Doing Poorly: A Lateral Section

he economic lives of the poor have been rendered invisible in three ways. At a popular level, there are prejudices about what the poor must be like as economic agents. Many believe that given half a chance, the poor will drink themselves to death rather than plan for their future. But what do we really know about the poor? What do they save? What are their needs and preferences? What are their beliefs? Many of our assumptions about these questions have simply no evidence to back them.

This thicket of assumptions is often countered by the abstractions of standard economic theory. On this view the poor are, plausibly, viewed as economic agents just like anyone else. They respond to incentives, they rationally plan for their future, they assess risk and reward. But these are just abstractions. Saying that economic agents respond to incentives is one thing. Describing what kinds of concrete incentives actually matter to them is another thing altogether. The third set of abstractions that render the poor invisible is the penchant for large schemes. The poor must be the objects of our beneficence. The state provides a set of services for them: health, education, credit and so forth. But how do the poor respond to these services? How do we evaluate their impact? Do the poor follow the script on which this state beneficence is constructed? Or is the architecture of these services orthogonal to their needs?

Poor Economics is a tour de force that cuts through many of these abstractions. It vividly, sensitively and rigorously brings alive the dilemmas of the poor as economic agents in a variety of contexts, whether as consumers or risk-takers. There are splendid chapters on a variety of topics that affect the poor: food, health, education, savings, micro-credit, insurance, risk and even some cursory observations on political behaviour. Unifying their rich analysis is one big insight about the structural situation of the poor. The poor inhabit what Banerjee and Dufflo call a poverty trap zone. Most privileged economic actors can assume that investments today will yield augmented returns tomorrow. The poor inhabit a horizon where their future incomes over a life cycle are likely to be lower.

Part of this has to do with threshold effects. Even if the poor experience incremental changes, they are not often sufficient to change the trajectory of their life’s expectations. A lot of the argument in the book turns on this psychological insight. The poor often behave the way they do because “they think that any change that is significant enough to be worth sacrificing for will simply take too long”. So, much of the book focuses on the kinds of small, well-designed interventions that can lift the poor out of these poverty traps. The focus is not on large structures or abstract theory, but on the behaviour of the poor. What do the poor do? The sheer richness of its observations on the economic lives of the poor (and by implication everyone else) makes it a landmark. It is gracefully written, full of humanity and insight.

Poor Economics has an astonishing wealth of arguments. The chapter on food points out, for example, that the real challenge of poverty now is not hunger, but nutrition. As incomes for the poor rise, instead of consuming more calories or more nutritious food, they consume more expensive calories. Another astonishing puzzle they pose is that of a Chennai fruit vendor. He borrows a thousand rupees each morning at an interest of 4.69 per cent a day (just think of the compounded annual interest they pay). But also consider this fact. If the vendors simply decided to drink fewer cups, it would save them five rupees a day. Seems like a small amount. But they would have to borrow five rupees less, and if ploughed back into the business, could make them debt-free in about three months. As Banerjee and Dufflo put it, “the vendors are sitting under what appears to be as close to a money tree as we are likely to find anywhere. Why don’t they shake it a bit more?”

It turns out that many of the barriers to saving arise from human psychology. Apart from the usual challenges that all agents face, the difficulty of self-control, the prevalence of time inconsistency in our preferences, saving is often less attractive to the poor because the goal of escaping poverty seems very far away. As Adam Smith had argued, taste and imagination, rather than need and necessity, often shapes patterns of consumption, even amongst the poor.

r think of another example: Simple chlorination of water would reduce disease considerably. But is a system that relies on individuals to do it every day likely to work, compared to systems where someone did it for you? Preventive healthcare is much cheaper than post facto cures. But how many of us have not experienced the effects of procrastination? In addition, there is another problem: how does one know whom to trust? The authors chart in insightful detail why the poor don’t have reason to trust several institutions we impose upon them, from health clinics to poorly run schools. The poor inhabit lives that are primarily marked by something we take for granted: systems of institutionalised trust. Much in line with recent behavioural economics, Banerjee and Dufflo argue that these psychological complexities need to be taken into account in designing effective programs. Small “nudges” can be designed that makes the poor more effective agents.

How do we know what nudges are effective? Banerjee and Dufflo’s core claim is that the effectiveness of policy interventions can be tested by a method akin to that used in medicine: the use of randomised trials. This is a complex process. But to simplify somewhat, the idea is that just as we test the effectiveness of drugs by doing randomised trials, we can test the effectiveness of policy interventions by conducting “experiments”. While it is widely regarded as a tool for evaluating programmes, the idea itself is very controversial for several powerful reasons. Randomised trials work in medicine because physical reality is thought to be invariant across time and space. Social reality is not. Even valid results of experiments yield what the great anthropologist Clifford Geertz once called generalisations within cases, rather than across cases.

Second, there is still a massive problem of identifying the mechanisms that explain the results of these trials. But their zealous advocacy of randomised trials has made Banerjee and Dufflo the hottest economists on the planet; the stuff Nobel Prizes are made of.

But this book might, paradoxically, provide more grist to the mill of those who are sceptical about how much randomised trials can actually tell us. Randomised trials can improve some interventions. But the obsession with evaluation can also be a trap that inhibits real experimentation and learning by doing, something so many of the great ngos in this book—Seva Mandir, Pratham or Spandana—excel at. Dufflo and Banerjee eschew big structural questions. They see themselves as designing interventions that do not require massive changes in political economy or large revolutionary transformation. They are great believers in the cumulative effects of small interventions. Whether this is the only standpoint to adopt in development is an open question. But Dufflo and Banerjee have demonstrated the power of paying attention to what the poor do.

The interesting parts of the book do not rely on any evidence gathered from randomised trials. The authors use a whole range of evidence, from econometric studies, to history, to anthropological observation. The results from the randomised trials they report are interesting, but not overwhelming. And they certainly do not add up to a case for using randomised trials as uniquely compelling pieces of evidence. In fact, the most powerful moments in the book are almost touchingly old-fashioned. In the chapter on education, there is a poignant moment that tells you more about the ways in which our education system fails the poor than any randomised trial would. This is the moment where one of their interlocutors uses the phrase “children from homes like ours..,” highlighting a persistent problem of treating the poor as another species. Banerjee and Dufflo indict the system for its low expectations of what poor students can accomplish; these low expectations constitute the poverty trap the poor are trying to escape. Non-economists may have an interest in exaggerating this aspect. But the qualities of research that stand out most vividly in this book are not the randomised trials, but the richness with which Dufflo and Banerjee bring the poor into the conversation. We are grateful to randomised trials because they have turned economists into first-rate anthropologists.