Stefan Dercon, Gambling on Development: Why Some Countries Win and Others Lose. London: Hurst & Company, 2022. 360 pages.

Born in Belgium, Stefan Dercon now lives in the U.K. and is a professor at Oxford University. He is a worthy addition to the list of academics, mostly economists, trying to understand why some countries escape the World Bank’s “low income” category and reach at least “lower-middle income” status; why some stall at “lower middle” and others graduate to “upper middle.”¹

Table 1 lists the countries Dercon discusses at length. I have added Canada to drive home the fact that Dercon’s book concerns those in the first three categories – countries with per capita GDP far below that of high-income countries. Despite its phenomenal manufacturing prowess over the last half century, China’s current per capita GDP is only one third of Canada’s; India’s is one seventh. Dercon devoted his early career to Ethiopia which, until its post-2020 civil war, had a growth rate second only to China’s among the countries in table 1. However, after a quarter century of rapid growth, Ethiopia’s per capita GDP in 2019 was only 5 per cent of Canada’s.

I quote from Dercon’s preface:

A decade or so ago, I came to a realisation. I needed to radically rethink how development comes about. By then, I had one foot in academia – as an Oxford professor – and one in government – … as chief economist and the most senior technocrat in the UK’s Department for International Development. My moment of truth came not long after I first flew into Beijing … I never saw in China the extreme deprivation I had seen in Burkina Faso, Ethiopia, or India … In the 1980s China was the country with the largest number of people living in such deprivation … By the time I was travelling around China, more than half a billion Chinese had risen from those depths.

Dercon’s epiphany came while travelling across China. At the time of Mao’s death in 1976, per capita incomes in China and India were similar. By 2019, China’s per capita income was two and a half times India’s. Whatever criticisms might be made of World Bank estimates, there is no denying, Dercon insists, that China has lifted 500 million people, the equivalent of half Africa’s population, out of “extreme deprivation.” When Deng Xiaoping assumed power, he had no Soviet-style five-year plans to implement. What underlies the phenomenal development success of post-1978 China, Dercon concludes, is a pragmatic agreement among Communist Party elites to pursue economic growth.

Dercon’s epiphany is the importance of elite agency: “The political and economic elites have much more agency than is usually allowed by the historical approach to institutions.” (I discuss the institutionalists in the accompanying box.) Among “low income” countries, Dercon is not much interested in democracy and honest elections, constitutions with ringing statements on human rights, the ideal of politically neutral courts and politicians subject to rule by law, or efficient implementation of taxes. On all these dimensions, successful developing countries vary widely in the scope accorded to government or courts. What the successful countries share is “a development bargain … an underlying commitment to growth and development among members of the country’s elite.”

Dercon defines a successful “development bargain” as having three shared components:

  • A durable political and economic deal: Development requires long-term horizons among the elites, which implies that they agree not to undertake violent intra-elite conflict.
  • A sensible state: The scope of the state differs widely among successfully developing countries. It may tolerate corruption and criminality in some sectors, provided it manages other important sectors well. What the state does take on, it performs “sensibly.” If, for example, the elites decide the state should provide primary and secondary schools for all, it does so with attention to school learning outcomes. In South Asia, the Sri Lankan government has – despite the corrupt incompetence of the Rajapaksas over the last 15 years – managed good public schools, reflected in the fact that the illiteracy rate among children aged 10 to 14 is 15 per cent, compared to the 58 per cent average in South Asia as a whole.
  • The ability to acknowledge errors and correct course: Dercon acknowledges that Bangladesh displays corrupt “volatile rent-seeking politics,” but the state does ultimately correct major obstacles to economic development. Bangladesh is the world’s second largest garment exporter. For decades, Bangladeshi governments tolerated massive violation of health and safety regulations in garment factories. For example, a political insider blatantly flouted all regulations and constructed a large building housing seven garment factories on a swamp. In 2013 the building collapsed: 1,100 workers died and more than 2,000 suffered injuries. The owner was captured and shot as he fled to India. Over the next five years, the government markedly improved safety regulations.

Dercon’s definition of the three components is admittedly fuzzy and omits some important factors that determine per capita GDP growth rate – education in particular. More on this later.

Yes, minister

About the time he visited China, Dercon found himself with a new minister in his department. She told the press, “I didn’t come into politics to distribute money to people in the Third World.” He provided her with a large pile of books, written mostly by economists, on economic development – whether she wanted them or not. Among the authors are four winners of the Nobel economics award. He insisted that she spend an afternoon in a tutorial discussing the authors’ conflicting analyses. Implicitly, he does the same for readers of his book (see Box).

