In his new book, The End of Poverty, New York: Penguin Press, 2005, and other more recent writings, Jeffrey Sachs makes a persuasive case for ending extreme poverty (that is, the proportion of population living on less than $1 a day) by 2025, by honouring the Monterrey Consensus of 2002 of raising the rich countries' share of Overseas Development Assistance in GNP from 0.2 per cent to 0.7 per cent.

As of 2002, the aid was $53 billion, just 0.2 per cent of the rich-world GNP. If the Monterrey target is reached, the aid would be $175 billion a year. For the US, foreign aid would rise from around $15 billion in 2004 (0.14 per cent of GNP) to around $75 billion (0.7 per cent of GNP).

Based on the estimates of the UN Millennium Project, net annual ODA flows required to achieve the Millennium Development Goals (specifically, of halving the proportion of poor by 2015) would range from $135 to $195 billion over the period 2005-2015.

The important point is that the required ODA flows would be 0.44 to 0.54 per cent of the rich-world GNP, well below the 0.7 per cent commitment made in Monterrey in 2002.

Vigorous advocacy by Tony Blair at the recent G-8 Summit at Gleneagles renewed the hope of additional annual aid flows of $50 billion, of which at least half will be directed to African countries.

Despite foot-dragging by the US and refusal of Japan and Canada to commit to the Monterrey time table, this is a step forward by rich countries towards eliminating the scourge of extreme poverty. But, as argued below, it is just that.

That the Monterrey target of 0.7 per cent of rich world GNP is affordable is not in question, as it would mean, for instance, an extra tax of 0.55 per cent of US GNP or a cut in US military spending.

The US spent 30 times more on the military than on foreign assistance in 2004, $450 billion compared with $15 billion. So the question is not, asks Sachs provocatively, whether the rich can afford to spend more on the poor but whether they can afford not to.

He then unveils his vision of development in which extending assistance for reconstruction and development, and humanitarian relief are "acts of enlightened self-interest" (Sachs, p.331), as these would end extreme poverty, spread democracy, and harness science and technology to advance human well-being.

While there is much in this vision that is bold, imaginative and compelling -- specifically, the role of rich countries in ending extreme poverty by 2025 -- it is somewhat dismissive of institutional constraints and what developing countries must do to overcome them.

While conceding the need for improving the quality of governance, he is emphatic that the concern for it is exaggerated. He asserts that higher incomes and literacy will improve governance in two ways -- by empowering civil society, and improving the quality of public administration.

A recent study (Gaiha, R K Imai, and M A Nandhi: "Millennium Development Goal of Halving Poverty in Asia and the Pacific: Progress, Prospects and Priorities", Discussion Paper 1, Rome: Asia and the Pacific Division, IFAD, 2005) points to a reversal of this causality, that is, instead of growth improving the quality of institutions, it is the latter that accelerates growth and, thus, reduces poverty.

The econometric model is estimated using a data base compiled from the World Bank, FAO and United Nations.

Five institutional quality indicators are used, comprising voice and accountability, political stability and absence of violence, control of corruption, rule of law, and an aggregate index of institutional quality (Kaufmann, D A Kraay and P Zoido-Lobaton: "Governance Matters III: Updated Governance Indicators for 1996-02", Washington DC: World Bank, 2003).

Some important findings are cited here. One is that institutional quality impacts poverty through income and not directly. In line with the work of Dani Rodrik and others, after allowing for variation in the index of openness due to institutional quality, geographical factors (for instance, the share of coastal population) and the size of a country, openness ceases to have any effect on income.

Finally, even modest improvements in institutional quality are associated with significant positive effects on income and consequently on poverty. For instance, with the voice and accountability index assumed to take on the average value of the top 30 performers in the sample, the headcount index will register marked reductions.

To illustrate, the dollar poor index for India in 2015 with greater voice and accountability will be 10.3 per cent, as against 14.7 per cent without any improvement; in Bangladesh, it will be 8.6 per cent, relative to 21.2 per cent; and in China, the head-count ratio will be as low as 1.1 per cent, relative to 6.1 per cent in the baseline scenario.

So, while modest goals of improvements in institutional quality may suffice, all too often (Sachs included) the constraints arising from elite domination and "capture" of local institutions (village panchayats, for instance) are overlooked.

Careful attention must, therefore, be given to strengthening the voice and accountability mechanisms (by, for instance, facilitating access to official records). Also, given the pervasiveness of co-ordination failures, how partnerships could be better forged between public and private agencies is perhaps equally important (for instance, between governments and markets).

In conclusion, while rich countries undoubtedly have a key role in ending extreme poverty, it would be a mistake to overlook the equally important role of developing countries in carrying out institutional reforms. In the absence of the latter, it would be over-optimistic to expect larger aid flows to translate into substantial reduction of extreme poverty.