Most empirical economics in this hub falls apart on close inspection. The Acemoglu-Johnson-Robinson research program on institutions did not. Twenty-five years of critique, replication, and counter-critique have left the core claim standing. The 2024 Nobel Prize confirmed it. Here is what makes this program empirically robust.

On October 14, 2024, the Royal Swedish Academy of Sciences announced that the Sveriges Riksbank Prize in Economic Sciences in Memory of Alfred Nobel would be awarded jointly to Daron Acemoglu (MIT), Simon Johnson (MIT), and James A. Robinson (University of Chicago) “for studies of how institutions are formed and affect prosperity.” The official citation is short. The body of work it recognizes is enormous --- spanning roughly a hundred academic papers, two widely read books, and a quarter-century of methodological argument inside development economics.

The question they have been working on is one of the oldest and most consequential in all of social science: why are some nations rich and others poor? Why did parts of the world that were, in 1500, far ahead of Europe in population, urbanization, and what looks from historical records like productive capacity --- the Aztec heartland, the Mughal Empire, the Inca civilization --- end up by 1900 vastly poorer than countries that were comparative backwaters five centuries earlier? Why does the per-capita income gap between the richest and poorest countries today exceed thirty-to-one, when in the year 1000 it was perhaps two-to-one?

The standard story, before Acemoglu, Johnson, and Robinson, was largely a story about geography (Jared Diamond’s “Guns, Germs, and Steel” being the most-cited popular version) or about culture (variants going back to Max Weber). The story they proposed instead --- that the proximate cause of the modern income distribution is differences in the political and economic institutions that countries inherited, and that those institutional differences themselves trace back to decisions made by European colonizers between roughly 1500 and 1900 --- is now the dominant explanation in modern development economics.

This article is an anti-example in a hub full of takedowns. It exists because the same standards of evidence that I have used to dismantle weaker claims elsewhere in this hub --- the ego depletion meta-analysis, the Reinhart-Rogoff debt-and-growth Excel error, the marshmallow test confounded by family income, the Card-Krueger minimum-wage natural experiment that turned out to be more contested than its supporters claimed --- those same standards leave the Acemoglu-Johnson-Robinson program in much better shape than almost anything else I have written about. The methodology has been attacked from multiple directions for twenty-five years. The findings have moved, but they have not collapsed. That distinction matters, and it is what an anti-example exists to illustrate.

For strategists who need to evaluate claims of the form “this institutional reform will boost growth in country X” --- and that includes anyone working in international development, frontier-market investing, sovereign-risk analysis, or even domestic policy framed around “we need a better business climate” --- understanding what this program robustly demonstrates, and what it does not, is the difference between calibrated decisions and political talking points.

What AJR 2001 Actually Demonstrated

The foundational paper is Acemoglu, D., Johnson, S., & Robinson, J. A. (2001). “The Colonial Origins of Comparative Development: An Empirical Investigation.” American Economic Review, 91(5), 1369-1401. DOI: 10.1257/aer.91.5.1369.

The problem AJR set out to solve was an old one in econometrics: institutions and incomes are correlated in cross-country data, but correlation does not establish causation. Maybe good institutions cause growth (the AJR hypothesis). Maybe growth causes good institutions, as countries get richer and demand better governance (the reverse-causation hypothesis). Maybe both are driven by some omitted third factor like geography, culture, or human capital (the confounding hypothesis). Any of these stories fits the raw correlation.

What AJR needed was an instrumental variable --- a source of variation in institutional quality that affected institutions but did not affect modern income through any channel other than institutions. If they could find such an instrument, they could isolate the causal effect of institutions on income.

Their proposed instrument was the mortality rate of European settlers during early colonization. The argument runs in three steps. First, where European settlers faced low mortality rates --- because the disease environment was favorable, malaria was absent, yellow fever was rare --- Europeans settled in large numbers and built institutions that protected property rights, constrained the state, and created political conditions for broad economic participation. Call these inclusive institutions. Examples: the United States, Canada, Australia, New Zealand. Second, where European settlers faced high mortality rates --- malaria-ridden coastal West Africa, the Congo basin, parts of South Asia --- Europeans did not settle in large numbers. Instead, they set up extractive institutions designed to move resources from the colonized population to the European metropole, with minimal protection of property rights for locals and minimal political constraints on the colonizing state. Examples: the Belgian Congo, much of the Spanish New World, the British Raj at its worst. Third, those institutional arrangements persisted long past the end of formal colonialism. Modern nations inherited the institutions that colonizers built, and the institutions colonizers built were a function of whether or not Europeans could survive there.

