Rising or Falling Global Income Inequality?

Stalemate in an Academic Debate

Chowpatty Inequality
Rising or Falling Global Income Inequality? : Stalemate in an Academic Debate - Claire Le Barbenchon


The academic debate on whether world income inequality is rising or falling has reached a stalemate: parties are unable to agree on the analysis of these economic world trends. This commentary examines recent research that supports the opposing “convergence” and “divergence” camps to examine the origins of the debate, and to determine why consensus is so difficult to reach. While this analysis concludes that the main driver of disagreement is calculation methods and data, that conclusion poses a key question: Is studying world income inequality useful?


The debate about global economic inequality has become particularly prevalent in recent years, as evidenced by the 2006 World Development Report, titled “Equity and Development”.[1]However dry this topic may seem, no debate in development economics may be as contentious as this one (or include as many dissenting footnotes and self-references).[2] The issue of income inequality likely prompted the World Bank to write in their 2006 report that “[w]hen it comes to judging inequality, intelligent people can disagree…”[3]

However, consensus on whether income inequality is rising or falling in the world is elusive. This is because many academics and development specialists utilize the notion of inequality to point to the benefits or drawbacks of globalization. Each economist that studies the issue cherry-picks methodology and data to present either the rise or fall of inequality, based on their stance. This paper will delve into the main debates surrounding inequality, focusing on those authors who discuss convergence or divergence in global/international inequality, and explore the potential motivations behind certain arguments. It will argue that authors’ conclusions on inequality are based on their calculation method and form of data, which is a point of contention in the literature itself.[4] It begins with an overview of authors that promote convergence theory (Bhalla, Sala-i-Martin, and recently, Bourguignon), followed by a look at some proponents of divergence theory (Milanovic, Pritchett and Collier). This paper does not aim to offer an exhaustive account of inequality theories, but a comprehensive overview of the debate through an examination of principal contributing authors.

The method used to calculate convergence or divergence has an important impact on the results. Branco Milanovic (2005) shows that when countries are taken as a unit, the divergence between countries has been increasing, particularly since 1980. However, when countries are weighted by population, inequality has been decreasing since the 1970s.[5] This difference has fueled the debate, with anti-globalization advocates opting for the first method, and their opponents using the second.[6]

Authors tend to align themselves with one of the two camps based on their own judgment, ethical stance, or personal agenda. Several scholars in the convergence camp have tried to dismantle contrary evidence by using new calculation methods. Surjit Bhalla criticizes the World Bank’s method, which uses survey sample means for calculating expenditure, and instead opts for a combination of survey and national-level data, approximating the Lorenz curve for each country. Using this method, he concludes not only that poverty relative to growth has decreased, but also that the Millennium Development Goal of halving poverty (those living under 1$ a day in 1985 dollars) had already been met in 2000.[7] His method came under fire from Ravallion, former director of research at the World Bank, who published a rebuttal supporting the World Bank’s calculation method and stating that “[Bhalla’s finding] would be good news, if it was right.”[8]

Why did Bhalla’s calculations create such controversy? Bhalla maintains that the World Bank is pursuing self-interested goals, overstating poverty to preserve the institution’s importance. Both camps staunchly defend their statistical methods. Bhalla promotes convergence based on surveys mixed with National Data Accounts, and then goes on to look at the declining divergence between the 20 richest countries and the 20 poorest countries. Because the latter includes China (19th poorest in 1900), Ravallion argues, Bhalla’s estimate gives an inaccurate picture of inequality.[9]

This particular disagreement is representative of much of the academic debate over inequality. However, this form of academic exchange overlooks the importance of improving living conditions for the world's poorest citizens. If the correct method of measurement is unclear, it is surely better to err on the safe side and assume that there is much work to be done, instead of defending one’s position tooth and nail to gain some ground in the debate.