Ideally, Dercon should have tutored his minister for more than one afternoon and offered his conclusion on the numerous debates among the development economists. However, his tutoring was cut short. In an interview, the Prime Minister (David Cameron at the time) said that Daron Acemoglu and James A. Robinson’s Why Nations Fail was his favourite book on development. So be it. Dercon decided his duty as a senior civil servant was to design policy consistent with the Prime Minister’s view of the world. This was not particularly difficult. Dercon, like Acemoglu and Robinson, is among the institutionalists, with the qualification that he thinks governing elites have more “agency” than most institutionalists acknowledge.

Dercon’s zoo

To say Dercon has travelled widely is a serious understatement. Very few academics, journalists or diplomats have visited as many countries. He describes meetings as a journalist with politicians and senior economic advisers in Kinshasa and conversations with peasants in rural Ethiopia. When I was in Bangladesh this past summer as a guest of BRAC, the country’s largest NGO, my hosts discussed Dercon’s visit in 2020, just before the COVID pandemic. As a policy analyst, he discusses in detail various successful programs pursued by Meles Zenawi, leader of the Tigrayan coalition that, in 1991, toppled Ethiopia’s Marxist dictator. Meles was not a democrat; he led a meritocratic government in the tradition of Deng Xiaoping in China and Lee Kuan Yew in Singapore. Dercon has a high regard for Meles, who died in 2012. (As noted above, Ethiopia ranks second only to China in per capita GDP growth among the countries in table 1.)

Dercon is, obviously, an economist. He alludes to other goals, but his principal criterion in evaluating countries is growth rate of per capita GDP. He began his academic career teaching in Addis Ababa and participated enthusiastically in early initiatives of Meles’s government. The optimism of youth has mellowed with age:

With the end of the Ethiopian civil war (in 1991), it was hoped that peace would launch a new period of change and progress across the country and on the continent as a whole … I did not see the genocide coming (in Rwanda). I was not alone. For many of us, it dashed our hopes that peace and open politics would bring rapid prosperity to the continent … Political systems seemed to change, but in too many places the change only ensured that everything stayed the same. Improvements in people’s lives were barely evident in the data we researchers collected and analysed, even though we looked hard for them.

Four economically successful countries in east and southeast Asia – South Korea, Taiwan, Hong Kong and Singapore – are often labelled as the Asian tigers (Hong Kong is no longer a country, but would very much like to be independent of mainland China). To leaven his discussion of economic progress – or lack of progress – among the countries he discusses, Dercon creates an imaginary zoo. China, the world-changing outlier, is a dragon; India is a peacock, very showy with promises but short on fulfilling them. Indonesia is a tiger cub, which may or may not mature into a full-fledged Asian tiger. He reaches the same conclusion with respect to Bangladesh.

Many African countries are compared to hippos in a lake. The visitor sees the eyes above water, but the political action is below the surface, hence invisible. As hippo examples, he discusses Sierra Leone and Malawi. He is more hopeful for Kenya, Uganda and Ghana, which may achieve economic takeoff. He has no faith in either Nigeria or the Democratic Republic of Congo escaping low-income status. Their governing elites engage in continuous zero-sum games that benefit themselves – but no one else.

The worst Ethiopian catastrophe of the 20th century was not the Italian invasion or Haile Selassie’s “traditional military and aristocratic” government; it was the economic incompetence of the Marxist regime that ran the country from 1976 to 1991. It did little to limit the 1984–85 famine, estimated to have killed up to a million people. After 1991, Tigrayans dominated the Ethiopian government for a quarter century. By 2018, resentment of their political and military dominance came to the fore. In 2018, a non-Tigrayan became prime minister. Subsequently, Tigray attempted to secede, and the ensuing civil war has shattered Ethiopia’s “development bargain.”² Biafra’s attempted secession from Nigeria half a century ago is an ominous precedent.³

Dercon labels several countries as hyenas. The secession of South Sudan from Sudan has yielded negligible benefits for citizens of both countries: “The South Sudanese state and its leaders could have been lions, but they chose to be clans of hyenas fighting over the scraps.” Other hyenas are Afghanistan, Nepal, Lebanon and Somalia. At various times over the last half century, these countries have experienced anarchic violence. At present, Lebanon, Afghanistan and Nepal are nominally at peace. But even in Somaliland, the most stable region in Somalia, no one places much faith in ever realizing a successful development bargain.

Paul Kagame of Rwanda is similar to Meles inasmuch as they both saved their countries from hyena status and introduced economic and social policies far more efficient than those of their neighbours. Like Meles, Kagame is a meritocratic dictator (Rwanda’s per capita GDP growth in table 1 is close to that of Ethiopia). Had Dercon published his book before COVID, he would probably have labelled post-1991 Ethiopia as an African lion (lions being the symbol of the Ethiopian empire). He would have done the same for post-1994 Rwanda. Now, faced with ethnic conflict in Ethiopia erupting into civil war, he adds a question mark to both. It is to be hoped that ethnic tensions in Rwanda and eastern provinces of the Democratic Republic of Congo won’t reignite the civil war of 1994.