This logic gave AJR an instrument: settler mortality during the colonial period, measured from historical records of European soldiers, clergy, and bishops. The instrument satisfied the relevance condition --- it strongly predicted modern institutional quality, as measured by indices of expropriation risk. And AJR argued it satisfied the exclusion restriction --- that historical settler mortality affected modern income only through the channel of inherited institutions, not directly through any other path (modern disease burden, modern geography, modern climate).

The two-stage least-squares estimates were large. After instrumenting for institutional quality, AJR found that moving from the institutions of, say, Nigeria to the institutions of Chile would more than triple per-capita income. The estimated effect of institutions on income was substantially larger than the effect found in ordinary least-squares regressions, which AJR interpreted as evidence that OLS was attenuated by measurement error in institutional quality.

The paper was instantly influential. It has been cited more than fifteen thousand times. It became the entry point for an entire subfield of empirical development economics. And it set up the debate that would unfold over the next two decades.

The 2002 Reversal of Fortune Paper

A year later, AJR followed up with Acemoglu, D., Johnson, S., & Robinson, J. A. (2002). “Reversal of Fortune: Geography and Institutions in the Making of the Modern World Income Distribution.” Quarterly Journal of Economics, 117(4), 1231-1294. DOI: 10.1162/003355302320935025.

The Reversal of Fortune paper attacked the same question from a different angle, and the new angle was specifically designed to test the AJR institutional story against the geographic-determinism story most associated with Jared Diamond.

The geographic story says: rich countries today are rich because their geography is favorable to economic activity --- temperate climate, navigable rivers, productive agriculture, low disease burden. Poor countries today are poor because their geography is unfavorable. Under this story, the correlation between geography and modern income is direct.

AJR pointed out that if the geographic story were correct, then countries with favorable geography in 1500 should have been the richest countries in 1500, and the same countries should be the richest countries today. Geography does not change much over five centuries. So the ranking of countries by income should be stable.

The Reversal of Fortune paper documents that the ranking is not stable. Among the countries that Europeans colonized, the relative ordering of incomes flipped between 1500 and the present. In 1500 --- as best AJR could measure it, using urbanization rates and population density as proxies for economic productivity, since per-capita GDP data does not exist for that period --- the most prosperous regions of the colonized world were the Aztec heartland (modern Mexico), the Inca regions of Peru and Bolivia, the Mughal heartland in India, and parts of West Africa with sophisticated trade networks. By 1900, these same regions had become substantially poorer in relative terms than thinly populated, agriculturally marginal areas that Europeans had settled in large numbers --- North America, Australia, New Zealand.

This reversal, AJR argued, is not consistent with the geographic-determinism story. The geography did not flip. What flipped was the institutional arrangement that colonizers imposed --- because Europeans had stronger incentives to set up extractive institutions in the densely populated, already-organized regions (where there was a large local population to extract from and pre-existing economic surplus to capture) and stronger incentives to set up inclusive institutions in the thinly populated regions (where they had to settle themselves and so wanted property rights, courts, and political voice for the settler population).

The Reversal of Fortune paper is methodologically tighter than the 2001 paper in some ways --- the documentation of the reversal does not require an instrumental-variable argument, just careful historical data --- and looser in other ways, because urbanization-rate and population-density proxies for 1500 income are noisy. But the pattern is large enough to survive the noise, and the empirical horse-race between geography and institutions has, in the literature that followed, generally come out in favor of institutions as the dominant proximate cause, with geography mattering primarily through the institutional channel that AJR proposed.

The 2012 Why Nations Fail Popularization

The decade after the foundational papers saw AJR (mostly Acemoglu and Robinson, with Johnson focused on related but somewhat different work on financial institutions and political economy) extend the research program in many directions: comparative slavery, the rise of European democracy, the political economy of state capacity, the dynamics of revolutions, the conditions under which extractive elites do or do not transition to more inclusive arrangements.