Another prominent advocate of convergence is Xavier Sala-i-Martin. He uses a combination of national accounts and survey data to compare GDP per capita and country income distributions to approximate inequality. He obtains a Gini coefficient for each country, which he then places in a world coefficient. He finds decreasing inequality over time, particularly since the 1980s, which he aptly calls, “convergence, period!”[10] Interestingly, he removes the principle drivers of divergence and convergence (the US, Africa and China) and finds the same results.[11] His findings, like Bhalla’s, have not escaped criticism from World Bank researchers, who have dismantled his methodology in order to discredit his results. Branco Milanovic, a former chief economist at the World Bank, has disputed Sala-i-Martin’s methods. Milanovic suggests that his approximations of income distribution are off the mark, and that his data is faulty due to missing values and estimations.[12] Furthermore, Milanovic objects to the fact that Sala-i-Martin keeps China in the calculations, which he contends leads to bias in the results. Sala-i-Martin’s 2006 paper, using the same method, does remove China, which indeed causes inequality to increase, but then removes the US as well, which again shows decreasing world inequality. This example illustrates that removing only one main driver skews the results.[13]Furthermore, Milanovic claims that this method is not really an estimate of world inequality, but instead closer to weighted international inequality, which he deems the “least interesting [concept of inequality]."[14] 

This debate, perhaps more than that between Bhalla and Ravallion, is based on methodological disputes. Milanovic, being the first economist to use survey data to calculate world inequality, promotes his method as the most accurate. Sala-i-Martin, in turn, criticizes the use of incomplete survey data that often diverges from national accounts.[15]

A final important advocate of convergence theory is François Bourguignon, an influential voice in the debate over inequality. In his 2012 book La mondialisation de l'inégalité, he finds decreasing inequality between coutries for the past two decades, weighted for population, which he attributes to the growth of emerging economies.[16] Furthermore, he demonstrates decreasing world inequality since the 1980s using an 80/20 ratio calculation, adjusted for PPP. Bourguignon’s book has made waves in the inequality debate, particulary since Bourguignon and Morrison’s 2002 study was a major influence on the convergence theory. Curiously, Bourguignon’s book has not yet come under fire for flawed methodology or biased judgment. Given that the World Bank is firmly ensconced in the divergence camp, his theory would seem to be a prime target for their rebuttals. However, his book has remained largely free from such criticism. Is this because the World Bank considers itself to be in a sufficiently strong position that they need not challenge these assertions, or does this lack of criticism reflect a growing consensus that inequality is rising after all?

As for the divergence paradigm, many authors maintain that inequality is increasing. Milanovic, as stated above, explains that inter-country unweighted inequality is increasing, and that weighted inequality increases once we remove China and India’s bias.

Other important theorists include Lant Pritchett, who in 1997 affirmed that there was “Divergence, Big Time.” While his calculations stop in 1990, he demonstrates slower per-capita growth in non-OECD countries vs. OECD countries, noting that there is convergence of developed countries’ growth numbers and divergence with and between developing countries.[17] Interestingly, Sala-i-Martin’s conclusions are framed as a direct response to Pritchett’s argument, each using different methodology (Pritchett focuses on growth rates, whereas Sala-i-Martin focuses on the Gini and other measures of inequality). This disagreement illustrates the complexity of this debate.

Paul Collier’s book The Bottom Billion presents the idea that the approximately one billion people contained in the world’s poorest countries are getting poorer as their countries experience slower and negative growth.[18] This argument has remained somewhat removed from the main debate on inequality trends. However, this kind of analysis should perhaps be at the center of the inequality debate.

In fact, focusing on the poorest countries in order to target the individuals most in need of development aid, as Collier proposes, would seem to be the most logical and altruistic goal. In that case, why is the debate on world inequality so fervent? First of all, as explained above, there is a possibility that economists involved in the debate have an agenda, if only academic pride, when they dismiss calculations of others that run contrary to their own. Whether or not the World Bank has a stake in showing high inequality as Bhalla states, there appears to be a trend in the World Bank’s calculations. This is perhaps linked to the fact that the leading economists working on inequality are at the World Bank and have access to the most wide-reaching and accurate data. Whether inequality is rising or falling is a matter of perception based on the methodology, data and calculations performed.

