Paul Collier

The Bottom Billion

If governments were being ordered about by donor agencies, whom should an electorate blame if things went wrong? Governments were quick to exploit the full potential for evading responsibility. In the week when the government of Zimbabwe launched economic reforms in 1998, its minister of information told the local press, 'They're not our reforms, they're the IMF's. We had to do them.' That sort of statement not only shifted the responsibility but made the reforms very easy to reverse. And the government of Zimbabwe most surely reversed them.

Policy conditionaluy as then practiced depended upon the government promising to make changes. This, for those of you who prefer to think in Latin, is known as ex ante conditionality The economics of getting money on the basis of a promise is known as the time consistency problem. Unless incentives are properly aligned, governments will promise, take the money, and then do what they like. To give you a real-life example, the government of Kenya promised the same reform to the World Bank in return for aid five times over a fifteen-year period. Yes, five times it took the money and either did nothing or made token, reforms that it then reversed. The amazing thing is that the money kept coming. How did Kenyan government officials manage to keep straight, sincere faces as for the fifth time they made the same commitment? How did officials of the agency manage to delude themselves into thinking that adherence this time was likely? But aid agencies have very little incentive to enforce conditions: people get promoted by disbursing money, not by withholding it. Eventually the World Bank and other donor agencies realized this limitation and largely switched to disbursing aid on the basis of the attained level of policies rather than on promises of improvement. For the Latin-speakers, this is ex post conditionality It was more consistent with the research evidence that suggested that whether aid worked well depended upon the level of policies rather than on how they were changing, and it avoided the need for promises. The only problem was that it squeezed aid out of the very countries that have the biggest problems. On the most favorable interpretation, it was a realistic recognition of the limits of aid to help in these environments. On a less favorable interpretation, it was giving up on the very environments where the agencies were most needed. Anyway, forget ex ante policy conditionality as a way of inducing policy improvement in failing states: it just doesn't work.

I take a very different view of governance conditionality. The key objective of governance conditionality is not to shift power from governments to donors but to shift power from governments to their own citizens. The struggle for this transfer of power took around two hundred years in Europe, and we should indeed want to speed it up in the bottom billion. External pressure was vital in the European struggle. The most common account of that struggle goes as follows. The threat of war forced governments to defend themselves with big armies. To pay for these armies the governments needed to tax. To get compliance for high taxation they had to concede representation and scrutiny We cannot; go through that process in the bottom billion. In Europe the threat of war turned into a reality sufficiently often for the whole process to have been murderous, and it would probably be so again. It was also slow. But the purely internal processes by which citizens force governments to accept scrutiny are probably pretty weak. External pressure is needed. And it is entirely legitimate. Why should we give aid to governments that are not willing to let their citizens see how they spend it?

Governance conditionality, in its ex post form, is gaining in popularity. Most dramatically, U.S. president George W. Bush launched his new Millennium Challenge Account based largely on allocation criteria of attained levels of governance. He wisely chose not to allocate the additional American aid money through the established American aid agency, for over the years USAID has been captured by congressional commercial lobbies, Voting line by line on USAID's budget. Congress has diverted spending so as to benefit particular American exporters, unrelated to African needs. However, somewhat surprisingly, no agency is doing ex ante governance conditionality One advantage of such an approach is that it would be much clearer what a government had to do, and on what time scale, in order to be rewarded by extra aid. And the aid could be targeted to countries that initially had weak governance—so it would focus on the bottom billion.