The synthesis came in 2012 with Acemoglu, D., & Robinson, J. A. (2012). “Why Nations Fail: The Origins of Power, Prosperity, and Poverty.” Crown Business.

Why Nations Fail is a popularization, not a primary research contribution. It restates the institutional-economics framework for a general audience, draws on the academic papers underneath it, and applies the framework to a sweep of historical cases --- the Spanish conquest of the Americas, the British Industrial Revolution, the divergence between North and South Korea, the failure of post-independence African states. The book sold widely and became, for a time, the most-recommended single-volume introduction to development economics for non-economists.

For purposes of this article, the book matters in two ways. First, it is the place where the AJR framework moved out of the journals and into the broader policy conversation --- the language of “inclusive versus extractive institutions” became part of how journalists, think tanks, and policy analysts talked about development. Second, the book is, like all popularizations, somewhat looser than the underlying academic work. Some of the case-study material does not pin down causation as tightly as the AJR cross-country instrumental-variable analysis does. Critics fairly point out that picking a few illustrative cases and telling an institutional-economics story about them is not the same as demonstrating that institutions caused the outcomes in those specific cases. The book is rhetorically powerful, but the empirical work that supports the broad thesis lives in the journal articles, not in the chapters about the Glorious Revolution.

This matters for how to evaluate the Why Nations Fail story specifically: the core cross-country result on institutions and growth is well-supported; the application of the framework to any particular country’s history is a narrative interpretation, not a tested causal claim, and should be read with the same skepticism one would apply to any single-case argument.

The Glaeser 2004 Critique --- Challenging the IV Assumptions

The first major academic critique of the AJR research program came from a quartet of Harvard economists in Glaeser, E. L., La Porta, R., Lopez-de-Silanes, F., & Shleifer, A. (2004). “Do Institutions Cause Growth?” Journal of Economic Growth, 9(3), 271-303. DOI: 10.1023/B:JOEG.0000038933.16398.ed.

Glaeser and coauthors made three connected criticisms.

First, they questioned what AJR were actually measuring when they used indices of expropriation risk and constraints on the executive as proxies for “institutions.” These indices, Glaeser et al. argued, are outcome measures --- they reflect how the political system is performing right now, not the underlying institutional rules. A country that is currently being well governed by a benevolent dictator might score high on these indices despite having weak underlying institutional constraints. A country with strong written constitutional protections might score low because the current government is violating them. If the measures track current performance rather than durable institutional structure, then regressing growth on these measures tells us less about institutions than the AJR interpretation claims.

Second, Glaeser et al. argued that human capital might be the deeper cause. In their reading, European colonizers brought not just institutions but people --- and people brought human capital, education, productive skills. When Europeans settled in large numbers in North America and Australasia, they brought a large stock of human capital with them, and the institutions that arose were partly a consequence of that human capital, not a cause of growth independent of it. In the OLS analysis Glaeser et al. ran, they found that human capital does a lot of the work that AJR attributed to institutions.

Third, Glaeser et al. challenged the exclusion restriction of the settler-mortality instrument. The AJR argument requires that settler mortality affected modern income only through inherited institutions. But settler mortality is correlated with disease environment, which is correlated with modern human capital accumulation (disease burden affects schooling and productivity), which is correlated with modern income. If settler mortality affects modern income through a human-capital channel as well as an institutional channel, then the instrument is invalid and the AJR estimates are biased.

Glaeser et al.’s preferred OLS specification, controlling for human capital, suggested that the institutional effect was much smaller than AJR found, and that poor countries get out of poverty mostly through good policies that build human capital --- often pursued by dictators --- with improved political institutions following rather than preceding the income gains.

This was a serious critique from serious economists, and it shifted the conversation. AJR responded with additional empirical work refining their institutional measures, showing that their results held when using narrower measures more clearly tied to durable structure (constraints on executive authority, for instance, rather than expropriation-risk indices), and showing that the human-capital story does not fit several of the cases the institutional story handles well. The debate did not end, but the AJR program absorbed the critique and the underlying conclusion --- that institutions matter substantially for long-run development --- survived in a form that engaged with the Glaeser critique rather than ignoring it.

