What drives growth and development? Five leading scholars looked at this elusive topic during a February 16 symposium that concluded the IMF Research Department’s conference on “Macroeconomic Challenges in Low-Income Countries.” The panelists were Abhit Bannerjee (Massachusetts Institute of Technology), Tim Besley (London School of Economics), Simon Johnson (IMF staff on leave from MIT), Dani Rodrik (Harvard University), and John Williamson (Institute for International Economics), and Arvind Subramanian (IMF staff) moderated the exchange. Participants generally agreed on the importance of quality institutions for long-term growth but disagreed on the role institutions play in igniting growth and whether microexperi-ments can be scaled up to explain big differences between countries.
Simulating an in-flight conversation between an “impatient macroeconomist” and a “self-righteous microeconomist,” Abhit Bannerjee took an entertaining approach to defending his main theme that microexperiments enhance macro learning. In his view, there is much that macroeconomists can learn from microexperimental studies. Microexperimental estimates can provide answers to questions too narrow and specific to be studied in aggregate data; specific need not mean devoid of abstraction; and micro evidence may motivate macro-economists to research certain issues and be alert to the potential for bias. By the time Bannerjee’s fictitious plane had begun its descent for landing, the impatient macroeconomist had become convinced that macro regressions and microexperiments should be seen as complementary and mutually enhancing.
But the fact that global income and inequality are accounted for largely by differences across, rather than within, countries, Dani Rodrik argued, “has some important methodological implications.” He discussed his doubts about whether micro-experiments can be scaled up to actually help in understanding big differences between countries at opposite ends of the development spectrum—for example, explaining why Kenya is not more like Sweden. Rodrik also questioned whether these randomized experiments could really be extended in a linear fashion. Bannerjee responded that extensive uncertainty in understanding differences in growth among countries provides the context for continuing to do microexperiments, because these “are the only levers we have.”
Institutions and development
Improving the quality of government is now considered one of the central problems of development. But is this recent emphasis on institutions yet another fad? Tim Besley said he was somewhat disturbed by recent efforts to study institutional rules, supporting structures, and institutional change using crosscountry comparisons. He pointed to India as an example of a single country with enormous variation in institutional design, across both space and time. While it is clear that institutions change significantly over time, it is also important to recognize that in the transition from autocracy to democracy, there is often a surprising amount of continuity. A study of democracies, he said, will yield enormous heterogeneity. The good news, however, is that the right questions are being asked, and there is now a wealth of data and experience that, until recently, had been ignored and that could help lead to a more nuanced understanding of development.
By contrast, Simon Johnson chose to deemphasize the variation within countries of legal, social, and economic structures supporting formal institutions, arguing that these “details of institutions do not matter so much.” Institutions matter, for both the long and short term, he argued, because they are major determinants of economic performance and key to vast cross-country differences in prosperity. History, he said, reveals that weak institutions eventually lead to instability and, over very long periods, lower income levels.
Johnson particularly underscored the importance of strengthening property rights, while simultaneously developing some form of democracy to uphold these property rights. Johnson agreed with Besley that institutions can, and do, change frequently. Referring to the “dramatic wave of democratization” over the past 20 years, Johnson argued that, generally, political institutions have become more democratic while economic institutions have improved, but by less. He saw little hope, however, for “outsiders” having a positive effect on institutions, noting that although their impact historically has been significant, it has often been imposed as part of colonization.
Change takes time
Rodrik disagreed with Johnson, arguing “you don’t need good governance to get growth going.” He disputed Johnson’s view on institutional change, arguing that this takes time and in its absence over the short term, countries can still grow. Differences in “institutional quality” across countries are the primary determinant of growth, Rodrik said, but only in the long term. And since “nobody really knows how to get institutions ‘right,’” he added, this finding is “surprisingly unhelpful” for policy. The good news is that empirical evidence shows growth accelerations do not seem to require large-scale institutional change.
Rodrik proposed a three-step agenda for designing growth strategies: undertake diagnostics to determine the most binding constraints on growth; design and target policies to best alleviate constraints; and institutionalize the diagnostic process, since the nature of the constraints changes over time. None of this is heterodox, he concluded, since this approach is based on empirical evidence and standard economic theory.
Williamson agreed with fellow panelists that evidence since the 1990s shows that “institutions matter.” Widely known for coining the term “Washington Consensus” to describe a list of policies that he saw as being advocated for Latin America by Washington-based institutions in the late 1980s, Williamson said that he never meant the term to serve as a “cookbook or laundry list” for policy reform. The policy agenda should not be taken as static, he emphasized, suggesting that it might now be updated to include institutional reform alongside countercyclical macro policies and measures to improve income distribution and address labor market rigidities.
Rodrik, Johnson, and Besley debated at some length the relative importance for growth of institutional reform and Rodrik’s proposed diagnostic agenda. Asked by Johnson to explain how his proposed framework prescribes how to sustain growth, as well as what is needed to ignite growth, Rodrik responded that the current obsession with institutional reform and governance is a distraction from what has been shown to work and what can actually be done. If institutional reforms were as critical for igniting growth as Johnson suggested, Rodrik argued, reforms would be accompanied by significant changes in the relative power and wealth of insiders (those protected by previous institutional arrangements). Yet the growth booms of the past 40-50 years show instead that insiders have always benefited. Besley remarked that the issue is complicated since entrenched elites can be growth-promoting—citing the example of Britain in the 19th century—as well as growth-retarding.
Beware of “magic bullets”
Prompted by the audience, panelists also shared their perspectives on IMF and World Bank approaches to growth and development. While Williamson voiced “no great dissatisfaction” with the current approach of the organizations, he did express concern in the area of exchange rate policies, where he thought it important “to avoid excessive overvaluation,” particularly for countries with floating exchange rate regimes. Rodrik expressed concern with the “current obsession” with inflation targeting, central bank independence, and floating exchange rates.
Shifting the focus to the World Bank, Johnson said he was shocked that so much donor aid is distributed with no attempt to randomize the allocations. Johnson commended the Bank’s “tremendous progress” in measuring institutions and thinking about what can be operationalized, but he saw this work as disparate pieces rather than as an attempt to devise one metric, diagnostic tool that makes sense. For his part, Bannerjee lamented the World Bank’s move away from specific projects and specific successes toward program financing and broad budgetary support. He further stressed the importance in the development field of continuing to experiment to find what works and what doesn’t work, while acknowledging it is a difficult agenda. In a similar vein, he later cautioned against searching for “magic bullets” to solve the world’s economic growth and development problems. Microcredit, he said, is an example of a “great idea” that should have first been evaluated properly and then, pending the outcome, possibly scaled up.
Webcasts of the symposium presentations and papers presented during the preceding sessions of the conference on “Macroeconomic Challenges in Low-Income Countries” are posted on www.imf.org/external/np/res/seminars/2005/macro/index.htm.