What Bill Gates Got Wrong About Why Nations Fail

Did the Microsoft founder even read our book before he criticized it?

Our recent book, Why Nations Fail: The Origins of Power, Prosperity, and Poverty, received the harshest reviews from those who see geography and culture as the root causes of poverty, and enlightened leaders -- or even more enlightened outside donors and organizations -- as the keys to economic development. Perhaps unsurprisingly, given his dedication to international aid, billionaire foundation chief Bill Gates falls into this category: His Feb. 26 review of our book was particularly uncharitable. Unfortunately, however, it was also dead wrong on many counts.

Gates's review is disappointing, but not just because he disagrees with us. As academics, we expect that. Research is all about arguing and contradicting, finding new pieces of evidence, developing new concepts and perspectives, and getting closer to the truth. Alas, Gates fails in this endeavor. His inability to understand even the most rudimentary parts of our thesis means that his review fails to invite constructive argument. Nonetheless, we feel compelled to respond because of the undue attention the review has generated.

To start with, Gates makes some pretty baffling statements about our book, such as his assertion that "important terms aren't really defined." Actually, all of the major concepts we use in the book are defined; one just needs to read the book. Other assertions demonstrate not only that Gates is unfamiliar with the academic literature, which is understandable, but that he actually did not bother to consult the bibliographic essay and the references at the end. He writes, "The authors ... attribute the decline of Venice to a reduction in the inclusiveness of its institutions. The fact is, Venice declined because competition came along ... Even if Venice had managed to preserve the inclusiveness of their institutions, it would not have made up for their loss of the spice trade."

This is just bad history. Venice didn't decline because of the loss of the spice trade. If that were the case, the decline should have started at the very end of the 15th century. But the decline was already well underway by the middle of the 14th century. More generally, research by Diego Puga and Daniel Trefler shows that Venice's fortunes had nothing to do with competition or the spice trade.

Likewise, Gates seems to think that the Maya declined because of the "weather." Though there is certainly scholarly dispute over why Maya civilization decayed, to our knowledge no reputable scholar argues that it was due to the weather. Instead, most scholars emphasize the role of inter-city warfare and the collapse of Mayan political institutions. Nor does the book, as Gates would have it, "overlook the incredible period of growth and innovation in China between 800 and 1400." We discuss that period, and explain why it didn't translate into sustained economic growth (see Chapter 8, in particular, pp. 231-234).

Gates also says at one point that our book "refers to me in a positive light." Sorry, we do no such thing. We point out that Gates, just like Mexican telecom mogul Carlos Slim, would have loved to form a monopoly. He tried and failed. What our book shows in a positive light are the U.S. institutions, such the Department of Justice, that stopped Gates and Microsoft from cornering the market. We say, "sadly there are few heroes in this book." Bill Gates was not one of them.

On a related note, Gates writes that that our book is "quite unfair to Slim." Mexico, he contends, is "much better off with Slim's contribution in running businesses well than it would be without him." But once again, this reveals a lack of understanding of our main thesis, which isn't that Carlos Slim is evil and the root cause of Mexico's problems. We argue that ambitious entrepreneurs like Gates or Slim will do good for society if inclusive institutions constrain them, and that they will mostly serve their own interests otherwise. So the right counterfactual to Slim isn't no Slim, but a Mexico in which people like Slim (and hundreds of other talented would-be entrepreneurs who never got the opportunity to flourish because of the country's poor education system or because of its terrible competition laws) operate within the context of inclusive economic institutions and therefore enrich their society to a much greater extent.

For the record, however, before cheerleading Slim, Gates might want to read the OECD's 2012 report on telecommunications policy and regulation in Mexico, which estimates the social costs of Slim's monopoly at U.S. $129 billion and counting. (The latest Forbes list of the world's richest people puts Slim's net worth at U.S. $79 billion). So in what way is Mexico better off exactly?