The Albouy 2012 Critique --- Data Quality and the AJR Reply

The most pointed empirical attack on the AJR settler-mortality data came from Albouy, D. Y. (2012). “The Colonial Origins of Comparative Development: An Empirical Investigation: Comment.” American Economic Review, 102(6), 3059-3076. DOI: 10.1257/aer.102.6.3059.

Albouy went after the underlying data construction itself. AJR had assembled their settler-mortality dataset from historical records of European soldiers, clergy, and bishops --- but the records were spotty, and for many countries in the sample, AJR did not have direct mortality estimates and had to impute mortality rates from neighboring or comparable countries. Albouy’s central claim was that 36 of the 64 countries in the AJR sample had mortality rates assigned to them from other countries, and that the assignments were sometimes based on questionable historical evidence and sometimes adjudicated in ways that conveniently favored the AJR hypothesis. Beyond the assignment problem, Albouy argued that AJR combined mortality rates from incompatible populations --- laborers, bishops, soldiers on campaign --- in ways that further biased the analysis.

When Albouy reanalyzed the data using only directly observed mortality rates, dropped the imputed assignments, and used what he argued was a more conservative treatment of incompatible source populations, the AJR result lost robustness. The first-stage relationship between settler mortality and modern expropriation risk weakened. The instrumental-variable estimates of the institutional effect on income became much less precise, with confidence intervals that in some specifications became uninformative.

This was a significant attack. If it had stood unanswered, the foundational AJR paper would have been substantially damaged.

AJR responded in the same issue: Acemoglu, D., Johnson, S., & Robinson, J. A. (2012). “The Colonial Origins of Comparative Development: An Empirical Investigation: Reply.” American Economic Review, 102(6), 3077-3110. Their reply made several points. The most consequential was that Albouy’s reconstruction effectively dropped all the Latin American data and most of the African data --- amounting to roughly 60 percent of the original sample --- on the grounds that the mortality data for those regions was assigned rather than directly observed. AJR argued that this was the wrong methodological choice because considerable mortality information for European populations in those regions does exist in the historical record, and that the assignment procedures, while imperfect, were defensible. When AJR reran their analysis with their original data, including Latin America and Africa, and incorporated some of Albouy’s more defensible technical points (around how to weight different source populations), the core result --- that historical settler mortality predicts modern institutions, and that instrumented institutional quality predicts modern income --- continued to hold, with somewhat wider confidence intervals than the original 2001 paper but at magnitudes that were still economically large.

The exchange was, from a meta-scientific perspective, what good empirical economics is supposed to look like. A critic identified a specific data-construction concern, did the work of reanalyzing the data under different assumptions, and published the result. The original authors responded with their own reanalysis, defended the methodological choices they had made, conceded the points worth conceding, and showed the result was robust to a wider range of reasonable specifications than the original paper had presented. Subsequent literature has largely concluded that the AJR result is more robust than Albouy’s most aggressive specification suggested, and less precisely estimated than the original 2001 paper claimed. That is roughly what an honest reading of a productive empirical exchange should produce.

What Makes This Empirically Robust --- Multi-Methodology Convergence

What distinguishes the Acemoglu-Johnson-Robinson program from the weaker findings in this hub is not that it survives every individual critique unchanged. It does not. The original 2001 specification has been modified, the original data has been challenged and partly repaired, the original confidence intervals have widened. What distinguishes it is that the core claim --- that political and economic institutions are a major proximate cause of long-run cross-country income differences --- has been independently supported by multiple research designs that do not share the AJR methodology and do not share its specific data.

Sub-national studies have shown that within a single country, regions that historically had more extractive institutions imposed by colonial powers remain poorer today, even after controlling for geography and modern policy. Studies of the regression-discontinuity variety at colonial-era borders --- where one side of a line was governed by Britain and the other by Spain, or one side by direct colonial rule and the other by indirect rule through local intermediaries --- have repeatedly found that the institutional inheritance shows up in modern income.

Historical economic studies using natural experiments other than settler mortality have produced compatible results. The literature on the “Mita” forced-labor system in colonial Peru and Bolivia, on the historical division of Korea, on the variation in Indian colonial land tenure systems --- different methodologies, different time periods, different geographies, all pointing in the same direction.

