CONGRESS OF THE UNITED STATES CONGRESSIONAL BUDGET OFFICE The 2019 Long-Term Budget Outlook Percentage of Gross Domestic Product 6 Actual Projected 4 2 0 Primary −2 Deficit −4 Net −6 Interest −8 Total −10 Deficit −12 2004 2009 2014 2019 2024 2029 2034 2039 2044 2049 JUNE 2019 At a Glance Each year, the Congressional Budget Office publishes a report presenting its budget projections for the next 30 years. Those extended baseline projections generally reflect current law. This report is the latest in the series. •• Debt in CBO’s Extended Baseline Projections. Large budget deficits over the next 30 years are projected to drive federal debt held by the public to unprecedented levels—from 78 percent of gross domestic product (GDP) in 2019 to 144 percent by 2049. That projection incorporates CBO’s central estimates of various factors, such as productivity growth and interest rates on federal debt. CBO’s analysis indicates that even if values for those factors differed from the agency’s projections, debt several decades from now would probably be much higher than it is today. •• Other Possible Outcomes. The agency’s projections of debt are highly sensitive to changes in the factors underlying them. For example, if the growth of total factor productivity in the nonfarm business sector was one-half of one percentage point lower each year than CBO projects, all else being equal, debt in 2049 would be 185 percent of GDP; if such growth was one-half of one percentage point higher, debt that year would be 106 percent of GDP. If interest rates were one percentage point higher each year than CBO projects, debt in 2049 would be 199 percent of GDP; if they were one percentage point lower, debt that year would be 107 percent of GDP. •• Debt Under Alternative Scenarios. If lawmakers changed current laws to maintain certain major policies now in place—most significantly, if they prevented a cut in discretionary spending in 2020 and an increase in individual income taxes in 2026—then debt held by the public would increase even more, reaching 219 percent of GDP by 2049. By contrast, if Social Security benefits were limited to the amounts payable from revenues received by the Social Security trust funds, debt in 2049 would reach 106 percent of GDP, still well above its current level. •• Interest Costs. The projected increase in federal borrowing would lead to significantly higher interest costs. In CBO’s extended baseline projections, net outlays for interest more than triple in relation to the size of the economy over the next three decades, exceeding all discretionary spending by 2046. •• Noninterest Spending. Mainly owing to the aging of the population, spending for Social Security and the major health care programs (primarily Medicare) is projected to rise as a percentage of GDP over the coming decades. The growth of spending for Medicare and the other major health care programs is also driven by rising health care costs per person. •• Revenues. Measured as a percentage of GDP, revenues are projected to be roughly flat over the next few years, rise slowly, and then jump in 2026 because of the scheduled expiration of certain provisions of the 2017 tax act. Thereafter, they continue to rise—but they do not keep pace with growth in spending. The factor contributing most to the long-term growth in revenues is the increasing share of income that is pushed into higher tax brackets. •• Comparison With Last Year’s Projections. Debt as a percentage of GDP in 2048 is 11 percentage points lower in this year’s extended baseline projections than it was in last year’s. That difference is largely driven by spending projections that are lower than they were last year. www.cbo.gov/publication/55331 Contents Visual Summary 1 1 The Budget Outlook for the Next 30 Years 5 Overview5 Consequences of High and Rising Federal Debt 9 Demographic and Economic Trends Underlying CBO’s Long-Term Projections 14 Projected Spending Through 2049 18 Projected Revenues Through 2049 28 Uncertainty of CBO’s Long-Term Projections 30 BOX 1-1. SENSITIVITY OF BUDGET PROJECTIONS TO CHANGES IN UNDERLYING ECONOMIC FACTORS 32 Changes From Last Year’s Long-Term Budget Outlook 35 2 The Long-Term Budget Outlook Under Alternative Scenarios 37 Overview37 Budgetary and Economic Effects of an Alternative Fiscal Scenario 38 Budgetary and Economic Effects of a Payable-Benefits Scenario 41 The Size and Timing of Policy Changes Needed to Meet Various Goals for Deficit Reduction 44 A CBO’s Projections of Demographic and Economic Trends 49 Demographic Factors 49 Economic Factors 53 B Changes in Long-Term Budget Projections Since June 2018 65 Changes in Projected Spending 66 Changes in Projected Revenues 70 Changes in Social Security’s Projected Finances 70 Changes in Analyzing Uncertainty 71 Changes in Alternative Scenarios for Fiscal Policy 72 List of Tables and Figures 74 About This Document 75 Notes The Congressional Budget Office’s extended baseline projections show the budget’s long-term path under most of the same assumptions that the agency uses, in accordance with statutory requirements, in constructing its 10-year baseline projections. Both sets of projections incorporate the assumptions that current law generally remains unchanged, that some mandatory programs are nevertheless extended after their authorizations lapse, and that spending for Medicare and Social Security continues as scheduled even if their trust funds are exhausted. Unless this report indicates otherwise, all projections shown in the figures are extended baseline projections. Unless the report indicates otherwise, the years that it refers to are federal fiscal years, which run from October 1 to September 30 and are designated by the calendar year in which they end. Budgetary values, such as the ratio of debt or deficits to gross domestic product, are calculated on a fiscal year basis; economic variables, such as gross national product or interest rates, are calculated on a calendar year basis. When October 1 (the first day of the fiscal year) falls on a weekend, certain payments that would have ordinarily been made on that day are instead made at the end of September and thus are shifted into the previous fiscal year. All budget projections in this report have been adjusted to exclude the effects of those timing shifts. For the dollar amounts of payments that are shifted in CBO’s 10-year baseline budget projections, see Congressional Budget Office, Updated Budget Projections: 2019 to 2029 (May 2019), Table 2, www.cbo.gov/publication/55151. Numbers in the text, tables, and figures may not add up to totals because of rounding. Unless the report specifies otherwise, Medicare outlays are presented net of offsetting receipts, which reduce outlays for the program. As referred to in this report, the Affordable Care Act comprises the Patient Protection and Affordable Care Act; the health care provisions of the Health Care and Education Reconciliation Act of 2010; and the effects of subsequent judicial decisions, statutory changes, and administrative actions. Data and supplemental information files—the data underlying the figures in this report, supplemental budget projections, and the demographic and economic variables underlying those projections—are posted along with the report on CBO’s website (www.cbo.gov/ publication/55331). Previous editions of this report are also available on the website (https://go.usa.gov/xmezZ). Visual Summary Each year, the Congressional Budget Office issues a set of long-term budget projections—that is, projections of what federal spending, revenues, deficits, and debt would be for the next 30 years if current laws generally did not change. CBO calls them extended baseline projections. This year’s projections of federal debt are slightly lower than last year’s, mainly because CBO has reduced its projections of discretionary and net interest spending. Those reductions are partially offset by a small reduction in projected revenues. Debt and Deficits Percentage of Gross Domestic Product 160 Actual Projected 120 World War II In CBO’s projections, federal debt held by the 80 public totals 144 percent Great of gross domestic Depression product (GDP) in 2049, an 40 World War I unprecedented level. Civil War 0 1790 1810 1830 1850 1870 1890 1910 1930 1950 1970 1990 2010 2030 2050 See Figure 1-1 Percentage of Gross Domestic Product 6 Actual Projected 4 2 Deficits grow from 0 4.2 percent of GDP in 2019 −2 Primary Deficit to 8.7 percent in 2049, driving up debt. Net spending −4 for interest on debt accounts −6 Net Interest for most of the growth in total deficits. −8 Total Deficit −10 −12 2004 2009 2014 2019 2024 2029 2034 2039 2044 2049 See Figure 1-4 2 THE 2019 LONG-TERM BUDGET OUTLOOK JUNE 2019 Debt and Deficits (Continued) Percentage of Gross Domestic Product 30 Actual Projected Outlays 20 Revenues Deficits grow because growth in spending outpaces growth in revenues. 10 0 2004 2009 2014 2019 2024 2029 2034 2039 2044 2049 See Figure 1-3 Revenues In CBO’s projections, federal revenues increase as a percentage of GDP—from 16.5 percent in 2019 to 19.5 percent in 2049. Percentage of Gross Domestic Product 15 Actual Projected Individual 10 Income Taxes Increases in receipts from individual income taxes Payroll Taxes account for most of the rise 5 in total revenues. Other Revenues Corporate 0 Income Taxes 2004 2009 2014 2019 2024 2029 2034 2039 2044 2049 See Figure 1-3 Percentage of Gross Domestic Product 4 Increases in Federal Revenues Other Factors The largest source of 3 Tax on High-Premium growth in tax revenues is Health Insurance Plans real bracket creep—the 2 Expiration of process in which, as income Major Provisions of rises faster than inflation, a the 2017 Tax Act larger proportion of income 1 becomes subject to higher Real Bracket Creep tax rates. 0 2019 2024 2029 2034 2039 2044 2049 See Figure 1-13 VISUAL SUMMARY THE 2019 LONG-TERM BUDGET OUTLOOK 3 Spending Federal spending grows from 20.7 percent of GDP today to 28.2 percent in 2049. Percentage of Gross Domestic Product 15 Actual Projected Spending increases, as a 10 Major Health percentage of GDP, for net Care Programs interest, the major health Social Security care programs, and Social Security. That spending Net Interest growth is partially offset by 5 Discretionary declining spending for other Spending programs. Other Mandatory Spending 0 2004 2009 2014 2019 2024 2029 2034 2039 2044 2049 See Figure 1-3 Millions of People 400 Actual Projected Age 65 or Older Much of the spending growth for Social Security and 300 Medicare results from the aging of the population. As baby boomers age and as 200 Ages 20 to 64 life expectancy continues to increase, the percentage of the population age 65 or 100 older will grow significantly, boosting the number of Ages 0 to 19 beneficiaries of those programs. 0 2004 2009 2014 2019 2024 2029 2034 2039 2044 2049 See Figure 1-5 Percent Annual Growth of Medicare Annual Growth of Medicaid 6 Costs per Beneficiary Costs per Beneficiary Growth in spending on Medicare and the other major 5 health care programs is also driven by rising health care 1.1 1.6 costs per beneficiary, as it 4 0.7 Excess Cost Growth 1.1 has been in the past. Excess cost growth is the extent to 3 which growth in health care costs per person, adjusted 2 3.8 3.8 to remove the effects of 3.4 3.4 Growth of aging, exceeds growth in 1 Potential GDP potential GDP (the economy’s per Person maximum sustainable output) 0 per person. 1985−2017 2019−2049 1985−2017 2019−2049 See Figure 1-11 4 THE 2019 LONG-TERM BUDGET OUTLOOK JUNE 2019 Uncertainty Percentage of Gross Domestic Product 180 Actual Projected 160 The economic and demographic variables 140 Debt Held by used to construct CBO’s projections are 120 the Public Extended Baseline uncertain. But even if their values differed 100 from those underlying the extended 80 baseline projections, in 20 years, federal 60 Middle Two-Thirds of the debt would probably be much higher than it 40 Range of Possible Outcomes is today, if current laws generally remained 20 0 unchanged. 2004 2009 2014 2019 2024 2029 2034 2039 See Figure 1-15 Alternative Scenarios Percentage of Gross Domestic Product 220 Actual Projected 219 Extended Alternative In relation to the extended 200 Fiscal Scenario baseline projections, debt 180 would be greater under an 160 extended alternative fiscal 140 144 Extended Baseline scenario (in which certain 120 major policies now in place 100 106 Payable-Benefits Scenario would be maintained) and 80 less under a payable-benefits 60 scenario (in which Social 40 Security benefits would be 20 limited to the program’s total 0 annual revenues after 2032). 2004 2009 2014 2019 2024 2029 2034 2039 2044 2049 See Figure 2-1 Various Goals for Deficit Reduction Percentage of Gross Domestic Product The reduction in each year’s primary deficit needed to Starting Year make federal debt held by the public in 2049 equal . . . 1.8 . . . its current share of GDP (78 percent) 2020 2.9 . . . its 50-year average (42 percent) The longer policymakers 2.2 waited to reduce primary 2025 deficits, the larger those 3.5 reductions would have to be. 2.7 2030 4.4 0 1 2 3 4 5 See Figure 2-3 CHAPTER 1 Chapter 1 The Budget Outlook for the Next 30 Years Overview represent the agency’s best assessment of future spending, By the end of this year, federal debt held by the public is revenues, deficits, and debt under the assumption that projected to equal 78 percent of gross domestic prod- current laws generally remain unchanged. In doing so, uct (GDP)—its highest level since shortly after World they give lawmakers a point of comparison from which War II. If current laws generally remained unchanged, to measure the effects of proposed legislation. growing budget deficits would boost federal debt dras- tically over the next 30 years, the Congressional Budget How Federal Debt Has Grown in Recent Years Office projects. Debt would reach 92 percent of GDP Debt held by the public is the amount that the federal by the end of the next decade and 144 percent by 2049 government has borrowed in financial markets by issuing (see Table 1-1). That level of debt would be the highest Treasury securities to pay for its operations and activi- in the nation’s history by far, and it would be on track to ties.2 Debt as a percentage of GDP is a useful measure increase even more. Although long-term projections are for comparing amounts of debt in different years because very uncertain, in CBO’s assessment, even if a key set of it removes the effects of changes in prices, population, factors, including productivity growth and interest rates, output, and income—all of which affect the nation’s abil- were favorable for the fiscal situation over the next three ity to finance the debt. That measure places the effects of decades, debt as a share of GDP would most likely rise if potential adjustments to the budget within the context current laws remained unchanged. If lawmakers changed of the nation’s resources. Examining whether debt as a current laws to maintain certain policies now in place— percentage of GDP is increasing is therefore a simple and most significantly, if they prevented a cut in discretionary meaningful way to assess the budget’s sustainability. spending in 2020 and an increase in individual income taxes in 2026—the result would be even larger increases Federal debt held by the public has ballooned over the in debt (see Chapter 2). The prospect of such high and past decade. At the end of 2007, federal debt stood at rising debt poses substantial risks for the nation and pres- 35 percent of GDP, but deficits arising from the 2007– ents policymakers with significant challenges. 2009 recession and subsequent policies caused debt to grow sizably in relation to the economy over the next What CBO’s Projections Represent five years. By the end of 2012, debt as a share of GDP The long-term projections of federal spending, revenues, had doubled, reaching 70 percent. The upward trajectory deficits, and debt in this report are consistent with the has generally continued since then, and debt is projected 10-year baseline budget projections that CBO published to be 78 percent of GDP by the end of this year—a in May 2019 and the economic forecast that the agency very high level by historical standards. (Over the past published in January 2019.1 They extend most of the concepts underlying those 10-year budget projections 2. When the federal government borrows in financial markets, it for an additional 20 years, and they reflect the macro- competes with other participants for financial resources and, economic effects of projected fiscal policy over that in the long term, crowds out private investment, thus reducing 30-year period. Together, those long-term projections economic output and income. By contrast, federal debt held by constitute the agency’s extended baseline projections. trust funds and other government accounts represents internal transactions of the government and does not directly affect financial markets. (Together, that debt and debt held by the CBO’s 10-year and extended baseline projections are public make up gross federal debt.) For more discussion, see not predictions of budgetary outcomes. Rather, they Congressional Budget Office, Federal Debt and Interest Costs (December 2010), www.cbo.gov/publication/21960. Several 1. See Congressional Budget Office, Updated Budget Projections: factors not directly included in the budget totals also affect the 2019 to 2029 (May 2019), www.cbo.gov/publication/55151, and government’s need to borrow from the public. They include The Budget and Economic Outlook: 2019 to 2029 (January 2019), fluctuations in the government’s cash balance as well as the cash www.cbo.gov/publication/54918. flows of the financing accounts used for federal credit programs. 