This week the U.S. reached a grim milestone in the war against Covid-19, with more than 200,000 deaths officially confirmed by the Centers for Disease Control and Prevention (CDC) [1].
Fig. 1
The 200,000 Americans dying of Covid-19 has occurred in the same year as the 75th anniversary of another grim event in American history: the world’s first atomic bomb test in a remote New Mexico desert that evoked words from the Baghavad Gita “I am become Death, the destroyer of worlds.”
While no one believes Covid-19 will destroy the world, it certainly has caused far more carnage than many had predicted just three months ago. The aggregate social cost of America’s first 200,000 COVID-19 deaths lies between $197 billion and $1.01 trillion. What follows is how I arrived at these figures.
The Value of Averting An Average Covid-19 Death in the U.S.
In an earlier post, I explained the various methods that economists used to calculate the dollar value of a human life. The chart below is an updated version of my earlier chart, showing how the value of preventing a “typical” Covid-19 death varies (by a factor of more than six!) across these various methods [2].
Fig. 2
The averages shown here take into account the fact that roughly 80% of Covid-19 deaths are among seniors age 65 and older (with nearly one third among the “oldest old” persons age 85 and over).
This average value takes into account both productivity losses (lost earnings or the value attached to unpaid work such as child care or homemaking) as well as the intrinsic value of life itself. Behind these averages are age-specific values attached to different ages over the lifespan. Some methods assign the identical value to all years of life, whereas others take into account the reality that the value of earnings and consumption generally rise during adult years but decline in senior years.
The Social Losses from the First 200,000 Covid-19 Deaths in the U.S.
I have used the preceding figures to estimate the loss to the U.S., which ranges from a low of $197.0 billion using the most conservative cost/QALY (quality-adjusted life year) approach to a high of $1.01 trillion using the most generous age-adjusted VSLY (value of statistical life year) method.
Fig. 3
The highest figure exceeds $1 trillion. It is based on the same sort of willingness-to-pay figures that are used by various federal agencies such as EPA, Department of Transportation and Consumer Product Safety Commission. Such agencies typically use an average value of statistical life (VSL) figure of roughly $10 million in determining how much health and safety regulation is sensible from a cost-benefit point of view. But that estimate also is based on allowing the value of a statistical life year (VSLY) to vary by year of age, which not all agencies allow.
The BMGF (Bill and Melinda Gates Foundation) VSLY estimates are most analogous to the figures used by federal agencies, as they are based on an average VSL figure for the entire population of $10.5 million and rely on the identical VSLY figure for all years of life without any adjustment for quality (the latter adjustment would lower the value attached to each year of life for seniors since quality of life generally declines with age).
The lowest value is under $197 billion. It is based on “optimal” willingness-to-pay per QALY but readers can see at the bottom of the table that this value ($127,750) is less than half the figure that undergirds all the VSLY methods. Moreover, VSLY methods apply a monetary value to each full year of remaining life expectancy. In contrast cost/QALY methods first “shrink” each year by multiplying by an average quality-of-life value that ranges from 0.964 for children under age 5 to 0.801 for seniors age 85 and over. Finally, any future years of life expectancy are, in this most conservative approach, discounted 7% a year.
Implications
These figures may be of interest in their own right to give readers a better sense of the impact (so far) of COVID-19 upon Americans. But they may be useful for other purposes that other analysts might explore.
Evaluating the Benefits and Costs of Lockdowns. Tallying all the benefits and costs of various lockdowns in the U.S. is well beyond the scope of this modest blog post. But as one example, a recent study published in JAMA estimates that school closures this past spring “may have been associated with approximately 1.37 million fewer cases of Covid-19 over a 26-day period and 40,000 fewer deaths over a 16-day period.” Others can debate the merits of the JAMA study. My analysis above simply implies that the value to society of averting those deaths is somewhere in the range of $40 to $200 billion. It is an open question whether the costs of the lockdown exceeded these benefits.
