Scoring CBO’s Scores: Ten of the Worst CBO Blunders of the 21st Century So Far

Overview

Since 1974, the Congressional Budget Office (CBO) has been charged with providing Congress with “objective, nonpartisan information” in order to support the federal budget process through the fiscal scoring of legislative priorities.¹ Each year, CBO produces roughly 600 to 800 cost estimates, in addition to offering technical assistance to members of Congress as they develop legislation.² CBO has provided fiscal estimates on major pieces of legislation, from the 2009 stimulus package to ObamaCare to COVID-era relief measures and more.

Unfortunately, CBO is far from perfect. In fact, the office’s fiscal estimates have been stunningly incorrect in many cases. This has unfortunately led members of Congress to make misinformed decisions about legislation since federal policymakers rely on the accuracy of CBO’s estimates when passing bills. In addition, CBO has also routinely misforecasted underlying assumptions relating to economic growth and budgetary trends, further compounding their mistakes.

Inaccurate predictions have piled up from CBO over the still-young 21st century. But a survey of some of CBO’s biggest missteps over the past two decades should give federal policymakers pause about the reliability of CBO moving forward.

MEDICAID EXPANSION ESTIMATES ARE IN ERROR

As part of ObamaCare, states were given the option to expand Medicaid to a new class of able-bodied adults earning up to 138 percent of the federal poverty level beginning January 1, 2014.³ While this policy change was initially mandatory, in the 2012 NFIB vs. Sebelius case, the U.S. Supreme Court ruled states may opt in to ObamaCare’s Medicaid expansion.⁴

In 2010—prior to the ruling that Medicaid expansion was optional—CBO predicted that just 13 million people would sign up once fully implemented.⁵

However, CBO got it wrong. 

By 2019—the year before the onset of the COVID-19 pandemic—only 65 percent of people who met the eligibility criteria for Medicaid expansion lived in states that had actually expanded Medicaid.⁶ Yet, actual Medicaid enrollment in 2019 reached roughly 19.5 million—well above the original estimate of 13 million in all states.⁷Notably, this 19.5 million enrollment figure was hit even though, at the time, 16 states had not implemented Medicaid expansion.⁸

Accounting for the fact that CBO originally predicted enrollment of 13 million able-bodied adults under expansion if all states expanded, the actual enrollment of 19.5 million (with just 65 percent of potential enrollees living in expansion states) suggests an astonishing error rate of more than 130 percent..⁹

Leading up to 2019, CBO recognized its embarrassing error. In a 2017 report, it noted: “In its May 2013 projection, CBO underestimated such [expansion] spending mainly because it underpredicted the number of people who would enroll in Medicaid. Fewer states expanded eligibility for Medicaid under the ACA than CBO anticipated, but many more people enrolled in Medicaid in expansion states than CBO expected.”¹⁰

Put simply, CBO grossly underestimated Medicaid expansion enrollment, which likely contributed to lawmakers sending America down a misguided course of one of the largest expansions of welfare in recent history.

MANDATORY AND SUBSIDIZED OBAMACARE EXCHANGE PLANS WERE EVEN LESS ATTRACTIVE THAN CBO PREDICTED

Another major facet of ObamaCare was the creation of federal (and state-based) health exchanges on which individuals would purchase health insurance that conformed to new regulations and requirements.

After the Sebelius decision, CBO predicted that by 2016 roughly 22 million Americans would be signed up for insurance via these exchanges, with enrollment ultimately peaking at between 24 and 25 million in the out-years.¹¹

However, for most of the decade, exchange enrollment has hovered around just 12 million—well below CBO’s predictions of 22 to 25 million.¹²

Even with the individual mandate still intact—a tax that was zeroed out beginning in 2019 by the Tax Cuts and Jobs Act of 2017—actual enrollment was well below CBO’s projections.¹³

It wasn’t until major federal policy changes were made—such as the dramatic expansion of insurance premium tax credits (PTCs) in the American Rescue Plan Act (ARPA)—that enrollment in the exchanges exceeded 21 million.^14-15 However, these more generous benefits did not exist at the time CBO made its original estimates.

As a result, CBO’s early estimates of exchange enrollment were off by a factor of close to 100 percent.

