By Otaviano Canuto and Tiago Ribeiro dos Santos

Daron Acemoglu and James Robinson caught policymaker’s attention to the critical role of institutions for development. Their work gives too much emphasis to the prospects of revolution, however. A reading of the World Bank’s World Development Report of 2017 points to directions that all actors involved in the process, whether domestic or international, elite or non-elite, can take to improve societies.

The field of development economics is defined by a puzzle: why are some societies so much richer than others? For a long time, people didn’t even realize there was a puzzle to begin with. Rich societies were rich because they had accumulated resources for longer and at higher rates than poor societies. What poor societies should do is gather resources, and they would be rich. But then Robert Solow demonstrated with a simple equation that, adopting basic premises, one could not explain persistent and large differences in wealth with differences in resources. Poor societies should be catching-up, but they were not. This meant that the bulk of the differences would have to be explained by the most elusive factor of the equation – productivity. In the oft-quoted words of Paul Krugman, “Productivity isn’t everything, but, in the long run, it is almost everything.”

Several attempts have been made to explain these large differences but while many have their share of truth, none seemed quite enough. In the past half-century, years of education in developing countries have increased significantly, as has urbanization, but many of those countries have not seen a correspondent increase in productivity (although some most definitely have). New institutionalists made a convincing case that it was institutions that mattered before all the other contributors to productivity. This is illustrated by the fact that, although cross-border flows of knowledge and technology have followed globalization, only countries with appropriate institutions have been able to use those flows to build local production and innovation capabilities.

Institutions are, in Douglas North’s classic definition“the rules of the game in a society, or more formally, are the humanly devised constraints that shape human interaction”. These can be some of the most central rules, such as a choice between a hereditary absolutist monarchy or an electoral democracy, to more mundane ones, such as how late can one be to a business meeting (which may differ a lot in countries like Brazil or the United States). 

The authors of “Why Nations Fail”, Daron Acemoglu and James Robinson (heretofore A&R) were decisive for advancing this view which highlights the importance of institutions in a series of papers (often with the contribution of Simon Johnson). “Why Nations Fail” is their attempt at conveying to a broad audience the main conclusions of this research. The work has received wide praise and has been extremely influential for the view of development economics among both academics and policymakers. But there is an urgent need to go beyond A&R – and this is what the World Development Report of 2017, “Governance and the Law” (WDR17)does. Comparing the texts, we see that while they share a belief that institutions are central to development, the treatment of what to do with that information is fundamentally different. 

A&R are not focused in identifying in a very specific way the characteristics of inclusive institutions, their work is more descriptive. They say that there must be political centralization and pluralistic institutions, but they suggest these may be in tension: “Political centralization is likely only when one group of people is sufficiently more powerful than others to build a state. In Somalia, power is evenly balanced, and no one clan can impose its will on any other. Therefore, the lack of political centralization persists”. 

The vagueness is not an issue for the authors because in their view a lack of understanding about which institutions do or do not promote development is no obstacle. They seem to assume common knowledge of those institutions and a lack of interest by elites in promoting these institutions. 

When A&R offer some specificity on institutions conducive to economic development – “enforce property rights, create a level playing field, and encourage investments in new technologies and skills”, we see that the common knowledge assumption cannot be true. The advice sounds exactly like the beginning of a policy note by a Washington-based institution addressing much-needed reforms in a developing country. And this sentence alone could generate a lot of backlash. Taken literally, it would suggest a “market fundamentalism” that does not have general support in all societies. Taken less literally, it would open the field for all the debates that occur in development economics: what is the role for health initiatives, education, infrastructure, anti-corruption, industrial policy, cash transfers? These are hotly debated issues by academics who dedicate their life to these questions and have access to the best tools and data to test different hypotheses. To simply assume that elites would know what to do if they wanted is a too strong belief.

The WDR17 sees an important role for information as a contributor to better institutions. As stated in page 23, “[L]eaders change the incentives of other elites by taking into consideration who wins and who loses over time. By overcoming information and coordination challenges through political strategy, they can help find areas of agreement among conflicting parties without necessarily shifting norms or preferences.” In that spirit, the WDR goes on to list and detail drivers of effectiveness, levers for change, and drivers of change. We’ll examine below some of the differences between these two texts that deal with institutions.

Pure competition vs potential mutual benefits

In “Why Nations Fail”, extractive elites are in a Pareto frontier with the rest of the population. Any change which could bring about an improvement in the situation of the population would necessarily mean a worsening of the elite’s conditions. This is one of the main themes of the book and is laid out clearly already on page 2 : “Egypt is poor precisely because it has been ruled by a narrow elite that have organized society for their own benefit at the expense of the vast mass of people”.

This assertion does not pass an introspection test. As this text is written, Queen Elizabeth II of the United Kingdom has completed 66 years of reign. Heads of State of the United Kingdom have come from her family for centuries. It may be true that the British royal family does not have any say in governmental issues in the UK, but arguably they are in an enviable position for the vast majority of elites around the world, whether in terms of finances or prestige. Leaders of failed states, on the other hand, are often killed, imprisoned, exiled, or ostracized. The statues they erect for themselves are frequently toppled.