To be more concrete, a study released in late March found "in our benchmark case, the extensive social distancing measures currently underway amount to spending an average of $5.62 million per life saved." My figures suggest that regardless of which approach is adopted to value COVID deaths averted, the extensive measures adopted early in the pandemic may have imposed costs larger than their benefits.
Restitution from China. Late last March, the U.S. Defense Intelligence Agency released an updated assessment of the origin of the novel coronavirus to reflect that it may have been accidentally released from an infectious diseases lab in Wuhan. Consequently, for months, some have discussed how to hold China accountable for Covid-19. This too is a very complicated question, not one that can be resolved here. But were restitution to be pursued by any means, then surely the value of lost lives would have to be part of the calculation. But here it is worth emphasizing what the mortality losses estimated above do not include:
· Non-Covid-19 Mortality Losses. The above figures only focus on deaths attributable to Covid-19. In late May, Scott Atlas—a senior fellow at Stanford University’s Hoover Institution (recently appointed to serve as an advisor on the White House Coronavirus Task Force)—and others calculated that each month of lockdown produced about 65,000 additional deaths [5]. In addition, delays in seeking care and foregone care result in the loss of 500,000 lost life-years for every month of lockdown [6]. If all of their estimates were monetized, it would add roughly $800 to $900 billion to the direct Covid-19 mortality totals per month of lockdown [7].
A lower-bound estimate is provided by economists James Broughel (Mercatus Center) and Kip Viscusi (Vanderbilt University) who more recently estimated that just the economic slowdown alone (apart from any fears patients may have had in seeking care due to concerns about infection) would conservatively result in 9,200 expected deaths for every $1 trillion loss of income. Using the $2.7 trillion net loss of income cited below implies a loss of at least 18,630 additional lives; this would add $214 billion in lockdown mortality losses through late September, exclusive of the 500,000 lost life-years estimated by Dr. Atlas [8].
· Morbidity Losses. Evidence from China suggests that nearly one-fifth of COVID-19 patients experience lung injury that made breathing difficult, while others experienced cardiomyopathy and catastrophic arrhythmias. U.S. studies are currently underway to better understand the long-term health effects of those who survive Covid-19, so it may be years before we actually know the full magnitude of such impacts in either health or dollar-equivalent terms. That said, economists Thomas Kniesner (Claremont Graduate School) and Ryan Sullivan (Naval Postgraduate School) have estimated using CDC projections through November that non-fatal Covid-19 losses amount to anywhere from $2.2 trillion to about $5.7 trillion.
· Economic Impact of Lockdown. University of Chicago economist Casey Mulligan—who continuously updates his estimate each day—has estimated the cumulative cost of the Covid-19 pandemic in the U.S. through September 23 at $2.7 trillion [8].
· Impact of School Learning Losses. Hoover Institute Eric Hanushek, the Paul and Jean Hanna Senior Fellow at the Hoover Institution of Stanford University, has estimated that the lifetime economic cost of learning losses associated school shutdowns in the U.S. alone already exceeds $14 trillion.
In short, mortality losses of $200 billion to $1 trillion evidently are only the tip of a sizable iceberg.
Note: the author is grateful for capable research assistance from Deanna Bucy-Anderson.
Update #1: September 29, 2020
Thanks to Tom Kniesner and Ryan Sullivan who pointed me to several papers I had not seen. I have updated the estimates for non-Covid-19 mortality losses and morbidity losses based on these studies.
As well, Lisa Robinson, Senior Research Scientist at the Center for Health Decision Science and Center for Risk Analysis, Harvard T.H. Chan School of Public Health and a co-author of the BMGF study I cited, pointed out that “The BMGF paper that you cite is focused on developing estimates for low- and middle-income countries (not the U.S.), and the VSLY is recommended solely for use in sensitivity analysis.” I probably should have explained more clearly in my original article or here that using the BMGF formula produces a VSL value ($10.5 million) nearly identical to the $10.63 central VSL threshold used by Department of Health and Human Services. So for all practical purposes, what I have labelled “BMGF VSLY” is equivalent to the VSL regulatory standard currently in use by DHHS. In mid-July, Ms. Robinson and Ryan Sullivan have co-authored their own exploration of how Covid-19 costs might vary by age.
