Since the outbreak of the Covid-19 pandemic I have closely been monitoring the data for the number of deaths and infected across different countries and I have spend considerable time trying to estimate statistical models to explain variations in deaths and infects across different countries.

It quickly became clear to me that relative few factors could explain this variation and back in April I wrote a blog post in which I claimed that ONE factor could explain most of the variation in Covid-19 deaths across countries.

That factor was age or rather the number men older than 80 years as share of the male population.

There really wasn’t anything overly surprising about that as it fast became clear that very few young people or children died from Covid-19 and the average age of the Covid-19 victims was around 80 years old in most countries.

However, I was really never satisfied with this explanation. There had to be more to the story.

The next factor I looked at – but never published my results of – was obesity. Here the results also were pretty clear. The share of the population who are obese seems to be a fairly strong indicator of Covid-19 mortality. This is also confirmed from numerous ‘micro’ studies looking at individual hospitals or cities (see here for an discussion).

But that still isn’t enough to explain the Covid-19 ‘mystery’ as we can also note what we in economics would call ‘stylized facts’ about Covid-19:

• There seems to be a strong – flu-like – seasonality in Covid-19 cases, which seems to be linked to latitude (see here and here).
• There is a significant over-representation of Covid-19 cases and deaths among blacks in the US (see here) and the UK (see here) and among black immigrants (mostly Somalis) in for example Sweden and Denmark (see here), while blacks in the Southern Hemisphere does not seem to be overly hard hit by the Covid-19 pandemic.
• A majority Covid-19 deaths in the developed countries seem to have happened in nursing homes (see here).

I have no medical training and certainly was not familiar with the academic literature on the importance of vitamin D deficiency in general health, but I have been catching up fast and it is now pretty clear to me that if we look at all of these groups – the elderly, particularly those in nursing home who spend a lot of time indoors, blacks living in Northern hemisphere and the obese they all tend to suffer from significant problems with vitamin D deficiency. The same of course is the case with the seasonality – and there is a well-established relationship between the seasonality in flu and seasonal variation in vitamin D (see for example here).

It is for example a well-established fact that particularly the elderly in Spain and Italy suffer from vitamin D deficiency (see here) and numerous studies have shown that African immigrants in Scandinavia also suffer from vitamin D deficiency (see here).

So the only logical thing would be to look at variation in Covid-19 deaths across countries and try to explain that with the share of the population who suffers from vitamin D deficiency.

However, this is where we run into problems – it is very hard to come across comparable data on this across countries. There is some, but there is still too much problems with the data to do a proper statistical study of enough countries (see here for an example nonetheless).

So instead I decided to look at something else – the variation in Covid-19 deaths across US states and the share of the population in each who are African-American.

The reason for this is that numerous studies have shown that as many as 80% of all African-Americans suffer from vitamin D deficiency (see here) so if vitamin D deficiency really is a key explanatory variable in terms of explaining Covid-19 mortality then we should expect that Covid-19 mortality rates should be higher in US states with a larger share of population who are black.

You can judge for yourself by looking at the graph below.

As we see there is a very strong correlation between the share of the African-American population and Covid-19 mortality rates across US states.

In fact it is by far the strongest statistical relationship I have been able to find among all the variables I have been looking at (including age, obesity, population density and longitude).

Obviously this relationship can be due to a number of factors – among them numerous socio-economic factors – but to me at least this is further indication that vitamin D deficiency is an extremely important variable in understanding variation in Covid-19 mortality across different groups of people and across countries.

And this leads me to the conclusion that maybe we should do less testing for Covid-19 and more testing for vitamin D deficiency and a significant part of the public health response to the Covid-19 pandemic should be to focus on protecting groups with vitamin D deficiency and treating it.