Poor Economics: The Challenges of Defining Development

Poor Economics
By Simon Halliday

We all agree that development is a good thing, but how do we know if or when a given policy works? Poor Economics enters the debate by detailing how randomised controlled trials (RCTs) help us to evaluate microeconomic development policy, while simultaneously underselling the benefits of alternative methods. Nevertheless, the book is insightful and provides a more nuanced picture of the issues of poverty and development than is typical of the popular literature on aid and development.

In each chapter, the authors interrogate narrow questions examining studies on topics ranging from hunger and health to political empowerment. Their approach becomes familiar: What do we think we know? What was the mainstream approach? What can “the poor tell us”? What are the “counterintuitive results”?

The authors refer predominantly to evidence from RCTs. On page eight, they write: “The cleanest way to answer such questions is to mimic the randomised trials that are used in medicine to evaluate the effectiveness of new drugs”.

But a social policy is not a drug. Doctors can assume that human bodies respond similarly for medical trials and thus, given a large enough sample size, a medical RCT in India should reproduce results found in Britain, Uruguay or Kenya. RCTs of human behaviour do not necessarily translate into generalisations because of different cultures, languages, political contexts, and institutions.

Nevertheless, Poor Economics teaches us that textbooks are ineffective at helping kids in Kenya learn, maybe get them de-wormed instead. Teachers target the top of the class, so stream (“track”) classes to improve education. Microlending makes a small difference because it is structured against default, thus against failure and taking risks (required for successful entrepreneurship).

Banerjee and Duflo present many results like these, always trying to set the studies in their local contexts, considering the lives of the subjects, and trying to tease out why the solutions worked. To this end the results of RCTs become something more - the underpinning evidence for “Five Key Lessons”:

1. The poor lack information and believe untrue things.

2. The poor bear responsibility for too many aspects of their lives (e.g. they must fetch and chlorinate water rather than receiving piped water).

3. There are good reasons some markets are missing for the poor (e.g. savings, health and insurance are costly and the poor won’t purchase them at market prices).

4. Poor countries are not doomed because they are poor, or because of unfortunate histories.

5. Expectations can turn into self-fulfilling prophecies (e.g. kids who are not taught to their level choose to give up, consequently teachers believe it was not worth the effort).

In trying to draw out these lessons, however, Banerjee and Duflo employ non-economic interpretations of why and how people conduct their lives.  While the authors focus, perhaps too greatly, on RCTs, they neglect the methods of the non-economic social sciences such as sociology, psychology and anthropology. Though RCTs are valuable, they are not the only way to study development, and they cannot always help us to tackle “big picture questions” or questions that cannot be answered by randomisation (so-called “FUQs”). Moreover, though economics is a useful tool, it cannot tell us everything we need to know about cultural dynamics, change, or the specifics of a time, a place and a people.

To examine people the authors populate the book with several animating characters, the interviewees: Ben Sedan, Ibu Tina, Xu Aihua. Are there no poor white people for economists to talk to? Poor people, it seems, live in Africa and Asia, but not in North America and Europe. But the book is called Poor Economics not The Economics of Poor People Outside the Developed World. If something unites the poor internationally, the authors should conduct or refer to equivalent experiments in the developed world. Moreover, as Andy Sumner reminds us, the majority of the poor no longer live in least developed countries – the majority of the poor live in middle income countries which calls into question what exactly differentiates the poor in the developing world from the poor elsewhere.

The book examines a valuable stream of literature that has emerged in development microeconomics over the past two decades the effects of which have begun to be felt internationally in governments, NGOs and board rooms. The agenda is crucial: providing high quality evidence to establish whether an intervention works. But the method’s advocates (so-called “randomistas”) need to be humbler about its applications and to recognise the many other disciplines that can inform development.

The book’s subtitle claims the work as “A Radical Rethinking”, but instead its content and advice are simply sensible. The authors’ advice harmonises with that suggested by others across the discipline (such as Eric Beinhocker and Paul Seabright), and it certainly offers interesting, evidence-based, and politically challenging advice, as applied to development. Ultimately, it will please the left because the authors argue that governments need to invest in and subsidise the poor. It will please the right because it emphasises that individuals often facilitate their own success. But it may frustrate both sides because of its particularism: sometimes markets work, sometimes they fail; sometimes incentives succeed, sometimes they backfire; sometimes successful representation calls for affirmative action, sometimes it just needs an electronic voting machine.


Simon Halliday is completing his PhD in Economics at the University of Siena, Italy. He has previously worked with the Southern Africa Labour and Development Research Unit as an Assistant Manager for the Quality of Life Survey of Land Reform in South Africa, and as a Lecturer in Economics in the School of Economics at the University of Cape Town. Simon's PhD research focuses on experimental economics and economics education in South Africa. 


10 September 2011