Book Review: Poor Economics
Economics, as it is traditionally taught, is preoccupied with analyzing processes and institutions that shape modern, industrialized nations. By contrast, scant attention is given to the economic decisions of the world’s poor, presumably because their choices seem so constrained that there simply isn’t much worth studying. Abhijit Banerjee and Esther Duflo, recipients of the 2019 Nobel Prize in Economics, beg to differ.
A familiar sight in countries outside the narrow cluster of WEIRD nations, especially in impoverished rural areas, is that of unfinished, unoccupied buildings. They aren’t just unfinished in the sense of being under construction, but rather they are abandoned, uninhabitable, at times at the verge of collapsing. Nor are these the remnants of a once-successful region that first ran out of natural resources and then of workers, or urban mega-projects left behind by overambitious investors. In fact, their owners typically live nearby and have a very real interest in seeing them completed, only that they are too destitute to bring the project to a timely closure. Many of them belong to what economist Paul Collier calls the “bottom billion”: people who live on less than $2 a day, and who are consequently facing existential danger day in, day out. Given their very limited means, these construction projects seem crazy at first glance: Exposed to the elements, unfinished buildings are deteriorating quickly, and money sunk into brick and mortar cannot easily be transformed back if an emergency comes up — and there is always one lurking just around the corner.
Except for a few conservative outposts, the notion that the poor themselves are largely to blame for their predicament — typically because of short-sightedness, impulsiveness and/or plain old stupidity — has lost almost all of its credit, certainly among academics and intellectuals. Alternative theories, whether of the postcolonial “structural injustice” or of the technocratic “lack of infrastructure and education” variety, have since taken over, which are less patronizing but, alas, also less plausible when it comes to explaining what drives so many poor people to invest in ill-fated construction projects. It would take a massive stretch of the imagination to make the case that these decisions are all the result of the long shadow of colonialism.
Enter Poor Economics, published in 2011 by a French-Indian couple, Esther Duflo and Abhijit Banerjee, who would become giants in the field of development economics. Their work suggests an altogether different angle to make sense of the observed phenomenon: As it turns out, once you get up from your desk and do something as revolutionary as simply asking people why they chose to invest whatever little savings they manage to put aside in this way, you’ll sooner or later realize that they understand perfectly well what they are doing. Making their assets less liquid, even in face of dramatic (and predictable) devaluation, is seen as the only way to accumulate any wealth at all, as there will always be more urgent demands. It’s the same reason why millions of well-off people in the West pour their money into accounts that penalize early withdrawal: Foreclosing options for immediate gratification is how we prepare for old age.
Poor Economics is, at its core, a highly empathic attempt to come to grips with the seemingly paradoxical and self-defeating everyday choices of the world’s poor. It’s also a prime example of a work that engages those who disagree with its perspective in good faith. Known mostly for its advocacy of randomized controlled trials (RCTs) as a means to establish causal relationships in economics — which, among other things, killed off the hype around microfinance — , it can also be read as a defense of the much-ridiculed rational choice theory. Perplexing as it may seem at first glance, I would claim that the subjects of their book all display behaviors best described as bounded rationality, rather than bias or sheer ignorance. That seems like a rather idiosyncratic interpretation, so let me back it up.
Before I begin, an obvious disclaimer: As anyone who ever took an introductory class in economics surely knows, a collection of rational individuals does not imply that society as a whole acts rationally, too. At least to a first approximation, an army composed of perfectly rational, selfish soldiers will flee because it makes sense for each individual to run off (hence the harsh punishment for déserteurs). A world inhabited by rational, selfish individuals will see everyone polluting the atmosphere to their heart’s content. Similarly, what is rational for the individual poor person may well result in a worse outcome for all. So there is no contradiction in upholding that each person acts rationally, and yet,as a collective, fail to coordinate their actions to improve their lot.
To see what I mean, let’s look at food. There’s a widespread notion in the West that people who live on less than $2/day are all on the brink of starvation — literally unable to afford enough food to survive. Yet this is not the case: Absent famines, even the wretched of the earth typically do not suffer from a sheer lack of calories. That does not mean that their diet is particularly healthy, and in fact it’s often lacking important nutrients that can’t be found in staple crops. Yet attempts to educate villagers about the benefits of a more balanced diet, or to offer allowances that would help them supplement the usual servings of rice, bananas or manioc have met with little success. Instead of using the extra money to purchase “healthy” food, it was more likely to be spent on sugar and other indulgences. Irrational? Not necessarily, say Banerjee and Duflo. Think of the many occasions on which you were unable to resist a bag of chips or a generous serving of ice cream, and then try put yourself in the shoes of a penniless Indonesian who is supposed to say no to a little guilty pleasure: How sure are you that you could muster the willpower to pick the healthy option instead — especially given that the pay-off, if it ever materializes, lies far in the future?