Unfortunately, CBO’s mistakes don’t end there. When Congress was considering rolling back the enhanced tax credits that were set to expire per ARPA, CBO told lawmakers that roughly 2.2 million people would be without insurance if these provisions were rolled back.¹⁶

However, that 2.2 million-person estimate rested on failed assumptions. For example, CBO’s 2022 baseline assumed the number of uninsured Americans would be 28 million in 2022.¹⁷ In reality, it was just 25.4 million—an overestimate of 10.2 percent.¹⁸

CBO also assumed that the number of individuals covered by employment-based insurance would be 155 million.¹⁹But by 2022, the number of people on employer-based coverage was 162.7 million—an underestimate of 4.7 percent.²⁰

Since the underlying assumptions about the uninsured population, the employer-sponsored coverage population, and the marketplace coverage population were fundamentally flawed, it is reasonable to assume that CBO’s assertion that 2.2 million individuals would lose coverage if the enhanced PTCs expired was also incorrect. As a result, policymakers were misled down the path of expanding subsidies based on bad projections.

HANDCUFFED TO MORE FLAWED MEDICAID ASSUMPTIONS

In March 2020, Congress passed the Families First Coronavirus Relief Act, which contained a series of provisions related to the response to the COVID-19 pandemic. One of these measures was an additional 6.2 percentage-point boost in states’ Federal Medical Assistance Percentages (FMAPs), which is the share of Medicaid expenditures paid for by the federal government.²¹ In exchange for the increased federal money, states were required to keep enrollees on the Medicaid program through the duration of the public health emergency, regardless of whether or not they continued to be eligible.²²

This became known as the “Medicaid handcuffs,” as states were locked in to maintaining coverage for even ineligible enrollees.²³

There were two direct costs that CBO considered: the cost of the 6.2 percentage-point FMAP boost itself and the cost associated with increased Medicaid enrollment as a result of the handcuffs. In April 2020, CBO indicated that the cost of both components would be $50 billion between 2020 and 2022.²⁴ This was a drastic underestimate.

The costs incurred by the federal government for just the first component—the 6.2 percentage-point FMAP bump—was $71.9 billion over the 2020 to 2022 period, a substantial overage compared to the $50 billion estimate for both components by CBO.²⁵

While it is impossible to quantify the precise amount spent on ineligible enrollees who were locked in, the magnitude of CBO’s error can be gleaned from their flawed Medicaid spending predictions.

In the March 2020 baseline, CBO estimated approximately $1.7 trillion in Medicaid expenditures from FY2020 to FY2023, plus another $50 billion attributable to the Medicaid handcuffs.²⁶ But actual spending over that period was roughly $2.1 billion—a roughly 20 percent overage.²⁷

Yet again, the CBO’s Medicaid predictions were wildly off base, contributing to misinformation about the cost of significant policy provisions.

ENHANCED CHILD TAX CREDITS WERE EVEN MORE “ENHANCED” THAN CBO THOUGHT

As part of ARPA, the Child Tax Credit was dramatically expanded. The enhanced credit saw increased payments on a monthly basis (instead of annual), a hiked refundability threshold, additional eligibility, and more.²⁸

In scoring ARPA, the CBO predicted the changes to the Child Tax Credit would decrease revenues by $21 billion and increase outlays by $88 billion over the budget window, at a cost of about $109 billion.²⁹

However, when calculating the possibility of extending the credit, its new cost, according to CBO, suddenly spiked to an astonishing $185 billion over the budget window—a 70 percent increase.³⁰

CBO even admitted that its revenue estimates were wrong, noting that “estimates of the portion of refundable tax credits recorded as outlays were too low by $74 billion. The Child Tax Credit accounted for $34 billion of that amount.”³¹ In other words, the Child Tax Credit alone accounted for nearly half of the CBO’s massive $74 billion refundable tax credit error.

While Congress did not extend the enhanced Child Tax Credit, the inaccurately low projections made prior to the passage of the law may have contributed to the adoption of bad public policy.