As a matter of fact, despite the thesis of the book, Acemoglu himself does not believe there are no conceivable mutually beneficial arrangements between elites and the people. In an earlier paper, “Why not a political Coase Theorem? Social conflict, commitment, and politics”, the author tries to answer “why do politicians and powerful social groups not make a deal with the rest of the society to choose the policies and institutions that maximize output or social welfare, and then redistribute part of the gains to themselves? Put even more strongly; why do powerful groups not predate efficiently”? He admits there are numerous states in which the population in general would have many more benefits, enough to compensate the elites better than in their situation. The reason, in his view, for these arrangements not being struck, is commitment problems.

Here is a classic illustration of a commitment problem. Suppose a masked criminal tries to rob a man. During the attack, the criminal loses his mask, and he also realizes he is robbing a local mafia boss, who is unarmed. The robber points the gun to the mafia boss, who promises to make his assailant rich for his life being spared. The criminal has no reason to doubt that the mafia boss would prefer to live and give away a small portion of his wealth instead of dying. What he does have reason to doubt, however, is that once the mafia boss is able to keep his promise of making him rich, he would not just decide to have the criminal killed instead.

For Acemoglu in the “Why not a political Coase Theorem” paper (but not in his book with Robinson), elites and the population suffer from severe commitment problems which hinder mutually beneficial arrangements. If, in the new arrangement, elites retain control over the resources, they cannot credibly promise to the population that they will not expropriate the newly generated wealth. If, instead, the population takes control of the resources, they also cannot credibly promise to share the new wealth with the old elite.

The WDR recognizes commitment problems as one of the main issues that get in the way of good governance. The difference with respect to the Acemoglu paper is that it does not assume these commitment problems are insurmountable. Institutions of governance, control of clientelism, elite bargaining are all factors that contribute to circumvent such issues.

Revolution vs evolution

Considering that A&R are skeptical that mutually beneficial arrangements can be struck for elites and the population, it seems natural that they believe that real change happens not through a series of reforms, but only after major events that at one fell swoop put countries in the right track, or at least in face of a threat of revolution.

A closer look at the evidence, however, shows this an inadequate generalization. As John Congleton convincingly documents in “Perfecting Parliament”, the story of inclusive institutions isn’t one of revolutions alone, but mainly of evolution. The powers of kings were gradually checked by parliaments in the different countries where these institutions did appear. According to Congleton, the revolutionary aspect of English reforms is overstated. Even “the great reforms of 1660 (the Restoration) and 1689 (the Glorious Revolution) can best be understood as reversions to England’s long-standing medieval constitution” (page 297). 

Congleton describes a history of near constant constitutional bargaining which have significantly expanded enfranchisement and the power of parliaments in European countries, most pronouncedly since the 19thcentury. This conclusion is in line with the research by Erica Chenoweth and Maria J. Stephan, documented in the article“Why Civil Resistance Works”. They rely on data from 1900 to 2006 to observe that nonviolent campaigns are much more likely to achieve success than violent ones, even when authoritarian regimes were being challenged. Nonviolent movements can gather more supporters and undermine loyalty among the repressive forces.

Nonviolence can gather more supporters through two mechanisms. First, it decreases the sense of risk of participants. Since nonviolent protests are repressed with less vigor than violent ones, more people may be willing to join in a movement. Second, a nonviolent protest is less prone to arguments of moral equivalence between protesters and repressive forces. Given that demonstrations are pacific, any repression is seen as unjust and may generate a backlash effect. 

Exactly because of this perceived legitimacy, nonviolent demonstrations can undermine the morale of repressive agents, whether in civil service or in the police or military forces. As Chenoweth and Stephan put it, “Nonviolent challenges should be more likely to evoke loyalty shifts in the opponent’s security forces, whereas armed resistance is more likely to encourage a closing of the ranks against the insurgency.”

Revolutions, on the other hand, are no guarantee for the establishment of inclusive institutions. As A&R recognize, very often they substitute one elite for another without changing the fundamental arrangements of society. Almost always, however, they generate a lot of bloodshed. This creates a rational fear of revolutions which undermines many reformist proposals because they are automatically associated with revolutions.

New evidence from the United States civil rights movement seems to confirm this thesis.

WDR17 recognizes that threats of revolution can sometimes promote inclusive changes, but this is not a necessary condition. “Even without a direct threat from below, many democratic transitions are initiated from a position of strength to ensure maximum benefits for empowered elites. In a ‘conceding-to-thrive’ scenario, the ruling coalition recognizes a future threat to the regime, but it maintains enough strength relative to the opposition to not fear losing an election.” This means that reform-minded actors may work for institutional improvements without the necessary trade-off of risking a revolution and the potentially disastrous consequences for welfare they may entail. 