Footnotes
[1] The CDC updated its official tally to 200,275 at 12:21 pm EDT Wednesday, September 23. The New York Times maintains its own database, reporting more than 200,000 deaths by Tuesday, September 22. The Wall Street Journal likewise reported hitting this threshold on Tuesday. This is a remarkable consensus given the fierce controversy since March over whether CDC’s official tally was under- or over-counting COVID-19 deaths.
[2] My figures have been updated in three important ways:
First, I am relying on a more recent breakdown of COVID-19 deaths by age (this alters the average value per COVid-19 patient death averted since all the figures shown in my chart are weighted averages based on the current distribution of deaths).
Second, in addition to an approach where the value of a statistical life year (VSLY) is identical across all years of life, I have added two other approaches that use age-adjusted VSLYs.
The Murphy-Topel approach is found here: Murphy, Kevin M., and Robert H. Topel. The Value of Health and Longevity. Journal of Political Economy 114, no. 5 (2006): 871–904.
The Aldy-Viscusi approach is here: Aldy, Joseph E., and W. Kip Viscusi. Adjusting the Value of a Statistical Life for Age and Cohort Effects. Review of Economics and Statistics 90, no. 3 (2008): 573–81.
Third, instead of using $100,000 per quality-adjusted life year (QALY), a somewhat arbitrary choice within the $50,000 to $200,000 range used by ICER in doing clinical cost-effectiveness studies, I used an optimal willingness-to-pay figure proposed by Phelps: Phelps, Charles E. A New Method to Determine the Optimal Willingness to Pay in Cost-Effectiveness Analysis. Value in Health 22, no. 7 (2019): 785–91.
This figure—$127,500—is nearly identical to the $125,000 per QALY spent by the U.S. Medicare program for dialysis on end-stage renal disease patients (all calculations here). Since these patients otherwise would have died but for dialysis, $125,000 can be viewed as the implicit willingness-to-pay per QALY under Medicare and lends further weight to use of the Phelps figure.
[3] I well recognize that a full-blown analysis would need to account for the medical costs averted due to fewer infections and well as assign a dollar value to morbidity losses that may have been prevented from school closure. But presumably mortality losses would be the dominant factor in any full-blown calculation.
[4] The full study: Thunström, Linda, Stephen C Newbold, David Finnoff, Madison Ashworth, and Jason F Shogren. “The Benefits and Costs of Flattening the Curve for COVID-19,” March 25, 2020.
[5] Prior research has shown that statistically, every $10 million to $24 million lost in U.S. incomes results in one additional death. The authors calculated that the lost economic output in the U.S. due to lockdowns was estimated to be 5 percent of GDP, or $1.1 trillion for every month of the economic shutdown. They divided this by $17 million to obtain 65,000. If we substitute the Broughel-Viscusi figure of one additional death per $108.5 million in lower income, this would lower monthly lockdown losses to 10,100 deaths per month.
[6] The authors took into account deaths from missed strokes due to fewer evaluations, reductions in cancer treatments (including missed chemotherapy treatments, but also delays in detecting new cancers due to fewer screenings), reductions in live-liver transplants, and delays in childhood vaccinations.
[7] If 65,000 non-Covid deaths are valued at a cost of $11.5 million apiece—the current VSL used by the Environmental Protection Agency (note 13)—this would add about $750 billion to the total. 500,000 life-years multiplied by a VSLY of $127,500 to $289,430 adds another $64 to $144 billion to this total.
[8] Mulligan estimates the total cost at $3.1 trillion, but 14 percent of this represents his own estimate of mortality losses (monetized at $4.3 million per VSL), so I have excluded the latter to avoid double-counting.
READ CHRIS’ BOOK, The American Health Economy Illustrated (AEI Press, 2012), available at Amazon and other major retailers or as a pdf at AEI.
Follow @ConoverChris on Twitter, and The Apothecary on Facebook. Or, sign up to receive a weekly e-mail digest of articles from The Apothecary.