Even Easterly and Levine, who in Easterly, W., & Levine, R. (2003). “Tropics, Germs, and Crops: How Endowments Influence Economic Development.” Journal of Monetary Economics, 50(1), 3-39 initially framed their work as a challenge to the AJR institutional story, ultimately concluded that geographic endowments affect modern income primarily through institutions, with little evidence of a direct geographic channel once institutional quality is controlled for. This is, broadly, the AJR thesis under a different methodology.

This pattern --- multiple independent research designs, using different data, different identification strategies, different control variables, converging on the same qualitative conclusion --- is what empirical robustness looks like. It is the opposite of what we see in the failed-replication studies that fill the rest of this hub, where each new attempt to test the original finding either fails or shrinks the effect dramatically.

The data and code from the AJR papers have been publicly available for the full lifetime of the research program. Critics have been able to reanalyze. AJR has been able to respond. The exchange has been adversarial in the productive sense, not in the defensive sense. This too is part of what robust empirical economics requires.

What’s Honest About The Current State of Institutions-Growth Research

I want to be careful here. The institutional-economics literature is robust on what it has actually demonstrated. It is much less robust on some of the claims that get attributed to it, including by enthusiastic readers of Why Nations Fail.

What the literature robustly demonstrates: institutions, broadly defined as the formal and informal rules that govern political and economic interaction, matter substantially for long-run cross-country differences in per-capita income. Countries that inherited institutional arrangements which protect property rights, constrain executive authority, allow broad political participation, and provide functional dispute resolution tend, over decades and centuries, to grow faster and reach higher income levels than countries that inherited institutions which concentrate power, expropriate broadly, suppress political voice, and provide weak property rights. This conclusion is supported by multiple research designs and survives the most aggressive published critiques.

What the literature does not robustly demonstrate: that specific institutional reforms, undertaken today in country X, will produce a specified amount of growth in country X. The aggregate cross-country relationship between institutional quality and income tells us that the average country with better institutions is richer. It does not tell us how to engineer a transition from worse institutions to better institutions, how long the transition takes, what the political-economy preconditions are, or whether reforms imposed without those preconditions will produce the predicted gains. The empirical evidence on specific institutional-reform interventions --- from World Bank attempts to improve governance, to constitutional reform efforts, to attempts to transplant Western property-rights regimes into post-colonial contexts --- is much weaker and much more mixed than the aggregate institutions-matter literature.

Acemoglu and Robinson have themselves been generally careful about this distinction. The Why Nations Fail framework explicitly emphasizes that institutional transitions are political processes shaped by the existing distribution of power, and that the literature does not provide a recipe for inducing such transitions externally. Some readers of the framework have been less careful. The implicit policy claim --- “country X is poor because it has bad institutions, so reform its institutions and it will become rich” --- is a step beyond what the empirical work supports.

The honest summary is: institutions matter, the magnitude is large, the mechanism operates over decades to centuries, and the policy implications are humble. Strategists evaluating reform proposals should anchor on those four claims and treat anything more specific with the skepticism it deserves.

The 2024 Nobel Prize --- What It Specifically Recognized

The Nobel committee, in its official statement of October 14, 2024, awarded the prize “for studies of how institutions are formed and affect prosperity.” The committee’s longer scientific background document emphasized three contributions specifically.

First, the empirical demonstration that institutional differences are a major proximate cause of cross-country income differences, with the colonial-origins research program as the centerpiece evidence.

Second, the historical-comparative analysis of how institutions are formed and how they persist or change, including the political-economy theorizing in the broader Acemoglu-Robinson body of work on why some societies transition to inclusive institutions while others remain trapped in extractive arrangements.

Third, the framework for understanding the divergence between inclusive and extractive institutional paths --- the conceptual vocabulary that has become the standard language for talking about the institutional foundations of long-run prosperity in modern development economics.

The committee did not, importantly, claim that AJR have resolved every empirical question in the institutions literature. The Nobel statement is careful about the open questions in the field. What it endorses is that the research program has substantially advanced our understanding of one of the most consequential questions in economics, and that the methodological innovations (the use of historical instrumental variables, the comparative-historical analysis, the integration of formal political-economy theory with empirical cross-country work) constitute a major contribution to how empirical economics is done.