Gates also complains in his review that we "ridicule modernization theory." We don't. We try to articulate an alternative theory of extractive growth -- which takes place under extractive, authoritarian political institutions -- where countries grow  because their leadership controlling these extractive institutions feels secure and able to control and benefit from the growth process. This occupies a large part of our book because it is a central feature of economic and political development over the last several thousand years. Our theory suggests why extractive growth doesn't automatically lead to more inclusive institutions: Growth is made possible, at least in most cases, by the leaders and dominant elites' belief in their relative security.

Gates is right that there are examples like South Korea (which we discussed in the book) that have transitioned to more inclusive institutions following a period of extractive growth. But South Korea's transition to democracy in the 1980s was in no way automatic. It came about as a result of protests by students and workers against the military regime, and only after the repression by the military failed to quell the unrest. More importantly, as a cursory look at our bibliographic essay would have shown, our dismissal of modernization theory isn't based on a few case studies or a gut feeling, but on careful econometric evidence. See, for example, our papers titled "Income and Democracy" and "Re-evaluating the Modernization Hypothesis, both jointly authored with Simon Johnson and Pierre Yared.

At another point in his review, Gates contends that economic growth is "strongly correlated with embracing capitalistic economics." Yet it is far from clear what he means by "capitalistic economics." Were Egypt's economic institutions during the presidency of Hosni Mubarak -- after he liberalized the economy and reduced the role of the state -- capitalistic? Most people refer to this as "crony capitalism," but perhaps this is all part of capitalist economics? Or consider the long dictatorship of Porfirio Diaz in Mexico in the 19th century, which eradicated many of the remaining restrictions of the Spanish colonial system, established an economy based on private enterprise (especially of his cronies), and "freed" markets (including the creation of the market for coerced labor). Was that capitalistic? What about South Africa under apartheid, based on private enterprise by whites, but disempowering and exploiting the majority blacks? Perhaps Gates himself should have more carefully defined his terms.

The concept of capitalism doesn't feature in our book for good reason. It muddies the waters. Our point, by which we stand strongly, is that what distinguishes societies isn't whether they are centrally planned or capitalist, but whether they are extractive or inclusive. Though centrally planned economies are by their nature extractive, so are many "capitalist" economies.

Finally, Gates takes issue with our supposedly "huge attack on foreign aid," citing in particular our "misleading" claims about Afghanistan. But again, he would have benefited from looking at the bibliography. The finding that about 10 percent of foreign aid goes to intended recipients isn't from Afghanistan, as he seems to think, but from Uganda, which was not a war zone but a peaceful country at the time of the 2004 study we cite. More importantly, there is now considerable evidence showing that foreign aid in the postwar era has had little positive impact on economic development, which Gates chooses to ignore (see, for example, William Easterly's White Man's Burden). Denying this is really putting your head in the sand.

But even sadder is the fact that we don't even argue against foreign aid. What we argue in the book is that aid -- the little of it that reaches its target -- does a lot of good for poor people. But it is not the solution to the real problems of development. Instead of endlessly asserting empirically untenable positions, we all need to move on and find more constructive ways to engage with poor countries. Foreign aid should certainly be part -- but not all -- of this engagement.

Gates does correctly point out that much is missing from the framework in our book. Even if underdevelopment isn't just a problem of bad leadership, and even if its solution won't come from enlightened leaders, a more complete framework should indeed integrate the behavior of leaders that play an important role in state building, organizing collective action, and articulating visions for social change. Examples of such leaders include Tunisia's Habib Bourguiba and Singapore's Lee Kuan Yew, both of whom undoubtedly influenced the course of their country's development. But we chose to emphasize institutions in our book because for leadership to have a lasting impact, it must become institutionalized via inclusive political institutions. After several decades of promoting education and developing a Tunisian national identity, for example, Bourguiba, who ran Tunisia as a dictator, was elbowed out of power by a very different sort of strongman, Zine El Abidine Ben Ali, who was far more interested in using his power to loot the country's resources. But Gates doesn't seem to be interested in such subtleties, preferring instead to criticize every aspect of Why Nations Fail.