6 The 2019 Long-term budget outlook june 2019 Table 1-1 . CBO’s Extended Baseline Projections Percentage of Gross Domestic Product Projected Annual Average 2019 2020–2029 2030–2039 2040–2049 Revenues Individual income taxes 8.2 8.9 9.9 10.4 Payroll taxes 5.8 5.9 5.9 5.9 Corporate income taxes 1.2 1.4 1.3 1.3 Other a 1.3 1.3 1.4 1.6 Total Revenues 16.5 17.5 18.5 19.2 Outlays Mandatory Social Security 4.9 5.5 6.2 6.2 Major health care programs b 5.2 6.0 7.5 8.8 Other 2.6 2.4 2.2 2.1 Subtotal 12.7 13.9 15.9 17.1 Discretionary 6.3 5.3 5.0 5.0 Net interest 1.8 2.6 3.5 4.9 Total Outlays 20.7 21.9 24.3 27.1 Deficit -4.2 -4.3 -5.8 -7.9 Debt Held by the Public at the End of the Period 78 92 113 144 Memorandum: Social Security Revenues c 4.4 4.5 4.5 4.4 Outlays d 4.9 5.5 6.2 6.2 Contribution to the Federal Deficit e -0.6 -1.1 -1.7 -1.8 Medicare Revenues c 1.4 1.5 1.6 1.7 Outlays d 3.6 4.4 5.8 7.0 Offsetting receipts -0.6 -0.8 -1.0 -1.3 Contribution to the Federal Deficit e -1.5 -2.1 -3.2 -4.1 Gross Domestic Product at the End of the Period (Trillions of dollars) 21.3 31.0 45.2 66.5 Source: Congressional Budget Office. This table satisfies a requirement specified in section 3111 of S. Con. Res. 11, the Concurrent Resolution on the Budget for Fiscal Year 2016. The extended baseline projections generally reflect current law, following CBO’s 10-year baseline budget projections through 2029 and then extending most of the concepts underlying those projections for the rest of the long-term projection period. a.Consists of excise taxes, remittances to the Treasury from the Federal Reserve System, customs duties, estate and gift taxes, and miscellaneous fees and fines. b.Consists of spending for Medicare (net of premiums and other offsetting receipts), Medicaid, and the Children’s Health Insurance Program, as well as outlays to subsidize health insurance purchased through the marketplaces established under the Affordable Care Act and related spending. c.Includes all payroll taxes for the program other than those paid by the federal government on behalf of its employees (which are intragovernmental transactions). Also includes income taxes paid on Social Security benefits, which are credited to the trust funds. Excludes interest credited to the trust funds. d.Excludes discretionary outlays related to administration of the program. e.The contribution to the deficit shown here differs from the change in the trust fund balance for the program because it excludes intragovernmental transactions, interest earned on balances, and outlays related to administration of the program. CHAPTER 1: THE BUDGET OUTLOOK FOR THE NEXT 30 YEARS The 2019 Long-term budget outlook 7 Figure 1-1 . Federal Debt Held by the Public Since 1790 Percentage of Gross Domestic Product 160 Actual Projected 120 World War II Growing deficits are projected to drive federal debt held by 80 the public to unprecedented levels over the next 30 years. By Great 2049, debt is projected to reach Depression 144 percent of gross domestic 40 product. Civil War World War I 0 1790 1810 1830 1850 1870 1890 1910 1930 1950 1970 1990 2010 2030 2050 Source: Congressional Budget Office. 50 years, such debt has averaged 42 percent of GDP.) It the 2007–2009 recession, and the other four followed a has exceeded 70 percent of GDP during only one other double-dip recession in the early 1980s.) From 2019 to period in U.S. history—from 1944 to 1950 follow- 2029, projected deficits average 4.3 percent of GDP— ing the surge in federal spending that occurred during nearly one-and-a-half times the average over the past World War II (see Figure 1-1). 50 years. Why Debt Is Projected to Grow In CBO’s projections, mandatory spending increases The total amount of debt is projected to increase each from 12.7 percent of GDP in 2019 to 14.9 percent in year as the government runs budget deficits. If current 2029. In contrast, discretionary spending decreases in laws generally remained unchanged, federal budget defi- relation to the size of the economy over that period— cits would grow substantially over the next 30 years (see from 6.3 percent of GDP in 2019 to 5.0 percent in Figure 1-2). In CBO’s projections, that increase occurs 2029. Revenues increase from 16.5 percent of GDP in because mandatory spending—in particular, outlays for 2019 to 18.3 percent in 2029. (A large portion of that Social Security and the major health care programs—and increase is attributable to the expiration of nearly all of interest payments on federal debt grow faster than reve- the individual income tax provisions of the 2017 tax act, nues (see Figure 1-3 on page 10). Public Law 115-97.) 2019 to 2029. Deficits (adjusted to exclude the effects of As a result of those changes in spending and revenues, shifts in the timing of certain payments) are projected to primary deficits (deficits excluding net spending for increase from 4.2 percent of GDP in 2019 to 4.5 percent interest) shrink in CBO’s projections, falling from of GDP by 2029—a level that has been exceeded in only 2.4 percent of GDP in 2019 to 1.6 percent in 2029. But eight years since 1946.3 (Four of those years followed growing debt and rising average interest rates on federal 3. When October 1 (the first day of the fiscal year) falls on a and 2029 to fiscal years 2022, 2023, and 2028. Those shifts will weekend, certain payments that would have ordinarily been made noticeably boost spending and the deficit in fiscal years 2022 and on that day are instead made at the end of September and thus 2028 and reduce spending and deficits in fiscal years 2024 and are shifted into the previous fiscal year. Over the next decade, 2029. No adjustments were made for timing shifts after the first certain payments will be shifted from fiscal years 2023, 2024, decade of the projection period. 8 The 2019 Long-term budget outlook june 2019 Figure 1-2 . The Federal Budget in 2019 and 2049 Percentage of Gross Domestic Product 2019 2049 30 Net Interest Deficit 20 Other Noninterest Other Revenuesc If current laws generally remained Spendinga Corporate Income Taxes unchanged, spending would grow faster than revenues over the next Payroll Taxes 30 years, causing deficits to increase Major Health Care 10 Programsb substantially. Individual Income Taxes Social Security 0 Outlays Revenues Outlays Revenues Source: Congressional Budget Office. a.Consists of all federal spending other than that for Social Security, the major health care programs, and net interest. b.Consists of spending for Medicare (net of premiums and other offsetting receipts), Medicaid, and the Children’s Health Insurance Program, as well as outlays to subsidize health insurance purchased through the marketplaces established under the Affordable Care Act and related spending. c.Consists of excise taxes, remittances to the Treasury from the Federal Reserve System, customs duties, estate and gift taxes, and miscellaneous fees and fines. debt increase net interest costs from 1.8 percent of GDP on high-premium health plans that is scheduled to take in 2019 to 3.0 percent in 2029. The resulting increase in effect in 2022. net outlays for interest more than offsets the decrease in primary deficits. As a result of those developments, primary deficits increase over the last two decades of the projection 2029 to 2049. Deficits continue to grow beyond the period, reaching 2.8 percent of GDP in 2039 and first 10 years in CBO’s extended baseline projections, 3.0 percent by 2049 (see Figure 1-4 on page 12). And rising from 4.5 percent of GDP in 2029 to 6.8 percent because in CBO’s projections federal debt is already high by 2039 and 8.7 percent by 2049 (an amount exceeded at the end of the next decade and interest rates continue only in 2009, following the last recession). In the last to rise, net outlays for interest increase from 3.0 percent two decades of the projection period, deficits average of GDP in 2029 to 5.7 percent in 2049, adding substan- 6.9 percent of GDP. tially to projected deficits. After 2029, mandatory spending continues to increase How CBO Analyzes the Uncertainty of Its Projections faster than economic output, reaching 16.6 percent of Long-term projections are very uncertain. CBO there- GDP in 2039 and 17.5 percent in 2049, whereas discre- fore examined the extent to which federal debt would tionary spending increases only slightly, to 5.1 percent differ from the extended baseline projections if a set of in 2049. Revenues also rise, although not as quickly as key factors—several demographic and economic fac- spending. They increase because of real bracket creep (the tors as well as the growth of health care costs—deviated process in which an ever-larger proportion of income from the paths underlying those projections. In CBO’s becomes subject to higher tax rates as income rises faster assessment, there is about a two-thirds chance that fed- than inflation) and because of collections from the tax eral debt would be between 71 percent and 175 percent of GDP in 2039. That range indicates that if current CHAPTER 1: THE BUDGET OUTLOOK FOR THE NEXT 30 YEARS The 2019 Long-term budget outlook 9 laws generally remained unchanged, in 20 years federal •• Increase the likelihood of less abrupt, but still debt—which is already high by historical standards— significant, negative economic and financial effects, would probably be much higher than it is today.4 such as expectations of higher rates of inflation and more difficulty financing public and private activity In addition to estimating that likely range by simulating in international markets. budgetary outcomes when the values for all of the key factors varied simultaneously, the agency examined the In addition, high debt might cause policymakers to feel sensitivity of its projections to higher or lower values for restrained from implementing deficit-financed fiscal some of those factors individually. For example, if growth policy to respond to unforeseen events or for other of total factor productivity in the nonfarm business sec- purposes, such as to promote economic activity or tor was 0.5 percentage points faster than CBO’s central strengthen national defense. estimate, in 2049 federal debt held by the public would be 106 percent of GDP; if such growth was 0.5 percent- Not all effects of the projected path of debt are negative. age points slower, debt would be 185 percent of GDP. In addition to allowing policymakers to achieve goals for Or if federal borrowing rates were 1.0 percentage point spending and revenue policies under current law, that lower than CBO’s central estimate, in 2049 debt would path would cause interest rates to be higher than they be 107 percent of GDP; if they were 1.0 percentage otherwise would be, giving the Federal Reserve more point higher, debt would be 199 percent of GDP. flexibility in implementing monetary policy. (Higher interest rates would also have adverse economic effects, as Consequences of High and described below.) Rising Federal Debt If federal debt as a percentage of GDP continues to rise If policymakers understand the potential effects of high at the pace that CBO projects it would under current and rising debt, they may be better equipped to weigh law, the economy would be affected in two significant the consequences of fiscal policy under current law ways: against those of proposed changes to law. In all likeli- hood, if policymakers postponed fiscal tightening and •• That debt path would dampen economic output over debt as a share of GDP continued to rise, the changes time, and necessary to stabilize debt would place an even greater burden on future generations. •• Rising interest costs associated with that debt would increase interest payments to foreign debt holders Effects Incorporated in CBO’s and thus reduce the income of U.S. households by Extended Baseline Projections increasing amounts. The path of federal borrowing in CBO’s extended base- line projections would have negative economic conse- That debt path would also pose significant risks to the quences over the longer term. CBO projects that rising fiscal and economic outlook, although those risks are not debt would crowd out the resources available for private currently apparent in financial markets. In particular, investment, reducing the growth of economic output that path would have the following effects: and income. In addition, rising interest payments would result in increasingly large payments to foreign investors •• Increase the risk of a fiscal crisis—that is, a situation and thus further dampen domestic income. in which the interest rate on federal debt rises abruptly because investors have lost confidence in the Crowding Out of Private Investment. The projected U.S. government’s fiscal position—and path of federal borrowing would reduce output in the long run. When the government borrows, it borrows 4. The range of likely outcomes that CBO’s models produce is less from people and businesses whose saving would other- informative after 20 years because the key parameters governing wise finance private investment in productive capital, the economic effects of fiscal policy in the agency’s models such as factories and computers. Although an increase are based on the nation’s historical experience with federal in government borrowing strengthens the incentive to borrowing. At the high end of such a range for 30 years in the future, projections of debt as a percentage of GDP would grow to save—in part by increasing interest rates—the resulting amounts well outside historical experience. rise in private saving is not as large as the increase in 10 The 2019 Long-term budget outlook june 2019 Figure 1-3 . CBO’s Budget Projections in Brief Percentage of GDP 200 Actual Projected 150 Federal Debt Held by the Public In CBO’s projections, federal debt 100 held by the public rises . . . 50 0 2004 2009 2014 2019 2024 2029 2034 2039 2044 2049 30 Actual Projected Outlays 20 . . . because growth in total Revenues spending outpaces growth in total revenues, resulting in larger budget deficits. 10 0 2004 2009 2014 2019 2024 2029 2034 2039 2044 2049 Continued government borrowing; national saving, or the amount national saving does in response to government deficits, of domestic resources available for private investment, however, because the higher interest rates that are likely therefore declines.5 Private investment falls less than to result from increased federal borrowing tend to attract more foreign capital to the United States. 5. In CBO’s assessment, another reason that an increase in government borrowing strengthens the incentive to save is that If investment in capital goods declined, workers would, some people expect that policymakers will raise taxes or cut on average, have less capital to use in their jobs. As a spending in the future to cover the cost of paying interest on the result, they would be less productive, their compensation additional federal debt. As a result, some of those people increase their saving to prepare for paying higher taxes or receiving less in benefits. For further discussion of that effect and the on National Saving and Private Domestic Investment, Working estimated effect of federal borrowing on private investment, see Paper 2014-02 (Congressional Budget Office, February 2014), Jonathan Huntley, The Long-Run Effects of Federal Budget Deficits www.cbo.gov/publication/45140. CHAPTER 1: THE BUDGET OUTLOOK FOR THE NEXT 30 YEARS The 2019 Long-term budget outlook 11 Figure 1-3. Continued CBO’s Budget Projections in Brief Percentage of GDP Outlays 15 Actual Projected Spending on certain components of the 10 Major Health Care Programsa budget—Social Security, the major health care Social Security programs, and net interest—is projected to Net Interest rise in relation to GDP; 5 Discretionary Spending other spending, in total, is projected to decline. Other Mandatory Spending 0 2004 2009 2014 2019 2024 2029 2034 2039 2044 2049 Revenues 15 Actual Projected Increases in individual Individual Income Taxes income taxes account for 10 most of the rise in total revenues relative to GDP. Receipts from all other sources, taken together, Payroll Taxes are projected to be slightly 5 higher in 2049 than they are today. Other Revenuesb Corporate Income Taxes 0 2004 2009 2014 2019 2024 2029 2034 2039 2044 2049 Source: Congressional Budget Office. GDP = gross domestic product. a.Consists of spending for Medicare (net of premiums and other offsetting receipts), Medicaid, and the Children’s Health Insurance Program, as well as outlays to subsidize health insurance purchased through the marketplaces established under the Affordable Care Act and related spending. b.Consists of excise taxes, remittances to the Treasury from the Federal Reserve System, customs duties, estate and gift taxes, and miscellaneous fees and fines. would be lower, and they would thus be less inclined to extended baseline projections. As a result, GDP would work. Those effects would increase over time as federal be 4.3 percent higher in 2049 than it is in the extended borrowing grew. As an example of the benefits of lower baseline projections, and GDP per person in 2049 would debt, in CBO’s estimate, budgetary changes that entailed be about $4,200 higher (in 2019 dollars). steadily reducing debt over 30 years to 42 percent of GDP (its average over the past 50 years) would, all else Rising Interest Payments. The projected increase in fed- being equal, boost economic growth each year by about eral borrowing would also drive up interest costs, increas- 0.1 percentage point in relation to growth in the agency’s ing the burden of interest outlays in the federal budget. 