There is a useful symposium on Africa at the Boston Review revolving around Ted Miguel's essay, including contributions by Bob Bates, Paul Collier, and David Weil. The questions are: what is driving Africa's recent growth, and how sustainable is it? Is growth the result of good policies, democratic governance, Chinese investment, or high commodity prices? And is it a blip, or something more permanent?
Collier offers the simplest answer of all: African policy makers have been learning from their mistakes.
But the growth we are seeing today is not just a result of commodity booms. I don’t think that is the key to Kenya’s pre-election economic success. There is a process at work that does not depend on democracy and is so simple that analysts generally miss it: learning from mistakes. Since 1970 African societies have accumulated a huge stock of experience in how not to manage an economy. For example, from the mid-1970s until the mid-1980s Tanzania adopted regulatory policies that proved to be ruinous. The knowledge they gained through failure is valuable. Tanzania is now one of the best-managed of all Africa’s economies. The European society with the best record of containing inflation over the past sixty years is Germany. It has the best record because it used to have the worst: the experience of hyperinflation immunized Germans from macroeconomic folly.
Miguel endorses this view, but spins it differently, focusing on the advantaged of randomized impact evaluations.
While it’s natural to focus on such success stories, randomized evaluations don’t always produce positive results about program impacts. But information on failures is just as useful; it allows policymakers to shift funding from the projects that don’t work toward those that do. This is at the heart of the learning agenda that ... is the key to Africa’s economic future. Democracies like those emerging in Africa are particularly good learning environments, settings where impact evaluations can be carried out, their fruits widely distributed, and governments held accountable for applying their lessons to policy. In nations with weaker governance, rigorous program evaluations can themselves serve as a form of political accountability, empowering decent government officials to push for reform.
With impact-evaluation results in hand, policymakers in poor countries will increasingly be able to rely on hard evidence when deciding how to use their scarce resources. We now know the benefits of anti-parasitic drugs in improving school attendance in Busia, and as a result the Kenyan national government has included mass school-based deworming in its official school health plan for the country. Word has spread, and other African countries have expanded their own school deworming plans. In Ghana, over four million children received anti-parasitic drugs at school in 2007.
Learning about deworming is a small step forward on its own. But it will be through many such small lessons—in areas as diverse as health, education, agriculture, governance, and foreign aid—that African countries might learn to sustain and possibly augment their recent economic growth, even after the inevitable fall in global commodity prices.
I think Collier would say that the learning that comes from having followed disastrous macro, trade, and financial policies is much more important than any learning from randomized evaluations--even though the former hardly qualifies as "hard evidence" in the sense that Miguel uses the term.
I would add that randomized impact evaluations rarely generate usable "hard evidence" for policy makers. As soon as we generalize, scale up, or extrapolate we are in the business of transforming hard evidence into soft evidence. I wish Miguel were right when he writes: "With impact-evaluation results in hand, policymakers in poor countries will increasingly be able to rely on hard evidence when deciding how to use their scarce resources." Alas, we are doomed to live in a world of soft evidence. And pretending otherwise may itself end up skewing where resources go.