China’s Coronavirus Stats Can’t Be Trusted

By John Banks

The coronavirus data coming out of China is manipulated, and possibly outright false. We know this, in part, because a Chinese health official has admitted as much. There have also been questions and red flags around China’s data reporting from very early on.

This matters because China’s coronavirus statistics — the number of cases, the number of fatalities, and the slowing of new infections — are frequently used as a benchmark and a source of hope.

It has been noted that the United States now has more cases than China, for example. But what if China’s true numbers are as much as 40 times higher, as the British government reportedly believes?

China’s ability to “slow the spread,” and then stop it cold (apart from inbound travelers), is also in doubt if the numbers are bogus.

This is a real problem, in part, because the media tends to cling to any data port in a storm, and China’s reported numbers are a big part of the coronavirus story. But if you can’t trust the numbers, any conclusions based on a close read of the numbers could be wrong.

This is also bad news for the hardcore optimists — those who still want to believe “it’s not that bad” and show a bias for sunny data points to support such a view.

If China is faking the data, its real caseload and fatality count are likely to be far grimmer than previously known. That would make it harder for willful optimists to sustain their “it’s not that bad” stance (which should have been chucked a long time ago, in our humble opinion, but people are stubborn).

With regard to admitting fakery, an act of deception was clarified in a Wall Street Journal scoop titled, “China’s Coronavirus Count Excluded Infected People With No Symptoms.”

As the WSJ reports:

“China said more than 1,500 people who were infected with the coronavirus but haven’t shown symptoms weren’t included in its tally of confirmed cases.

…Chang Jile, a top Chinese health official, said the country would report the number of infected people who aren’t showing symptoms beginning Wednesday.”

Not reporting asymptomatic carriers is an act of deception, full stop.

It was known from very early on that a fair portion of infections resulted in mild to negligible symptoms. It has also been known for quite a while that the “super-spreader” infection rate is high, which has a natural connection to symptomless carriers.

China has presented this as a simple oversight. Our hunch is that the Chinese government is being caught out by reality on the ground. To the extent their previously reported case numbers were too small, dealing with ongoing outbreak issues is becoming a problem.

The prior deception may have been exposed — sort of, by a partial amount — in order to keep the fakery in line with some plausible semblance of reality.

Then, too, China’s reported coronavirus figures were setting off alarm bells six weeks ago or more.

In a mid-February piece, Barron’s highlighted the impossibly smooth nature of China’s coronavirus fatality stats.

With the proper training in statistics, it is possible to tell the difference between real data sets and artificial (fake) data sets. This is because humans are bad at faking randomness: When a human tries to artificially create a natural sequence, they subconsciously introduce a level of order and regularity that Mother Nature would leave out.

To put it another way, artificial data sets tend to be “smooth” while real data sets are “noisy.” If you conduct a regression analysis on a data set and it comes out smooth as glass, statistically speaking, you are probably dealing with fake data.

That is how China’s early fatality numbers felt — artificially smooth. Those with the tools to measure such things had a sense some bureaucrat was either massaging the numbers or making them up entirely.

This could be why the British government is “furious with China’s handling of the coronavirus,” according to Reuters and the U.K.’s Mail on Sunday, with scientists telling Prime Minister Boris Johnson China could be downplaying its coronavirus stats “by a factor of 15 to 40 times.”

Last but not least, there is the anecdotal evidence of urn deliveries in Wuhan.

Hedge funds have long been known for creative and aggressive data gathering techniques, like using satellite imagery to count cars in parking lots. If, say, a big box retailer consistently has a higher number of cars in the parking lot, that is a good sign for same-store sales, and so on.

With the coronavirus, similar techniques have been deployed, but in a much darker way.

Iran, for example, is another country where the true caseload and fatality rate is probably off by multiple orders of magnitude. In terms of anecdotal evidence, the Washington Post has reported on “coronavirus burial pits so vast they’re visible from space.” (Via WaPo, you can see the satellite images here.)

The Wuhan urn deliveries are a comparable piece of evidence — not hard data, but hard to ignore. As Bloomberg reported last week:

“The long lines and stacks of ash urns greeting family members of the dead at funeral homes in Wuhan are spurring questions about the true scale of coronavirus casualties at the epicenter of the outbreak, renewing pressure on a Chinese government struggling to control its containment narrative.

Outside one funeral home, trucks shipped in about 2,500 urns on both Wednesday and Thursday, according to Chinese media outlet Caixin. Another picture published by Caixin showed 3,500 urns stacked on the ground inside. It’s unclear how many of the urns had been filled.

People who answered the phone at six of the eight funeral homes in Wuhan said they either did not have data on how many urns were waiting to be collected, or were not authorized to disclose the numbers. Calls were not answered at the other two.”

China’s official coronavirus fatality rate is 3,300 as of March 31 — below the current mark in the United States, which is 3,400 and climbing.

But what is the real number? If the reported “15 to 40 times” estimate put forth by scientists advising the British government is something to go by, it could be as high as 130,000.

This puts us in very strange territory.

It is plausible, though not confirmable, that unreported coronavirus deaths out of China, Iran, and other misreporting or nonreporting countries — Russia’s numbers are likely bogus, too, for example — could exceed the “official” death toll by an order of magnitude. 

This matters because the human species is now engaged in World War V — where “V” is for virus — and fatality data massaged to look less bad, though it might make people feel better, is like failing to grasp the true measure of the enemy.

We need an accurate picture of COVID-19 and the damage it can do — and has already done — in order to accurately calibrate the response. And as we have noted before, this war against the virus is nowhere close to over — we are still in the early stages.