This is the relevant frame for thinking about what the Nobel signals. It is not a stamp of approval on every specific finding. It is recognition that the research program met the standards the field uses to judge important empirical contributions --- rigorous methodology, sustained engagement with critics, demonstrated robustness across multiple research designs, and a substantial accumulation of evidence on a question that matters.

What This Anti-Example Tells Us About Robust Empirical Economics

I am writing this article in a hub where most of the case studies are failures. Ego depletion, power posing, money priming, the marshmallow test in its strong form, Reinhart-Rogoff’s debt-and-growth cliff, the original Card-Krueger minimum-wage finding in some of its more contested specifications, half the social-priming literature, large parts of the implicit-bias literature --- all of these have either failed to replicate, shrunk dramatically on close examination, or turned out to depend on specifications and data choices that did not hold up.

The Acemoglu-Johnson-Robinson program is different, and the difference is instructive. The pattern that distinguishes robust empirical economics from fragile empirical economics, as far as I can tell from looking at many cases, has roughly these elements.

The methodology is novel and identifies a specific source of exogenous variation rather than relying on observational correlation alone. The use of historical settler mortality as an instrumental variable for modern institutions was a genuine methodological innovation. It made the identification argument explicit and testable, in a way that “institutions are correlated with growth” alone could not be.

The data is shared, the code is shared, and replication by critics is possible. AJR have made their data publicly available throughout. When Albouy and others did reanalysis, the reanalysis was technically possible because the materials were there.

Critics engage substantively and the original authors respond substantively. The Glaeser-Albouy exchanges with AJR are the model. Critics did not just write skeptical commentary; they did the empirical work and showed their results. AJR did not just defend their original findings; they responded with their own additional analysis, conceded what was worth conceding, and showed what was robust to legitimate methodological refinements.

The core finding is supported by independent methodologies. The institutions-cause-growth conclusion is not held up by the AJR papers alone. It is supported by sub-national studies, by regression-discontinuity studies at colonial borders, by historical natural experiments with different data and different research designs. The convergence across methodologies is what allows the conclusion to survive even though individual specifications have been challenged.

The original authors are calibrated about what their work demonstrates and what it does not. AJR have been generally careful, in their academic writing if not in every popularization, about the difference between the aggregate institutions-matter finding (well-supported) and specific institutional-reform recommendations (much more uncertain). This calibration is part of what allows the research program to persist as critics chip away at specific claims --- the broader framework is not staked on any particular contested finding.

This is what good empirical economics looks like. It is rare. Most published economic findings, like most published findings in any social-science field, do not have these properties. When you find a research program that does, it is worth treating its conclusions with substantially more weight than the average paper in the literature.

What This Means For Strategists Evaluating “Institutional Reform” Claims

The practical application for strategists --- in international investing, sovereign-risk analysis, frontier-market evaluation, development consulting, or domestic policy framed around “we need a better business climate” --- comes down to calibration.

The claim “institutional quality matters enormously for long-run national prosperity” should be treated as well-supported. The empirical evidence behind this claim is, by the standards of cross-country economics, as good as it gets. When evaluating a country as an investment destination, as a sovereign credit, as a destination for productive capital, the institutional environment --- property rights enforcement, constraints on arbitrary state action, functioning courts, basic rule of law --- deserves substantial weight. Countries with weak institutions can grow for years on commodity booms or favorable demographics, but the long-run trajectory of countries with strong institutions tends to be substantially better. This is what the AJR program demonstrates.

The claim “country X will boost growth by N percent if it adopts reform Y” should be treated with much more skepticism. The aggregate cross-country relationship does not provide reliable point estimates for the effects of specific reforms in specific places. A country can adopt the formal trappings of property rights, courts, and constitutional constraints without changing the underlying political economy that makes those institutions function or fail. The literature on institutional transplantation is much weaker than the aggregate institutional-economics literature, and the gap between “country X has weak institutions” and “here is the specific reform package that will fix that” is much larger than enthusiastic policy advocates typically admit.

