ambulance
There are two 'official' death tolls on the Government's COVID-19 dashboard. 138,852 is the number of deaths within 28 days of a positive test. 162,620 is the number of deaths with COVID-19 on the death certificate.

The main reason the latter is larger than the former is lack of testing during the first wave. In the spring of last year, about 15,000 people in whose death COVID-19 was a contributing factor died without being tested.

So is 162,620 the pandemic's true death toll? No. And that's because it includes a large number of deaths that probably would have happened anyway.

How do we know this? Because if we calculate the excess deaths - the number of deaths in excess of what we'd expect based on previous years - we get a much lower number.

The official death toll for England and Wales, based on death certificates, is 147,031. Yet if we add up all the deaths since the start of March 2020, and subtract the average over the last five years, we get a figure of 117,476 (about 20% lower).

What's more, due to population ageing, the average over the last five years understates the expected number of deaths. Hence the true number of excess deaths is about 15% lower. Taking this into account, the pandemic's total death toll in England and Wales is about 100,000.

However, when it comes to events like pandemics, estimating the total death toll isn't the best way to gauge the impact on mortality. Consider an example.

Japan and Mexico have about the same population, but there are more deaths each year in Japan. How can this be, when everyone knows Japan is a very long-lived country? The reason is simple: there are more elderly people in Japan, so there are more people at high-risk of dying each year.

A better way of comparing the level of mortality in Japan and Mexico is to use the age-standardised mortality rate or life expectancy. Both of these measures take into account the risk of dying at different ages, as well as the age-structure of the population. (In 2019, Japan's life expectancy was 84, whereas Mexico's was only 76.)

Last year, the U.K.'s age-standardised mortality rate rose by 12.8%. Although this is the largest one-year change since 1940 (the first year of the Blitz), the level to which mortality rose was lower than in 2008. And even the change should be put into context: 2019 was a year of unusually low mortality.

I previously estimated that the life expectancy in England and Wales last year was 80.4 - down from 81.8 in 2019. (Other researchers have reported similar figures.) So despite tens of thousands of excess deaths, life expectancy was still around 80.

The reason life expectancy didn't fall further is that the vast majority of excess deaths were to people in their 70s and 80s. If there had been 100,000 excess deaths of people in their 20s and 30s, the drop in life expectancy would have been far greater. (In 1918, Spanish life expectancy fell by a staggering 12 years.)

Using data on life expectancy and population from the World Bank, we can calculate the percentage of the world's population that lives in countries with a life expectancy lower than 80.4. For 2019, it comes out as 91%. This means that, in 2019, nine out of ten people lived in countries with a higher level of mortality than Britain experienced last year.

In many countries, of course, life expectancy is pulled down by the high level of infant mortality. Yet there are still large cross-country differences in life expectancy at age 10. For example, it's 46 in Lesotho, compared to 72 in the U.K.

All this means that the level of mortality Britain experienced last year was neither exceptionally high by historical standards, nor by international standards.

As the ONS noted in a report last July: "The highest mortality rate observed during a 'normal' winter in Bulgaria has historically been greater than the highest mortality rate observed during the 'abnormal' coronavirus pandemic in England."

For the last few paragraphs, I've been talking about 2020. But wasn't the second wave, which peaked in January of 2021, even more deadly than the first? In short, no.

Although the number of deaths with COVID-19 on the death certificate was higher in the second wave, the age-adjusted excess mortality was lower. In fact, the overall level of mortality in the first eight months of 2021 was lower than in 2018.

The reason is that many of those who died with COVID-19 on the death certificate in the winter of 2021 would probably have died anyway. In addition: some deaths that would have occurred a few months later were 'brought forward' by the pandemic.

Overall then, the pandemic was a major event, resulting in the largest one-year drop in life expectancy since the Second World War. On the other hand, mortality only rose to the level of 2008, and was still lower than the normal level of mortality in most countries around the world.