By David Apgar, the co-founder of GoalScreen LLC, which has a free cell phone app in testing at www.goalscreen.com that helps you test what you think drives results in your job, fitness program, love life, investments, and upcoming sports events in order to determine which assumptions are helping you succeed and which ones are undermining you
Jeff Sachs has always been the most outspoken advocate of development aid so it would be out of character if he were not outspoken about becoming the head of the World Bank. But there has always been a lingering concern about his projects and his approach to development. And it sheds a lot of light on where development, economics, and politics are heading even if it’s the wrong concern.
The concern is that because Sachs is so sure he has identified the best places to apply development aid – and so outraged aid offered to date has failed to meet his standards even for obvious causes such as health – he brushes aside risks like corruption and mistakes in program design. Worse, he sometimes seems to accuse aid critics of bad faith. And yet there are cases like Ethiopia where government military offensives have indeed followed periods of drought that saw big increases in assistance – at the urging of Sachs, among others.
In his defense, however, corruption in Africa is probably nowhere as vicious, privileged, and pervasive as what he saw in post-Soviet Russia. In Russia, furthermore, Sachs was only tangentially involved with Jonathon Hay’s conflict-ridden Harvard Project. And aid critics in the US routinely level baseless accusations of corruption, most recently at the Global Fund to Fight AIDS.
A deeper question for Sachs is whether impatience is really always a good thing in development. It’s a natural response, of course, when rich countries squabble about assistance levels in the middle of an acute famine in poor ones. Beyond the emergencies, however, impatience with the provision of funds for development is hard to separate from impatience with the often very slow process of learning about which interventions work. And his impatience puts Sachs in the company of the glassy-eyed optimists who have an enduring belief in the simplicity of development solutions.
The Millennium Development Goals, for which Sachs served as Special Advisor at the UN under Kofi Annan, represent a stark belief in the simplicity of development objectives if not solutions. The Goals focus on health, education, and poverty reduction. It is no coincidence that these are exactly the areas that allow statistical tests of effectiveness. You can jab a random sample of the kids coming into a clinic to see if they avoid tuberculosis but you cannot start dam construction near random villages up a river to see which ones supply the best labor.
Partly as a result, a cottage industry has sprung up, centered on MIT’s Jameel Poverty Action Lab, to test development effectiveness in the kinds of programs that admit randomized trials. Even as these trials shed light on health and education interventions, however, they are having the unintended effect of marginalizing the vast majority of programs that do not let you set aside a control group.
Money for metrics is the new mantra. On its own merits, it is a very powerful mantra. Monitoring and evaluation have transformed development assistance into a learning enterprise over the past ten years. And if formal impact evaluation is limited to health and education – well, those are not bad places to start. The trouble is the background assumption that everyone needs to measure the same thing. After all, that’s what statisticians do. It’s certainly what the Millennium Development Goals do.
And it’s true that indicators defined inconsistently across countries or aid projects can ruin an otherwise brilliant PhD dissertation. But what matters to a government trying to tackle endemic poverty or a region fighting a collapse in its water supply or a village with too many three-year-olds dying is just this – whether someone measures what people are doing about the problem with enough consistency to tell from results whether it works. Health and education programs that are amenable to control groups will benefit from faster learning. But learning does not stop in areas where control groups are impractical. It just slows down, as in much of the rest of life.
There is something that those governments, regions, villages, and aid organizations need to measure. They must measure outcomes of the most precisely described components and risk factors of the programs or strategies that they are pursuing. The more precise the strategy description, the more information actual outcomes will provide about whether and how it works. Just don’t expect those components and risk factors to mesh with other projects in other places attacking related problems.
In one sense, this cautionary note is not new. The Millennium Development Goals have drawn criticism precisely because their uniformity contrasts with the growing sense that every country has a separate path to development. The newer concern is the idea that you just cannot tell whether those paths are heading in the right direction unless you describe them precisely.
Precisely described strategies, however, are not widely shared. The challenge for brilliant, articulate development economists with an enduring belief in the simplicity of development solutions is not that simple solutions, be they free-market or statist, are inherently ineffective. The challenge is that solutions must be at least as complex as the problems they are to solve. And the simple problems have already been solved.