UNEMPLOYMENT BONUSES BUST CBO’S BEST GUESS

As part of the 2009 stimulus package—otherwise known as the American Recovery and Reinvestment Act, or ARRA—various provisions related to unemployment compensation were changed. This included a temporary $25 weekly unemployment insurance (UI) bonus, a temporary provision to allow states to expand certain eligibility requirements, the suspension of income taxes on the first $2,400 of UI, and more.³²

In February 2009, CBO predicted the UI portions of the ARRA would cost $39.2 billion over the 10-year budget window.³³ However, a revised score by CBO six years later indicated that the actual costs for the UI provisions would be $64 billion—or roughly 63 percent more than originally estimated.³⁴

CBO’s substantial underestimation of the costs of these UI provisions contributed to Congress adopting a massive economic stimulus that, unfortunately, contributed to the slowest economic recovery in modern American history.³⁵

FROM OUTLAYS TO ENROLLMENT, CBO GETS FOOD STAMP DATA WRONG

In addition to increases in unemployment compensation, ARRA also included an increase in monthly food stamp benefits.

CBO originally estimated the food stamp changes in ARRA would cost $20 billion over the ten-year budget window (almost all of which would occur within the first five years).³⁶ However, a 2019 evaluation by CBO of its own estimates concluded that: “CBO now estimates that ARRA increased total spending on SNAP by $43 billion over the 10-year projection period. That is, the ARRA-related increase in spending on SNAP was more than double the amount CBO estimated in 2009.”³⁷

Had Congress not sunset the ARRA food stamp increases in 2010, CBO’s underestimate would have been off by $43 billion—or roughly double their original error.^38-39

Part of the problem was that CBO used its unemployment rate projection to construct its estimate of food stamp outlays over the budget window, which led to an underestimate when the unemployment rate climbed higher and stayed there longer than the CBO assumed—leading to an underestimate of food stamp enrollment.⁴⁰

Once again, CBO’s ARRA score misinformed lawmakers by significantly underestimating food stamp enrollment and spending—a mistake that would be repeated again a decade later.

FOOD STAMP EMERGENCY ALLOTMENTS TURN OUT TO BE A LOT MORE

During the early stages of the COVID-19 pandemic, Congress approved “emergency allotments” for the food stamps program that boosted benefits so that everyone eligible for the program would receive the maximum benefit for their household’s size, regardless of their income.⁴¹

In 2020, CBO predicted these emergency allotments would cost a combined $18.5 billion between 2020 and 2021.⁴² However, not only did these emergency allotments not expire until February 2023, but they also cost $97.4 billion—more than five times CBO’s estimate.⁴³

Even when constraining the allotments to the 2020–2021 forecast window, actual expenditures far exceeded CBO’s projections.⁴⁴

Unfortunately, CBO’s mistake gave Congress erroneous guidance that perpetuated a program that was far more expensive than originally anticipated.

REALITY LET THE AIR OUT OF THE BALLOON ON CBO’S INFLATION REDUCTION ACT’S “DEFICIT REDUCTION”

President Biden’s so-called “Inflation Reduction Act” was initially supposed to reduce the deficit by approximately $58.1 billion over the 2022–2031 budget window, plus approximately $200 billion in additional revenues from increased IRS enforcement efforts.⁴⁵

However, a February 2024 outlook substantially revised the portion of the law pertaining to energy tax credits. According to CBO, a variety of factors led to a projected increase in the cumulative deficit over the 10-year budget window by $428 billion, more than half of which—$224 billion—is “from revised projections of amounts claimed for clean vehicle tax credits and of revenues from excise taxes on gasoline” contained in the Inflation Reduction Act.⁴⁶ Another $204 billion in deficit increases is due to other energy-related tax provisions, such as incentives for wind, solar, battery manufacturing, and more.⁴⁷

While some of these factors—such as a new Environmental Protection Agency rule and guidance from the Treasury Department—were outside the scope of CBO’s ability to forecast, others were not.⁴⁸ For example, CBO did not predict key developments in certain green energy markets related to manufacturing, contributing to rising deficits.⁴⁹

This act—originally expected to reduce the deficit—has been recategorized as a major deficit-hiking series of initiatives, which the CBO failed to anticipate.

REVENUE OUTLAYS HAVE BEEN OVER AND UNDER

CBO has routinely mis-estimated revenues and outlays in a more general sense.