Education matters

Many academics in developing countries are content to blame European imperialism for their countries’ poverty. The argument rings true, for example, when discussing the problems posed by arbitrary national borders, which oblige multiple ethnic communities to accept rule by other ethnic elites. However, we cannot rewrite history. Dercon argues that blaming the colonizers is intellectually lazy: it minimizes the responsibility of governing elites to undertake “development gambles.”

I am sympathetic to his argument, with one major exception. His discussion of basic education as an element in a development bargain is superficial. Whatever the errors of Chinese policy in Mao’s reign, by the time Deng Xiaoping adopted “sensible” policies and persuaded his politburo colleagues not to “cancel” one another, the majority of Chinese adults could read at a basic level and do basic arithmetic; hence, they could migrate to cities and become productive industrial workers. By the early 1990s, according to UNESCO’s (excessively) modest criteria, nearly 80 per cent of Chinese could read, whereas less than 50 per cent of Indians could do so. This literacy gap is a relevant factor in any explanation for why India’s per capita GDP is currently only 40 per cent of China’s – and countries in sub-Saharan Africa rank much lower relative to China.

In 2021, the World Bank launched a “learning poverty” statistic to estimate the national share of children aged 10 to 14 who cannot read short Grade 2–level stories comparable across countries (see table 1). In the specified age range, children are expected to be in an upper primary grade. A country’s learning poverty rate is the sum of (a) children in this age cohort who have dropped out or never enrolled in a school and are assumed not to be literate and (b) children who are in school but unable to read a simple story.⁴

In South Asia, for example, the proportion no longer in school is about 15 per cent; the proportion of students in school but unable to read is 50 per cent. In a random sample of 100 South Asian children aged 10–14, about 15 cannot read having dropped out or never enrolled; half of the 85 in school cannot read simple stories. The average learning poverty rate in South Asia is 58 per cent unable to read; in China the equivalent share of children aged 10–14 is only 18 per cent. In sub-Saharan Africa it is 87 per cent. In Dercon’s “home” country, Ethiopia, he discusses in detail Meles’s sensible macroeconomic and agriculture policy. He doesn’t discuss Ethiopian schools. In Ethiopia, the learning poverty rate is 90 per cent, slightly worse than the sub-Saharan average.

Table 1 illustrates national poverty rates based on the World Bank’s suggested poverty threshold for lower-middle-income countries (US$3.65 per person per day). For a family of four living at the poverty threshold, annual income is equivalent to US$5,330. There is very high correlation between high 2019 per capita GDP and low poverty rate. There is also very high correlation between high 2019 per capita GDP and low learning poverty rate. The implications are that more prosperous countries achieve lower poverty rates, and that producing literate children contributes to high per capita GDP. Disentangling the role of education from other relevant factors in explaining per capita GDP is an interesting statistical exercise. Here it suffices to say that education matters.

Finally, there is very high correlation between a low learning poverty rate and a low poverty rate. The implication is that, independent of per capita GDP, literacy contributes to low poverty rates. Relative to illiterate women, for example, women able to read are more likely to earn income for their families, and usually exercise more authority on issues such as limiting the number of children. In addition to Dercon’s zoo, I grade table 1 countries in terms of achieving a low poverty rate. The “A” grade countries are Sri Lanka, China and Indonesia (poverty rate 14–22 per cent). The countries deserving a “B” grade are Nepal, Pakistan, India, Ghana and Bangladesh (40–52 per cent). The “C” grade countries are Kenya, Nigeria, Sierra Leone and Ethiopia (60–65 per cent); “D” grade countries are Uganda, Tanzania, Zambia and Rwanda (72–78 per cent). The unambiguous failures are Democratic Republic of Congo, Malawi and Somalia (88–90 per cent).

Dercon’s Reading List for his Minister

Jeffrey Sachs, an economist at Columbia University, wrote The End of Poverty: How We Can Make It Happen in Our Lifetime (Penguin, 2005). He emphasized two factors to explain the persistence of poverty in “low income” countries. First, most such countries are prone to a high incidence of debilitating tropical diseases, which reduces productivity. Sachs was the chief economist associated with the Millennium Development Goals (MDGs), the ambitious United Nations campaign between 2000 and 2015 to achieve eight social policy goals. One MDG set a target for reduced under-five mortality, another for reduced maternal mortality. Second, per capita GDP in these countries is so low that it is unrealistic to expect them to save and invest in needed infrastructure and quality social services. His basic advice: high-income countries should give lots of aid.

