By Adam Gaffney, M.D., M.P.H.; David Himmelstein, M.D.; and Steffie Woolhandler, M.D., M.P.H.
Health Affairs Blog, February 16, 2021
For the first time in a quarter century, the Congressional Budget Office (CBO) has undertaken an economic analysis of single-payer health care reform, also known as Medicare for All. The more than 200-page working paper, released last month, includes a rich explanation of methodology together with cost projections for 2030 and will no doubt serve as an important reference for years to come.
The report makes many sound assumptions but also some questionable ones that are overly pessimistic. Yet, overall, its bottom-line estimates should reassure those concerned about the economic feasibility of single payer: The CBO projects that such reform would achieve universal coverage, bolster provider revenues for clinical services, and eliminate almost all copayments and deductibles—even as overall health care spending fell.
The CBO models costs under five different variants of single payer. The first four envision universal coverage of all services other than long-term care, while the fifth incorporates a large expansion of long-term services and supports (LTSS) for people with disabilities of all ages. The scenarios vary by patient cost sharing and provider payment level. Low cost-sharing has no copays or deductibles for medical services and minimal cost sharing for prescription drugs; under high cost-sharing, patients with incomes above 150 percent of poverty would bear about 7.5 percent of costs out of pocket. Low payment rates to providers would set rates slightly higher than Medicare’s; high rates would be equivalent to the current average of the rates paid by private insurers and government programs. The five variants (or scenarios) are summarized in exhibit 1.
Exhibit 1: Congressional Budget Office’s five single-payer scenarios
Source: Congressional Budget Office’s (CBO’s) Single-Payer Health Care Systems Team. How CBO analyzes the costs of proposals for single-payer health care systems that are based on Medicare’s fee-for-service program. Washington (DC): CBO; 2020 Dec.
We discuss below the CBO’s estimates of single payer’s overall effects on national health spending, the implications of the estimates for providers, and the concerns the analysts raise about worsened “provider congestion” under a single-payer health care system. Throughout, we point out instances where the CBO’s assumptions differ from previous, widely publicized analyses, or from provisions included in the Medicare for All legislation currently in Congress.
CBO’s Estimates Of Single Payer’s Effects On National Health Expenditures
The CBO projects that variants 1–4 of single-payer reform would reduce national health expenditures (NHE) despite substantial increases in the use of care triggered by expanded and upgraded coverage. If a vast new program covering LTSS for all US residents were included (scenario 5), the CBO estimates that NHE would rise by 4.4 percent above currently projected spending levels. Exhibit 2 summarizes the CBO’s spending estimates for the three low cost-sharing scenarios, which are similar to the Medicare for All bills in Congress.
Exhibit 2: Congressional Budget Office estimates of effect of single payer on health spending in 2030: low cost-sharing scenarios
Source: Authors’ analysis of the Congressional Budget Office’s (CBO’s) Single-Payer Health Care Systems Team. How CBO analyzes the costs of proposals for single-payer health care systems that are based on Medicare’s fee-for-service program. Washington (DC): CBO; 2020 Dec. Note: “Net change in provider payments” reflects combined effect of rate setting and increased use.
How can the use of care rise even as spending falls? Mostly, according to the CBO, through greater administrative simplicity. As the CBO notes, traditional Medicare’s administrative overhead accounts for approximately 2 percent of its total revenue, compared to the 12 percent overhead of private insurers. Under single payer, the CBO projects, administrative spending would fall accordingly; overall overhead for the Medicare for All system is estimated by the CBO at below 2 percent. As shown in exhibit 2, this translates into around $400 billion annually (more than $1,000 per capita) in savings under all of the single-payer variants.
Of note, these evidence-based projections clash with the analyses by the Urban Institute and the RAND Corporation, which perplexingly assumed much higher administrative overhead under a single-payer system than under the current traditional Medicare program or under universal systems abroad.
Implications For Provider Overhead And Revenue
The CBO appropriately projects substantial administrative savings for providers, again unlike many previous analyses. US hospitals and physicians waste money and time contending with multiple payers, each with its own complex and varying coverage rules and payment procedures, formularies, and so forth. The CBO projects that the share of revenues that hospitals spend on administration would fall from 19 percent at present to 12 percent under single payer; that physicians’ administrative overhead would fall from 15 percent to 9 percent; and that the administrative expenses of other medical providers (for example, dentists, home health agencies, and hospices) would fall from 9 percent to 6 percent. In addition, it estimates that physicians and nurses would spend less time on administrative activities, freeing up 4.8 percent of physicians’ work hours and 18.4 percent of nurses’ work time. These assumptions build on a large evidence base showing high administrative overhead among US health care providers relative to other nations.