Here’s another example: It’s a well-known fact that fertility rates are typically highest in very poor countries — in 2021, Niger, Somalia, Chad and the DRC came out on top, all with more than six children per women. In some sense, this is extremely counter-intuitive: Shouldn’t it be precisely the other way round, in the sense that one would expect a strong correlation between wealth and number of children per family, who can now afford additional offspring? Since this is clearly not what we observe, many theories have been put forth to explain how the poor end up having very big families: Subpar or non-existent sanitation plus lack of access to proper health care result in many children dying in infancy, and having more babies is simply a strategy to cope with this terrible reality. Or maybe they cannot purchase, or do not know how to use, contraceptives — or if they do, they don’t possess enough self-control to avoid unwanted pregnancies. Alternatively, there may be strong social norms in place that, in one way or another, urge the members of the community to be fruitful and multiply.
No doubt these factors play a role, but what Poor Economics argues is that even after controlling for all these factors, there’s still a huge gap left to explain. And when interviewed, it becomes clear that subjects understood rather well what they were doing, and that having lots of children strained their already very limited resources. But, they say, it’s like rolling a dice: Try once, and chances are it’ll land on a low number. Yet if you keep playing, eventually you’re bound to produce a six — the equivalent of which would be a child that can support you in old age.
If this analysis is correct, what does it imply for the development community? Debates in these circles have often focused on big questions, which has resulted in two irreconcilable camps with almost diametrically opposed views and policy prescriptions. Banerjee and Duflo think such perspectives obscure more than they illuminate, and recommend we focus on the concrete context, rather than trying to devise one-size-fits-all solutions:
“This book will not tell you whether aid is good or bad, but it will say whether particular instances of aid did some good or not. We cannot pronounce on the efficacy of democracy, but we do have something to say about whether democracy could be made more effective in rural Indonesia by changing the way it is organized on the ground, and so on.” (p. 4–5)
Their preferred analytical tool is the concept of poverty traps: Roughly speaking, such traps imply that below a certain level of wealth or income, for every step forward there are two steps back. Or, in a language more suitable to the context, there are negative returns for every dollar invested. This may mean bank accounts with negative interest rates, a small business that’s never becoming profitable (or even losing money on average), or choosing not to invest in their children’s education because of the belief that the investment is more expensive than what they can get out of it. The question then becomes whether a specific situation is indeed best characterized as a poverty trap (so that it would make sense to “push” the poor past the inflection point, after which compound effects kick in) or not (in which case they might often be able to pull themselves out without any external support). However, it’s not enough to merely identify those traps, it’s also crucial to understand their nature: How present income compares to future income is massively shaped by expectations, and those might be much harder to alter than simply providing goods or extra cash.
It would speak to a certain lack of ambition if the book simply tried to convince policymakers to step out of their ideological straitjackets, avoid reducing every problem to the same set of general principles, and instead approach each situation with an open mind and a sharp eye. And although Poor Economics attempts to avoid any kind of sweeping conclusions, the authors cannot resist the urge to draw at least some broadly applicable lessons from their study. When all is said and done, some general patterns do emerge:
- The poor are often unaware or misinformed about basic things, such as that simple oral rehydration solutions (ORS) are effective treatments for diarrhea. Tragically, despite ORS being cheap and widely available, an estimated 1.6 million people die each year of dysentery. It’s easy to blame superstition for this, but such a pronouncement ignores the fact that reliable knowledge is much harder to come by if you grow up in a small village in sub-Saharan Africa. Think of how many things that guide our everyday decisions are informed by knowledge that we couldn’t possibly verify, and what it takes to build institutions that allow us to outsource this process to trustworthy authorities!
- Not only is there a massive pool of knowledge that people in rich countries can readily tap into (without having to suspect at every turn that they are being lied to), the sheer range of decisions we have to make is, in an important sense, much smaller. If you want drinkable water, well, you just open the tap! If you want to save money, there’s a wide range of banks happy to take you on as a customer. Want you children to learn how to read and write? You don’t need to lose any sleep over whether the local elementary school will get them there. In other words, if you are well-off, there are lots of things you never even cross your mind, as the “right” decisions are already made for you. Not so for the world’s poor, who are responsible for far too many aspects of their lives.
- Why do we sometimes spend considerable time trying to decide which car insurance to select, but never trying to find an insurer in the first place — any insurer at all, really? In most walks of life, rich countries are home to reasonably competitive, well-functioning markets, but many of these simply don’t exist in the developing world. Insurance is a classic case — for the bottom billion, premiums are simply too low to make such a business worthwhile. Sometimes these markets can be brought to live through clever designs, as e.g. Alvin Roth has demonstrated in the case of kidney donations, but it remains an uphill battle.