The claim “we know from the Acemoglu-Robinson research which institutional package to recommend” should be treated as substantially overstated. The research program has identified the importance of institutions; it has not identified a recipe for institutional improvement that can be applied externally to specific countries with predictable results. Reform packages should be evaluated on their own evidence, in their own contexts, with appropriate humility about the gap between cross-country aggregate findings and country-specific causal claims.

The right disposition is roughly: take the institutional environment seriously as a structural determinant of long-run prosperity; be skeptical of confident claims that a particular reform will produce a particular growth gain; recognize that the most robust finding in development economics is also one of the hardest findings to act on directly.

This is the kind of calibrated takeaway that distinguishes useful empirical economics from political talking points. The Acemoglu-Johnson-Robinson program supports it. It is what the 2024 Nobel Prize is recognizing. And it is what makes this an anti-example worth including in a hub otherwise full of cautionary tales.

Sources

  • Acemoglu, D., Johnson, S., & Robinson, J. A. (2001). “The Colonial Origins of Comparative Development: An Empirical Investigation.” American Economic Review, 91(5), 1369-1401. DOI: 10.1257/aer.91.5.1369. https://www.aeaweb.org/articles?id=10.1257/aer.91.5.1369
  • Acemoglu, D., Johnson, S., & Robinson, J. A. (2002). “Reversal of Fortune: Geography and Institutions in the Making of the Modern World Income Distribution.” Quarterly Journal of Economics, 117(4), 1231-1294. DOI: 10.1162/003355302320935025.
  • Acemoglu, D., & Robinson, J. A. (2012). “Why Nations Fail: The Origins of Power, Prosperity, and Poverty.” Crown Business.
  • Glaeser, E. L., La Porta, R., Lopez-de-Silanes, F., & Shleifer, A. (2004). “Do Institutions Cause Growth?” Journal of Economic Growth, 9(3), 271-303. DOI: 10.1023/B:JOEG.0000038933.16398.ed.
  • Albouy, D. Y. (2012). “The Colonial Origins of Comparative Development: An Empirical Investigation: Comment.” American Economic Review, 102(6), 3059-3076. DOI: 10.1257/aer.102.6.3059.
  • Acemoglu, D., Johnson, S., & Robinson, J. A. (2012). “The Colonial Origins of Comparative Development: An Empirical Investigation: Reply.” American Economic Review, 102(6), 3077-3110.
  • Easterly, W., & Levine, R. (2003). “Tropics, Germs, and Crops: How Endowments Influence Economic Development.” Journal of Monetary Economics, 50(1), 3-39.
  • The Royal Swedish Academy of Sciences. (2024, October 14). “The Sveriges Riksbank Prize in Economic Sciences in Memory of Alfred Nobel 2024.” https://www.nobelprize.org/prizes/economic-sciences/2024/press-release/

FAQ

Is the Acemoglu-Johnson-Robinson research the only robust finding in development economics?

No, but it is one of a relatively small set. The broader empirical-development literature has produced other reasonably robust findings --- the importance of human capital, the conditional convergence result from cross-country growth regressions, the consistent finding that openness to trade is positively correlated with growth in most specifications. The institutional finding is among the most influential because it speaks to the deepest question (why are some nations rich) and because the methodological apparatus has held up to particularly aggressive critique. But it is not unique. What distinguishes the genuinely robust findings from the fragile ones, across the literature, is the pattern this article identifies: multiple independent methodologies, productive engagement with critics, transparent data, calibrated claims.

Is Why Nations Fail accurate?

The broad framework is well-supported. The specific historical case studies in the book are interpretations, not tested causal claims, and should be read as such. Most economic historians find some of the specific applications more persuasive than others. The Korea case --- two countries with shared geography, shared culture, shared starting point, divided into different institutional arrangements, with vastly different subsequent trajectories --- is the cleanest natural experiment in the book and the one most economists find most convincing. Other cases, particularly the Glorious Revolution chapter and parts of the colonial Americas chapters, have been challenged by economic historians on specific factual and interpretive grounds. The book is rhetorically powerful and the framework is sound; the individual chapters are narrative arguments, not empirical demonstrations.

What about specific institutional reforms --- can we use this research to recommend policy?

