As America remained effectively shut down for the last week, federal, state, and local officials continued to tighten the vice on workers and businesses almost daily.
Texas Gov. Greg Abbott said Sunday he would “applaud” local officials who adopt “shelter-in-place” orders akin to those California and New York have adopted to combat the coronavirus threat. Indications are that this lockdown could continue for as long as three months, shuttering much of the American economy.
The economic and human costs of job losses and bankruptcies caused by this policy will exceed what most Americans can imagine. Data points to an apocalyptic surge in initial unemployment claims from 281,000 to 2.25 million in only a week, more than three times the previous one-week record of 695,000 (1982). The president of the St. Louis Federal Reserve bank predicts 30 percent unemployment. Goldman Sachs is projecting GDP loss for the second quarter to be 24 percent. To put those numbers in perspective, during the Great Depression, the highest unemployment rate was 24.9 percent, and the largest annual GDP loss was 12.9 percent.
The potential human cost is immense. Tens of millions of Americans who were small business owners, college graduates working their first job to pay off student loan debt, or just trying to make ends meet will struggle to keep their home or find their next meal. Government relief efforts will cause inflation that will eviscerate retirement savings. Millions of parents will be deprived of the ability to care for and feed their children.
Even if the most pessimistic forecasts of the virus’ effects are correct, this is not an acceptable or sustainable solution. It is akin to burning your house down to fight a termite infestation. What’s worse is that policymakers do not even have the data necessary to make an informed decision.
The justification for an extended shutdown is predicated on statistical modeling of scenarios predicting the effects of the virus and its spread. But models are only as good as the data and assumptions upon which they are based. The current death and hospitalization rates are based upon identified cases. The numerator in these formulas is apparent: patients who are experiencing serious symptoms present for medical care and those who die can be counted. But mild or asymptomatic cases will likely go uncounted, and we simply do not know how many of these exist.
John P.A. Ioannidis, a professor of medicine, epidemiology, population health, biomedical data science, and statistics at Stanford University called the U.S. government response “a fiasco in the making,” stating that “[t]he data collected so far on how many people are infected and how the epidemic is evolving are utterly unreliable.”
Ioannidis pointed specifically to the selection bias in over-counting severe cases and the lack of a randomly tested sample. He analyzed data from the only situation where an entire, closed population was tested—the Diamond Princess cruise ship and its quarantine passengers—to more accurately estimate the virus threat.
The Diamond Princess had a case fatality rate (CFR) of 1 percent. But this was a largely elderly population, in which the death rate from the coronavirus is much higher. Adjusting the Diamond Princess mortality rate to the age structure of the U.S. population, and adjusting further to account for the risk that some passengers could die later and that tourists may have different frequencies of chronic diseases than the general population, he estimated CFR in the general U.S. population to be between 0.05 and 1 percent—a best-case scenario comparable to the seasonal flu with a worst-case scenario well below the current 1.5 percent U.S. CFR including only confirmed positives as the denominator.
Our pandemic data modeling has been wrong before. In 2008-10, H1N1 swept through North America and was declared a pandemic by the World Health Organization (WHO). Early data indicated it could be a threat as severe as coronavirus.
A July 9, 2009, WHO report shows a worst-case estimated CFR of 1.5 percent, which is exactly the U.S. CFR as of March 19 for the Wuhan coronavirus. Other studies estimated nearly 1 percent CFR for H1N1. Hospitals were reported as being “pushed to their limits,” and healthy adults in their 30s and 40s were spending months in ICU. Models predicted the healthcare system could be overwhelmed. A .02 percent CFR for H1N1 wasn’t determined until years later when serological studies sampling the proportion of the population that developed antibodies to the virus were conducted and found that many mild and asymptomatic cases were not being counted. It would be negligent not to at least consider the possibility we are making the same mistake again.
Policymakers can address the coronavirus threat in a manner that better takes into account the risks posed by the virus and the costs and benefits of the proposed solutions.
First, we should dedicate resources to obtain random sample testing from affected areas and work toward developing a serological test to better assess virus risk and the propriety of mitigation measures.
Second, we can emulate steps taken in South Korea proven to suppress the spread of the virus while keeping the economy moving: ensure widespread testing, focus quarantines on the sick and at-risk populations, and establish rules for social behavior.
Third, we should ensure an adequate supply and remove legal barriers to the use of hydroxychloroquine, which looks to be a promising treatment to lessen the spread, duration, and severity of illness.
Our government need not create one disaster to address another. With a sound policy going forward, we can overcome the coronavirus threat without engaging in economic self-immolation.
This is a commentary submitted and published with the author’s permission. If you wish to submit a commentary to Texas Scorecard, please submit your article to firstname.lastname@example.org.