12 The 2019 Long-term budget outlook june 2019 Figure 1-4 . Total Deficit, Primary Deficit, and Net Interest Percentage of Gross Domestic Product 6 Actual Projected 4 2 Although rising revenues and 0 shrinking discretionary spending −2 Primary Deficit are projected to decrease primary deficits as a percentage of gross −4 domestic product from 2019 to −6 2029, total deficits remain large Net Interest because of rising net spending for −8 interest. Total Deficit −10 −12 2004 2009 2014 2019 2024 2029 2034 2039 2044 2049 Source: Congressional Budget Office. Primary deficits or surpluses exclude net spending for interest. In CBO’s extended baseline projections, net interest out- from 2.4 percent in 2019 to 4.2 percent by 2049. The lays grow from 1.8 percent of GDP in 2019 to 3.0 per- additional interest costs resulting from that increase in cent in 2029 and then continue to increase over the next interest rates accounts for roughly one-quarter of the two decades to 5.7 percent by 2049. Moreover, because increase in debt as a share of GDP over the next three foreign investors hold a significant portion of Treasury decades in CBO’s extended baseline projections; the securities, the increase in outlays represents an increase cost of financing the primary deficits projected over that in payments to foreign investors and thus a reduction in period at current interest rates accounts for the remain- domestic income relative to total U.S. economic output. der of that increase. If, for example, debt was reduced to 42 percent of GDP by 2049, gross national product—which, unlike GDP, That interest rate projection reflects the relatively muted includes income that U.S. residents earn abroad and rise in interest rates over the past decade, which has excludes income payments to nonresidents—would be generally surprised CBO, other government agencies, 5.8 percent higher than it is in CBO’s extended base- and many private-sector forecasters. CBO’s projections line projections. (That increase is 1.5 percentage points of interest rates also reflect the trajectory of federal greater than the percentage increase in GDP that would debt in the agency’s baseline, prices in financial mar- result from that path for debt.) GNP per person in 2049 kets that indicate expectations of future interest rates, would be about $5,500 higher (in 2019 dollars) than it is and other factors. Although factors such as slower labor in the extended baseline projections. force growth are projected to put downward pressure on interest rates, CBO expects rates to rise because of CBO projects a substantial increase in interest costs in such factors as an increase in inflation, faster growth of part because of a projected increase in interest rates. productivity, increased demand for investment in emerg- Although the agency does not expect interest rates to ing economies, and increases in federal borrowing (see rise as much as it previously anticipated, the projected Appendix A). increase in debt from an already high level means that even moderate increases in interest rates would lead to Still, even as the outlook for federal borrowing has wors- significantly higher interest costs. CBO now projects ened over the past decade, financial markets have shown the average interest rate on federal debt to increase few signs of adverse effects, and interest rates on Treasury CHAPTER 1: THE BUDGET OUTLOOK FOR THE NEXT 30 YEARS The 2019 Long-term budget outlook 13 securities have remained relatively low. CBO has revised enough to cause some financial institutions to fail. A its projections of interest rates downward several times in fiscal crisis could thus lead to a financial crisis. Because response. For example, from 2030 to 2035, the average the United States plays a central role in the international rate on federal debt is now projected to be 3.5 percent, financial system, such a crisis could spread globally. 1.7 percentage points lower than the agency projected for that period in June 2010. Similarly, the average real Policymakers’ options for responding to a fiscal crisis (inflation-adjusted) interest rate on federal debt is now would each have negative economic consequences, and projected to be 1.1 percent, 1.6 percentage points lower choosing among them would involve difficult trade- than the 2010 projection. Those downward revisions offs. Such options include using monetary policy to have reduced the projected costs of federal borrowing raise inflation, thereby reducing the burden of financing under current law and reduced the estimated changes in outstanding securities; restructuring the debt (that is, fiscal policy that would be necessary to stabilize debt as a modifying the contractual terms of existing obligations); share of GDP. or dramatically cutting spending or increasing taxes. Although the government has benefited from persistently The risk of a fiscal crisis depends on many factors low interest rates, which have dampened the costs of beyond the level of federal debt. Among those factors federal borrowing, those low rates can also have negative are investors’ expectations about the budget and eco- implications, including their potential to constrain the nomic outlook, which can shift over time, and domestic implementation of monetary policy. Persistently low and and international financial conditions, including global declining interest rates could affect the Federal Reserve’s interest rates. Furthermore, the relationships between ability to use monetary policy to respond sufficiently those many factors and the risk of a crisis are uncertain to a negative shock—such as a sudden worsening in and can shift over time depending, in part, on the state international conditions or abrupt and unexpected fiscal of the economy. In CBO’s assessment, the debt-to-GDP tightening—because monetary policy would be less able ratio has no set tipping point at which a crisis becomes to support economic growth once short-term interest likely or imminent. Indeed, CBO cannot reliably quan- rates were lowered to zero. In the long run, less effec- tify the probability that a fiscal crisis will occur. Thus, tive monetary policy would reduce national income, on the distribution of possible outcomes that the agency average. The current path of debt helps mitigate those considered in preparing its baseline projections does not potential negative effects by keeping rates from being include the potential budgetary and economic outcomes even lower. of a fiscal crisis. Risk of a Fiscal Crisis At this time, financial markets show little indication of High and rising federal debt increases the likelihood of a the risk of a fiscal crisis in the near future. Yet, markets fiscal crisis because it erodes investors’ confidence in the do not always fully reflect risks on the horizon, and more government’s fiscal position and could result in a sharp important, the risk of a fiscal crisis is subject to sudden reduction in their valuation of Treasury securities, which change in the wake of unexpected events. Moreover, all would drive up interest rates on federal debt because else being equal, the risk increases as the debt level rises, investors would demand higher yields to purchase which it is projected to do under current law; if certain Treasury securities. For example, concerns about the tax increases and discretionary spending cuts do not U.S. government’s fiscal position could lead to a sudden take place as scheduled during the next few years, the increase in inflation expectations, fear of a large decrease debt level would rise even more than it does in CBO’s in the value of the U.S. dollar, or a loss of confidence in extended baseline projections. the federal government’s ability or commitment to repay its debt in full. An economic downturn could heighten the risk of a fiscal crisis. In a downturn, the economy shrinks and In a fiscal crisis, dramatic increases in Treasury rates automatic stabilizers boost federal spending and reduce would reduce the market value of outstanding govern- tax liabilities (and thus revenues). As a result of those ment securities, and the resulting losses incurred by hold- developments, deficits and debt (measured as a share of ers of those securities—including mutual funds, pension GDP) would be larger than they are in CBO’s extended funds, insurance companies, and banks—could be large baseline projections. Moreover, policymakers would 14 The 2019 Long-term budget outlook june 2019 face heightened risk that a fiscal crisis would result from position over the long term as they worked to determine elevated debt during circumstances that in the past have whether the uptick in rates was temporary or signaled led them to enact new policies that increased deficits a long-run trend. Alternatively, a lower borrowing rate and in situations in which the Federal Reserve has less would result in smaller interest costs than those in CBO’s flexibility in implementing monetary policy. The effect of extended baseline projections. the increase in federal borrowing on interest costs would be mitigated to some degree if interest rates fell during High debt and large deficits might also create constraints the downturn, as they have in the past. But deficits and for policymakers as they contemplate making changes debt that were larger than CBO projects could make to fiscal policy. As the federal government increases its investors more likely to drastically reduce their valuations borrowing, ever larger cuts in primary deficits would be of Treasury securities, which would lead to significantly required to achieve particular deficit or debt targets. In higher interest rates on those securities. Those factors addition, as a result of the outlook for federal borrow­ suggest lawmakers could avoid certain risks to the ing, policymakers could feel restrained from using economy by reducing deficits in times of relatively strong deficit-­ nanced fiscal policy to respond to unforeseen fi economic growth. events or for other purposes, including to promote eco- nomic activity or to further other goals. High debt could Risks of Other Disruptions also undermine national security if policymakers felt Even in the absence of an abrupt fiscal crisis, high and constrained from increasing national security spending rising debt could generate persistent negative effects to resolve an international crisis or to prepare for such a on the economy beyond those incorporated in CBO’s crisis before it began. extended baseline projections, including a gradual decline in the value of Treasury securities and other Demographic and Economic Trends domestic assets. High and rising debt could lead to Underlying CBO’s Long-Term Projections moderate but ongoing increases in inflation expectations. Demographic and economic projections are key determi- Increases in federal borrowing could also lead to an ero- nants of the long-term budget outlook. Through 2029, sion of confidence in the U.S. dollar as an international the projections in this report are the same as those that reserve currency. Among other effects, such develop- underlie CBO’s 10-year baseline budget projections; ments would make it more difficult to finance public and for later years, the agency projects conditions on the private activity. Moreover, the increased dependence on basis of its assessment of long-term trends. The agency foreign investors that would accompany high and rising uses a model with four components to integrate demo- debt could pose other challenges, such as making U.S. graphic and economic changes into its long-term budget financial markets more vulnerable to a change in valua- projections.6 tion of U.S. assets by participants in global markets. •• A demographic model is used to project the size of The projected level of debt creates the risk that interest the population by age and sex. costs would be substantially greater than projected— even without a fiscal crisis—if interest rates were higher •• A microsimulation model is used to project year- than those underlying CBO’s extended baseline projec- to-year changes in demographic characteristics and tions. For example, if unexpected changes in financial economic outcomes for individuals in a representative factors caused the average borrowing rate to be 1 per- sample of the population. centage point higher every year than the rate underly- ing the agency’s extended baseline projections but all •• A long-term budget model is used to project federal other aspects of the economy were unaffected, then the outlays, revenues, deficits, and debt beyond CBO’s government’s net interest costs would amount to about standard 10-year budget period. 10 percent of GDP 30 years from now, CBO projects. That amount is equal to about half of federal revenues projected for that year. Moreover, under those circum- stances, federal debt would equal almost 200 percent of 6. See Congressional Budget Office, An Overview of CBOLT: The GDP, CBO estimates. If interest rates jumped, investors Congressional Budget Office Long-Term Model (April 2018), could become concerned about the government’s fiscal www.cbo.gov/publication/53667. CHAPTER 1: THE BUDGET OUTLOOK FOR THE NEXT 30 YEARS The 2019 Long-term budget outlook 15 Figure 1-5 . Population, by Age Group Millions of People 400 Actual Projected Age 65 or Older 300 The percentage of the population Ages 20 to 64 age 65 or older is projected to 200 rise over the coming decades, maintaining a long-standing historical trend. 100 Ages 0 to 19 0 2004 2009 2014 2019 2024 2029 2034 2039 2044 2049 Source: Congressional Budget Office. Actual data are shown through calendar year 2016, the most recent year for which such data are available. •• A model of economic growth is used to simulate how rate is slower than the average annual growth rate of the demographics, fiscal policy, and economic factors past 50 years (0.9 percent). The share of the population affect the U.S. economy and, in turn, the federal that is 65 or older also rises over the coming decades, budget. continuing a long-standing historical trend. By 2049, 22 percent of the population would be age 65 or older, Those four components interact in a variety of ways. For whereas today that share is 16 percent (see Figure 1-5). example, the economic projections reflect the effects that To estimate the growth of the U.S. population, CBO increases in spending and revenues in the extended base- projects rates of fertility, immigration, and mortality. line projections—in particular, increased federal borrow- ing and rising effective marginal tax rates—would have Fertility. The total fertility rate is calculated as the sum on the economy. Such effects would result in a smaller of fertility rates for women between 15 and 49 in a given labor supply, a smaller stock of capital, and less output year and represents the average number of children that a than would otherwise be the case. (Appendix A describes woman would have in her lifetime.7 In general, that rate CBO’s demographic and economic projections.) In turn, tends to decline during recessions and rebound during the budgetary outcomes in the extended baseline projec- recoveries. Instead of rebounding after the 2007– tions reflect those economic effects. 2009 recession, however, the fertility rate fell. In 2007, the rate was 2.1 births per woman, but it has steadily Demographic Projections declined since then, falling to 1.9 children per woman in The size and age profile of the U.S. population affect the 2010 and to 1.8 children per woman in 2017 (the most federal budget and the nation’s economy. For example, the composition of the population influences the size of the labor force and the number of beneficiaries of Social Security and other federal programs. In CBO’s projections, the U.S. population increases from 333 mil- 7. The total fertility rate can also be defined as the average number lion at the beginning of 2019 to 388 million in 2049, of children that a woman would have if, in each year of her life, she experienced the birth rates observed or assumed for that year expanding by 0.5 percent each year, on average. That and if she survived her entire childbearing period. 