Published on Data Blog

Is COVID-19 increasing global inequality?

This page in:

No country has been able to fully avoid the economic consequences of the global pandemic, but some have been hit more than others. Do the countries hardest hit tend to be situated in the bottom of the global income distribution? If so, the pandemic would be increasing inequality between countries. In this blog post, we explore this topic by comparing the income losses at different places of the global income distribution.

The global income distribution includes every person in the world ranked using their own income vis-à-vis everyone else’s income. To construct the global income distribution, we bring the most recent household income or consumption surveys available for each country forward to 2019 using the same methodology as the World Bank uses for estimating global poverty (see PovcalNet). For the pandemic affected years, while we do not have adequate household surveys, we do have country-level growth forecasts. We use the June-2021 growth forecasts from the Global Economic Prospects (GEP) to project each person’s income into 2020 and 2021. Note that these projections of the survey data assume distribution-neutral growth within countries—this implies that every person within a country changes their income by the same per capita growth associated with her country. In doing so, we hold the inequality within a country constant. Hence, while we capture within-country differences using the latest household surveys, the changes in the global income distribution thereafter are solely due to the changes between countries (adjusted for population). Within country changes are important for the global distribution as well. However, due to lack of data, at this point we are unable to say much about the changes within each country. On the flip side, holding the changes constant within each country gives us an unimpeded view of the between-country effect of the pandemic. We explore this below.

As we have done previously, we project the global income distribution in 2020 and 2021 both with and without the COVID-19 pandemic. For the latter, we rely on growth forecasts launched before the pandemic spread. Our method also assumes that household income or consumption broadly follows GDP growth (scaled by a pass-through factor) which is the only available information in the absence of new survey data.

Effect of COVID-19 on the global distribution

Figure 1 depicts the COVID-19-induced income loss for each percentile of the global income distribution in 2020 and in 2021. On the horizontal axis, individuals are ranked (in 100 bins or percentiles) from the poorest (left) to the richest (right). The vertical axis reports the percent of income lost due to the pandemic. Income loss is calculated as the difference in income between the global income distribution projected with the pandemic and one without. In other words, the figure shows how much less income each percentile has because of the pandemic. Importantly, the figure also shows the relationship between different parts of the distribution. For instance, the steeper the curve (from left to right) the greater the income gain for the top (richer) compared to bottom (poorer) of the distribution (and vice versa). In 2020, the richest two deciles (i.e., the top 20 percentiles together) on average lost about 5% of their expected income, while the poorest two deciles on average lost about 6% of their income. The curve becomes even steeper in 2021, suggesting that the divergence in the impact of COVID-19 is increasing over time. For instance, in 2021 the richest two deciles on average are expected to recover nearly half of their 2020 losses, while the poorest two deciles on average are expected to further lose 5% of their income. 

Figure 1: Global income losses due to the COVID-19 pandemic

This finding that global income inequality has increased due to the pandemic is not new. In fact, there was widespread belief that a pandemic such as the one we are living through now will undoubtedly affect the poorer regions worse than richer regions. Deaton (2021), using population-weighted per capita GDP, also found a similar effect in 2020. However, Deaton argues that this increase in population-weighted inequality between countries is largely explained by China (whose incomes rose in 2020) and India (whose incomes fell in 2020). Our findings are consistent with Deaton for 2020 (the paper does not include estimates for 2021).[1] In 2020, when we use all countries, we find that the losses for the bottom 40% of the distribution are larger than that for the top 60%, suggesting rising inequality. When China and/or India are excluded, the evidence is more mixed with losses fairly evenly distributed (when either China or India are excluded) and losses slightly larger for the top (when both China and India are excluded).

In 2021, however, we consistently find divergence between richer and poorer parts of the global income distribution, as shown in the figure above—on average, the bottom 40% has 2.8 percentage points larger loss than that of the top 60%. We find that this divergence holds even when we exclude China and/or India from the global sample. The evidence so far thus suggests that while the pandemic has affected everyone relatively evenly, the recovery has been less uniform. However, given the sensitivity of our exercise to changes in growth vintages, it is important to note that our projections might still change for 2021.

Putting the increase in inequality in perspective 

Earlier studies found a decline in between-country inequality over recent years (e.g. see the recent paper by Milanovic, as well as his earlier work and this World Bank report). The average incomes across countries have been converging for over two decades. Figure 2 shows the change in income inequality between countries, as measured by the Mean Log Deviation (MLD), from 1988 to 2021.[2] The first 5 years of the period saw little change in the part of global inequality that is due to differences between countries. From 1993 up to 2017, between-country inequality has been constantly falling. Interestingly, the largest reduction came in the period affected by another global shock, the financial crisis. Between 2008 and 2013, inequality decreased by over 16% (i.e., the between-country MLD decreased from 59.6 to 49.8 points). Overall, the inequality between countries has declined by 34% between 1993 and 2017. Put differently, a third of the gap between nations has been eradicated in roughly 25 years.  This convergence has been attributed to the economic development of mainly poor and populous countries in Asia.

We find that the gap between nations is expected to increase for the first time in a generation. The between component of the MLD is expected to rise by 1.2%. Had there been no pandemic, we project that between-country inequality would have decreased by 2.6%, in line with recent trends. Taken together, COVID-19 has directly offset the reduction in the gap between countries observed from 2013 to 2017.  It is important to stress that these estimates only tell us about the part of global inequality that is due to differences between countries. As the earlier work has shown, changes in within-country inequality over the past 25 years have somewhat counteracted the decline in between-country inequality. The available evidence also points to an increase in within-country inequality as a result of COVID-19. Once more survey data become available for this period, we plan to revise these estimates to get a fuller picture of overall global inequality.

Figure 2: Change in between-country inequality, 1988-2021

The authors gratefully acknowledge financial support from the UK Government through the Data and Evidence for Tackling Extreme Poverty (DEEP) Research Program and the Republic of Korea, acting through the Korea Development Institute School of Public Policy and Management (KDIS) and its Knowledge Partnership with the World Bank.

[1] It is important to note that while Deaton (2021) uses population-weighted per capita GDP to calculate the inequality changes (thus assuming zero within-country inequality), we have used the distribution of individual income from surveys. Since the within country differences are held fixed in both cases, the difference in approaches impacts the level of inequality but not the trend.

[2] The MLD is a measure of inequality, that can be decomposed into differences within and between countries (in contrast to the Gini index). Note that the periods from 1988-2013 are five-year intervals, while the periods from 2013-2021 are four-year intervals. This is because the last year with a global poverty estimate is 2017.


Nishant Yonzan

Economist, Development Data Group, World Bank

Christoph Lakner

Program Manager, Development Data Group, World Bank

Daniel Gerszon Mahler

Senior Economist, Development Data Group, World Bank

Join the Conversation

The content of this field is kept private and will not be shown publicly
Remaining characters: 1000