Germany can make an important contribution to Africa's development when it hosts the summit of the Group of Eight leading industrial nations in June; not by adding new promises, but by ensuring implementation of those already made.

The most important, by far, is the 2005 G8 promise to double aid to Africa from Dollars 25bn in 2004 to at least Dollars 50bn in 2010. Now the G8 needs to set out the timescale for increasing aid to each recipient country - conditional on reasonable governance - so that African countries can plan and undertake the multi-year investments that will make such a difference to their long-term development.

The current situation is a bit bizarre: African governments have been explicitly encouraged to increase their investments in disease control, water, sanitation, roads, power, schools, clinics and the like. They have been advised to drop user fees in education and health. They have been told that aid will double by 2010. But they have not been told how, when, or where the aid increases will take place, even though 2010 is only three years away. The result is a lose-lose situation for both donors and recipients.

Governments in Africa cannot plan even a year or two ahead, much less to 2015, when they are supposed to achieve the United Nations' millennium development goals. Public investments in health, education and infrastructure lie at the core of the goals. Children's lives are saved by preventative health investments such as anti-malaria bed nets and de-worming and by therapeutic health investments in clinical services. Children are educated when there are trained teachers, classrooms and midday school meals. Market economies are promoted when paved roads, power grids and functional ports allow a country to compete in world markets.

Each of these basic steps requires multi-year programmes for the physical investments, the training of skilled workers and the building of management systems. These practical investments cannot be financed by African countries alone - hence the pledges of aid. So far so good. But the investments also require timetables, strategies and multi-year planning. Predictability is essential.

Everybody recognises these needs, but here is the reality. Following the G8 pledge in 2005 there have been no proper timetables set out for African countries explaining to them how their aid levels will rise to meet the 2010 pledge. In country after country, African leaders and finance ministers are told by local representatives of the donor agencies that there is no information as to how or when aid levels will increase to meet the commitment. In some countries, recipients of US development aid have been advised that aid, other than that for HIV/Aids and malaria, is being cut as budgets are shifted to Iraq.

It is understandable that G8 finance ministers would like to leave themselves maximum flexibility in the timing and direction of aid, but this approach to development assistance ends up being self- defeating.

Instead of a rational process of raising critical investments, we have a guessing game. Will the aid be doubled or not? Will it be doubled as an accounting trick (for example, by cancelling unpayable debts and labelling each dollar of debt cancellation as a dollar of aid), or will it be an actual flow of commodities and cash? Will the aid come as fees for high-priced consultants or as funds for practical investment? Will the G8 wait until 2010 to increase actual cash flows, or will the aid increase step-by-step between now and then?

The basic arithmetic can be made clear. The tropical sub-Saharan region, which must get the lion's share of aid, has about 725m people, rising to about 775m by 2010. Perhaps 150m or more live in countries that will not qualify for much aid for one reason or another (such as poor governance, lack of plans, less need). Some 625m people in 2010 would then share most of the Dollars 50bn or more of aid, leading to about Dollars 80 per capita of aid by then in the well-governed, high-need countries, a level far higher than now.

The step-by-step increase to Dollars 80 or more should be made predictable and therefore the basis for training, investment and the building up of management. The International Monetary Fund could then use such predictable aid trajectories in working with the countries to set medium-term expenditure frameworks in the social and infrastructure sectors.

Germany is no stranger to careful fiscal policy and multi-year budgeting. It can rightly impress upon the rest of the G8 members that such straightforward and transparent budgeting is a mark of basic good governance on the side of the donor countries. With this simple step, the G8 could foster enormous trust that its promises to Africa are real.

Moreover, we would finally get onto the serious work of increasing the life-saving and growth-promoting investments that lie at the heart of public sector responsibility - a responsibility that finally would be adopted on both sides of the donor-recipient relationship.

The writer is director of the Earth Institute at Columbia University and author of The End of Poverty (Penguin)