Over the period ranging from FY10–FY18, CBO predicted the federal government would raise $30.1 trillion in revenue. In reality, it raised just $25.9 trillion—an astonishing difference of $4.2 trillion, or about 14 percent less than originally estimated.⁵¹

In the most recent fiscal year alone, by its own admission “CBO overestimated revenues by 11 percent and underestimated outlays by nine percent.”⁵²

CBO’s historical projections of revenues and outlays have been marked by mistakes and inaccuracies, obfuscating decisions for federal policymakers.⁵³

EVEN MORE ECONOMIC ERRORS

CBO has a tendency to miscalculate the growth in the overall economy, as well as incorrectly gauge key economic variables.

Broadly speaking, CBO’s projections of real gross domestic product (GDP), interest rates, and more, have been routinely incorrect.⁵⁴

For example, during the peak of the Great Recession, CBO predicted real GDP growth of 2.3 percent, when, in fact, GDP declined by 2.9 percent that year—a shocking difference of 5.2 percentage points.⁵⁵ These changes persisted in subsequent years, with CBO routinely overestimating the growth in real GDP.⁵⁶

CBO’s failure to appropriately anticipate macroeconomic factors is highly concerning given the implications these errors have on fiscal scores for major policy proposals.

THE BOTTOM LINE: CBO’s fiscal errors are highly concerning and should be viewed cautiously and critically by lawmakers.

As we approach the 50-year anniversary of the CBO, it is clear that its track record is not only far from perfect, but also highly concerning. The office’s failure to accurately predict the cost of major pieces of federal legislation, changes in macroeconomic variables, and overall trends in revenues and outlays has wreaked havoc on policymakers’ ability to make informed decisions. However, the CBO is still heavily relied upon despite these blunders.

Moving forward, lawmakers should take CBO projections with a high degree of uncertainty. If the past is any predictor of the future, CBO will continue to make serious mistakes in its fiscal estimates—a reality that policymakers need to incorporate in their decision-making when considering CBO scores.

REFERENCES

1. Congressional Budget Office, “10 things to know about CBO,” Congressional Budget Office (2024), https://www.cbo.gov/about/10-things-to-know.   

2. Ibid.

3. Jonathan Ingram and Nic Horton, “A budget crisis in three parts: How ObamaCare is bankrupting taxpayers,” Foundation for Government Accountability (2018), https://thefga.org/research/budget-crisis-three-parts-obamacare-bankrupting-taxpayers.  

4. See, e.g., National Federation of Independent Businesses v. Sebelius, 567 U.S. 519 (2012), https://www.supremecourt.gov/opinions/boundvolumes/567BV.pdf.

5. Sarah Masi et al, “CBO’s record of projecting subsidies for health insurance under the Affordable Care Act: 2014 to 2016,” Congressional Budget Office (2017), https://www.cbo.gov/system/files/115th-congress-2017-2018/reports/53094-acaprojections.pdf.

6. Kate Fritzsche et al, “Federal subsidies for health insurance coverage for people under age 65: 2019 to 2029,” Congressional Budget Office (2019), https://www.cbo.gov/system/files/2019-05/55085-HealthCoverageSubsidies0.pdf.

7. MACPAC, “MACSTATS: Medicaid and CHIP data book,” Medicaid and CHIP Payment and Access Commission (2021), https://www.macpac.gov/wp-content/uploads/2021/12/MACStats-Medicaid-and-CHIP-Data-Book-December-2021.pdf.

8. Ibid.

9. Authors’ calculations comparing 19.5 million actual enrollment to 13 million original predicted enrollment, adjusting for the 65 percent distribution of expansion-eligible individuals who actually resided in expansion states.

10. Sarah Masi et al, “CBO’s record of projecting subsidies for health insurance under the Affordable Care Act: 2014 to 2016,” Congressional Budget Office (2017), https://www.cbo.gov/system/files/115th-congress-2017-2018/reports/53094-acaprojections.pdf.

11. Congressional Budget Office, “Table 1. CBO’s May 2013 estimate of the effects of the Affordable Care Act on health insurance coverage,” Congressional Budget Office (2013), https://www.cbo.gov/sites/default/files/recurringdata/51298-2013-05-aca.pdf.

12. Centers for Medicare and Medicaid Services, “Open Enrollment Reports,” Centers for Medicare and Medicaid Services (2024), https://www.cms.gov/data-research/statistics-trends-and-reports/marketplace-products.

13. Marc Short and Brian Blase, “The fundamental error in the CBO’s health-care projections,” The Washington Post (2017), https://www.washingtonpost.com/opinions/the-fundamental-error-in-the-cbos-health-care-projections/2017/07/14/25f0d8a4-67ee-11e7-a1d7-9a32c91c6f40_story.html.