16 The 2019 Long-term budget outlook june 2019 recent year for which data are available).8 CBO expects over the second decade. Net flows of foreign-born people the total fertility rate to gradually increase to 1.9 children without legal status increase over the next five years in per woman by 2022 and to remain at that level for the CBO’s projections, from zero net flows in 2019 (mean- rest of the projection period. The lower fertility rates over ing that immigration is offset by emigration in this cat- the past decade result in slower growth of the population egory) to about 170,000 in 2024; thereafter, annual net age 16 or older in the future. That slow future growth flows remain about the same through 2039. The annual has noticeable effects on CBO’s projections of economic net increase of legal temporary residents is projected to growth in the second decade of the projection period. remain relatively steady, at approximately 80,000 per year, over the next 20 years. Immigration. With birth rates projected to remain low, net immigration flows become an increasingly important In its projections for years after 2039, CBO uses the part of overall U.S. population growth; in 2019, pro- same annual rate of growth for all categories of immi- jected net inflows account for approximately 45 percent grants. Specifically, CBO projects that the net number of overall population growth, but by 2049 that share is of new immigrants would grow at a rate equal to the nearly 87 percent. CBO projects three broad categories growth of the overall population in the previous year; of immigration: legal permanent residents (LPRs), legal that rate averages 0.4 percent annually through 2049. temporary residents, and foreign-born people without The share of the population that is foreign born is thus legal status.9 In the agency’s projections, the rate of net projected to grow from approximately 14 percent today annual immigration averages 3.1 immigrants per thou- to approximately 16 percent in 2049. sand people over the next 30 years, rising from 2.8 in 2019 to 3.1 in 2029 and staying at that level through Mortality. Life expectancy is projected to improve (that 2049. That rate, which accounts for all people who enter is, mortality rates are projected to decline) over the next or leave the United States in a given year, is slightly 30 years, on average. In CBO’s projections, mortality higher than the average net annual immigration rate rates, which measure the number of deaths per thou- since the end of the 2007–2009 recession. sand people in the population, decline at the same pace as the rates for each age and sex group declined from Of those three categories, annual net flows of LPRs are 1950 to 2015. Average life expectancy at birth increases largest, averaging approximately 860,000 people per year from 79.1 years in 2019 to 82.5 years in 2049 in CBO’s in the first decade and approximately 890,000 annually projections. Similarly, life expectancy at age 65 increases by 2.1 years over that period, from 19.4 years in 2019 to 21.5 years in 2049.10 8. See Brady E. Hamilton and others, Births: Provisional Data for 2017, Vital Statistics Rapid Release Report 4 (National Center for Health Statistics, May 2018), www.cdc.gov/nchs/nvss/vsrr/ Economic Projections reports.htm. The performance of the U.S. economy in coming decades will affect the federal government’s spending, 9. CBO uses the term “foreign-born people without legal status” to refer to foreign-born people other than LPRs, refugees, asylees, revenues, and debt accumulation. CBO makes its eco- and temporary residents and visitors. Most foreign-born people nomic projections by assessing trends in key economic without legal status either unlawfully entered the United States variables, such as the size and composition of the labor without inspection or lawfully entered the United States in a force, capital accumulation, productivity, inflation, and temporary status and then unlawfully remained in the country interest rates. The agency also considers ways in which after that temporary status expired. Some foreign-born people without legal status are beneficiaries under Temporary Protected fiscal policy influences economic activity. Status or under policies whereby the executive branch does not seek their immediate removal from the United States (Deferred Economic Growth and the Size of the Labor Force. In Action for Childhood Arrivals, for example); others are allowed CBO’s extended baseline projections, growth in poten- to remain in the United States while they await their removal tial (maximum sustainable) GDP in the future is slower proceedings in immigration courts. Many foreign-born people are authorized to work in the United States and can therefore apply for a Social Security number, which would also make them 10. Life expectancy as used here is period life expectancy, which is eligible for certain refundable tax credits. People are more likely the amount of time that a person in a given year would expect to to report employment income and pay the applicable income and survive beyond his or her current age on the basis of that year’s payroll taxes when they have a Social Security number. mortality rates for various ages. CHAPTER 1: THE BUDGET OUTLOOK FOR THE NEXT 30 YEARS The 2019 Long-term budget outlook 17 Figure 1-6 . Average Annual Growth of Real Potential GDP Percent 4 Historical Projected 3 2.2 0.7 Growth in potential 1.7 GDP is projected to 2.0 be slower than it has 2 been in the past. That 1.4 slowdown is driven mostly by slower 2.4 1.4 1.5 Potential growth of the labor 1 1.0 1.5 Labor Force 1.7 force. 1.6 Productivity 1.2 0.9 0.5 0.5 Potential 0.3 0.4 0 Labor Force 1951– 1975– 1983– 1992– 2003– 2009– 2019– 2030– 2040– 1974 1982 1991 2002 2008 2018 2029 2039 2049 Source: Congressional Budget Office. Real potential GDP is the maximum sustainable output of the economy, adjusted to remove the effects of inflation. The two contributing factors are the potential labor force and potential labor force productivity. The potential labor force is the labor force (that is, the number of people in the civilian noninstitutionalized population who are age 16 or older and who are either working or actively seeking work) adjusted to remove the effects of fluctuations in the business cycle. Potential labor force productivity is the ratio of potential GDP to the potential labor force. GDP = gross domestic product. than it has been over the past 50 years. Over the next 1.4 percent that such growth has averaged annually since 30 years, real potential GDP increases at an average rate 1950. Factors influencing that projection include slower of 1.9 percent per year, whereas from 1969 to 2018, it productivity growth over the past several decades (except grew at an average annual rate of 2.8 percent. That slower during a period of rapid growth in the late 1990s and growth is attributable to several factors—most notably, early 2000s), relatively modest growth in labor quality (a slower growth of the potential labor force (the labor force measure of workers’ skills), and a projected reduction in adjusted for fluctuations in the business cycle). In CBO’s federal investment as a share of GDP. projections, the potential labor force grows by 0.4 per- cent per year, on average, through 2049 (see Figure 1-6); Potential labor productivity in the entire economy— the average annual growth rate over the past 50 years was defined as real potential GDP per potential hour of 1.5 percent. That slower projected growth of the poten- work—is likewise projected to grow more slowly than tial labor force results mainly from slowing population it has in the past, reflecting the slower growth of total growth and the aging of the population. factor productivity and less private investment in capital goods. Since 1950, labor productivity has risen by Productivity. Total factor productivity in the nonfarm 1.7 percent per year, on average; through 2049, it is pro- business sector grows more slowly than its historical jected to increase by an average of 1.5 percent per year. average in CBO’s projections, increasing by 1.1 percent per year, on average, from 2019 to 2049. That rate, Interest Rates. As the economy continues to expand, which measures the growth of the average real output per interest rates rise in CBO’s latest economic projections unit of combined labor and capital services in the non- but remain lower than they have been historically. The farm business sector (which accounts for approximately interest rate on 10-year Treasury notes rises from 2.9 per- 75 percent of economic activity), is slower than the cent at the end of 2018 to 3.8 percent in 2029. That rate 18 The 2019 Long-term budget outlook june 2019 is projected to increase to 4.6 percent in 2049—1.2 per- Excluding net spending on interest, federal outlays aver- centage points below the 5.8 percent average recorded aged 18.3 percent of GDP from 1969 to 2018. Under over the 1990–2007 period. In CBO’s projections, current law, noninterest outlays are projected to rise from slower growth of the labor force and lower inflation than 18.9 percent of GDP in 2019 to 19.8 percent in 2029: in the past push interest rates down from their historical Mandatory spending (which includes spending on Social levels; the effects on interest rates of those two factors Security and the major health care programs as well as and others are projected to outweigh the effects of rising outlays for many smaller programs) is generally projected federal debt and other factors that tend to push interest to increase as a share of the economy, and discretionary rates up above their historical levels. spending is generally projected to decrease. The average interest rate on all federal debt held by After 2029, under the assumptions that govern the the public tends to be lower than the rate on 10-year extended baseline, noninterest spending relative to the Treasury notes. (Interest rates are generally lower on size of the economy would continue to rise, reaching shorter-term debt than on longer-term debt, and the 22.5 percent of GDP by 2049. (For a summary of the average term to maturity of federal debt has been less assumptions about spending and revenues that underlie than 10 years since the 1950s.) On the basis of projec- CBO’s extended baseline, see Table 1-2.) That increase tions of interest rate spreads, the average interest rate would mostly result from larger outlays for the two big- on federal debt is projected to be about 0.4 percentage gest mandatory programs: Social Security and Medicare points lower than the interest rate on 10-year Treasury (see Figure 1-7). notes after 2029. As a result, in CBO’s projections, the average interest rate on federal debt rises to 4.2 percent Under current law, net interest costs would, CBO pro­ by 2049. jects, rise from 1.8 percent of GDP in 2019 to 3.0 per- cent in 2029 as debt accumulates and interest rates Effects of Fiscal Policy. CBO’s economic projections increase from their currently low levels. By 2049, net incorporate the macroeconomic effects of projected interest costs would equal 5.7 percent of GDP, boosting changes in federal tax and spending policies under cur- total federal spending to 28.2 percent of GDP. Spending rent law. In particular, the agency projects that increased has exceeded that level only once, for a three-year period borrowing by the federal government under current law during World War II. In those years, when defense would crowd out some private investment in capital in spending increased sharply, total federal spending topped the long term. Less private investment in capital goods 40 percent of GDP. would, in turn, make workers less productive, leading to lower wages. Lower wages would reduce people’s incen- CBO projects that growth in spending for Social tive to work and thus lead to a smaller supply of labor. Security, the major health care programs, and interest would reshape the spending patterns of the U.S. gov- The agency also incorporates the economic effects of ernment (see Figure 1-8 on page 21). Net spending higher marginal tax rates in its extended baseline projec- for interest would account for a much greater portion tions. As more income is pushed into higher tax brackets of total federal spending in 2049 than it does today, and over time, labor and capital income face higher effective spending on Social Security and the major health care tax rates. Higher marginal tax rates on labor income programs would account for a much larger share of all would reduce after-tax wages and thus people’s incentive federal noninterest spending. Discretionary spending, to work, and the increase in the marginal tax rate on however, would account for a much smaller share of all capital income would reduce their incentive to save and federal noninterest spending in 2049 than it does today. invest. All told, less private investment and a smaller labor supply would lower economic output and income. Spending for Social Security and the Major Health Care Programs Projected Spending Through 2049 Mandatory programs have accounted for a growing share Spending for all of the government’s programs and activ- of the federal government’s noninterest spending over ities and for its net interest costs is projected to account the past few decades. Most of that growth has occurred for a larger percentage of GDP in coming years than it because Social Security and Medicare provide benefits has, on average, over the past 50 years. mainly to people age 65 or older, a group that has been CHAPTER 1: THE BUDGET OUTLOOK FOR THE NEXT 30 YEARS The 2019 Long-term budget outlook 19 Table 1-2 . Assumptions About Outlays and Revenues Underlying CBO’s Extended Baseline Projections Assumptions About Outlays Social Security As scheduled under current law a Medicare As scheduled under current law through 2029; thereafter, projected spending depends on the estimated number of beneficiaries and health care costs per beneficiary (for which excess cost growth is projected to move smoothly to a rate of 1.0 between 2030 and 2049) a Medicaid As scheduled under current law through 2029; thereafter, projected spending depends on the estimated number of beneficiaries and health care costs per beneficiary (for which excess cost growth is projected to move smoothly to a rate of 1.0 between 2030 and 2049) Children’s Health Insurance Program As projected in CBO’s baseline through 2029; thereafter, projected spending remains constant as a percentage of GDP Subsidies for Health Insurance As scheduled under current law through 2029; thereafter, projected spending depends on the estimated Purchased Through the Marketplaces number of beneficiaries, an additional indexing factor for subsidies, and excess cost growth for private health insurance premiums (which is projected to move smoothly to a rate of 1.0 between 2030 and 2049) Other Mandatory Spending As scheduled under current law through 2029; thereafter, refundable tax credits are estimated as part of revenue projections, and the rest of other mandatory spending is assumed to decline as a percentage of GDP at roughly the same annual rate at which it is projected to decline between 2024 and 2029 b Discretionary Spending As projected in CBO’s baseline through 2029; thereafter, projected spending remains roughly constant as a percentage of GDP c Assumptions About Revenues Individual Income Taxes As scheduled under current law Payroll Taxes As scheduled under current law Corporate Income Taxes As scheduled under current law Excise Taxes As scheduled under current law d Estate and Gift Taxes As scheduled under current law Other Sources of Revenues As scheduled under current law (remain constant as a percentage of GDP after 2029) Source: Congressional Budget Office. The extended baseline projections generally reflect current law, following CBO’s 10-year baseline budget projections through 2029 and then extending most of the concepts underlying those projections for the rest of the long-term projection period. For CBO’s most recent 10-year baseline projections, see Congressional Budget Office, Updated Budget Projections: 2019 to 2029 (May 2019), www.cbo.gov/publication/55151. Excess cost growth is the extent to which the growth rate of nominal health care spending per person (adjusted to remove the effects of aging) exceeds the growth rate of potential GDP per person. (Potential GDP is the maximum sustainable output of the economy.) GDP = gross domestic product. a.The payment of full benefits as calculated under current law is assumed to continue regardless of the amounts available in the program’s trust funds. b.In that projection, GDP includes the macroeconomic effects of the policies underlying the extended baseline projections. If it did not, the rest of other mandatory spending after 2029 would decline at the same rate at which it is projected to decline between 2024 and 2029 (excluding the decline in spending for the Supplemental Nutrition Assistance Program). c.In that projection, GDP includes the macroeconomic effects of the policies underlying the extended baseline projections. If it did not, discretionary spending after 2029 would remain the same (measured as a percentage of GDP) as the amount projected for 2029. d.The current-law assumption does not apply to expiring excise taxes dedicated to trust funds. The Balanced Budget and Emergency Deficit Control Act of 1985 requires CBO’s baseline to reflect the assumption that those taxes would be extended at their current rates. That law does not stipulate that the baseline include the extension of other expiring tax provisions, even if they have been routinely extended in the past. 20 The 2019 Long-term budget outlook june 2019 Figure 1-7 . Outlays and Revenues in Selected Years Percentage of Gross Domestic Product Outlays Major Health Other Noninterest Total Social Security Care Programsa Spendingb Net Interest Outlays 1969 2.7 0.8 13.9 1.3 18.7 1989 4.1 2.1 11.3 3.0 20.6 2019 4.9 5.2 8.8 1.8 20.7 2029 5.9 6.7 7.3 3.0 22.5 2049 6.2 9.3 7.1 5.7 28.2 Revenues Individual Total Income Taxes Payroll Taxes Corporate Income Taxes Other Revenuesc Revenues 1969 8.9 4.0 3.7 2.4 19.1 1989 8.0 6.5 1.9 1.5 17.8 2019 8.2 5.8 1.2 1.3 16.5 2029 9.6 5.9 1.4 1.4 18.3 2049 10.7 5.8 1.3 1.7 19.5 Source: Congressional Budget Office. a.Consists of spending for Medicare (net of premiums and other offsetting receipts), Medicaid, and the Children’s Health Insurance Program, as well as outlays to subsidize health insurance purchased through the marketplaces established under the Affordable Care Act and related spending. b.Consists of all federal spending other than that for Social Security, the major health care programs, and net interest. c.Consists of excise taxes, remittances to the Treasury from the Federal Reserve System, customs duties, estate and gift taxes, and miscellaneous fees and fines. growing significantly. In CBO’s extended baseline, the age 65 or older (including spending for Social Security, aging of the U.S. population continues to drive up Medicare, and Medicaid—the federal health care pro- outlays for Social Security and Medicare. Moreover, gram for people with limited income and resources) Medicare outlays also climb because, in CBO’s estima- would account for about half of all federal noninterest tion, health care costs per person will continue to rise. spending; today, that share is about two-fifths. By 2049, CBO projects, federal spending for people CHAPTER 1: THE BUDGET OUTLOOK FOR THE NEXT 30 YEARS The 2019 Long-term budget outlook 21 Figure 1-8 . Composition of Federal Outlays Percent Total Outlays Noninterest Spending 100 100 9 20 Net Interest Other 31 Noninterest 75 75 47 Spendinga 50 50 Major Health 91 41 Care Programsb 80 Noninterest 28 Spending 25 25 Social Security 26 27 0 0 2019 2049 2019 2049 Source: Congressional Budget Office. a.Consists of all federal spending other than that for Social Security, the major health care programs, and net interest. b.Consists of spending for Medicare (net of premiums and other offsetting receipts), Medicaid, and the Children’s Health Insurance Program, as well as outlays to subsidize health insurance purchased through the marketplaces established under the Affordable Care Act and related spending. Social Security. Created in 1935, Social Security is the the status of the program’s trust funds.11 That approach largest single program in the federal budget. Its two is consistent with a statutory requirement that CBO’s components pay benefits to 64 million people in all. 10-year baseline projections incorporate the assumption The larger of the two, Old-Age and Survivors Insurance that funding for such programs is adequate to make all (OASI), pays benefits to retired workers, their eligible payments required by law.12 (For analysis of a scenario in dependents, and some survivors of deceased workers. which benefit payments would be limited to the amounts The smaller program, Disability Insurance (DI), makes in the trust funds, see Chapter 2.) payments to disabled workers and their dependents until those workers are old enough to claim full retirement The Social Security program is funded by dedicated tax benefits under OASI. revenues from two sources. Currently, 96 percent comes from a payroll tax; the rest is collected from income taxes Under current law, CBO projects, spending for Social on Social Security benefits. Revenues from the payroll Security would increase noticeably as a share of the econ- omy, continuing the trend of the past five decades. CBO 11. The balances of the trust funds represent the total amount that projects that the number of Social Security beneficiaries the government is legally authorized to spend for those purposes. For more details about the legal issues related to exhaustion of a would rise from 64 million in 2019 to 97 million in trust fund, see William R. Morton and Barry F. Huston, Social 2049 and that spending for the program would increase Security: What Would Happen If the Trust Funds Ran Out? Report from 4.9 percent of GDP to 6.2 percent over that period for Congress RL33514 (Congressional Research Service, June 11, (see Figure 1-7 on page 20). Those projections reflect 2018), https://go.usa.gov/xEtaw. the assumption that Social Security will continue to pay 12. Sec. 257(b)(1) of the Balanced Budget and Emergency Deficit benefits as scheduled under current law, regardless of Control Act of 1985 (Deficit Control Act), P.L. 99-177 (codified at 2 U.S.C. §907(b)(1) (2016)). 22 The 2019 Long-term budget outlook june 2019 Table 1-3 . benefits at the end of the period.13 For Social Security, Summary Financial Measures for the that difference is traditionally expressed as a percentage of the present value of taxable payroll over 75 years.14 Social Security System With the trust funds’ revenues projected to grow more Projection Period Actuarial (Calendar years) Income Rate Cost Rate Balance slowly than their expenditures, the program would have a long-term actuarial deficit. Over the next 75 years, if As a Percentage of Gross Domestic Product current laws remained in place, the program’s actuarial 25 Years (2019 to 2043) 5.1 6.1 -1.1 deficit would be 1.5 percent of GDP, or 4.6 percent 50 Years (2019 to 2068) 4.8 6.1 -1.4 of taxable payroll, CBO projects (see Table 1-3).15 75 Years (2019 to 2093) 4.6 6.2 -1.5 According to CBO’s projections, it would therefore be As a Percentage of Taxable Payroll possible to pay the benefits prescribed by current law 25 Years (2019 to 2043) 14.6 17.6 -3.1 and maintain the necessary trust fund balances through 50 Years (2019 to 2068) 14.0 18.0 -4.0 2093 if payroll taxes were raised immediately and perma- 75 Years (2019 to 2093) 13.9 18.4 -4.6 nently by about 4.6 percent of taxable payroll, if sched- uled benefits were reduced by an equivalent amount, Source: Congressional Budget Office. or if some combination of tax increases and spending These projections incorporate the assumption that spending for Social reductions of equal present value was adopted.16 Security continues as scheduled even if its trust funds are exhausted. Through 2049, the projections incorporate macroeconomic feedback caused by rising federal debt and marginal tax rates. After 2049, they do not account for such feedback. 13. A present value expresses a flow of past and future income or payments as a single amount received or paid at a specific time. For programs such as Social Security that have both a trust fund and a The value depends on the interest rate, known as the discount dedicated source of revenue, a common measure of sustainability is the rate, used to translate past and future cash flows into current actuarial balance, which is the income rate over a given period minus dollars at that time. To account for the difference between a trust the cost rate over that period. The income rate is the present value of fund’s current balance and the desired balance at the end of the annual tax revenues plus the initial trust fund balance, divided by the present value of gross domestic product (GDP) or taxable payroll. The period, the balance at the beginning is added to the projected tax cost rate is the present value of annual outlays plus the present value revenues, and an additional year of costs at the end of the period of a reserve equal to a year’s worth of benefits at the end of the period, is added to projected outlays. divided by the present value of GDP or taxable payroll. (The present 14. Taxable payroll is the total amount of earnings (wages and value of a flow of revenues or outlays over time expresses that flow self-employment income) from employment covered by Social as a single amount received or paid at a given time. The present value Security that is below the applicable annual taxable maximum depends on the rate of interest, known as the discount rate, that is used ($132,900 in 2019). to translate the cash flow into current dollars.) 15. The 75-year projection period used here begins in calendar year 2019 and ends in calendar year 2093. The Social Security trustees have estimated that the program’s 75-year actuarial tax and the tax on benefits are credited to the Old-Age shortfall would be 2.8 percent of taxable payroll, which is about and Survivors Insurance Trust Fund and the Disability 1.8 percentage points less than CBO’s projection. For details on Insurance Trust Fund, which finance the program’s the trustees’ projections, see Social Security Administration, The benefits. In CBO’s extended baseline projections, dedi- 2018 Annual Report of the Board of Trustees of the Federal Old-Age and Survivors Insurance and Federal Disability Insurance Trust cated tax revenues for the combined trust funds remain Funds (June 2018), www.ssa.gov/oact/tr/2018. roughly constant through 2049 at about 4.4 percent of GDP. 16. A policy that either increased revenues or reduced outlays by the same percentage of taxable payroll each year to eliminate the 75-year shortfall would not necessarily place Social Security on A common measure of the sustainability of a program a permanently stable financial path. Estimates of the actuarial that has a trust fund and a dedicated revenue source is deficit do not account for revenues or outlays after the 75-year its estimated actuarial balance over a given period—that projection period. Because shortfalls are smaller earlier in the is, the sum of the present value of projected tax revenues 75-year projection period than they are later, such a policy would and the current trust fund balance minus the sum of the create surpluses in the next several decades but result in deficits later and leave the system financially unbalanced after calendar present value of projected outlays and a year’s worth of year 2093. Additionally, the calculation of the actuarial balance does not include the effects of any macroeconomic feedback that would result from an increase in taxes or a reduction in benefits. CHAPTER 1: THE BUDGET OUTLOOK FOR THE NEXT 30 YEARS The 2019 Long-term budget outlook 23 Another commonly used measure of Social Security’s sus- CHIP, combined with outlays for marketplace subsidies tainability is the trust funds’ dates of exhaustion. CBO and related spending, would grow by 1.0 percent of projects that under current law, the DI trust fund would GDP (see Figure 1-9).18 be exhausted in fiscal year 2028 and the OASI trust fund would be exhausted in calendar year 2032. If their bal- Causes of Growth in Spending for Social Security and ances were combined, the OASDI trust funds would be the Major Health Care Programs exhausted in calendar year 2032, CBO estimates. The aging of the population and rising health care costs per person are the primary reasons for the sharp rise in The Major Health Care Programs. Outlays for the major projected spending for Social Security and the major health care programs consist of spending for Medicare, federal health care programs over the next 30 years. The Medicaid, and the Children’s Health Insurance Program extent to which health care costs per person (adjusted (CHIP), as well as outlays to subsidize health insur- to remove the effects of aging) grow faster than poten- ance purchased through the marketplaces established tial GDP per person is known as excess cost growth. In under the Affordable Care Act (ACA) and related CBO’s extended baseline projections, spending for Social spending.17 Medicare, which provides health insurance Security and the major federal health care programs to about 61 million people (most of whom are at least grows from 10.7 percent of GDP in 2019 to 16.8 per- 65 years old), accounts for more than 60 percent of that cent in 2049 (see Figure 1-10).19 Spending for Social spending. Security grows from 4.9 percent of GDP in 2019 to 6.2 percent in 2049, and spending for the major federal CBO projects federal spending for the government’s health care programs grows from 5.9 percent of GDP to major health care programs for 2019 through 2029 10.7 percent. under the assumption that the laws governing those pro- grams will, in general, remain unchanged. As with Social If CBO had set the shares of the population by age Security, CBO assumes that Medicare will pay benefits as at today’s proportions and had set excess cost growth scheduled under current law, regardless of the amounts at zero when developing its projections, spending on in the program’s trust funds. For longer-­ erm projections, t those programs as a share of GDP in 2049 would have considerable uncertainty surrounds the evolution of been projected to be 10.7 percent—the same share as health care delivery and financing systems. That uncer- estimated for 2019.20 Aging accounts for an increase tainty led CBO to use a formulaic approach to prepare of 3.0 percentage points, or roughly half of the differ- projections beyond 2029: It combines estimates of the ence between 10.7 percent and 16.8 percent. Excess number of expected beneficiaries of the government’s cost growth accounts for the other half, an increase of health care programs with mechanical estimates of the 3.1 percentage points. For Social Security, aging accounts growth in spending per beneficiary. for more than the full increase in spending. For the major health care programs, aging accounts for 1.5 per- Over the past five decades, spending for the major centage points of the growth, and excess cost growth health care programs has steadily grown faster than the accounts for the remainder. economy, and that trend continues in CBO’s extended baseline. In 2019, net federal spending for the major The Aging of the Population. In CBO’s projections, the health care programs is estimated to equal 5.2 percent aging of the baby-boom generation and continued gains of GDP. If current laws generally remained in place, net outlays for those programs would increase to 9.3 percent 18. In CBO’s projections, the outlays for subsidies for insurance in 2049: Medicare spending, net of offsetting receipts purchased through the marketplaces and related spending are combined with outlays for Medicaid and CHIP. Federal subsidies (mostly premiums paid by enrollees), would grow by for health insurance for low- and moderate-income households 3.0 percent of GDP, and spending on Medicaid and account for most of those outlays. 19. This analysis of causes of spending growth includes gross 17. Spending related to subsidies for insurance purchased through spending on Medicare. the marketplaces includes spending for subsidies for insurance provided through the Basic Health Program and spending for the 20. If the effects of aging and excess cost growth were removed, risk-adjustment and reinsurance programs that were established spending on those programs as a percentage of GDP would be by the ACA to stabilize premiums for health insurance purchased slightly lower in 30 years than it is today, mainly because of the by individuals and small employers. scheduled increase in the full retirement age for Social Security. 24 The 2019 Long-term budget outlook june 2019 Figure 1-9 . Federal Outlays for the Major Health Care Programs, by Category Percentage of Gross Domestic Product 10 Actual Projected 8 Medicare spending, net 6 Medicarea of offsetting receipts, is projected to account for about three-quarters of the 4 increase in spending for the major health care programs over the next 30 years. 2 Medicaid, CHIP, and Marketplace Subsidiesb 0 2004 2009 2014 2019 2024 2029 2034 2039 2044 2049 Source: Congressional Budget Office. CHIP = Children’s Health Insurance Program. a.Refers to net spending for Medicare, which accounts for offsetting receipts that are credited to the program. Those offsetting receipts are mostly premiums paid by beneficiaries to the government. b.“Marketplace Subsidies” refers to spending to subsidize health insurance purchased through the marketplaces established under the Affordable Care Act and insurance provided through the Basic Health Program, as well as spending to stabilize premiums for health insurance purchased by individuals and small employers. in life expectancy increase the share of the population health care programs as a share of GDP over the 2019– that is age 65 or older from 16 percent to 22 percent 2049 period in CBO’s projections. between 2019 and 2049. Rising Health Care Costs per Person. Even though Aging accounts for all of the projected long-term increase growth in health care costs per person has slowed in Social Security spending as a percentage of GDP. recently, over the next 30 years such costs are projected Because the share of the population that is 65 or older is to continue to grow faster than potential GDP per growing, a larger segment of the population will receive person (see Figure 1-11). In CBO’s extended baseline Social Security benefits, increasing federal spending for projections, excess cost growth accounts for about two- the program. thirds of the increase in spending, measured as a share of GDP, for the major health care programs between 2019 Aging also contributes to the projected increase in spend- and 2049. ing, relative to GDP, for the major health care programs, particularly Medicare, which is the largest such program. Other Noninterest Spending Most beneficiaries qualify for Medicare at age 65. As that In CBO’s extended baseline projections, total federal group becomes larger and older, on average, Medicare spending for everything other than Social Security, the spending will increase, not only because the number major health care programs, and interest declines as a of beneficiaries will rise but also because people tend share of GDP to its lowest level in more than 70 years. to require more health care as they age. Aging explains Over the past 50 years, such spending has averaged about one-third of the increase in spending for the major 11 percent of GDP, but it has been as high as 14 per- cent (in the late 1960s and early 1970s) and as low as CHAPTER 1: THE BUDGET OUTLOOK FOR THE NEXT 30 YEARS The 2019 Long-term budget outlook 25 Figure 1-10 . Spending for Social Security and the Major Health Care Programs in 2019 and 2049 Percentage of GDP Social Security Major Health Care Programs Both 20 16.8 Attributable to 15 3.1 Excess Cost Growth 3.0 Attributable to the 10.7 10.7 Aging of the Population 10 3.1 6.2 5.9 1.5 4.9 1.5 10.7 Without Aging and 5 Excess Cost Growth 6.1 4.6a 0 2019 2049 2019 2049 2019 2049 Source: Congressional Budget Office. Spending for the major health care programs consists of gross spending for Medicare (which does not account for the offsetting receipts that are credited to the program), Medicaid, and the Children’s Health Insurance Program, as well as outlays to subsidize health insurance purchased through the marketplaces established under the Affordable Care Act and related spending. Those outlays have been adjusted to exclude the effects of shifting payments from one fiscal year into another so that those payments are not made on a weekend. Excess cost growth is the extent to which the growth rate of nominal health care spending per person (adjusted to remove the effects of aging) exceeds the growth rate of potential GDP per person. (Potential GDP is the maximum sustainable output of the economy.) GDP = gross domestic product. a.If aging and excess cost growth did not occur after 2019, spending on Social Security as a share of GDP would be lower in 30 years, mainly because of the scheduled increase in the full retirement age for Social Security. 