COVID-19 vaccination coverage varies starkly across countries. Figure 1 depicts the percentage of people fully vaccinated against COVID-19 by countries’ income group: For each income group and date, I divided the cumulative number of fully vaccinated individuals by the total population of that income group. Individuals are considered fully vaccinated if they completed the initial vaccination protocol (three doses in three-dose vaccines, two doses in two-dose vaccines, and one dose in one-dose vaccines). As of July 2023, 75% of people in high-income countries and 79% of people in upper-middle-income countries have been fully vaccinated. In lower-middle-income countries, the percentage of fully vaccinated individuals is around 59%. In contrast, in low-income countries, only 27% of the population has been fully vaccinated. The time profile of vaccinations also differs markedly across countries, with higher-income countries starting the vaccination process earlier than lower-income countries. Note that Figure 1 is based on countries’ official reports of vaccinations. The discontinuities in the time patterns of vaccination reflect discontinuities in the timing of reports.

Percentage of fully vaccinated individuals by income group. Source: Our World in Data [20]. The classification of countries by income is based on the 2021 World Bank definition of income groups

In 2021, the World Health Organization recommended that countries should have at least 70% of their population fully vaccinated by the end of June 2022 [26]. This target was set to guarantee global COVID-19 protection as 70% was considered a good estimate of the herd immunity threshold (i.e., the minimum proportion of COVID-19 vaccinations that would allow interruption of the chain of transmissions). As Figure 1 shows, only high-income and upper-middle-income countries achieved that goal. Even though COVID-19 herd immunity is now considered an elusive goal due to the continuous emergence of new variants that escape infection-acquired and vaccination-acquired immunity [27], the comparison with the 70% vaccination coverage level is still informative about the large inequality in vaccination coverage across countries.

Figure 1 looks at full vaccinations. Instead, Figure 2 depicts the number of COVID-19 vaccine boosters administered per 100 people across income groups. Note that the number of individuals who have received at least one booster is lower than the number of boosters administered since some people might have received more than one booster dose. In addition, the decision to administer booster doses, the timing of administration, and the eligible population differed across countries. Notwithstanding all these facts, the graph shows that the administration of booster doses is highly regressive. In low-income countries, the number of administered boosters covers less than 4% of the population, while in high-income countries, the number of administered boosters could cover 66% of the population. This comparison highlights an additional element of inequality in the global allocation of COVID-19 vaccines, as one could easily argue that some of the booster doses could have been used to increase primary vaccination rates in lower-income countries.

COVID-19 vaccine boosters administered per 100 people by country income group. Source: Our World in Data [20]

The previous figures speak about the inequality in global COVID-19 vaccination rates. To what extent is the unequal distribution of COVID-19 vaccines also inequitable? To answer this question, I take an ex-ante perspective and reflect on what would have been an equitable global allocation of COVID-19 vaccines in the initial phases of the vaccination campaign given the information available at that time concerning COVID-19 epidemiology and anticipated effectiveness of COVID-19 vaccines. In other words, suppose we are back at the beginning of 2021, and we have to decide how to globally allocate the first batches of COVID-19 vaccines. Which factors should we consider to determine an equitable global outcome?

The United Nations defines global vaccine equity as the situation where vaccines are “allocated across countries based on needs and regardless of their economic status” [28]. The crucial issue here is how to characterize “needs” for COVID-19 vaccines. I am going to take for granted that the populations with the largest need are those who are expected to suffer the most from COVID-19 if not vaccinated. However, harm can be measured in multiple ways. Direct harms of COVID-19 include deaths, disabilities, and hospitalizations caused by COVID-19 infections. Thus, countries with the largest need may be those that anticipate the largest health burden if their population is not vaccinated. But COVID-19 causes also indirect harms, including additional deaths and disabilities due to strains in the healthcare system, unemployment, business closures, educational gaps, and the risk of falling into poverty. Therefore, countries with the largest need may be those that anticipate the largest socioeconomic burden if their population is not vaccinated.

At least four additional factors complicate the definition and measurement of “needs” for COVID-19 vaccines. First, the lack of COVID-19 vaccination creates both health and non-health harms. This requires a methodology for weighing the relative importance of qualitatively different harms [29]. For instance, since the beginning of the pandemic, it was clear that mortality and severe health consequences were positively associated with age [30]. On the other hand, the socioeconomic burden of the pandemic fell largely on children and working-age adults from low socioeconomic backgrounds due to disruptions in the education system and labor markets [31,32,33]. In addition, less resilient economies are expected to suffer more from the pandemic (and less resilient economies typically coincide with younger and poorer populations) [34]. Should the distribution of COVID-19 vaccines aim at reducing the health burden first and foremost, or should it aim at saving livelihoods and reducing the economic burden? The former would justify sending the vaccines first to countries with an older population, while the latter would justify sending the vaccines first to countries that face the largest risk of poverty due to COVID-19. Note also that income reductions tend to have detrimental effects on population health (e.g., individuals forego prevention activities and treatments because of liquidity issues, thereby putting them at higher risk of morbidity and mortality in the future). Thus, there is not only a trade-off between saving lives and saving livelihoods but also between saving lives now and saving lives in the future.

Second, and related to the previous issue, overall harms can be reduced by increasing the resilience of healthcare systems and welfare programs. For example, a country with a large share of older adults but a good healthcare system may face overall lower mortality rates than a country with a smaller share of older adults but a less resilient healthcare system. Similarly, the coverage and generosity of welfare programs (e.g., unemployment subsidies) can substantially reduce the socioeconomic costs of the pandemic, thereby weakening the economic argument in support of vaccination prioritization. This calls for skewing the allocation of vaccines towards low-income countries, as they are less likely to be able to cope with the negative effects of a pandemic.

