Crossing the great divide: Convincing others of what works to address poverty

By Betty Wilkinson on Thu, 23 January 2014

Photo courtesy of the Abdul Latif Jameel Poverty Action Lab (J-PAL) Facebook page
Photo courtesy of the Abdul Latif Jameel Poverty Action Lab (J-PAL) Facebook page

The Abdul Latif Jameel Poverty Action Lab just marked its tenth year anniversary on the MIT campus outside Boston with a one day event which drew over 1200 participants from all over the world.  J-PAL started as a small group of people with “a vision for generating evidence that could be translated into action to improve the lives of the poor” (according to its brochure).  Along with its sister organization, Innovations for Poverty Action, J-PAL designs, implements, and provides reports of results from randomized and other forms of evaluation of projects and programs to address poverty.  The focus of these two groups is on evaluating the actions taken by implementing organizations of all kinds to improve the lives of the poor, and seeing what specific step is the ‘trigger’ for reducing poverty. This is big business. In ten years J-PAL has become a worldwide organization with 91 affiliated professors and 439 completed or ongoing  randomized evaluations in 54 countries. Scale-ups of proven approaches have reached 165 million people.  

Three things were important about this anniversary event. The first was the sheer brainpower in the room. Many people who are creative and innovative are working on this problem. Leading academics, politicians, and people who have been very successful in the private sector attended and talked about fundamental issues of poverty reduction and quality of life.  For example, the former Minister of Planning in Chile Mr. Felipe Kast and former Chairman of the US Council of Economic Advisors Mr. Alan Kreuger spoke on these issues.  The number of young people in attendance was also impressive, showing that a large number of the ‘Gen X’ers and those born even later are truly worried about the circumstances of the poor and the need for change. 

Second, there were presentations of evidence for specific actions that can reduce poverty. The lives of the poor are complex, and their motives are not always rational.  Increasingly both development practitioners and policymakers are engaged with researchers to design studies to see what specific interventions can work, and results are gradually emerging. Evaluations take time and money, and the fact that so many are being done collaboratively is useful. While there are limitations in randomized evaluations, they can help us see what specific interventions we can use, and how.  Simple things, like deworming for children to help their school performance, are often quite powerful.

The third and most fundamental aspect of the meetings was the need to get positive results implemented more broadly.  The co-founders of J-PAL expect to spend much of the next ten years on “scaling up” or replicating successful activities which help the poor at a larger level, either nationally or internationally.  This is difficult for three reasons― policymakers have different motives, timelines and agendas than researchers; it is not easy to ensure that national government resources are allocated for implementation of these results; and the public may be unconvinced that evaluation results from elsewhere are applicable to their own circumstances.

Scaling up is the key challenge and a critical next step.  The evidence produced by J-PAL and its sister organization across a number of countries on three continents is that with the right programs, for example, the ultra-poor can be helped over a three year period to substantially and sustainably improve their family well-being. But no country ―or donor― has yet stepped up to implement these programs at scale.  We need to find better ways of sharing evaluation results in ways which will enable and persuade country policymakers to take them seriously, to figure out how to implement them in their own countries, and to get help if and when needed.

Subjects: