$16 Trillion and $14 Trillion Are Answers to Different Questions
Put the two estimates in one table and the "disagreement" disappears. Only one line is common to both, and on that line they differ for a reason that is dated, not disputed.
| What is being counted | Cutler & Summers, JAMA (Oct 2020) | USC Schaeffer (May 2023) |
|---|---|---|
| Lost economic output (GDP) | $7,592 billion — the Congressional Budget Office's 2020 projection of output lost "during the next decade" | $14 trillion — modelled cumulative GDP 2020–2023 of $103 trillion against $117 trillion without the pandemic, in 2020 dollars ("nearly 14% higher") |
| Premature deaths, valued | $4,375 billion — an expected 625,000 pandemic-associated deaths through 2021 at a "conservative" $7 million per statistical life | Not counted — "we didn't estimate … lost years of work after an early death" |
| Long-term health impairment (long COVID) | $2,572 billion — about a third of severe/critical survivors impaired, valued as a 35% loss of quality-adjusted life expectancy | Not counted — "or a severe case of long-COVID-19" |
| Mental-health impairment | $1,581 billion — 80 million additional adults with depression or anxiety symptoms for one year at about $20,000 each | Not counted — "the many ways that the disease has affected the physical and mental health of the U.S. population" |
| Direct health-care spending | Not separated out | $20 billion (best case, 65,000 deaths) to $365 billion (worst case, 2 million deaths), "driven mostly by hospitalization costs" — inside the $14 trillion, and the reason the authors say "most economic losses were not due to these health care expenditures" |
| Total | $16,121 billion — "approximately 90% of the annual gross domestic product"; $196,475 for a family of four | $14 trillion — "twice the size of the Great Recession", "20 times greater than … 9/11" |
| Kind of number | A 2020 forecast, on "the optimistic assumption that it will be substantially contained by the fall of 2021" | A 2023 model fitted to the first 30 months of data, projecting to December 2023 |
The reconciliation in three sentences
First, $8.5 trillion of Cutler's total is not economic output at all — it is deaths, disability and distress converted to dollars with a value-of-statistical-life method that USC Schaeffer deliberately did not use. Strip that out and Cutler's economic figure is $7.6 trillion. Second, the two GDP figures differ because they were made three years apart with different tools: Cutler took CBO's October 2020 decade-long projection; Schaeffer built a general-equilibrium model on thirty months of observed data and found the four-year loss alone was $14 trillion — almost twice CBO's early guess for ten years. Third, neither figure is a bill. Both are output that never happened, spread across 331 million people, and the direct cost of treating the disease is a rounding error inside either — our economic impact page sets out what kind of number a model estimate is and shows the per-person arithmetic ($14 trillion ÷ 331.4 million ≈ $42,200).
So the honest one-line answer to "how much did COVID cost the US" is: about $14 trillion in lost output through 2023 (USC Schaeffer), plus health losses that Cutler and Summers valued at a further $8.5 trillion in JAMA in 2020 and that nobody has re-estimated with the same care since. Quoting $16 trillion as "the economic cost" overstates the economic part and understates it at the same time; quoting $14 trillion as "the total cost" leaves out every death.
What One COVID Hospital Stay Costs — by Insurance, Treatment and Period
The largest study is Kapinos and colleagues' analysis of 1,333,404 inpatient stays at 841 US hospitals — 97% of academic medical centres — from March 2020 to March 2022, published in JAMA Network Open in January 2024. It measures what the hospital spent to provide care, in January 2022 dollars, not what it billed or was paid.
| Cut | Mean cost per stay (Jan 2022 $) | Reading |
|---|---|---|
| Overall, unadjusted | $13,023 | The raw average of all 1.33 million stays |
| Overall, adjusted | $11,275 (Table 2 gives $11,432 under its own model specification) | The figure to quote; the gap from the unadjusted mean is the effect of outlier patients |
| Private / commercial insurance | $11,274 | Insurance status barely moves the hospital's cost — a spread of about $2,500 across six payer types — because this is what the hospital spent, not what it charged. What payers paid is a different number: studies of Medicare beneficiaries cited in the same paper report average payments of $21,752 to $24,033 per stay |
| Medicare | $11,609 | |
| Medicaid | $10,629 | |
| Military | $11,488 | |
| Self-pay (uninsured) | $11,085 | |
| Workers' compensation / automobile | $13,106 — the highest | |
| ICU stay | $12,490 (vs $10,724 without) | Treatment intensity is what moves the cost: a ventilated stay costs more than twice an unventilated one, and an ECMO stay more than three times the average. The authors attribute the rise over time partly to growing ECMO use |
| Invasive mechanical ventilation | $20,941 (vs $9,614 without) | |
| ECMO | $36,484 (vs $11,266 without) | |
| Discharged to a skilled nursing or rehabilitation facility | $14,005 | |
| Patient died in hospital | $9,580 | Lower than average — deaths were often early in the stay |
| End of March 2020 → end of March 2022 | $10,394 → $13,072 | "More than 5 times the rate of medical inflation" over the period |
The total, derived — and why it is small
Kapinos and colleagues open with the count that makes the total: "more than 6.2 million hospitalizations due to COVID-19 in the US" by the time of their study. Multiply by the adjusted mean and 6.2 million × $11,275 ≈ $70 billion of hospital cost to provide care through early 2022 — our arithmetic, not the paper's, and an underestimate because stays after March 2022 cost more and are not counted. USC Schaeffer's model brackets the same quantity from the other direction: $20 billion in direct health expenses in a scenario with 65,000 deaths, $365 billion in one with 2 million, "driven mostly by hospitalization costs"; the real death toll (about 1.1 million by 2023, per the same page) lands between. Whichever figure is used, hospital care is a few per cent of $14 trillion. That is the finding both sources share and neither headline conveys: the cost of COVID was the economy that stopped, not the patients who were treated.
Long COVID: The Line Only One Estimate Counts
The largest single difference between the two totals after lost GDP is Cutler's $2.6 trillion for long-term impairment, which Schaeffer explicitly leaves out. In May 2022 Cutler revisited it in JAMA Health Forum and concluded the estimate "seems very much on target". His inputs: CDC's count of roughly 81 million US cases by May 5, 2022; British population data putting 12% to 17% of infected people at three or more symptoms twelve weeks on, so "even the lower-end estimate … implies that 9.6 million people in the US may have developed long COVID — roughly 10 times the number of COVID-19 deaths"; a survey in which 44% of people with long COVID were out of the labor force and 51% worked fewer hours; and "more than 1 million people may be out of the workforce at any given time", a lost income of "more than $50 billion annually". Medical costs, if long COVID is treated like myalgic encephalomyelitis, "could be about $9,000 per person annually".
None of that is in the $14 trillion. It is the reason the two headline figures should be read as a floor and a ceiling rather than as rivals: $14 trillion is what the pandemic cost the economy through 2023 on the best available model; $16 trillion is what it cost the country once the deaths and the disability are priced, on a 2020 forecast that has since proved conservative on its GDP line. Our long COVID page covers the clinical evidence behind the prevalence figures.
What These Numbers Cannot Tell You
- A cost per case. COVID-19 is no longer nationally notifiable and reported cases never approximated infections, so any "per case" figure divides a real cost by a denominator nobody believes. The hub's cost-per-case table leaves that cell blank on purpose and gives cost per stay and per resident instead.
- What a patient paid. Kapinos measures the hospital's cost; Medicare's payments per stay were roughly double it; a privately insured patient's out-of-pocket share depends on their plan and on the cost-sharing waivers most insurers offered in 2020 and withdrew in 2021. No national figure exists for what patients themselves paid.
- A current per-stay cost. The study ends in March 2022, when the mean was $13,072 and rising faster than medical inflation. Stays since are fewer and, with ECMO use higher, probably dearer per stay; nobody has published the update.
- Whether $14 trillion is right. It is a model output — a counterfactual GDP path nobody observed — as our economic impact page explains at length. A different reasonable set of assumptions produces a materially different total. What the model does establish is the order of magnitude and the composition: output, not medicine.
Frequently Asked Questions
How much did COVID cost the US — $14 trillion or $16 trillion?
Both. Cutler and Summers' $16.1 trillion (JAMA 2020) is $7.6 trillion of lost output plus $8.5 trillion of deaths, long-term impairment and mental-health loss valued in dollars. USC Schaeffer's $14 trillion (2023) is lost output only, 2020–2023, from a model that excludes all three health lines. The one shared line is GDP, where Schaeffer's later, larger figure supersedes Cutler's 2020 CBO projection.