Considering that commitment problems are correctly identified as the main obstacle for mutually beneficial arrangements between elites and the general population, game theory suggests that an incremental approach would be a more guaranteed way of improving institutions. The defense of revolutionary situations involves an assumption of a sudden change from a negative to a positive equilibrium, which could happen. But revolutions are convoluted times, which means that the information available about the behavior of different groups competing for power is scarce. Also, there is a selection of the very most ambitious and sometimes ruthless leaders in toppling a well-established regime. This means the chances that the new institutions are particularly inclusive are slim.

Gradual changes, on the other hand, are better supported by theory, as the WDR explains. Repeated iterations of a prisoner’s dilemma are shown to enable cooperation in the long run. When  concessions are broken down into smaller bits, each side may be able to help the other with the expectation of reciprocation, but without completely risking their position in the face of defection. This is the constant bargaining to which Congleton refers.

Perhaps without realizing it, Acemoglu seems to contradict himself with respect to the elites’ capacity for reforming, if they are willing. In the ““Why not a political Coase Theorem”, he argues that commitment problems prevent them from doing so, but in “Why did the West extend the Franchise?”, in another collaboration with Robinson, he argues that the threat of revolutions caused elites to extend the franchise. Well, if they had the option of moving towards inclusion in the face of revolution, there is no reason to believe that commitment problems would be insurmountable without the threat. This is particularly true if the establishment of these inclusive institutions were what was in the way of failed nations to “get to Denmark”, in Pritchett and Woolcock’s expression.

What role for international actors?

In his review of Why Nations Fail, Peer Vries is incisive: “The scant attention for international political economy in Why nations fail is quite problematic”. As Vries mentions, the participation of external forces is neglected. Foreign powers only affect countries to the extent that they chose what institutions to promote in the colonies, but after independence, their role is minimal. To the extent that A&R deal with foreign actors, it is to dismiss aid for development as capable of accomplishing anything like it – and possibly making things worse by allowing the prolonged survival of extractive elites. We shall not once more engage in the targeted aid effectiveness debate but to declare A&R’s view as simplistic, since the debate is far from settled on this issue. A&R do suggest that foreign aid be more directed towards improving institutions and wide participation in politics, but they see only a marginal role for that.

This is an untenable position. A glance at the historical evidence shows the importance of the global environment for anything that can reasonably be called institutions – or the “rules of the game”. Dictatorships, communist regimes, democracies, all of them appear and recede in waves. Notably, the Glorious Revolution, the paradigmatic case of a revolution creating inclusive institutions in A&R’s view, can be seen as having been facilitated by a foreign intervention by the Netherlands.

As Lant Pritchett points out, the benefits of helping create sustained growth episodes are so large that even if international organizations or think tanks increase the probability of them happening by only a small amount, the return-on-investment for these organizations is extremely large. The WDR argues for an important role for these actors, through two primary instruments: sponsoring international rules and using development assistance to affect domestic institutions. These instruments help solve cooperation, coordination and commitment problems. The number of ways international actors can affect International actors may participate in domestic policies directly, provide alternative sites for contestation (such as arbitration agreements), empower citizens or elites through resources or shift the preferences or beliefs of citizens and elites.

While A&R may be too dismissive of the potential role of international actors, it is also true that even if foreign aid as is currently practiced does more good than harm, it usually does not go far enough to address the underlying institutions which, if improved, could either put countries in a sustained growth path or avoid that growth stall or recede. In this sense, one of the major shortcomings of the World Development Report of 2017 is that It has not of yet accomplished enough in affecting the priorities of the development community. Admittedly, knowledge with respect to how to improve institutions reliably is in its infancy. But one of the main tasks of the development community is exactly doing research on these issues. As Lant Pritchett mentions, however, there has been a concentration of efforts into targeted programs which can have their impact more easily evaluated than in big picture, growth accelerating efforts.

A look at the World Bank Annual Report of 2018 shows that improving governance still has not received the priority status that it deserves. The priorities seem much more aligned with a view that does not give institutions their full weight: energy, transport, digital technologies, water, etc., but no explicit mention of institutions. The same report shows that the share of resources devoted to public administration projects has fallen quite significantly in 2018 to less than 10% of total disbursements.


The research of Acemoglu and Robinson is seminal to the recognition of the importance of institutions for development, and the World Development Report of 2017 on Governance and the Law relies heavily on their research and its by-products. The WDR and the book “Why Nations Fail” present, however, important differences, and the WDR gives reason for much more optimism than the book. Unlike A&R would have us believe, there are numerous opportunities for collaboration between elites and the general population that benefits both sides. There is no reason to believe that institutions are incapable of evolving without a revolution or a threat of revolution. International actors may help significantly more than suggested by A&R, but they must bear in mind these findings when doing so.

Otaviano Canutois Principal at the Center for Macroeconomics and Development and a former vice-president and former executive-director at the World Bank. Tiago Ribeiro dos Santosis a Brazilian diplomat. All opinions expressed here are their own.