Paul Collier, an Oxford colleague of Dercon’s and former World Bank economist in Africa, wrote The Bottom Billion: Why the Poorest Countries are Failing, and What Can be Done about It (Oxford University Press, 2007). The book contains detailed advice to donor agencies when they are faced with corrupt host governments. The problem of corruption is particularly acute, Collier contends, in countries (such as Nigeria and the Democratic Republic of Congo) that have large resource rents from oil and minerals. The magnitude of such rents is an irresistible incentive for host-country politicians to concentrate on zero-sum conflicts that maximize income for those allied with the governing party – and provide only rudimentary services to the majority.

William Easterly, a “rogue” World Bank economist, wrote The White Man’s Burden: Why the West’s Efforts to Aid the Rest Have Done So Much Ill and So Little Good (Penguin, 2006). His critique goes further than Collier’s. If countries had democratic elections, property rights and the rule of law, and the governing elites let markets function, economic growth would take place. He cites Thailand and Japan as fortunate never-colonized countries that, before World War II, realized significant development without aid. Donors complain about corrupt governments, but in practice they subsidize bad behaviour. Hence, in most circumstances, aid has done more harm than good: it is a prop for corrupt governments and hampers the emergence of “sensible” governments.

Dambisa Moyo, author of Dead Aid: Why Aid is Not Working and How There is a Better Way for Africa (Farrar, Straus and Giroux, 2009), is a Zambian economist. She more or less agrees with Easterly, and emphasizes the potential advantage for host governments of pursuing international capital investments.

Joseph Stiglitz, Nobel economics prize recipient and yet another senior World Bank economist, wrote Globalization and its Discontents (W.W. Norton, 2002). He emphasizes the monopoly power of multinational firms. They capture far too large a share of revenue arising from their investments in developing countries. Decisions of international organizations such as the World Trade Organization, the World Bank and the International Monetary Fund function in a manner unduly favourable to multinational firms. Stiglitz wants major reform in all three. Nonetheless, he favours developing countries engaging in international trade. Why? Despite unequal income distribution, workers in tradable sectors benefit because they typically earn more than they could in traditional sectors (such as agriculture) and, over time, local entrepreneurs may replace multinational firms.

Ha-Joon Chang, a Korean professor at Cambridge University in the U.K., is the author of Bad Samaritans: The Myth of Free Trade and the Secret History of Capitalism (Bloomsbury, 2007). He fundamentally disagrees with Stiglitz. Chang wants developing countries to do as South Korea did: protect the domestic market until host countries mature and are able to compete effectively with multinational firms.

Abhijit Banerjee and Esther Duflo shared a Nobel economics prize for their many carefully implemented randomized control trials. The technique requires two samples identical in all relevant respects except that the treatment sample receives a policy intervention while the control sample does not. If allocation to the treatment sample is random, the only explanation for a significant difference in sample outcomes must be the treatment. Such trials have demonstrated many potentially productive policy interventions. Banerjee and Duflo jointly wrote Poor Economics (Penguin, 2011). The problem with randomized control trials is external validity: will the result obtain outside the confines of the trial? Critics of their work, such as Douglass North and the team of Daron Acemoglu and James A. Robinson, insist that randomized control trials ignore underlying institutional obstacles that prevent adoption of productivity-enhancing interventions.

This brings us to the institutionalists. Douglass North earned the Nobel economics award in 1990 for his critique of market “fundamentalists” (such as Easterly) and his rehabilitation of institutional analysis among economists. The first sentence of Institutions, Institutional Change and Economic Performance (Cambridge University Press, 1990) is as good a definition of institutions as any: “Institutions are the rules of the game in a society or, more formally, are the humanly devised (formal and informal) constraints that shape human interaction.” Acemoglu and Robinson are disciples of North. Their book Why Nations Fail: The Origins of Power, Prosperity, and Poverty (Random House, 2012) begins with a discussion of Nogales, a city straddling the U.S.-Mexico border. In the Mexican half, average income is one third that in Arizona. The fundamental explanation of this 3:1 ratio, they argue, is differences in U.S. versus Mexican institutions.


¹ The World Bank slots countries into four categories based primarily on per capita GDP: low income, lower-middle income, upper-middle income, high income.

² Jon Lee Anderson, The Palace Gates: Is the Prime Minister of Ethiopia Rebuilding His Country or Tearing It Apart?, New Yorker, October 3, 2022.

³ See my Nigeria’s Voltaire, or Maybe Dostoyevsky, Inroads, Summer/Fall 2022.

⁴ World Bank, The State of Global Learning Poverty: 2022 Update, June 23, 2022.