The CBO assumes that single payer would allow providers to keep these savings on administration and billing, and use the resources freed up to provide more care. Hence, providers’ administrative savings do not appear as savings in exhibit 2. In effect, the CBO interprets these efficiencies as reductions in providers’ costs that would enable them to deliver a greater quantity of services for a given amount of revenue.
Consequently, the “net change in provider payments” shown in exhibit 2 doesn’t fully account for the extra resources made available to hospitals, doctors, and other providers under Medicare for All. In contrast, exhibits 3 and 4 present our estimates of single payer’s financial impacts on physicians and hospitals, incorporating CBO projections of providers’ savings on administrative spending and physicians’ and nurses’ time.
As exhibit 3 demonstrates, payments to clinicians would rise in all five scenarios according to CBO estimates. We estimate this translates into an additional $39,816–$157,412 in revenue per practicing physician. At the same time, physicians’ practice overhead would shrink under single payer, increasing practices’ take-home income. Such a windfall, in our view, may be excessive, at least for some providers. (Disclosure: The authors are all physicians).
Exhibit 3: Modeling Congressional Budget Office’s low-cost sharing scenario: ramifications for physicians
Source: Authors’ analysis of the Congressional Budget Office’s (CBO’s) Single-Payer Health Care Systems Team. How CBO analyzes the costs of proposals for single-payer health care systems that are based on Medicare’s fee-for-service program. Washington (DC): CBO; 2020 Dec. Note: *Based on Association of American Medical Costs estimate of 840,000 practicing physicians in 2030 and the assumption that physician payments account for 77.78 percent of payments in the “Physician and other Clinical Service” category, as they did in 2018 according to the National Health Expenditure Accounts.
For hospitals, the CBO estimates that gross revenue would fall by $187 billion under the “low pay” scenario but rise by $144 billion under the high-payment scenario. Again, these figures do not give the full picture of the impact on hospitals’ bottom lines, given the CBO’s projections that single-payer reform would shrink hospitals’ spending on administration and would also free up substantial amounts of nursing time that is currently devoted to payment-related tasks. The CBO’s estimates suggest that hospitals would save $143 billion to $166 billion on administration and an additional $59 billion on freed-up nursing time, resources that could be redirected to clinical care (exhibit 4). Overall, building on the CBO’s projections, we estimate that hospitals’ clinical funding would change little (rising by $15 billion) under the low-payment scenario or increase substantially, by $369 billion, under the higher-payment one.
Exhibit 4: Modeling Congressional Budget Office’s low cost-sharing scenario: ramifications for hospitals’ revenues and clinical operating budgets
Source: Authors’ analysis of the Congressional Budget Office’s (CBO’s) Single-Payer Health Care Systems Team. How CBO analyzes the costs of proposals for single-payer health care systems that are based on Medicare’s fee-for-service program. Washington (DC): CBO: 2020 Dec. Notes: **Based on the CBO estimate that hospitals’ spending on administration (excluding registered nurse (RN)and licensed practical nurse (LPN) time spent on administration) would be reduced from 19 percent to 12 percent of hospital revenues. ***Based on (1) Bureau of Labor Statistics estimate of number of RNs and LPNs employed in hospitals and average RN and LPN wages in 2019; (2) the assumption that benefit costs equal 20 percent of wages; (3) the assumption that nursing costs would rise at the same rate as overall hospital costs; and (4) CBO’s estimate that RNs and LPNs devote 23 percent of time to administration and that single payer would reduce that time by 80 percent.
The large increase in hospital funding that the CBO projects under the high-payment scenario would, in our view, give an appropriate boost to many struggling rural and safety-net hospitals but would be excessive for hospitals that are currently realizing large annual profits.
Additionally, the fee-for-service hospital financing modeled by the CBO differs from what is envisioned in the Medicare for All bill in the House of Representatives. The House bill proposes paying hospitals, nursing homes, and other institutional providers using “global budgets,” that is, lump sums to cover all of their operating activities. This is the model used by Canada and the US Veterans Health Administration. Global budgeting could allow even greater reductions in administrative spending by eliminating the need for per-patient billing altogether. It would also allow payments to match current operating budgets and facilitate the gradual redistribution of funding among hospitals, directing more resources to communities most in need.
The global budgeting approach, however, would require separate financing of capital expenditures (as in Canada and some European nations), which again is an approach explicitly included in Medicare for All bills in Congress but not in the CBO report. Such regulated capital financing, in turn, could help regulate cost growth in the longer term by avoiding investments in duplicative but profitable high-tech facilities that encourage the delivery of low- or no-value care.
Hence, the CBO’s decision to model a system that used fee-for-service payments for hospitals and other institutional providers rather than global budgets with separate capital financing leads it to underestimate single-payer savings in the short term and to ignore potential longer-term savings and improvements in the distribution of hospital infrastructure. That caveat aside, an important takeaway from the CBO’s analysis is that, overall, providers would see stable or even increased funding for patient care under single payer, even while overall health spending fell.
While the CBO report foresees greatly improved financial access to care under single payer, it raises the specter of “provider congestion,” that is, greater difficulty making appointments or rising waiting times under single payer. Because such reform would newly cover the uninsured and improve coverage for most other people, the CBO projects large increases in the demand for care, and hence use. However, it projects that not all of the demand could be met because of supply constraints, for example, the finite number of doctors, nurses, and hospital beds. Consequently, the report predicts increased “provider congestion” that would cause some foregone care.
This prediction, however, is out of touch with clinical reality, an issue we recently explored in Health Affairs. The CBO’s approach to supply constraints is only half right. There is little question that, as the CBO assumes, a finite supply of health care providers constrain utilization increases after coverage expansions: As we have demonstrated, in nation after nation, universal coverage expansions have led to modest, or even no, societywide increases in use. Increased use by the newly covered has usually been partially or fully offset by small reductions in use among the well-covered or well-off.
Yet, these small reductions appear to be mostly due to the reduced provision of low-value care. Two econometric studies found that offsets after coverage expansion reflect reductions in the amount of wasteful care provided to the already insured. Similarly, studies by the Dartmouth Group indicate that the volume of elective and low-value care increases when the supply of doctors and hospital beds rise—a phenomenon called supplier-induced demand. In other words, faced with an increase in demand for care, providers prioritize their time and services, delivering more high-value care while reducing the provision of low- or no-value services.
Characterizing such offsets as “unmet demand” or “congestion” is hence misleading. It mirrors previous forecasts of “patient pileups” when Medicare was first implemented—forecasts that proved incorrect. In the US today, nearly one-third of all health care delivered is unnecessary or wasteful. Overprovision is common. If coverage expansion leads to some attenuation in supplier-induced demand, all the better.
A Notable Omission
Finally, while the CBO projects reductions in health spending in four out of five single-payer scenarios, as noted, it projects substantial increases in federal spending that would replace all private insurance premiums and nearly all out-of-pocket health care costs. Yet, there are also substantial savings for state and local governments with single-payer reform. The CBO’s brief mention of the savings likely to accrue to these governments (and their taxpayers) omits probable savings from no longer having to bear the costs of public employees’ health insurance (projected to total $318 billion in 2030), as well as about $162 billion in savings on other health programs. These, together with savings on Medicaid (which the CBO does remark on), would bring state and local governments’ total savings from single payer to about $800 billion in 2030 alone, reductions that provide important context for the CBO’s estimates of increased federal government expenditures.
The Economics Of Single-Payer Financing: CBO’s Bottom Line
Overall, the CBO report provides one of the most detailed explorations to date of the economics of single-payer financing. As we have noted, it makes many sound assumptions, particularly about payer- and provider-side administrative savings. At the same time, it adopts some unfavorable assumptions about the structure of single-payer reform (including some that conflict with key provisions of the Medicare for All bills in Congress), projects excessive windfalls for some providers, and asserts clinically nescient portrayals of “unmet demand.” Nonetheless, the bottom line of the CBO analysis—that universal coverage can be affordably achieved even as benefits are expanded and cost sharing all but eliminated—should reinvigorate debate over such reform.
Drs. Gaffney, Himmelstein, and Woolhandler have served as leaders of Physicians for a National Health Program, a nonprofit organization that supports single-payer health care reform.