8 percent (in the late 1990s and early 2000s). Other Over the past half-century, discretionary spending has noninterest spending is estimated to equal 8.8 percent diminished markedly as a percentage of GDP: Between of GDP in 2019. In CBO’s extended baseline projec- 1969 and 2018, it declined from 12.0 percent of GDP tions, such spending falls to 7.3 percent of GDP in 2029 to 6.3 percent. In CBO’s baseline projections, discre- and to 7.1 percent of GDP in 2049. Both discretionary tionary outlays equal 6.3 percent of GDP in 2019 and spending and other mandatory spending are projected to then decrease steadily over the coming decade, falling to decline in relation to GDP. 5.0 percent of GDP in 2029. Discretionary Spending. About half of all discretionary Through 2021, most discretionary funding is limited by spending is dedicated to national defense, and the rest is caps on annual discretionary appropriations that were for an array of federally funded investments and activi- originally specified in the Budget Control Act of 2011 ties, including education, transportation, housing assis- (P.L. 112-25, as amended). The decline in discretion- tance, veterans’ health care, health-related research and ary outlays relative to GDP over the next eight years public health programs, the administration of justice, in CBO’s projections reflects lower statutory limits on and international affairs. discretionary funding in 2020 and 2021 and CBO’s assumption (required by law) that discretionary fund- ing will grow at the rate of inflation—which is slower 26 The 2019 Long-term budget outlook june 2019 Figure 1-11 . Average Annual Growth of Health Care Costs per Beneficiary Percent Medicare Medicaid 6 5 1.1 1.6 4 0.7 Excess Cost Growth 1.1 3 2 3.8 3.8 3.4 3.4 Growth of Potential GDP 1 per Person 0 1985−2017 2019−2049 1985−2017 2019−2049 Source: Congressional Budget Office, using data from the Centers for Medicare & Medicaid Services. Excess cost growth is the extent to which the growth rate of nominal health care spending per person (adjusted to remove the effects of aging) exceeds the growth rate of potential GDP per person. (Potential GDP is the maximum sustainable output of the economy.) The averages of the components (excess cost growth and growth in potential GDP per person) are estimated separately using weighted regression, with twice as much weight placed on the last year of the period as on the first year. As a result, the components do not add up exactly to the average growth in health care costs per person (adjusted to remove the effects of aging), particularly over history, when growth rates varied substantially from year to year. GDP = gross domestic product. than the projected growth of GDP—beginning in (One exception occurred in 2009, when such spend- 2022. CBO’s extended baseline projections reflect the ing, referred to as other mandatory spending, spiked to assumption that after 2029, discretionary spending will 5.1 percent because of policies enacted in response to the remain roughly constant as a percentage of GDP (see severe recession.) Other mandatory spending includes Figure 1-12).21 retirement programs for federal civilian and military employees, certain veterans’ programs, the Supplemental Other Mandatory Spending. Since the mid-1960s, Nutrition Assistance Program (SNAP), Supplemental mandatory spending excluding that for Social Security Security Income, unemployment compensation, and and the major health care programs has generally refundable tax credits.22 remained between 2 percent and 4 percent of GDP. Other mandatory spending declines slightly as a share 21. CBO assumed that discretionary spending after 2029 would of the economy over the next 10 years in CBO’s projec- remain constant as a percentage of GDP before the agency tions. Such spending accounts for 2.6 percent of GDP accounted for the effect on the economy of the fiscal policies today and, if current laws generally remained unchanged, projected under the extended baseline. Because CBO estimates that those policies would dampen economic growth, projected discretionary spending would not grow at precisely the same rate as GDP. Although discretionary spending declines in relation to GDP from 2019 to 2029 in CBO’s projections, 22. Refundable tax credits reduce a filer’s overall income tax liability; historical evidence suggests that such a decline is unlikely to if the credit exceeds the rest of the filer’s income tax liability, persist: Discretionary spending has historically been a larger the government pays all or some portion of that excess to the share of economic output than it is projected to be in 2029. For taxpayer (and the payment is treated as an outlay in the budget). that reason, CBO did not assume that the share would decline See Congressional Budget Office, Refundable Tax Credits further. (January 2013), www.cbo.gov/publication/43767. CHAPTER 1: THE BUDGET OUTLOOK FOR THE NEXT 30 YEARS The 2019 Long-term budget outlook 27 Figure 1-12 . Other Federal Noninterest Spending Percentage of Gross Domestic Product 15 Actual Projected 10 Measured as a percentage of economic output, other federal noninterest spending declines between 2019 and 2049 in CBO’s projections, mainly because of a projected decrease 5 in discretionary spending over the Discretionary Spending next decade. Other Mandatory Spendinga 0 2004 2009 2014 2019 2024 2029 2034 2039 2044 2049 Source: Congressional Budget Office. a.Consists of all mandatory spending other than that for Social Security and the major health care programs. It includes the refundable portions of the earned income and child tax credits and of the American Opportunity Tax Credit. would decline to 2.3 percent of GDP in 2029.23 That Net Interest Costs decrease stems primarily from average benefits’ increasing Over the past 50 years, the government’s net interest more slowly than income (because benefit growth is gen- costs have averaged 2.0 percent of GDP, although they erally indexed to inflation measures that do not account have been as high as 3.2 percent and as low as 1.2 per- for real income growth). cent. In CBO’s extended baseline projections, net interest costs increase steadily as a share of the economy In CBO’s extended baseline projections, other manda- over the next decade—from 1.8 percent of GDP in tory spending falls to 2.0 percent of GDP by 2049. In 2019 to 3.0 percent by 2029—as greater federal borrow- part, that reduction is attributable to growth in income, ing boosts debt-service costs and as interest rates rise. which would reduce the number of people eligible for Those costs reach 5.7 percent of GDP by 2049—higher refundable tax credits. It also reflects the assumption that than they have ever been before (see Figure 1-7 on after 2029 other mandatory spending, excluding outlays page 20). If net interest costs followed that projected for such tax credits, would decline at roughly the same path, they would exceed other mandatory spending by rate as such spending is projected to fall between 2024 2023, exceed all discretionary spending by 2046, and and 2029.24 approach spending for Social Security by 2049. Deficits and debt rise in CBO’s projections because 23. Sec. 257(b)(2) of the Deficit Control Act, which governs CBO’s baseline projections, makes exceptions regarding current law for of the growing gap between spending and revenues, some programs, such as SNAP, that have expiring authorizations and higher interest costs are a major contributor to but that are assumed to continue as currently authorized. the growth of that gap. More than half of the increase 24. For the years after 2029, other mandatory spending was not in spending as a percentage of GDP from 2019 to projected in detail because of the number of programs involved 2049 results from higher net interest costs. Moreover, of and the variety of factors that influence spending on them. the 4.5 percentage-point increase in the federal bud- Instead, CBO used an approximate method to project spending get deficit over that period, only 0.6 percentage points for those programs as a group. Except for the outlays for refundable tax credits, after 2029 such spending is assumed to decline in relation to GDP (before any possible effects of fiscal which it is projected to fall between 2024 and 2029 (excluding policy on the economy are accounted for) at the same rate at the decrease in spending for SNAP). 28 The 2019 Long-term budget outlook june 2019 Figure 1-13 . Increases in Federal Revenues Percentage of Gross Domestic Product 4 3 Other Factors Tax on High-Premium Revenues are projected to Health Insurance Plans grow steadily over the next several years, rise sharply 2 Expiration of following the expiration of Major Provisions of the 2017 Tax Act certain temporary provisions of the 2017 tax act at the end Real Bracket Creepa of 2025, and then resume 1 steady growth. 0 2019 2024 2029 2034 2039 2044 2049 Source: Congressional Budget Office. a.Real bracket creep is the process in which, as income rises faster than inflation, a larger proportion of income becomes subject to higher tax rates. are attributable to the primary deficit—the rest of the nearly all of the individual income tax provisions of the increase is due to rising net interest costs. In large part, 2017 tax act. those rising interest costs stem from increases in interest rates that reflect long-term economic trends, which CBO For years after 2029, revenues are projected following projects would occur even if debt did not rise beyond the assumption that the rules for all tax sources will its current level. But greater federal borrowing places change only as scheduled under current law.25 Thus, in additional upward pressure on interest rates and thus on CBO’s extended baseline projections, revenues continue interest costs. Moreover, growth in net interest costs and to grow faster than GDP after 2029 and total 19.5 per- growth in debt reinforce one another: Rising interest cent of GDP in 2049. Increases in receipts from indi- costs boost deficits and debt, and rising debt pushes up vidual income taxes account for most of the projected interest costs. 3.0 percentage-­ oint rise in total revenues as a share of p GDP over the next three decades. Receipts from all other Projected Revenues Through 2049 sources combined are projected to increase slightly as a In CBO’s extended baseline projections, revenues share of GDP (see Figure 1-7 on page 20). measured as a share of GDP are generally higher than they have been, on average, in recent decades. As a share Over the entire 30-year period, the underlying causes of GDP, revenues have averaged 17.4 percent over the of the projected increase in total revenues as a share of past 50 years, but they have fluctuated between 15 per- GDP are real bracket creep in the individual income tax cent and 20 percent of GDP because of changes in tax system, expiring tax provisions and the tax on high-­ laws and interactions between those laws and economic premium health insurance plans, and other factors. conditions. If current laws generally remained unchanged, revenues 25. The sole exception to the current-law assumption during would increase in relation to GDP over the coming the 30-year projection period applies to expiring excise taxes decade, CBO projects. Revenues are projected to rise dedicated to trust funds. The Deficit Control Act requires steadily from 16.5 percent of GDP in 2019 to 17.4 per- CBO’s baseline to reflect the assumption that those taxes would cent by 2025 and then to grow more rapidly, reaching be extended at their current rates. That law does not stipulate that the baseline include the extension of other expiring tax 18.3 percent by 2029. The projected growth in revenues provisions, even if lawmakers have routinely extended them after 2025 is largely attributable to the expiration of before. CHAPTER 1: THE BUDGET OUTLOOK FOR THE NEXT 30 YEARS The 2019 Long-term budget outlook 29 Figure 1-14 . Shares of Income Taxed at Different Rates Under the Individual Income Tax System Percent Income Tax Rate 9 10 11 39.6 Percent 28 31 33 25–35 Percent The largest contributor to growth in revenues over the long-term projection period is real bracket creep. That is the process in which, as income rises faster than inflation, a larger 36 proportion of income becomes subject to 34 32 10–25 Percent higher tax rates. While the share of income taxed at higher rates grows, the share exempt from taxation shrinks. 26 25 24 0 2029 2039 2049 Source: Congressional Budget Office. Income here refers to adjusted gross income—that is, income from all sources not specifically excluded by the tax code, minus certain deductions. The income tax rate is the statutory rate specified under the individual income tax system. The lowest statutory rate is zero because of deductions and exemptions. Real Bracket Creep in the Expiring Tax Provisions and the Tax on Individual Income Tax System High-Premium Health Insurance Plans The largest contributor to the increase in total revenues The second largest contributor to the increase in reve- is real bracket creep (see Figure 1-13 on page 28). If nues is the expiration, after calendar year 2025, of nearly current laws generally remained unchanged, real bracket all provisions of the 2017 tax act that affect individual creep would continue to gradually push up taxes in income taxes. The expiration of those provisions would relation to income over the next three decades, CBO boost individual income tax receipts as a share of GDP projects. That occurs because most income tax brackets, by 0.8 percentage points by 2029, CBO projects. exemptions, credits, and other tax thresholds are indexed to inflation.26 When income grows faster than inflation, The third major source of the increase in revenues is a as generally happens during economic expansions, more tax on certain employment-based health insurance plans income is pushed into higher tax brackets and credits with high premiums that was originally enacted in 2010 phase out, thereby increasing tax receipts. Between 2029 and is scheduled to take effect in 2022.27 Although the and 2049, the share of income taxed at the top rate revenues raised by that tax would initially be small, rises by 2 percentage points—and the share of income rapid growth in health care costs is projected to drive up exempted from taxation falls by 2 percentage points— revenues from that tax over subsequent decades. CBO because of real bracket creep (see Figure 1-14). projects that the tax would bring in revenues equal to 0.7 percent of GDP by 2049. 27. Under the Affordable Care Act, employer-sponsored health benefits will be subject to an excise tax equal to 40 percent of the value of those benefits exceeding certain thresholds. That 26. Some parameters of the tax system, including the amount of the tax was originally scheduled to take effect in 2018 but has been child tax credit, are fixed in nominal dollars and are not adjusted delayed twice by legislation, most recently by the Extension of for inflation. Continuing Appropriations Act, 2018. 30 The 2019 Long-term budget outlook june 2019 Table 1-4 . rise (see Table 1-4). The effective marginal tax rate Effective Marginal Federal Tax Rates Underlying on capital would also rise by a small amount. Higher marginal rates can dampen economic activity: Increases CBO’s Extended Baseline Projections in the marginal tax rate on labor income reduce people’s Percent incentive to work, and increases in the marginal tax rate 2019 2029 2049 on capital income reduce their incentive to save.28 Marginal Tax Rate on Labor Income 26.8 30.3 32.0 Marginal Tax Rate on Capital Income 15.7 15.7 15.9 Uncertainty of CBO’s Long-Term Projections Budget projections are inherently uncertain. Even Source: Congressional Budget Office. if future tax and spending policies do not vary from The extended baseline projections generally reflect current law, those specified in current law, budgetary outcomes following CBO’s 10-year baseline budget projections through 2029 and will undoubtedly differ from those in CBO’s extended then extending most of the concepts underlying those projections for the baseline projections because of unexpected changes in rest of the long-term projection period. demographics, the economy, and other factors. To quan- The effective marginal tax rate on labor income is a weighted average of tify the uncertainty of budgetary outcomes over the long the percentage of an additional dollar of a taxpayer’s labor income that term, CBO examined the extent to which federal debt is paid in federal individual income taxes and payroll taxes. Weights are as a percentage of GDP would differ from its extended assigned to taxpayers on the basis of their labor income. The effective marginal tax rate on capital income is the percentage of the return on baseline projections if a set of key factors—several an additional dollar of investment made in a particular year that will be demographic and economic factors and the growth of paid in taxes over the life of that investment. The before- and after-tax health care costs—deviated from the paths underlying rates of return used to calculate that effective tax rate are weighted those projections. averages of the rates for every combination of asset type, industry, form of organization, and source of financing; the weights used are the values of the assets for each combination. CBO projects that there is a two-thirds chance that federal debt held by the public would be between 71 percent and 175 percent of GDP in 2039 if current Other Factors laws generally remained unchanged (see Figure 1-15). Many other factors affect revenues in the extended That range of outcomes indicates that federal debt held baseline projections. For example, earnings are projected by the public after 20 years could be as much as 42 per- to grow faster for higher-income people than for other centage points lower or as much as 62 percentage points people over the next 30 years. That trend would cause a higher than the agency’s extended baseline projection of larger share of individual income to be taxed at higher 113 percent of GDP. rates. The resulting increase in individual income tax revenues would be largely offset by a decrease of nearly To estimate that likely range, CBO simulated budgetary the same amount in payroll tax receipts, CBO projects, outcomes 30 years from now by varying all of the key because a greater share of earnings would be above the factors at once, but the agency also examined the sensi- maximum amount subject to Social Security payroll tivity of its projections to higher or lower values for some taxes. In addition, several other tax provisions are sched- of those factors, including productivity growth or inter- uled to expire over the coming decade, which generally est rates, in isolation (see Box 1-1). CBO’s analysis does pushes up revenues in CBO’s projections. not address certain sources of uncertainty in the budget projections, such as the risk of an economic depression Those factors would, under current law, cause the effects or a major war or catastrophe. Also, although the factors of the tax system in 2049 to differ substantially from considered here are some of the more important ones, the system’s effects today. Taxpayers across the income they are not the only ones. Nonetheless, the results distribution would, on average, pay more of their income show that the main implications of this report apply in taxes in 2049 than similar taxpayers do now if current under a wide range of possible values for key factors laws generally remained unchanged. In addition, a larger that influence federal spending and revenues. If current share of each additional dollar of income that households earned would go to pay taxes because under current law, 28. Although the marginal tax rate on capital income is projected to the effective marginal federal tax rates for labor would rise under current law, it would still be lower than it has been in recent years. CHAPTER 1: THE BUDGET OUTLOOK FOR THE NEXT 30 YEARS The 2019 Long-term budget outlook 31 Figure 1-15 . Uncertainty in CBO’s Projections of Federal Debt Held by the Public Percentage of GDP 180 Actual Projected 160 140 120 Extended Baseline There is a two-thirds chance 100 that federal debt held by 80 the public will be between 71 percent and 175 percent 60 of GDP by 2039. Middle Two-Thirds 40 of the Range of Possible Outcomes 20 0 2004 2009 2014 2019 2024 2029 2034 2039 Source: Congressional Budget Office. The extended baseline projections generally reflect current law, following CBO’s 10-year baseline budget projections through 2029 and then extending most of the concepts underlying those projections for the rest of the long-term projection period. To quantify the uncertainty of long-term budgetary outcomes, CBO estimated a distribution of outcomes using simulations of its long-term model, each with a unique set of values for key economic and demographic factors. Specifically, CBO assessed the consequences of alternative paths for the following factors: the growth rate of total factor productivity, interest rates on federal debt held by the public, the civilian unemployment rate, the rates of excess cost growth for Medicare and Medicaid spending, the fertility rate, the rate of mortality improvement, and the rate of immigration. The civilian unemployment rate is the percentage of people in the labor force who are unemployed. (The labor force is the number of people in the civilian noninstitutionalized population who are age 16 or older and who are either working or actively seeking work.) Excess cost growth is the extent to which the growth rate of nominal health care spending per person (adjusted to remove the effects of aging) exceeds the growth rate of potential GDP per person. (Potential GDP is the maximum sustainable output of the economy.) GDP = gross domestic product. laws generally remained unchanged, in 20 years federal agency also evaluated the extent to which some factors— debt—which is already high by historical standards— productivity growth, interest rates, and unemployment would probably be much higher than it is today. rates—have moved together over long periods of time. The Basis of CBO’s Uncertainty Analysis Using simulations that incorporated historical data, the If the size of the population, productivity growth, inter- agency projected potential future outcomes for each key est rates, unemployment rates, and excess cost growth factor.30 On the basis of those simulations, the agency for Medicare and Medicaid diverged from the paths constructed ranges that capture two-thirds of possible underlying CBO’s extended baseline budget projections, outcomes over the next two decades for each of those budgetary outcomes could differ markedly from those factors as well as for the budgetary outcomes resulting projections.29 To quantify the uncertainty of its budget from them. There is a two-thirds chance that an outcome projections arising from the uncertainty of those key will fall within the range estimated for it. The ranges factors, CBO assessed past trends of those factors. The reflect the uncertainty of the long-term trend of each 29. The civilian unemployment rate is the percentage of people in the 30. Details about the methods used in this analysis of uncertainty labor force who are unemployed. will be provided in a forthcoming publication. 32 The 2019 Long-term budget outlook june 2019 Box 1-1. Sensitivity of Budget Projections to Changes in Underlying Economic Factors How would the budget be affected if the economy ended Growth of Nonfarm Business Productivity up growing more quickly than it does in the Congressional CBO assessed average nonfarm business total factor produc- Budget Office’s projections? For example, what if productiv- tivity (TFP) growth over 30-year periods between 1950 and the ity grew more quickly than CBO expects? Or what if interest present.1 Over those periods, the 30-year average of produc- rates turned out to be higher than CBO’s central estimates? tivity growth varied by about 1 percentage point, indicating that For instance, what if investors were willing to take on more future outcomes would most likely fall within a 1 percentage-­ risk than projected and interest rates on federal debt rose in point range around the agency’s central estimates.2 The relation to interest rates on private securities more than CBO agency therefore projected economic and budgetary outcomes expects? How would such a development affect the budget? To help answer those questions and others, the agency examined the sensitivity of its budget projections to values for productiv- 1. Total factor productivity is the growth of real (inflation-adjusted) output per unit of combined labor and capital services. ity growth and for interest rates on federal debt that differed from its central estimates for those key factors. 2. See Congressional Budget Office, The 2016 Long-Term Budget Outlook (July 2016), Chapter 7, www.cbo.gov/publication/51580. Federal Debt If Total Factor Productivity Growth or Interest Rates Differed From the Values Underlying CBO’s Projections Percentage of GDP Total Factor Productivity Growth 200 Actual Projected 185 TFP Growth That Is 0.5 Percentage Points Lower 150 144 Extended Baseline 100 106 TFP Growth That Is 0.5 Percentage Points Higher 50 0 2004 2009 2014 2019 2024 2029 2034 2039 2044 2049 Interest Rates 200 Actual Projected 199 Interest Rates That Are 1 Percentage Point Higher 150 144 Extended Baseline 100 107 Interest Rates That Are 1 Percentage Point Lower 50 0 2004 2009 2014 2019 2024 2029 2034 2039 2044 2049 Source: Congressional Budget Office. GDP = gross domestic product; TFP = total factor productivity. Continued CHAPTER 1: THE BUDGET OUTLOOK FOR THE NEXT 30 YEARS The 2019 Long-term budget outlook 33 Box 1-1.Continued Sensitivity of Budget Projections to Changes in Underlying Economic Factors using rates of growth of nonfarm business TFP that were public would be 199 percent of GDP in 2049 rather than 0.5 percentage points higher and 0.5 percentage points lower the 144 percent in the extended baseline projection (see than the rate underlying the extended baseline projections.3 the figure). •• If nonfarm business productivity grew 0.5 percentage •• If, instead, federal borrowing rates were 1.0 percentage points faster each year than CBO projects, federal debt point lower each year than they are in CBO’s extended held by the public would be 106 percent of gross domestic baseline projections, federal debt held by the public would product (GDP) in 2049 rather than 144 percent, as it is in the be 107 percent of GDP in 2049. extended baseline projections (see the figure). Other Factors •• If, instead, nonfarm business productivity grew 0.5 percent- CBO has also examined the sensitivity of its budget projec- age points more slowly each year than projected, federal tions to other factors, as it has done in the past. Last year, for debt held by the public would be 185 percent of GDP in example, CBO examined the extent to which federal debt as a 2049. percentage of GDP would differ from amounts in the extended baseline projections if four key factors underlying its analysis Interest on Federal Debt Held by the Public varied by fixed amounts: the labor force participation rate, the CBO also examined the variation in interest rates on federal growth rate of total factor productivity in the nonfarm business debt held by the public over 30-year periods since 1949 to sector, interest rates on federal debt held by the public, and assess the extent to which unexpected changes in financial excess cost growth for Medicare and Medicaid spending.4 factors contribute to changes in those rates. (The agency esti- The degree of variation in each of those factors was based on mated how much of that variation could be explained directly historical movements and considered the effects of possible or indirectly by economic or budgetary factors; the remaining, future developments. Estimates of the budgetary outcomes of unexplained variation was the contribution of unexpected alternative paths for each of the four factors (including the two changes in financial factors.) On the basis of that analysis, discussed here), as well as estimates of the effects when all CBO determined that future outcomes are likely to fall within a four factors vary simultaneously, are presented in the supple- range of 2 percentage points around the agency’s central esti- mental data posted along with this report on CBO’s website mate. Thus, CBO projected economic and budgetary outcomes (www.cbo.gov/publication/55331). using interest rates on federal debt that were 1.0 percentage point higher (before accounting for macroeconomic effects) and 1.0 percentage point lower than the agency’s central estimates. 4. See Congressional Budget Office, The 2018 Long-Term Budget Outlook •• If federal borrowing rates were 1.0 percentage point higher (June 2018), pp. 23–35, www.cbo.gov/publication/53919. The labor force participation rate is the percentage of people in the civilian each year than CBO projects, federal debt held by the noninstitutionalized population who are age 16 or older and either working or actively seeking work. Excess cost growth is the extent to which the 3. CBO’s extended baseline projections generally reflect current law, following growth rate of nominal health care spending per person (adjusted to the agency’s 10-year baseline budget projections through 2029 and then remove the effects of aging) exceeds the growth rate of potential gross extending most of the concepts underlying those projections for the rest of domestic product per person. Potential GDP per person is the maximum the long-term projection period. sustainable output of the economy. factor as well as the uncertainty associated with the long- current trends, CBO estimates that there is roughly a term correlation among some factors. two-thirds chance that those factors would fall within the following ranges: Uncertainty of Demographic Factors CBO projected a range of outcomes for the total fertility •• The total fertility rate would be between 1.8 and rate, the rate of mortality improvement, and the rate of 2.0 births per woman, or between 5 percent lower net immigration, all of which affect the size of the popu- and 5 percent higher than the rate underlying the lation (see Table 1-5). On the basis of historical data and extended baseline projections. 34 The 2019 Long-term budget outlook june 2019 Table 1-5 . Middle Two-Thirds of the Ranges of Possible Outcomes for Key Factors Used to Quantify the Uncertainty in CBO’s Projections of Federal Debt Held by the Public Average Annual Value, 2019 to 2039 Low Extended Baseline High Demographic Factors Total fertility rate (Children per woman) 1.8 1.9 2.0 Rate of mortality improvement (Percent) a 0.6 0.9 1.2 Net immigration rate (Per 1,000 people in the U.S. population)  2.9 3.1 3.3 Economic Factors (Percent) Growth in total factor productivity in the nonfarm business sector b 0.6 1.1 1.6 Interest rate on 10-year Treasury securities 2.4 3.8 5.1 Civilian unemployment rate c 3.7 4.6 5.5 Excess Cost Growth (Percent) d Medicare 0.5 1.1 1.7 Medicaid 1.2 1.8 2.4 Source: Congressional Budget Office. The extended baseline projections generally reflect current law, following CBO’s 10-year baseline budget projections through 2029 and then extending most of the concepts underlying those projections for the rest of the long-term projection period. GDP = gross domestic product. a.The rate of mortality improvement is the average annual percentage improvement in mortality rates. A mortality rate is the number of deaths per thousand people; improved (that is, lower) mortality rates result in longer life expectancy. b.Total factor productivity in the nonfarm business sector measures the growth of the average real (inflation-adjusted) output per unit of combined labor and capital services. c.The civilian unemployment rate is the percentage of people in the labor force who are unemployed. Labor force data are restricted to the civilian noninstitutionalized population who are age 16 or older and are either working or actively seeking work. d.Excess cost growth is the extent to which the growth rate of nominal health care spending per person (adjusted to remove the effects of aging) exceeds the growth rate of potential GDP per person. (Potential GDP is the maximum sustainable output of the economy.) •• The rate of mortality improvement would be between 6½ percent less than and 6½ percent greater than the 0.6 percent and 1.2 percent, or as much as one-third rate underlying the extended baseline projections.31 lower or higher than the rate underlying the extended baseline projections. If rates of mortality improvement As a result of varying those key demographic factors, in fell within that range, life expectancy at birth would two-thirds of CBO’s simulations the civilian noninsti- be between 80.5 and 82.2 years by 2039, or between tutionalized population grew by an average of between 1 percent shorter and 1 percent longer than the 0.6 percent and 0.7 percent from 2019 to 2039. If the value underlying the extended baseline projections. growth of that population was within that range, by (Life expectancy at age 65 would be between 20.3 2039 that population would be between 293 million and 21.4 years, or 2½ percent shorter to 2½ percent and 297 million, or between 2 million (or 0.7 percent) longer than CBO’s baseline value.) smaller than and 2 million (or 0.7 percent) larger than the population underlying CBO’s extended baseline •• The average net immigration rate over the 20-year projections. projection period would be between 2.9 and 3.3 per thousand people in the U.S. population, or between 31. For some categories of immigrants, the number of people admitted in any year is restricted by caps. For this analysis, the net number of immigrants in those categories does not vary. Net inflows of foreign-born people without legal status are highly uncertain, and