14. Hayden Dublois, “Broken promises: Why expanded ObamaCare subsidies must expire on time,” Foundation for Government Accountability (2022), https://thefga.org/research/broken-promises-why-expanded-obamacare-subsidies-must-expire-on-time.

15. Centers for Medicare and Medicaid Services, “Marketplace 2024 open enrollment period report: Final national snapshot,” Centers for Medicare and Medicaid Services (2024), https://www.cms.gov/newsroom/fact-sheets/marketplace-2024-open-enrollment-period-report-final-national-snapshot.

16. Philip Swagel, “Re: Health insurance policies,” Congressional Budget Office (2022), https://www.cbo.gov/system/files?file=2022-07/58313-Crapo_letter.pdf.

17. Carolyn Ugolino et al, “Federal subsidies for health insurance coverage for people under 65: 2022 to 2032,” Congressional Budget Office (2022), https://www.cbo.gov/system/files/2022-06/57962-health-insurance-subsidies.pdf.

18. Jessica Hale et al, “Federal subsidies for health insurance: 2023 to 2033,” Congressional Budget Office (2023), https://www.cbo.gov/system/files/2023-09/59273-health-coverage.pdf.

19. Carolyn Ugolino et al, “Federal subsidies for health insurance coverage for people under 65: 2022 to 2032,” Congressional Budget Office (2022), https://www.cbo.gov/system/files/2022-06/57962-health-insurance-subsidies.pdf.

20. Jessica Hale et al, “Federal subsidies for health insurance: 2023 to 2033,” Congressional Budget Office (2023), https://www.cbo.gov/system/files/2023-09/59273-health-coverage.pdf.

21. Hayden Dublois et al., “Millions of ineligible Medicaid enrollees come at a high cost to states,” Foundation for Government Accountability (2022), https://thefga.org/research/ineligible-medicaid-enrollees-high-cost.

22. Ibid.

23. Ibid.

24. Philip Swagel, “Re: Preliminary estimate of the effects of H.R. 6201, the Families First Coronavirus Response Act,” Congressional Budget Office (2020), https://www.cbo.gov/system/files/2020-04/HR6201.pdf.

25. Centers for Medicare and Medicaid Services, “Medicaid CMS-64 FFCRA and CAA increased FMAP expenditure data collected through MBES,” Centers for Medicare and Medicaid Services (2024), https://www.medicaid.gov/medicaid/financial-management/state-budget-expenditure-reporting-for-medicaid-and-chip/expenditure-reports-mbescbes/medicaid-cms-64-ffcra-and-caa-increased-fmap-expenditure-data-collected-through-mbes/index.html.

26. Congressional Budget Office, “Medicaid—CBO’s baseline as of March 6, 2020,” Congressional Budget Office (2020), https://www.cbo.gov/system/files/2020-03/51301-2020-03-medicaid.pdf.

27. Centers for Medicare and Medicaid Services, “Expenditure reports from MBES/CBES,” Centers for Medicare and Medicaid Services (2024), https://www.medicaid.gov/medicaid/financial-management/state-expenditure-reporting-for-medicaid-chip/expenditure-reports-mbescbes/index.html.

28. Hayden Dublois, “‘Fixing’ what wasn’t broken: Why Biden’s Child Tax Credit scheme is a recipe for failure,” Foundation for Government Accountability (2021), https://thefga.org/wp-content/uploads/2021/08/Biden-Child-Tax-Credit-Failure.pdf.

29. Congressional Budget Office, “Reconciliation recommendations of the House Committee on Ways and Means,” Congressional Budget Office (2021), https://www.cbo.gov/system/files/2021-02/hwaysandmeansreconciliation.pdf.  

30. Philip Swagel, “Re: Budgetary effects of making specified policies in the Build Back Better Act permanent,” Congressional Budget Office (2021), https://www.cbo.gov/system/files/2021-12/57673-BBBA-GrahamSmith-Letter.pdf.

31. Barry Blom et al, “The accuracy of CBO’s budget projections for Fiscal Year 2022,” Congressional Budget Office (2023), https://www.cbo.gov/publication/58893.