All of this seems very plausible, if sometimes counter-intuitive, which makes it all the more important to consider where their approach might fall short. Now, it is of course perfectly possible to criticize the studies that inform the book, or challenge their interpretation of the result (e.g. because they are often underpowered), but I will focus on the bigger picture here. There are two main challenges that I see, both carrying substantial weight, and the success of research institutions such as the Poverty Action Lab they co-founded, or NGOs like GiveWell that put their insights into action, depends on how well they are able to respond to them.
The first challenge has to do with politics. Implicit in Banerjee and Duflo’s framework is the belief the development economics is, or should mostly be concerned with finding the right tools: “It is possible to improve governance and policy without changing the existing social and political structures” (p. 271). Anti-poverty programs, they argue, typically don’t fail because of corruption or other systemic issues, but because of design flaws or other “technical” reasons. And of course this is often true. Then again, we should not be fooling ourselves to what extent it is really possible to implement policies and yet remain completely apolitical. For as long as states have existed, people have tried to conceal their naked power grabs by masquerading as purely technical experts, putting in place only measures informed by science. As one example of how this plays out in practice, James Ferguson’s The Anti-Politics Machine makes a compelling case how aid agencies that are unaware of the political nature of their work can get co-opted by local governments with much less noble goals. I don’t see this point being addressed explicitly in the book.
A second challenge — one they anticipate in their book — is that the very nature of RCTs, which call for narrowly constrained, precisely controlled settings, excludes everything but the lowest-hanging fruits. Sure, it may be a fine thing to study whether chicken ownership or cash transfer are a more effective way to improve the lot of the world’s poor, but it takes a particular mindset to think running such a trial is “the best investment we could make to fight world poverty” (I am not making this up!). As Acemoglu and Robinson argue in Why Nations Fail, whatever positive effects may result from implementing the findings of an RCT, these net gains will be dwarfed by sustained economic growth. Put more pessimistically, unless and until political institutions are fixed, no other policies are going to make much of a dent in the lives of the poor. The question is really one of scale: Maybe some interventions have a fantastic return on the dollar but cannot, as a consequence of how they are designed, affect more than a handful of people. Lant Pritchett in particular has argued forcefully that we should focus our energies on interventions that, despite lower benefit-cost ratios, lead to (significantly) higher absolute welfare gains (see also his shorter blog post, and a response).
Without going too much into details, a big part of Banerjee and Duflo’s answer comes down to this: Sure, in theory and as a matter of the empirical track record, we know that certain institutions perform much, much better than others. The question remains, however, how we implement these institutions on the ground. It’s a trite cliché nowadays that nation-building doesn’t work, that functioning institutions need to grow organically rather than being imposed by outside forces, but what is certainly true is that two nominally identical institutions often work very different in practice. To take just one rather extreme example, both Germany and Switzerland are democratic countries. Yet the direct democratic model that works so well for the Swiss would, in my mind, prove disastrous if tried in the same form on the other side of Lake Constance. I have a hard time imagining my compatriots consistently voting against proposals that are in their narrow self-interest. And if that is the case for two countries with a common border and much cultural and linguistic overlap, say Banerjee and Duflo, how can we say anything with certainty about whether the abstract idea of property rights or free markets will spur economic growth in rural Uganda?
Finally, there’s also the question of who is to blame. This is a bit of an old-fashioned idea that we tend not to debate anymore, but I think it has some merit, if only because ignoring it makes it more likely to reach the wrong conclusions about what to do. We have made great strides in the fight against poverty over the last 200 years. Two centuries ago, a whopping 75% of the world’s population lived in abject misery, while today the figure is closer to 10% (that such a chart does not take into account population growth in no way lessens this achievement). But why have we not eradicated it? This is the elephant in the room: What if it is the poor’s fault after all? Have Banerjee and Duflo not demonstrated that, time and again, the poor make decisions that are clearly detrimental to their own long-term success?
I, for one, would be willing to bite that bullet — provided that I could be reasonably sure to that I would make the correct long-term decisions in their situation. And while such an experiment is fortunately unlikely to ever take place in real life, a little introspection can go a long way. Many of us have worked hard for their success, but that does not imply that we could have pulled off the same stunt under vastly less favorable circumstances. How much of what we own is a product of hard work alone, and how much of it is better attributed to the birth canal lottery? How far would perseverance and grit have gotten you if you were born in Mogadishu instead? At least for me the answer is obvious, and as a consequence, I have no troubles believing both that (a) because poor people often make bad decisions for themselves, and that (b) this does not free us from any obligation whatsoever to help them. How we can do so without deluding ourselves is a topic on which Poor Economics offers a fresh and enlightening perspective.