Poor Economics

Not as far behind the curve as usual (less than a year), I just finished reading Poor Economics by Banerjee and Duflo.  It’s well-written and certainly worth a close look by anyone interested in global poverty, which should be everyone.  Here are some reactions on my part:

1. At times the perspective of the book flirts with nominalism.  It seems to claim that only bottom-up perspectives and actions can meaningfully change the world: the search for a better “health policy” may be a chimera, for instance, while there is much we can do to improve the performance of this particular clinic in this region.  My guess is that the authors would not go this far and would say that they are simply redressing an imbalance—that previous analysts of health and other policies were too prone to make sweeping judgments at an altitude too far above ground level, and that it is time to become more fine-grained.  Fine, but I would like them to say this more clearly.  To put it a bit differently, interventions are needed at all levels, from the village (or even lower) to the international, and in an ideal world our experiences in each would inform how we go about the others.  To continue the example, just as national health policies should reflect what one has learned from the close study of a particular clinic, one would go about health policy at the village level differently depending on decisions taken higher up (for instance regarding what kinds of investments to make in training health professionals).  Maybe this wider view will show up in their next book.

2. The book works best in the later chapters that are more traditionally “economic”, such as those on insurance and credit; the early chapters are less successful.  The reason is not hard to find: the authors are rather narrowly trained as economists, and they do not cite much literature outside their own field.  An instructive example is their misuse of the term “iron law of oligarchy”, which dates to Michels’ study of the organizational structure of syndicalism.  It is not nitpicking to point this out: Michels is one of the canonical texts in the field of political sociology, and almost every sociologist knows his theory.  Inadvertently, B&D are telling us they do not have a background in social theory.  Of course, everyone cannot know everything, and we need a division of labor in the study of poverty, just as in any other domain.  But economists should know their limitations going in and make an effort to at least acquaint themselves with the thinking of other specialists on the topics they are studying.  The chapters on education and health, for instance, would have been much stronger if there were more attention given to what education and public health analysts have been saying and doing.

3. Even within economics the authors are rather narrow.  This is striking, because their spirit and instincts are generous, even though their analytical toolkit is not.  One telling example which runs through the entire book is their depiction of the “poverty trap”.  They present it as the hypothesis that there are increasing returns to some kinds of investment, so that at low levels of investment there is insufficient incentive to invest more.  This is good as far as it goes, but it doesn’t acknowledge other mechanisms.  If you think about a poverty trap as a kind of multiple equilibrium, you would expect three different varieties.  There are two kinds of static multiple equilibrium processes, increasing returns (the B&D hypothesis) and interaction (sufficiently strong cross-effects between the elements of an interactive system).  In addition, there is a third, dynamic basis, adjustment processes that alter the initial conditions (hysterisis).  The first of these, the B&D version, is the most widely “approved” within the economics profession; the third has made inroads beginning with the Sonnenschein-Debreu-Mantel intervention of the 1970s and the subsequent equilibrium unemployment literature, but is still on the margins; the second remains utterly heretical and is confined to cranks and misfits like me.  How does this translate to poverty?  Hysterisis is about how the coping mechanisms of poor people, which may be necessary to alleviate the worst effects of their situation, can also reinforce their situation.  Fatalism as a psychological strategy illustrates this.  Interactive traps would be those in which multiple economic, social and political factors combine to entrap individuals in poverty and are mutually reinforcing, so that it is not possible to make progress on any one front alone.  These are not idle speculations.

4. The authors are self-aware culturally but not politically-economically.  They are refreshingly honest about their preconceptions as outsiders from a much more prosperous background, and their learning process is one of the main themes of the book.  At the same time, there is no discussion at all about where their funding comes from, why some projects are funded and not others, who funds their local partners and why, etc.  This is important not only because there really are competing agendas in the world of development research and policy (surprise), but also because their approach is expensive.

5. Although B&D don’t seem to recognize it, their core recommendation regarding policy, that it be based on careful, ongoing assessment of outcomes, adheres to one of the major practitioner traditions of recent decades.  Its roots go back at least as far as Dewey’s “intelligent action”; another theoretical foundation is cybernetic theory, e.g. Stafford Beer.  (Schön’s reflective practitioner is not far from this either.)  The idea is that effective management requires a feedback loop incorporating information-gathering, analysis and plan revision, one that is built in from the beginning and operates continuously.  The idea has really taken hold in the “adaptive management” of natural resources, for instance in ecological restoration.  Academic readers of this blog may be familiar with it through the emphasis on assessment in effective teaching.  (You design your classes so that you get a continuous stream of information, in real time, on how students are learning, and you adjust your methods and content accordingly.)  I think the book would have benefitted enormously by making the connection between poverty policy and broader perspectives in management and the professions.