32. Congressional Research Service, “Unemployment insurance provisions in the American Recovery and Reinvestment Act of 2009,” Congressional Research Service (2010), https://crsreports.congress.gov/product/pdf/R/R40368.

33. Douglas Elmendorf, “Conference agreement for H.R. 1, the American Recovery and Reinvestment Act of 2009,” Congressional Budget Office (2009), https://www.cbo.gov/sites/default/files/111th-congress-2009-2010/costestimate/hr1conference0.pdf.

34. Molly Dahl et al, “Estimated impact of the American Recovery and Reinvestment Act on employment and economic output in 2014,” Congressional Budget Office (2015), https://www.cbo.gov/sites/default/files/114th-congress-2015-2016/reports/49958-ARRA.pdf.

35. U.S. Senate Committee on the Budget, “The economy since 2009: A chart book on economic life in working America,” U.S. Congress (2014), https://www.budget.senate.gov/newsroom/budget-background/the-economy-since-2009.

36. Douglas Elmendorf, “Conference agreement for H.R. 1, the American Recovery and Reinvestment Act of 2009,” Congressional Budget Office (2009), https://www.cbo.gov/sites/default/files/111th-congress-2009-2010/costestimate/hr1conference0.pdf.

37. Molly Dahl et al, “A review of CBO’s estimate of the effects of the Recovery Act on SNAP,” Congressional Budget Office (2019), https://www.cbo.gov/system/files/2019-01/54864-SNAP_ARRA.pdf.

38. Ibid.

39. Congressional Research Service, “Reducing SNAP (food stamp) benefits provided by the ARRA: P.L. 111-226 and P.L. 111-296,” Congressional Research Service (2010), https://crsreports.congress.gov/product/pdf/R/R41374/5.

40. Molly Dahl et al, “A review of CBO’s estimate of the effects of the Recovery Act on SNAP,” Congressional Budget Office (2019), https://www.cbo.gov/system/files/2019-01/54864-SNAP_ARRA.pdf.

41. Hayden Dublois, “Food stamp boosts are bankrupting taxpayers,” Foundation for Government Accountability (2023), https://thefga.org/research/food-stamp-boosts-are-bankrupting-taxpayers.

42. Philip Swagel, “Re: Preliminary estimate of the effects of H.R. 6201, the Families First Coronavirus Response Act,” Congressional Budget Office (2020), https://www.cbo.gov/system/files/2020-04/HR6201.pdf.

43. Food and Nutrition Service, “SNAP COVID-19 emergency allotments guidance,” U.S. Department of Agriculture (2023), https://www.fns.usda.gov/snap/covid-19-emergency-allotments-guidance.

44. Ibid.

45. Congressional Budget Office, “Estimated budgetary effects of Public Law 117-169, to provide for reconciliation pursuant to Title II of S. Con. Res. 14,” Congressional Budget Office (2022), https://www.cbo.gov/publication/58455.

46. Julie Topoleski  et al, “The budget and economic outlook: 2024 to 2034,” Congressional Budget Office (2024), https://www.budget.senate.gov/imo/media/doc/the_budget_and_economic_outlook_2024_to_2034.pdf?utm_source=Daily%20on%20Energy%20020824%20a_02/08/2024&utm_medium=email&utm_campaign=WEX_Daily%20on%20Energy&rid=24439411&env=03fbeee7e76969829ee40f5e1b5fcc8054fb.

47. Ibid.

48. Ibid.

49. Ibid.

50. Peter G. Peterson Foundation, “Decade in review: Looking back at CBO projections then and now,” Peter G. Peterson Foundation (2020), https://www.pgpf.org/blog/2023/03/decade-in-review-looking-back-at-cbo-projections-then-and-now.

51. Ibid.

52. Aaron Feinstein et al, “The accuracy of CBO’s budget projections for fiscal year 2023,” Congressional Budget Office (2023), https://www.cbo.gov/publication/59682.

53. Ammar Inayatali, “Always in motion is the future: The accuracy of historical CBO debt projections,” U.C. Berkley (2019), https://www.econ.berkeley.edu/sites/default/files/Inayatali_Ammar_Thesis.pdf.

54. Chris Edwards, “CBO forecast accuracy,” Cato Institute (2012), https://www.cato.org/blog/cbo-forecast-accuracy.

55. Ibid.

56. Ibid.