trends, and that income inequality can rise quickly in unstable economies.

9. Reporting on the Tanzania study, led by Graham Eele of OPM with REPOA of Dar es Salaam, Stephen Ackroyd emphasised data problems - paradoxically more constraining in a single-country enquiry - and the extent of policy change over the last decade. Recent figures suggest declining poverty incidence but rising inequality in Tanzania. The implications of inequality are similar to those in the global study. On cautiously optimistic assumptions based on improving recent growth performance, the income poverty target is met. On the other hand, the infant and under-five mortality, and even the primary-enrolment, targets seem unlikely to be met in view of current levels, and there are major uncertainties about maternal mortality.

10. For Uganda, Rosemary McGee reported a similarly mixed picture. The targets for economic well being did not seem attainable, given that they would require real growth of consumption of 4.9 per cent - higher than recent rates which represent growth from a very low base. Infant and under-five mortality achievements are likely to fall short of targets on account of AIDS. Universal primary education has been declared but quality is expected to decline sharply - perhaps a case of a target distorting priorities. And gender equity, even when narrowly conceived in terms of parity of enrolments, is not to be expected by the relevant target year (2005). There are clear messages for donors, including giving greater priority to the North and East, and to the rural subsistence sector, and resolving conflicts more swiftly.

11. Susan Joekes reported that the Pakistan case presents particular challenges to any study that involves extrapolations from past performance. Sharp changes in government policy and the external environment have taken place since Pakistan last achieved substantial reductions in poverty in the 1980s. Economic growth rates have been historically volatile, and the relationship between poverty and summary measures of inequality like the Gini coefficient is complex. Paradoxically in view of its poor human development record, Pakistan looks set to do better on improving educational access and reducing gender disparities than in improving economic well being. But this is partly a "fluke" arising from a massive injection of donor funds to address these deficiencies.

12. The Uganda country study underlines the point that national policies are being judged quite properly against nationally agreed targets, which cover the same range of issues but are not the same as the international ones. Donors need to bear this in mind when discussing results such as these in partner countries.

13. In response to questions, the speakers made several clarifications:

• Yes, the relationship between inequality and the growth/poverty elasticity is statistically robust, so policy implications do need to be drawn.
• Yes, population projections were taken into account (using ILO figures).
• No, debt relief was not considered in the econometrics, although it was important in Tanzanian and Ugandan poverty-reduction planning.
• Using a worst-case scenario for per capita growth that is below, rather than the same as, the historic trend would not have made a big difference, as the historic trend for sub-Saharan Africa was already zero.

14. Other points raised in the discussion included:

• A very large part of the story is going to have to do what happens to HIV/AIDS, as also highlighted in recent information from UNICEF.
• The other big question raised by this work is: what kinds of economic or social policies can lead to the low levels of inequality needed to produce poverty-reducing growth?
• The apparent paradox that in Tanzania and Pakistan there have been times when poverty has been reduced while inequality, as measured by the Gini coefficient, is worsening, points to the fact that this summary measure has serious limitations. There is a need to look at what is happening to the full range of the income distribution.
• What contribution can donors make? While domestic policies and institutions are clearly more important than anything else, donor funding and influence have already been important in Pakistan and Uganda, as mentioned. And there is considerable scope for more effort to change the policy-related factors influencing the spread of HIV.

David Booth