It would be fair to say that, over the last two years, Imperial College has not built a reputation for unimpeachable epidemiology. For those who've forgotten, the university is home to Professor Neil Ferguson - the scientist who, perhaps more than any other, convinced our leaders to impose lockdown.

The latest study to come out of Imperial College isn't the work of Neil Ferguson. But its conclusions have the just same ring of implausibility. Published in Lancet: Infectious Diseases, the study estimates that - from December 8th 2020 to December 8th 2021 - the vaccines averted 19 million deaths, including 500,000 in the U.K.

While both these figures are implausible, I want to focus on the latter - just because there are so many more unknowns when dealing with global-level Covid statistics.

The authors' conclusion is based on a mathematical model, which was applied to national data on excess deaths. I'm not going to pretend I've taken the time to understand the details of their model. But I don't have to. The 500,000 number is just too implausible, and I'll explain why.

Before doing so, how many actual excess deaths were there in the UK during the relevant time period? If we take the numbers for England and Wales, and ignore age-standardisation, there were about 60,000. Add 14% for the other UK nations, and we get to 68,000. So what the Imperial College researchers are saying is that, in the absence of vaccines, excess deaths would have been eight times higher.

There are several ways of showing this figure couldn't possibly be right, and the true number of deaths averted must be much lower.

First, how many excess deaths were there up to 8 December of 2020, when no vaccines were available? Using the same (highly simplified) methodology as above, there were about 76,000. The Imperial researchers want us to believe that, in the absence of vaccines, excess deaths would have been more than seven times higher in the second year of the pandemic, by which time a decent chunk of people had already caught Covid.

Second, how much excess mortality (in percentage terms) would the UK have had in 2021 if the Imperial researchers are correct? The average number of deaths from 2015 to 2019 was about 605,000. So cumulative excess mortality up to 8 December of 2021 would have been 106% (as opposed to just 24%). To see why this figure is implausible, let's compare Bulgaria and the UK.

Bulgaria has the lowest vaccination rate in Europe. By February of this year, only one third of over 60s had been double-vaccinated. By contrast, around 90% of British over 70s were already double-vaccinated in March of last year.

As of December 8th 2021, cumulative excess mortality in Bulgaria was 51%. (And at that time, the elderly vaccination rate was even lower.) Yet the Imperial estimates imply Britain would have had more than double that level of excess mortality in the absence of vaccination - despite being a much richer country, with higher quality healthcare.

So where did the researchers go wrong? As Bhaskaran Raman notes, they made numerous assumptions that exaggerate the effectiveness of the vaccines. Perhaps their most questionable assumption was that the mRNA vaccines are 90% effective against death - when real world evidence suggests their effectiveness is much lower.

The vaccines did save lives in Britain. But I'm not convinced the number is anywhere close to 500,000.