Some say that all publicity is good publicity, and we should be thrilled to have Bill Gates review our book. Publicity is nice. But we spent more than 15 years researching, writing, and thinking about these topics, and we would be thrilled if the reviewers actually read and understood the book in the first place. Then we could have a constructive debate about the root causes of poverty in the world.

Win McNamee/Getty Images


Our Experience in Luvungi

International Medical Corps responds to Laura Heaton's investigation of mass rape in a small Congolese town.

The March/April issue of Foreign Policy magazine features an opinion piece by Laura Heaton titled "What Happened in Luvungi? On rape and truth in Congo." The article questions accounts of an attack on villages around Luvungi, a town in North Kivu, Democratic Republic of Congo (DRC), that occurred during four days in late July and early August of 2010, and questions the attention that sexual violence in the DRC receives in the wider international community.

In the article, Heaton alludes to a conspiracy among the U.N., international humanitarian organizations, researchers, and Congolese community leaders to fabricate or inflate stories of rape in DRC. What might motivate this range of actors to collectively lie about sexual violence is left unclear, and Heaton's allegations of distortion are largely based on "wildly disparate reports" of rape during the Luvungi attacks. But these simply do not exist. Every humanitarian organization reporting on the Luvungi response, and multiple human rights investigations all found that hundreds of women and girls reported being raped during those four days. In July 2011, the U.N. Office of the High Commissioner for Human Rights released a Final Report of U.N. Fact-Finding Missions into both the Luvungi incidents and the widely criticized U.N. peacekeeping response to the incidents. The U.N. confirmed that at least 387 civilians were raped, many by multiple perpetrators, and a great number in front of multiple witnesses.

This weight of evidence is not counterbalanced by the suspicions of one anonymous health worker whom Heaton interviewed, nor by her own inexpert and offensive suggestion that a "psychological element seemed to be missing" in the three survivors she encountered. Sadly, the author missed an opportunity to really explore what happened in Luvungi -- and the truth about rape in Congo -- when she chose to disregard available information (including information provided by both of us) that wasn't compatible with a premise she was determined to pursue.

In early August 2010, at the request of the Walikale Health Zone authorities and civil society leaders, we sent a joint Ministry of Public Health-International Medical Corps team to respond to the Luvungi incidents. On August 6, three days after the attacks ended, we began transporting people to the nearest health center, roughly 8 kilometers from Luvungi. While sexual violence is a too-common feature of conflict in DRC, we did not expect to receive such large numbers of patients reporting rape. As days and weeks passed, more and more women came forward. For survivors who arrived at the health center there was never any financial or material incentive provided. Far from encouraging increased reports, we were concerned about our ability to provide adequate treatment to those who needed it, as our supplies of emergency contraception and antibiotics were limited and we struggled to restock.

There are undoubtedly many important lessons to learn from the Luvungi incidents. For us, the priority has been determining how to better prepare response teams to meet the needs of survivors. We have since conducted numerous trainings on appropriate care and have pre-positioned emergency stocks of essential drugs in several health zones in North and South Kivu.

International Medical Corps has worked in DRC since 1999, providing health, nutrition, and other services to conflict-affected populations. In recent years, we have worked with communities to increase access to medical, psychosocial, legal, and livelihoods services for survivors of violence and vulnerable women. These programs have not replaced or come at the expense of other important initiatives, as Heaton suggests; on the contrary, programming to prevent and respond to violence against women and girls aims to increase the overall health and wellbeing of communities. Our programs, like most others in DRC, are carefully designed to not exclusively benefit survivors. This reduces a risk of further stigmatizing survivors and also eliminates a possible incentive for false reporting.