The next question should be: Does it matter? The answer is a cautious yes. The debate is often framed in terms of pro- or anti-globalization advocacy, either confirming or refuting the argument that increased interconnectedness has improved welfare. Looking at country-level aggregated income data can be important in explaining the benefits and/or downfalls of globalization. However, this data does not provide a comprehensive picture, and relying solely on this methodology could render the argument irrelevant. The disadvantages of increased income inequality are numerous, and include higher crime rates, decreased biodiversity, and reduced standards of living. However, these are linked more closely to within-country income inequality, or even regional inequality. The debate has shifted towards these two factors, as illustrated by the 2006 World Development Report, which focused on within-country inequality as a main concern. With this in mind, and considering the fact that consensus on whether global income inequality is rising or falling is elusive, the debate appears to be more focused on petty academic exchanges than working towards improving human welfare. 

In conclusion, this article has explored the two principle sides of the world inequality debate: convergence and divergence. Authors from each camp have come under fire for their theories, and particularly the methodologies that led to their results. However, in this debate, it is important to consider the motivations of these academics and the perceived benefits of this form of research in terms of the questions they address and the policy implications of this research. It is not clear that current literature on world income inequality can bring fruitful change or development to the poorest segments of the population. In order for this debate to mature, it is important that academics consider the realities of those on the ground who are most in need, and are the true victims of inequality. In light of recent publications, such as the 2006 World Development Report, which includes evidence that high within-country inequality is linked to negative outcomes, the debate should be refocused on reducing within-country inequality, a problem that is proven and not simply theoretical.

Notes & References

  1. World Bank, World Development Report 2006: Equity and Development, (Washington DC: World Bank, 2006).
  2. This is particulary obvious in the debate between Bhalla and Ravaillon about the attainment of the Millenium Development Poverty targets.
  3. World Bank, World Development Report 2006, 2006.
  4. M Ravaillon, E Thorbecke and L Pritchett. Competing Concepts of Inequality in the Globalization Debate [with Comments and Discussion]. (Brookings Trade Forum, Washington DC: Brookings Instituion Press, 2004).
  5. Branco Milanivic, Worlds Apart, (New Jersey: Princeton University Press, 2005), 39.
  6. Ravaillon, et al., Competing Concepts, 2004.
  7. Surjit Bhalla. “Crying Wolf on Poverty: Or How the Development Goal on Poverty has already been Reached,”Economic and Political Weekly 38 (2003): 2844.
  8. Martin Ravallion, “Have We Already Met the Millennium Development Goal for Poverty?” Economic and Political Weekly 37 (2002): 4638.
  9. Ravallion, “Millennium Development Goal”, 4639.
  10. Xavier Sala-i-Martin, “The World Distribution of Income: Falling Poverty and ... Convergence, Period,”The Quarterly Journal of Economics 121, (2006): 386.
  11. Sala-i-Martin, “The World Distribution of Income,” 387.
  12. Branco Milanovic. “The Ricardian Vice: Why Sala-i-Martin’s calculations of world income inequality are wrong,”Mimeo (2003): 2-6.
  13. Sala-i-Martin, “The World Distribution of Income,” 387.
  14. Branco Milanivic, Worlds Apart, (New Jersey: Princeton University Press, 2005), 10.
  15. Sala-i-Martin, “The World Distribution of Income,” 356.
  16. François Bourgignon, La mondialisation de l'inégalité, (Paris: Seuil, 2012).
  17. Lant Pritchett, “Divergence, Big Time,” The Journal of Economic Perspectives (1997).
  18. Collier, Paul, The Bottom Billion, (London: Oxford University Press, 2007).