Statistics and the pandemic

CAMBRIDGE — “There are three kinds of lies,” Mark Twain famously wrote. “Lies, damned lies and statistics.” Too often, the COVID-19 crisis has lent support to the suspicions Twain’s bon mot expresses.

Data are critical to fighting the pandemic, but cross-country comparisons have focused too much on the wrong sort. This has given some political leaders a strong incentive to downplay the pandemic, arguably contributing to millions of deaths.

Given widespread mistrust of experts and mainstream media, it is important to emphasise the most informative and reliable COVID-19 data possible. Most attention has centred on countries’ officially recorded infections and deaths. The first statistical priority here is pretty obvious: Divide cases and deaths by population size in order to express them in per capita terms.

But even when reported on a per capita basis, official totals usually grossly underestimate the true number of infections and deaths. Coronavirus carriers count as infected only if they test positive or are hospitalised, while the death toll comprises only those with COVID-19 listed on their death certificate.

Estimates of the extent of undercounting of infections and deaths vary. The Economist’s excess-mortality model calculates that COVID-19 has so far killed 7-13 million people globally, about three times the current official death toll of 3.5 million. The undercount tends to be more extreme in lower-income countries. Egypt, for example, has 13 times as many excess deaths as the number officially attributed to COVID-19.

For richer OECD countries, the overall undercount is smaller, at an estimated 17 per cent. In the United States, the estimated discrepancy was running at only 7 per cent in March and April of this year. In July 2020, by contrast, professional estimates of the true number of US infections were twice to six times higher than the official count.

Focusing on official reports of COVID-19 cases and attributed deaths does not merely understate the gravity of the crisis. It can lead politicians to reduce testing and suppress bad news, for example, by encouraging doctors to list co-morbidities (such as diabetes) or complications (such as pneumonia) as the cause of death, rather than COVID-19.

The media and the public should instead pay more attention to other indicators that are more informative and less vulnerable to political influence. For starters, the rate of positive coronavirus tests as a percentage of all tests administered is much more instructive than the rate relative to a country’s population. India’s current positivity rate is an alarming one in four. If Prime Minister Narendra Modi had focused on this indicator a few months ago, perhaps he would have known not to declare victory against the virus prematurely. Instead, he facilitated super-spreading mass political rallies and religious celebrations.

In contrast, the US positivity rate has recently fallen below 3 per cent. This presumably reflects a vaccination rate that is higher than in the European Union and most other countries.

Paying more attention to the positivity rate would have encouraged governments to increase testing. Instead, the focus on verified infections as a percentage of the population gave policymakers a dangerous incentive to reduce testing. “If we didn’t do any testing, we would have very few cases,” former US president Donald Trump said in May 2020, falsely claiming that the US was testing more than most other countries.

Trump was not alone. Brazilian President Jair Bolsonaro limited testing for the sake of appearances and actually cut back when COVID-19 cases accelerated sharply. Other leaders who have sought to downplay the seriousness of the virus include Mexican President Andrés Manuel López Obrador, Russian President Vladimir Putin, UK Prime Minister Boris Johnson, and now-deceased Tanzanian president John Magufuli.

To recognise the degree of undercounting is to understand that the pandemic is even worse than we thought. But applying a historical perspective offers at least some encouragement — provided we handle the statistics properly.

For example, a recent New York Times article claimed that the 2020 jump in the US death rate was not only the worst in decades, but also supposedly surpassed even that caused in 1918 by the global influenza pandemic. It would seem surprising if, with all the advantages of modern medical science and greater economic resources, Americans had suffered more in 2020 even than they did in 1918. In fact, they didn’t. Yes, the federal and state governments could have done far better in combating COVID-19. But the outcome was not worse than the so-called Spanish flu. The Times story was misleading in its description of the comparison of excess deaths between 1918 and 2020.

According to the US Centres for Disease Control and Prevention, from 2019 to 2020, the estimated age-adjusted US death rate increased from 715.2 to 828.7 deaths per 100,000 population — an increase in mortality of 0.114 percentage points. This was a tragically large rise that reflected the pandemic. But is it useful to describe it as a 16 per cent increase in the death rate?

In comparison, the age-adjusted death rate rose from 2,276 per 100,000 in 1917 to 2,542 per 100,000 in 1918. That represents an increase of 0.266 percentage points, more than twice the magnitude of that in the 2020 crisis. It is misleading to talk of a 12 per cent jump in mortality in 1918 and interpret it as less severe.

The point is that the 2020 increase in the death rate looks worse than the 1918 spike only when compared to other causes of death, which have declined by much more over the last century. What is most striking in a graph of the US death rate over time — even more than the exceptional leaps of 1918 and 2020 — is the otherwise gradual but strong downward trend from 1910 to the present. Even the awful reversal of 2020, from 0.7 per cent to 0.8 per cent, still left the death rate below what it had been as recently as 2000.

Steven Pinker’s 2018 book “Enlightenment Now” is a good antidote to the impression we all get that the world is going to hell. Data on health, life expectancy, literacy, personal security, and poverty reduction all show favourable long-term trends. Such historical progress is not automatic but — as Pinker argues — reflects faith in “reason, science, and humanism”. Holding leaders accountable also is essential. For that, as the pandemic has shown, good statistics are crucial.

Jeffrey Frankel, professor of Capital Formation and Growth at Harvard University, previously served as a member of president Bill Clinton’s Council of Economic Advisers. He is a research associate at the US National Bureau of Economic Research. ©Project Syndicate, 2021. 

www.project-syndicate.org