Globally, rape is considered to be one of the most underreported violent crimes, and DRC is no exception. Survivors of violence often feel ashamed, afraid their partners or families will abandon them, worried their communities will ostracize or isolate them. Many sense that reporting is futile, expecting that police and healthcare providers will not believe their stories. We train health and psychosocial workers to respect survivors' testimonies, and to provide services they seek without subjecting them to suspicion or inquiry. This approach is critical to increasing access to care. Last year, health facilities we support in eastern DRC treated an average of 128 survivors of rape every month, and psychosocial support staff received an average of 143 individuals reporting violence each month. Half of those cases reportedly involved rape, with more than 80 percent identifying armed actors as their perpetrators.

The brutality of sexual violence in eastern DRC is difficult to fathom, and all of us should feel outraged by the large numbers of women and girls, as well as men and boys, who are victimized by widespread rape. This outrage has led to increased support of organizations like ours that respond to the needs of survivors. In her article, Heaton points to increased funding for sexual violence response as evidence that "the advocacy works." This ignores the fact that current funding remains woefully inadequate to cover the tremendous needs in a large country, with limited infrastructure and continued insecurity in the east. Some of the facilities International Medical Corps supports in North Kivu cannot be reached by car or motorbike; our staff must walk for up to four days, with no telephone access, through terrain that often erupts in conflict. Responding to sexual violence in eastern Congo is difficult, dangerous, and expensive work. Organizations engaged with this work understand the great impact of sexual violence for survivors, families, and communities, and are consequently committed to advocating for increased protection and support of affected populations. It is deeply cynical to suggest that humanitarian organizations would misrepresent the toll of sexual violence for the purpose of increasing funding.

Sexual violence remains a salient feature of the conflict in eastern DRC. Some may be of the opinion that this is a tired story, or one that has been exaggerated. From our experience working with survivors and communities, we contend that the issue of sexual violence in DRC deserves increased attention and action. And we believe that in any debate over rape in eastern DRC, Congolese women and girls who have been most affected deserve the final say on truth.

Laura Heaton responds: 

On March 5, International Medical Corps (IMC), the American aid group who first responded to a violent incident in the eastern Congolese village of Luvungi in 2010, published a response to my article in the March/April issue of Foreign Policy. Our diverging accounts, and IMC's description of the challenges its team faced in the immediate aftermath of the attack, underscore the complexity of providing appropriate assistance to civilians in volatile, isolated communities -- which are all too common in eastern Congo, particularly now.

There are many important questions that the Luvungi incident raises, which is why I set out to examine it, knowing full well when I started that definitively pinning down all the details would never be possible. But IMC's rebuttal suggests that my reporting is based on a single source and a hunch; in fact, nearly two years of research led me to many different sources who had varying levels of knowledge of the incident and diverse perspectives on how and why the established narrative about Luvungi was inaccurate. I stand by my sources and the editorial decisions about which accounts to present.

I also take issue with IMC's characterization of my piece as suggesting a "conspiracy" between various actors. Nothing in the Foreign Policy article asserts anything of the sort. What the article does say is fact: that the U.N. and NGO teams that subsequently investigated the rapes in Luvungi drew on IMC's original account -- and stated so in their various reports and in interviews with me. The inquiries all these groups undertook in the months following the rapes were carried out by seasoned investigators and experts in working with traumatized victims. But as time passed, sources told me, establishing a precise count of sexual violence victims became increasingly difficult.

By writing this story I hoped to catalyze further discussion and investigation into the causes of sexual violence and how it ties into the broader understanding of the conflict in Congo. I have never disputed -- nor have I sought to detract from -- the fact that sexual violence is a tragic and alarming problem in Congo; academic research and survivor narratives have established irrevocably that the incidence of sexual violence is exceptionally high and can be horrifyingly brutal. But what is less often highlighted is this: most sexual violence in Congo is committed by civilians, and not just demobilized soldiers.

Finally, the authors of IMC's rebuttal claim that I "chose to disregard available information (including information provided by both of us) that wasn't compatible with a premise [I] was determined to pursue." I did, in fact, have several conversations with IMC personnel, who declined to comment on the record except for an emailed statement provided to me by IMC communications director Margaret Aguirre, from which I quoted in the article.