He is also the associate producer of the nationally-syndicated “Stacy on the Right” talk radio show. In addition to writing that has been featured in The Washington Times, Townhall.com and by FreedomWorks, Demetrius is the author of the book Preservation and Purpose: The Making of a Young Millennial and a Manifesto for Faith, Family and Politics.
I talked about the Corona virus a few weeks back, and I mentioned that it could turn out to be nothing significant, or a real problem, or a global pandemic. Clearly one of those three options is no longer on the table.
I’m not a medic, and I’m certainly not an expert on infectious diseases, so I don’t want to comment on something I don’t know much about, you can get that from any barstool bore. But I do know a bit about statistics, and I think that I’ve spotted something that isn’t being reported, or at least not reported very widely.
After China, two of the worst-hit countries are South Korea, and Italy. That’s handy because these countries are quite alike in economic and population terms, and that allows us to make comparisons, but they are very different in social terms. I know that Iran is reporting a similar number of cases, but Iran is a closed country with a very different economy, and it’s hard to get reliable information so I’m not including it here.
South Korea and Italy have a lot of similarities, they have roughly the same population, 50 million and 60 million, they are both wealthy developed countries, both in the OECD, with advanced economies. And they’re both reporting high levels of Corona virus – about 15 per hundred thousand in South Korea, about 30 per hundred thousand in Italy.
But there is one statistic where they diverge sharply. The death rate. In Italy, the death rate currently stands well over seven per cent. In South Korea it’s 0.9 per cent. Get that, Italy is reporting a death rate from the disease, the same disease, that is eight times higher than in South Korea.
To understand what’s going on, I want to go to a book called Risk by Professor John Adams. Checking it on Google here, I see that Adams has put the entire book online for free, which is nice of him, I’ll link to that on the website. The book is about risk, and he tries to do things like understand the risks of driving, for which he needed an accurate measurement of road crashes. Where do you get that? Police reports?
Adams analyzed road crash data in the UK – he’s British – and he found interesting anomalies. For example, the death rate for reported road crashes is much higher in rural Scotland than it is in the centre of London. In fact he could demonstrate a dose-response effect where the further you were from built-up areas, the more likely a crash was to be fatal. Why? Well, could be that when you are far from the rescue services, you are more likely to die of your injuries? Or that people drive differently in isolated rural areas, causing more fatalities?
That’s possible, but that didn’t explain a different phenomenon. On days that there was a terrorist attack in London, the total number of road crashes reported in the city dropped sharply. What could cause that? People being extra careful? People staying at home on days that a terrorist attack was in the news? Maybe, but probably not, because although the rate of crashes apparently dropped, the rate of fatal road crashes didn’t change at all.
Adams goes through the figures in much more detail, and his conclusion is that the number of crashes doesn’t change at all, what changes is the reporting rate. On days that police are called away from their stations in great numbers, and the ones that remain are incredibly busy, people don’t bother them by reporting minor road crashes. But the more serious the crash, the more likely it is to get reported anyway, and of course if someone dies, then it is almost certain to be reported.
Similarly, if you live in a remote area of rural Scotland, you could be hours away from the nearest police station, and therefore you are less likely to report minor fender-benders, but the more serious the crash, the more likely you will be to make the effort. So the apparent effect of crashes being more likely to be fatal in Scotland isn’t true at all, it’s an artifact of a totally different statistical effect.
I think we could be seeing the same effect in Italy now. Remember, South Korea and Italy, similar countries, similar reported rates of Corona virus, but Italy has eight times the death rate.
Italy is a country that has a serious problem with the efficacy of public institutions. I know, I’ve lived there. There are the things that are clichéd about Italy, but those clichés have a strong basis in reality. Seat-belt laws and motorcycle helmet laws are regularly flouted. There’s a constant battle between the tax authorities and the public over tax compliance.
Even a passing experience with Italian bureaucracy will convince you how the entire system can be set up for the convenience of the people who work there – or, let’s say, are employed there – rather than the service users.
And the culture in Italy is very different to South Korea. In brief, conforming with the rules is highly prized in Korean culture; not conforming with the rules is prized in Italian culture.
For that reason, I’m not in the least convinced that the corona virus death rate in Italy is eight times higher than in South Korea. I think that it’s much more likely that the corona virus rate in Italy is eight times higher than is being reported.
That has very serious implications for the containment of this dangerous disease. But the whole of Italy is on lockdown now; they are at last taking it seriously, now that far more than a thousand people are dead.
But how can this information help elsewhere? There is one terrifying statistic about this disease. The USA at the moment has an incredibly low rate of reported infections, as I write on Saturday 14 March; it’s only 0.6 infections per hundred thousand. That’s hardly one-fiftieth the rate in Italy. But hold on, that’s not the terrifying statistic.
Remember South Korea? They are testing 10,000 people for the virus every day. 10,000 a day. That’s roughly the number that the United States has tested in total. And that’s not the terrifying statistic either. If you look at Washington State, the epicenter of the outbreak in the US, they are reporting a death rate of 6.5 per cent, almost the same as Italy. That’s Italy, where I think there is good reason to suspect that the reporting of infections is drastically underestimated.
That could mean that there are thousands of infected people in Washington State – and other states – a large portion of them who are un- or underinsured and afraid of the cost of going to the doctor, who are showing little or no symptoms, and afraid to miss a shift at a job they need to put food on the table. That 6.5 per cent death rate, that’s a canary in the coalmine.
One last point – Iceland is a tiny country, a third of a million people, but it is very rich, it is a highly cohesive society, people have a very strong allegiance to authority, and in turn the authorities are highly professional. Iceland has a free, high-quality healthcare system that covers all residents, and Iceland has the highest infection rate in the world. Or, the highest reported infection rate in the world, more almost 40 per hundred thousand, or more than 60 times higher the reported rate in the US.
Iceland hasn’t had any deaths; it is too small for that to be statistically significant. Its people are rich, and travel quite a bit, but it’s an island in the middle of the Atlantic, it isn’t anything like as connected as Europe where people regularly get in their car and drive across borders to do their weekly shopping.
I think that when the history of this is written, there’ll be a big fat chapter on the quality of the statistical reporting.