Cautiously. The aggregate cross-country result that institutions matter for long-run growth does not translate cleanly into specific reform recommendations for specific countries. Reform packages need to be evaluated on their own context-specific evidence, with attention to the political economy of implementation and the gap between formal institutional change and underlying institutional function. The literature on attempts to externally promote institutional reform --- through aid conditionality, through technical assistance, through international pressure --- is much weaker than the aggregate institutional-economics literature, and the historical record is mixed at best. Use the AJR framework as background, not as a policy recipe.

What about cultural factors --- doesn’t culture also matter?

Yes, and the literature has increasingly recognized that the dichotomy between institutions and culture is somewhat artificial. Informal institutions --- norms, expectations, social capital --- are themselves cultural phenomena, and the AJR framework explicitly includes both formal and informal rules as part of what it means by “institutions.” The more interesting question is which cultural factors are themselves shaped by historical institutional arrangements (and so are downstream from the AJR story) versus which are independent of institutional history. There is real work being done on this --- including by Acemoglu and Robinson themselves --- but the broad finding that the institutional channel is large does not require denying that culture also matters; it requires saying that the institutional channel exists and is substantial, which is a weaker and more defensible claim.

What about other Nobel Prizes in economics --- are they all this robust?

No. The Nobel Prize in economics has been awarded to a wide range of research programs over the decades, and not all of them have held up equally well. Some prize-winning work has been substantially revised in light of subsequent evidence. Some has been narrowed in its applicability. The Nobel is a recognition of important contributions to the field, not a guarantee that every specific finding will survive future scrutiny. Readers should evaluate the empirical strength of any specific research program on its own merits, not on the basis of its prize history. The Acemoglu-Johnson-Robinson program holds up particularly well by current empirical standards. Some other Nobel-winning programs would not survive the same scrutiny applied in this hub.

Does the AJR critique of geographic determinism mean Jared Diamond is wrong?

Not exactly. Diamond’s argument in Guns, Germs, and Steel was about why certain regions of the world developed agriculture, urbanism, and complex states earlier than others --- and his account of that question, focused on the geographic distribution of domesticable plants and animals, has substantial support. What AJR challenge is the further claim that the same geographic factors directly explain the modern income distribution. The Reversal of Fortune evidence suggests that within the colonized world, the modern income ranking is not the same as the pre-colonial development ranking, which means something else --- AJR argue institutions --- is doing the proximate work for the modern distribution. The Diamond-style geographic story may be correct about pre-modern development; the AJR institutional story is better-supported as an account of modern cross-country income differences.

How should I evaluate a sovereign-credit or frontier-investment decision in light of this research?

Treat institutional quality as a major structural variable. Countries with robust property-rights regimes, functional courts, and meaningful constraints on executive arbitrariness tend to have substantially better long-run trajectories than countries without them, even controlling for current GDP, current growth rate, and commodity exposure. This is one of the most well-supported findings in development economics. Be skeptical of confident short-term predictions tied to specific reform announcements --- the gap between announced reforms and implemented reforms is large, and the gap between implemented reforms and economic outcomes is even larger. The institutional channel operates over decades, not over single news cycles. Position size accordingly.

Is this finding politically convenient for any particular ideological position?

It is politically inconvenient for several positions, which is part of what makes it interesting. The finding is uncomfortable for strong geographic-determinism views (associated with some strands of evolutionary thinking and with some popular development writing). It is uncomfortable for strong cultural-essentialism views (which would predict that culture, not institutions, drives the differences). It is uncomfortable for some left-wing development economics that emphasizes structural global factors over national institutional ones, and it is uncomfortable for some right-wing positions that emphasize policy choices over deeper structural factors. The AJR framework cuts across the standard political dimensions in interesting ways, and the fact that it is empirically robust despite not flattering any particular political camp is itself a useful signal that the finding is tracking something real about the world.

replication-crisis acemoglu-robinson institutions-economics development-economics evidence-evaluation
Share this article
LinkedIn (opens in new tab) X / Twitter (opens in new tab)
Atticus Li

Experimentation and growth leader. CXL-certified CRO practitioner, Mindworx-certified behavioral economist (1 of ~1,000 worldwide). 200+ A/B tests across energy, SaaS, fintech, e-commerce, and marketplace verticals.

About LinkedIn Newsletter