Doing Development Better
Jim Yong Kim's appointment as World Bank president may have been predictable, given the long-standing tradition that renders the selection an American prerogative. But even the appearance of competition between Kim and the other candidates, Ngozi Okonjo-Iweala and José Antonio Ocampo, served to expose a deep fissure within the field of development policy, because Kim and his two rivals represented dramatically different approaches.
The vision for which Kim stands is bottom-up. It focuses directly on the poor, and on delivering services—for example, education, health care, and microcredit—to their communities. This tradition's motto could be, "Development is accomplished one project at a time."
The other approach, represented by Okonjo-Iweala and Ocampo, takes an economy-wide approach. It emphasizes broad reforms that affect the overall economic environment, and thus focuses on areas such as international trade, finance, macroeconomics, and governance.
Practitioners in the first group idolize NGO leaders like Mohammad Yunus, whose Grameen Bank pioneered microfinance, and Ela Bhatt, a founder of India's Self-Employment Women's Association (SEWA). The heroes of the second group are reformist finance or economy ministers such as India's Manmohan Singh or Brazil's Fernando Henrique Cardoso.
At first sight, this might seem like another dispute between economists and non-economists, but the rift runs within, rather than between, disciplinary boundaries. For example, recent work with field experiments and randomized controlled trials (RCTs), which has caught on like wildfire among development economists, lies strictly in the tradition of bottom-up development.Poverty is often best addressed not by helping the poor be better at what they already do, but by getting them to do something different.
The relative effectiveness of the two visions is not easy to determine. Proponents of the macro approach point out that the greatest development successes have typically been the product of economy-wide reforms. The dramatic reductions in poverty achieved by China over the span of a few decades, as well as by other East Asian countries like South Korea and Taiwan, resulted largely from improved economic management (as much as earlier investments in education and health may have played a role). Reforms in incentives and property-rights arrangements, not anti-poverty programs, enabled these economies to take off.
The trouble is that these experiences have not proved as informative for other countries as one might have wished. Asian-style reforms do not travel well, and, in any case, there is significant controversy about the role of specific policies. In particular, was the key to the Asian miracle economic liberalization or the limits that were placed on it?
Moreover, the macro tradition vacillates between specific recommendations ("set low and uniform tariffs," "remove interest-rate ceilings on banks," "improve your 'doing business' ranking") that find limited support in cross-country evidence, and broad recommendations that lack operational content ("integrate into world economy," "achieve macroeconomic stability," "improve contract enforcement").
Development specialists in the bottom-up tradition, for their part, can deservedly claim success in demonstrating the effectiveness of education, public health, or microcredit projects in specific contexts. But, too often, such projects treat poverty's symptoms rather than its causes.
Poverty is often best addressed not by helping the poor to be better at what they are already doing, but by getting them to do something altogether different. This calls for diversification of production, urbanization, and industrialization, which in turn require policy interventions that may lie at considerable distance from the poor (such as fixing regulations or targeting the value of the currency).
Moreover, as with macro-level economic reforms, there are limits to what can be learned from individual projects. An RCT conducted under specific conditions does not generate usable hard evidence for policymakers in other settings. Learning requires some degree of extrapolation, converting randomized evaluations from hard evidence into soft evidence.Development reformers are suspicious of "best practices" and universal blueprints. They look instead for policy innovations, small and large.
The good news is that there has been real progress in development policy, and, beneath the doctrinal differences, there is a certain convergence—not on what works, but on how we should think about and do development policy. The best of the recent work in the two traditions shares common predilections. Both favor diagnostic, pragmatic, experimental, and context-specific strategies.
Conventional development policy has been prone to fads, moving from one big fix to another. Development is held back by too little government, too much government, too little credit, the absence of property rights, and so on. The remedy is planning, the Washington Consensus, microcredit, or distributing land titles to the poor.
By contrast, the new approaches are agnostic. They acknowledge that we do not know what works, and that the binding constraints to development tend to be context-specific. Policy experimentation is a central part of discovery, coupled with monitoring and evaluation to close the learning loop. Experiments do not need to be of the RCT type; China certainly learned from its policy experiments without a proper control group.
Reformers in this mold are suspicious of "best practices" and universal blueprints. They look instead for policy innovations, small and large, that are tailored to local economic circumstances and political complications.
The field of development policy can and should be reunified around these shared diagnostic, contextual approaches. Macro-development economists need to recognize the advantages of the experimental approach and adopt the policy mindset of enthusiasts of randomized evaluation. Micro-development economists need to recognize that one can learn from diverse types of evidence, and that, while randomized evaluations are tremendously useful, the utility of their results is often restricted by the narrow scope of their application.
In the end, both camps should show greater humility: macro-development practitioners about what they already know, and micro-development practitioners about what they can learn.
© 2012 Project Syndicate. Republished with kind permission.blog comments powered by Disqus