Third, the optimal allocation of COVID-19 vaccines is complicated by the potential effectiveness of the vaccine in reducing transmission risks. Given the network of relations and contacts among countries and populations, protecting the needs of the most vulnerable may call for vaccinating first less vulnerable, but highly connected populations. For example, debates about the optimal allocation of COVID-19 vaccines within a country were shaped by the trade-off between vaccinating the older people first (high risk, low contacts) or the working-age population first (lower risk, high contacts) [35]. The same argument can be extended at the global level: Should vaccines be allocated to countries that are facing the largest harm (however defined) or to countries that are more likely to transmit the virus to the rest of the world? Even if one believes that poorer countries have less need for COVID-19 vaccines due to their demographics (i.e., low share of older people), distributing vaccines in low-income countries could in principle prevent the surge of new variants and new waves of infections, whose negative impacts will likely be transmitted to wealthier countries.

Finally, countries’ relative “need” for COVID-19 vaccines is not a static concept, but it changes over time depending on the proportion of people already infected or vaccinated (i.e., the proportion of people that are presumed to be protected at least in the short-term), on the characteristics of the dominant virus variant (e.g., its transmissibility and lethality), on the expected future trends of virus transmission channels, and on the capacity of a country to sustain new pandemic waves from both socioeconomic and healthcare perspectives (e.g., the experienced degree of disruption of the healthcare system and the state of public finances after prolonged public economic support).

Determining the optimal and equitable global allocation of COVID-19 vaccines is beyond the scope of the paper. A few contributions have considered this issue and proposed frameworks to guide the distribution of vaccines and to judge its reliance on principles of fairness. A noteworthy proposal is the “fair priority model” that suggested replacing the proportional allocation (by population size) from COVAX with an allocation based on the urgency of needs [36,37,38]. The fair priority model envisions three phases of vaccine allocation, with the goals, respectively, of reducing premature deaths, reducing serious economic and social deprivation, and reducing community transmission.

Among the factors to consider in determining the equitable global allocation of COVID-19 vaccines, I listed also the potential effectiveness of vaccines in reducing transmission. The importance of the transmission-reducing goal in COVID-19 vaccine allocation is nowadays an open question. Although existing COVID-19 vaccines were found to be partially effective at preventing transmission of the initial virus strains [39], vaccine-associated reductions in transmission of the new variants are considerably lower and rapidly wane over time [40], leading to larger numbers of breakthrough infections. On the other hand, even if not perfectly shielded from infectiousness, vaccinated individuals appear to be less infectious than unvaccinated individuals and present reduced and faster-disappearing infectious viral load [41,42,43]. All this considered, should differences in transmission risk across countries matter in determining the largest need for COVID-19 vaccines? I argue that a retrospective evaluation of the inequity in global COVID-19 vaccinations should reflect the information available in the initial phases of vaccine distribution when the potential effects of COVID-19 vaccines on reducing transmission were deemed to be positive. Instead, prospective evaluations (i.e., how COVID-19 vaccines or other pandemic vaccines should be distributed in the future) should account for the most recent information on the characteristics of vaccines.

Although inequality is not necessarily synonymous with inequity, the stark differences in vaccination rates across country income groups, the large health and economic toll suffered by unvaccinated populations, and the current lack of constraints in global vaccine supply all indicate that the unequal distribution of COVID-19 vaccines and vaccination rates has indeed been inequitable and suboptimal. To reinforce this argument, Figure 3 shows the number of fully vaccinated individuals as a percentage of the population aged 65 and over by country income group. The share of older people is an imperfect metric of the population at highest risk of COVID-19 hospitalization and death. Mortality from COVID-19 is correlated not only with age, but also with the presence of comorbidities, pollution levels, and living arrangements, among other factors [45,46,47]. Due to the nature of their job, frontline healthcare workers and other essential workers were also more likely to get infected and suffer severe health consequences independently of their age [48]. In addition, as previously discussed, COVID-19 mortality and morbidity encompass only a fraction of the possible harms caused by the pandemic.

Number of fully COVID-19 vaccinated individuals as a percentage of people aged 65 and over. Source: Numbers of fully vaccinated individuals are from our World in Data [20]. The total population aged 65 and over is from United Nations World Population Prospects 2022 (2021 data) [44]

The figure shows that, by the end of April 2021, high-income countries had distributed enough vaccines to cover the population aged 65+ (i.e., presumably the population at highest risk). In contrast, low-income countries reached that goal only 8 months later, and middle-income countries about 3 months later. Considering the smaller proportion of older people in lower-income countries (3.1% in low-income countries and 6.0% in lower-middle-income countries compared to 18.9% in high-income countries and 11.8% in upper-middle-income countries), the difference in protection is remarkable. Of course, not all vaccine doses were administered preferentially to older people. Some of those initial vaccine doses were given to frontline healthcare workers, individuals with comorbidities, and other individuals who were considered essential or worthy of prioritization. Still, the graph shows that, for many months, the number of fully vaccinated people in lower-income countries was substantially lower than the number of people considered at most risk. Data from COVAX confirm that by August 2021, only 33 million COVID-19 vaccine doses were delivered to low-income countries compared to 1.6 billion in high-income countries [22], thereby suggesting that the slow vaccine uptake in low-income countries in the first half of 2021 was foremost due to supply issues.