How much does one COVID hospitalization cost?
$13,023 unadjusted, $11,275 adjusted, per stay, as the hospital's cost to provide care across 1.33 million stays in 2020–22 (Kapinos et al., JAMA Netw Open 2024). Insurance status changed it by only about $2,500 across payers; ventilation ($20,941) and ECMO ($36,484) changed it far more. Medicare's payments per stay were $21,752–$24,033.
What did all the hospitalizations cost together?
About $70 billion by our arithmetic — 6.2 million stays × $11,275 — against USC Schaeffer's scenario range of $20–$365 billion in direct health expenses. A few per cent of the $14 trillion either way.
What does long COVID cost?
Cutler's $2.6 trillion (2020), which he judged "very much on target" in 2022: perhaps 9.6 million Americans affected, more than a million out of the workforce, over $50 billion a year in lost income, about $9,000 a year in medical costs per person if it is treated like chronic fatigue syndrome. Schaeffer's $14 trillion excludes it.
Why is there no cost per case?
No clean denominator: COVID-19 is not nationally notifiable and reported cases never captured infections. The figures with real denominators are cost per hospital stay and cost per US resident (about $42,200).
Medical and financial disclaimer. This page reconciles published economic estimates and reproduces a hospital-cost study. It is not medical advice, and it is not a guide to what any individual will be billed for COVID-19 care, which depends on insurance, hospital and treatment; nothing here should delay anyone from seeking care.
Sources & References
- Cutler DM, Summers LH. The COVID-19 pandemic and the $16 trillion virus. JAMA 2020;324(15):1495–6 — Table: lost GDP $7,592 billion; premature death $4,375 billion; long-term health impairment $2,572 billion; mental health impairment $1,581 billion; total $16,121 billion; $196,475 per family of four; ≈90% of annual GDP. Value of a statistical life $7 million; 625,000 expected deaths; "optimistic assumption that it will be substantially contained by the fall of 2021". doi:10.1001/jama.2020.19759 · PubMed 33044484
- Hlávka J, Rose A. COVID-19's total cost to the U.S. economy will reach $14 trillion by end of 2023. USC Schaeffer Center, May 16, 2023 (republished from The Conversation) — $103 trillion vs $117 trillion cumulative GDP 2020–2023 in 2020 dollars; direct health expenses $20 billion to $365 billion across death scenarios; exclusions quoted. schaeffer.usc.edu
- Kapinos KA, Peters RM Jr, Murphy RE, Hohmann SF, Podichetty A, Greenberg RS. Inpatient costs of treating patients with COVID-19. JAMA Netw Open 2024;7(1):e2350145 — 1,333,404 stays, 841 hospitals, March 2020–March 2022; unadjusted mean $13,023; adjusted $11,275 (Table 2 model: $11,432); all payer, treatment, discharge and period figures as tabulated; Medicare payment studies $21,752–$24,033; "more than 6.2 million hospitalizations". doi:10.1001/jamanetworkopen.2023.50145 · PubMed 38170519
- Cutler DM. The costs of long COVID. JAMA Health Forum 2022;3(5):e221809 — 81 million cases and 994,187 deaths (CDC, May 5, 2022); 12–17% with three or more symptoms at 12 weeks; 9.6 million; 44% out of the labor force and 51% working fewer hours; more than 1 million out of the workforce, more than $50 billion a year; about $9,000 per person per year; "$2.6 trillion … seems very much on target". doi:10.1001/jamahealthforum.2022.1809
- U.S. Census Bureau. 2020 Census apportionment population, 331,449,281 — the denominator of the per-resident figure on our economic impact page. census.gov
- Centers for Disease Control and Prevention. COVID Data Tracker — the case and hospitalization counts cited by Cutler 2022 and Kapinos 2024; no longer a nationally notifiable condition. cdc.gov/covid
Continue Reading
Economic Impact
What kind of number the USC Schaeffer estimate is, and the per-person and per-federal-dollar arithmetic.
Long COVID
The clinical evidence behind the prevalence figures that drive Cutler's long-COVID estimate.
Treatment
What a hospital stay involves — from oxygen to ventilation to ECMO.
Cost Per Case, Virus by Virus
COVID's per-stay cost beside Ebola, norovirus, influenza and measles on the hub.
