Herd mentality: don’t be complacent about immunity

The media has seized on new research from the University of Oxford that includes the startling possibility that the majority of the UK’s population has already been infected by the coronavirus.

In one sense, this is obviously alarming. But in another, it is reassuring. If so many people in the UK have already caught the disease, we may be closer than previously thought to achieving herd immunity and the outbreak may have peaked.

This would be good news for investors as well as the nation: it would imply we are at or close to the market’s nadir, recent stimulus-inspired rallies notwithstanding, and could therefore begin to reallocate to risk assets.

We would not be so confident. Looking at the Oxford study in more detail, we see that it has used an identical model to our own one – with one crucial exception, the mortality rate.

In the Oxford team’s base case, which leads them to the mass infection conclusion, the mortality rate is set to 0.014%. However, the overwhelming scientific consensus is that this rate is instead somewhere between 0.5% and 1%.

Put simply, if you start by assuming a very low mortality rate, the implied number of infections from the number of observed deaths must be very high. If one person has died from an infectious illness with a mortality rate of 1%, perhaps another 99 have caught it and lived. But if one person dies from a virus with a mortality rate of 0.1%, you could presume another 999 contracted it.

The academics say they chose 0.014% because it best fits the number of deaths in the UK and Italy during the first 15 days of the virus in each country. This is a problematic starting point for several reasons:

• A sample period of 15 days is far too small.

• Many of the early cases and deaths would have originated overseas, such as with those travelling back to Italy, which their model cannot capture.

• Calibrating an exponential model based on just the first few days of an outbreak leads to small mistakes growing exponentially too.

It is not clear how they chose their baseline, either: their research also shows that higher mortality rates – which imply no herd immunity – fit well with the observed death patterns too. There is no empirical evidence for starting with a mortality rate set at 0.014%.

Based on our own work on this topic – which is in line with most serious estimates of a fatality rate around 0.5% – and putting these more conservative numbers into the Oxford model, we believe we are only at the start of the pandemic, not near the end.

If this is the case, the economic damage is likely to be much more severe than we have seen so far and we may not have reached the bottom of the market yet.