Why were so many people hanging around in the hospital at Kamenge, my wife wanted to know? The answer was that in Burundi, having a clean bill of health has taken on a very particular meaning: unless and until you have paid for your hospital treatment, you simply can't leave, you are in effect a captive. Being detained and being discharged have developed a close and special relationship in the health care system of this small and struggling African state.

In the wake of a World Bank/IMF package, the government felt compelled to impose what is euphemistically known as a 'cost recovery' system for medical care – in plain English, you have to pay directly and on the spot. Medical staff acknowledge the cruelty and absurdity, but cannot otherwise provide anything at all.

This is where the arguments that have washed backwards and forwards about the recent G8 summit suddenly become immediate. Aid, trade, good government, debt relief, food security – where does the virtuous cycle of development really begin?

But on the ground there is little time to sort out those arguments. Images from Niger have brought that fact vividly home to everyone. I first saw them on the way back from a post-Gleneagles trip to Burundi. Burundi is next after Niger in the catalogue of the world's poorest states. Those images served to intensify an impatience that could not be stifled by the rational recognition that Gleneagles could have been a lot worse.

The trouble with plenty of talk in and around the G8 summit was that it created an abstraction – the Problem of Poverty in Africa. Generalisations abounded – about the benefits or otherwise of trade liberalisation, about governmental corruption and conditionality in debt remission.

But the truth is an uncomfortable one for summits, commentators and international agencies alike: development is a finely woven net of endless small changes. It is delivered by equipping millions of individuals at countless different levels. And it needs to happen without too much dogma about where the cycle 'ought' to begin.

Burundi is an overpopulated country of rather more than six million people, recovering from twelve years of murderous civil strife. In recent months, the first successful election in nearly a generation has taken place. The president has agreed to stand down. A former rebel leader, brought into the administration as Minister for Good Governance, is the likely next president. Almost every rebel group has signed up to the peace accord. The country is, to all appearances, poised for change.

But this is also a moment of high risk. Expectations are huge. Thousands of refugees are planning to return to family lands vacated a decade ago. However, soil degradation and drought mean that food security is on a knife edge in many areas. Those who want to go back to their land will find it cannot sustain them.

Jobs are few, industry almost non-existent. Somehow the army has to shed about a third of its personnel. There is no funding for new equipment or building: the condition of schools and hospitals is dire. The university library I visited had made no serious acquisitions for twelve years.
Without tangible change in the next eighteen months, Burundi could sink back into the chaos from which it has begun to emerge. As the outgoing president, Domitien Ndayizeye, told us: while he is trying to persuade the international community that Burundi has attained an 'acceptable' level of good government, growing popular frustration makes the continuance of that good government even harder.

A hectic week of trips and meetings left me feeling deep frustration at the Catch 22 situation that refuses investment until stability is confirmed and thereby tends to perpetuate the instability that creates the problem. Our thinking about the need for democracy and accountability before aid or investment is fine in the abstract. In practice, in places like Burundi holding back until all possible conditions are met may simply intensify the risks.

All this is compounded by a worrying level of hostility or cynicism in some educated circles about what might be called the 'development establishment'. There is resentment about people, who apparently enjoy standards and styles of life very remote from those of most others, imposing structures of development and levels of provision that do not belong and cannot be sustained.

It is possible, critics say, to set your short-term sights too high; for example, in creating health education services or fixing unrealistic expectations of teachers' salaries. And there is scepticism about a culture of 'capacity-building' conferences that do not connect adequately with real lives. The criticism may be unfair but it takes its toll.

Burundi urgently needs much higher levels of aid and investment; it needs outside sponsors to take advantage of what could be a brief window of stability. It needs a new regime for the funding of health care. But it also poses a bigger question about aid methods generally.

Who is able to deliver development goals at grass roots? Who can provide basic literacy and health education, especially for women; basic economic capacity, to move away from subsistence farming towards modest enterprise?

The answer is not unequivocally welcome to many governments and many NGO's. For churches and other faith groups are often the only networks that have sufficient credibility and reach to deliver. Real advances have to lie in more and better partnerships with such groups.

The Mothers' Union is one of those Anglican agencies regarded with a rather patronising affection by large sections of the British public – part of the warm beer, church bells and cricket-on-the-green image of a rather tired Englishness. This is an image that is ludicrously off the mark.

In much of Africa, it is quite simply the most effective agency for the empowerment of women. Burundi is a majority Roman Catholic country, with international Catholic aid agencies at work; but the Mothers' Union remains an important player. Its creation of a network of literacy schemes, of programmes for sex education and nutritional advice, of training in conflict resolution, of micro-finance projects and trading initiatives, all have immense potential. How is this potential to be developed?

It hardly needs saying that faith organisations are not problem free. Churches and faith-based groups in Africa have sometimes been slow to condemn the tribalism that scars so many societies; they have not been immune to corruption, and they have been defensive about outside challenges over transparency and accountability.
At the official level, they can still be alarmingly insensitive about the role of women. But they do have credibility; and no scheme that bypasses or ignores them can claim comparable scope. They continue to reach the parts other agencies don't.

Time is short, for many countries. Will governments and the aid establishment rise to the challenge of supporting the kind of grass roots development that does not depend on the elephantine slowness of negotiation about whether a country has arrived at a tolerable level of stability? If partnership with local faith-based projects can make even a small immediate difference to opening up real and practical advances, is it not something that the donor community should be looking at rather harder?

© Rowan Williams 2005