those net flows are varied for this analysis. CHAPTER 1: THE BUDGET OUTLOOK FOR THE NEXT 30 YEARS The 2019 Long-term budget outlook 35 Uncertainty of Economic Factors and the Changes From Last Year’s Growth of Health Care Costs Long-Term Budget Outlook CBO examined three economic factors and the growth of As a share of GDP, federal debt and deficits are now health care costs in its uncertainty analysis. In the agen- projected to be lower over the next three decades than cy’s assessment, there is roughly a two-thirds chance that CBO projected last year. In the agency’s current extended the following outcomes would occur over the next two baseline projections, debt is equal to 141 percent of GDP decades (see Table 1-5 on page 34): in 2048, which is 11 percentage points lower than the amount the agency projected last year. Projected deficits •• The average growth rate of total factor productivity (both primary and total) as a share of GDP in this year’s in the nonfarm business sector would be between report are smaller throughout the entire projection period 0.6 percent and 1.6 percent, or as much as than those in last year’s report. (See Appendix B for more 0.5 percentage points less than or greater than the rate information on changes in the long-term budget projec- underlying CBO’s extended baseline projections. tions since last year.) •• The average interest rate on 10-year U.S. Treasury The revised projections of debt and deficits resulted securities would be between 2.4 percent and primarily from a reduction in projected outlays, specif- 5.1 percent, or as much as 1.4 percentage points ically in discretionary spending and in net spending for lower or higher than the rate underlying the extended interest, which was partially offset by a small reduction in baseline projections. projected revenues. This year’s projections of discretionary spending are lower than last year’s projections because •• The average civilian unemployment rate would be appropriations for relief and recovery efforts related to between 3.7 percent and 5.5 percent, or as much as hurricanes and wildfires were smaller in 2019 than they 0.9 percentage points less than or greater than the rate were in 2018. (Projections for future years are based on underlying the extended baseline projections. the 2019 appropriations.) This year’s projections of net spending for interest are lower because less debt is pro- •• The rates of excess cost growth for Medicare jected to be accumulated and because CBO has revised and Medicaid spending would be as much as downward its projections of the average interest rate on 0.6 percentage points higher or lower than the rates that debt. Revenues are projected to be slightly lower than underlying the extended baseline projections. they were in last year’s projections because of new admin- istrative and tax data. Changes in those factors would affect the budget in important ways. For example, if the unemployment rate The 75-year actuarial deficit currently projected for was higher than projected, the economy’s output and tax Social Security is 1.5 percent of GDP (the same amount revenues would be less than they are in CBO’s extended that CBO estimated last year) or 4.6 percent of taxable baseline projections, and consequently, federal deficits payroll (which is slightly larger than last year’s estimate and debt would be greater than the agency projects. By of 4.4 percent). Those projections reflect several devel- contrast, if the rates of cost growth for Medicare and opments since last year. The actuarial deficit increased Medicaid were lower than projected, deficits and debt partly because CBO lowered its projections of payroll would be less than they are in the extended baseline pro- taxes. Also, the agency incorporated another year with jections, primarily because outlays would be smaller. a relatively large deficit into the analysis. Largely off- setting those increases, however, was a downward revi- sion that CBO made to its projections of outlays for Social Security. CHAPTER 2 Chapter 2 The Long-Term Budget Outlook Under Alternative Scenarios Overview •• Spending on Social Security benefits for older people This chapter expands on the analysis in Chapter 1 in would be greatly curtailed, leading to increases in the various ways. First, it shows how the federal budget and overall labor supply and private saving. the nation’s economy would evolve under an extended alternative fiscal scenario in which substantial tax •• That drop in benefits would induce beneficiaries to increases and discretionary spending cuts would not take reduce their spending, causing real GDP to be lower place as scheduled; instead, current law would change in the short term; but real GDP would be higher to maintain certain major policies that are now in place. in the longer term, when the reduction in federal Compared with outcomes in the Congressional Budget deficits would boost the funds available for private Office’s extended baseline projections, which generally investment. reflect current law, outcomes under the extended alterna- tive fiscal scenario would differ in the following ways: •• The risk of a fiscal crisis occurring would be lower over the longer run. In addition, the risk of negative •• Federal deficits and debt would be far larger. economic and financial effects that were less abrupt but still significant would be lower. •• Real gross domestic product (GDP) would be lower in the long run. (Real GDP is nominal GDP that has Third, the chapter examines the size and timing of policy been adjusted to remove the effects of inflation.) changes needed to meet various goals for deficit reduc- tion. (The policy changes examined here are illustrative, •• Federal spending would be higher, and most and the results do not reflect any particular assumptions taxpayers would pay less in taxes. about specific changes.) If lawmakers aimed for debt as a share of GDP in 2049 to fall to its 50-year average •• The risk of a fiscal crisis occurring would be greater through across-the-board fiscal adjustments of equal over the longer run. In addition, the risk of negative size (as a percentage of GDP) each year, for example, economic and financial effects that were less abrupt they could reach that goal by increasing revenues or by but still significant would be greater. decreasing spending by $1,900 per person in 2020, CBO projects. Second, this chapter presents an analysis under which Social Security benefits are limited to the amounts pay- Additionally, the timing of deficit reduction has impli- able from revenues received by the Social Security trust cations for its effects, in terms of costs and benefits, funds. Under that payable-benefits scenario, spending for on different generations of the U.S. population. CBO Social Security would be significantly lower than it is estimates that delaying policy action would require in the extended baseline projections. Other outcomes larger changes in revenues and outlays to reach a given relative to CBO’s extended baseline projections are the level of debt as a percentage of GDP by 2049. That is, following: making policy changes in 2025 or 2030 that aimed to achieve a target ratio of debt to GDP would require a •• Federal deficits and debt as a percentage of GDP greater percentage reduction in noninterest spending or a would be lower. larger percentage increase in revenues than making such changes in 2020. 38 THE 2019 LONG-TERM BUDGET OUTLOOK june 2019 Furthermore, delaying policy action would reduce the •• Certain temporary tax provisions that have recently well-being of younger and future generations while expired or are scheduled to expire in coming years, improving the well-being of older generations. Even including several trade preference programs, are though the burden of delaying policy action would be permanently extended. borne by future generations, income among those gener- ations is projected to be higher, on average, owing to the •• Certain postponed taxes established by the Affordable growth of the U.S. economy. Care Act are repealed. Budgetary and Economic Effects of an As a result, in 2029, discretionary outlays are projected Alternative Fiscal Scenario to total 5.7 percent of GDP, 0.7 percentage points CBO examined budgetary and economic outcomes under greater than they are in the extended baseline projections. an extended alternative fiscal scenario. Under that sce- Revenues are projected to total 17.0 percent of GDP, nario, current law would be changed to maintain certain 1.3 percentage points lower than they are in the extended policies that are now in place. As a result, deficits would baseline projections. be larger than they are in CBO’s extended baseline pro- jections. For example, the deficit would be $774 billion After 2029, projections of discretionary spending reflect larger in 2029—about 60 percent larger than the deficit the assumption that such spending would remain roughly in CBO’s baseline projections. Federal debt would equal constant as a percentage of GDP.2 By 2049, that amount 219 percent of GDP in 2049 and continue to rise in later would exceed outlays in the extended baseline projections years. by 0.9 percentage points. In the extended alternative fiscal scenario, spending and Extending the expiring tax provisions is projected to tax policies for the first 10 years are identical to those in lower revenues (relative to amounts in the extended CBO’s alternative fiscal scenario.1 The budgetary out- baseline projections) by an average of 1.5 percent of GDP comes differ, however, because in addition to the conven- each year between 2030 and 2049. tional estimates, this report incorporates estimated effects of the changes in fiscal policy on the economy and the Nevertheless, revenues as a share of GDP trend upward effects of those economic changes on the budget. Over under this scenario, mostly because of structural features the next 10 years, the extended alternative fiscal scenario of the tax code; they reach 17.6 percent of GDP in 2049. incorporates the following features: That upward trend differs from historical experience, however. Over the past 50 years, federal revenues as a •• The caps on discretionary appropriations currently percentage of GDP have fluctuated around their 50-year in effect through 2021 cease after 2019, and average of 17.4 percent with no evident long-term trend. appropriations instead grow at the same rate as inflation in each year. How CBO Analyzed Outcomes Under the Extended Alternative Fiscal Scenario •• The expiring revenue provisions of the 2017 tax Relative to the fiscal policy in place under current law, act—including provisions that specify tax rates and fiscal policy under this scenario would reflect significant brackets, the number of allowable deductions, the changes. Those changes are projected to have effects on size of the child tax credit and the portion that is the economy that would feed back to budgetary out- refundable, and the reach of the alternative minimum comes. CBO has not analyzed every way in which those tax—are extended. changes would affect the economy in the long term. Instead, for the simplified analysis presented in this •• The expansion of bonus depreciation for businesses report, CBO has analyzed three of those effects.3 deducting certain investments is held at 100 percent. 2. That assumption also underlies the extended baseline projections. See Table 1-2 on page 19. 3. For a general explanation of how CBO analyzes the effects of fiscal policies, see Congressional Budget Office, How CBO Analyzes 1. See Congressional Budget Office, Updated Budget Projections: the Effects of Changes in Federal Fiscal Policies on the Economy 2019 to 2029 (May 2019), www.cbo.gov/publication/55151. (November 2014), www.cbo.gov/publication/49494. CHAPTER 2: THE LONG-TERM BUDGET OUTLOOK UNDER ALTERNATIVE SCENARIOS THE 2019 LONG-TERM BUDGET OUTLOOK 39 •• Effective marginal tax rates on labor income would within U.S. borders (including the contributions of be lower under the extended alternative fiscal scenario foreign-owned capital and labor). GNP is important than they are in the extended baseline projections, because it is a more complete measure of the income encouraging people to work and save more and available to U.S. residents. (GNP differs from GDP by thereby increasing output.4 including the income that U.S. residents earn abroad and excluding the income that nonresidents earn from •• Effective marginal tax rates on income from most domestic sources.) Under the extended alternative fiscal types of capital would also be lower, which would scenario, the amount of federal debt owned by foreigners encourage saving and investment and again increase and the inflows of foreign capital are larger than they are output.5 in CBO’s extended baseline projections. As a result, the long-term negative effects of that debt on GNP are larger •• Federal debt would be greater under the extended than the negative effects on GDP. alternative fiscal scenario than it is in the extended baseline projections—drawing money away from Budgetary and Economic Outcomes Under the (or “crowding out”) investment in capital goods and Extended Alternative Fiscal Scenario services, reducing the stock of private capital, and Under the extended alternative fiscal scenario, CBO proj- making output smaller than it would be otherwise. ects, the primary deficit (which excludes interest costs) in 2049 would be 6.1 percent of GDP. (In the extended In addition to those three effects, any changes to fis- baseline projections, it is 3.0 percent of GDP.) Once the cal policy could alter people’s incentives in other ways, rising costs of debt service are added, the total deficit in possibly resulting in significant long-term changes to the 2049 would equal 15.5 percent, not the 8.7 percent of economy. For example, changes to tax policy might alter GDP it equals in CBO’s extended baseline projections businesses’ choices about how they were structured, and (see Table 2-1). those choices might then alter the effective marginal tax rate on capital income. Similarly, changes in the tax treat- CBO projects the following outcomes in 2049. ment of mortgage debt would affect households’ decisions (Amounts in the extended baseline projections are about how much to save. Because this analysis is simpli- shown in parentheses.) fied, it does not incorporate those effects. •• Net interest costs would be 9.4 percent of GDP CBO also analyzed short-term outcomes under the (rather than 5.7 percent). extended alternative fiscal scenario. Policies that increased spending or reduced revenues would boost overall •• Total spending excluding interest payments would be demand for goods and services over the next few years, 23.7 percent of GDP (rather than 22.5 percent). thereby making output and employment in the short term higher than they would be otherwise. •• Revenues would be 17.6 percent of GDP (rather than 19.5 percent). CBO estimated the effects of this scenario on both GDP and GNP (gross national product). Each of those mea- •• Debt held by the public would be 219 percent of sures is important for different reasons. GDP is import- GDP (rather than 144 percent). ant because by accounting for effects on domestic eco- nomic and income growth, it helps assess the productive The crowding out of private investment, the smaller capacity—and therefore the tax base—of the economy capital stock, and the larger supply of labor would, on balance, cause output to be lower and interest rates to 4. The effective marginal tax rate on labor income is the share of an be higher in the long term under the extended alterna- additional dollar of such income that is paid in federal individual tive fiscal scenario than they are in the extended baseline income taxes and payroll taxes—averaged among taxpayers, with weights proportional to their labor income. projections. In 2049, for instance, real GDP would be 2.5 percent lower (see Table 2-2). In addition, real GNP 5. The effective marginal tax rate on capital income is the share in 2049 would be 3.6 percent lower, and real GNP per of the return on an additional dollar of investment made in a particular year that will be paid in taxes over the life of that person would be about $3,400 lower (see Figure 2-1 on investment. page 42). Also, the interest rate on 10-year Treasury 40 THE 2019 LONG-TERM BUDGET OUTLOOK june 2019 Table 2-1 . the risk of a fiscal crisis and would limit lawmakers’ Budget Projections Under Three Scenarios ability to respond to unforeseen events. Negative eco- nomic and financial effects that were less abrupt but still Percentage of Gross Domestic Product significant—such as higher inflation expectations or an 2029 2049 increased burden of financing public and private activ- Revenues ity in international markets—would also have a greater Extended Baseline 18.3 19.5 chance of occurring under this scenario. Those effects Extended Alternative Fiscal Scenario 17.0 17.6 would worsen the consequences associated with high and Payable-Benefits Scenario 18.3 19.6 rising federal debt. Spending Excluding Interest Payments The policies underlying the extended alternative fiscal scenario would have short-term effects as well. Over the Extended Baseline 19.8 22.5 Extended Alternative Fiscal Scenario 20.6 23.7 next few years, greater federal spending would boost the Payable-Benefits Scenario 19.8 20.5 overall demand for goods and services, causing output to be higher than it otherwise would be. In CBO’s estima- Deficit (-) or Surplus, Excluding tion, real GDP would be 0.7 percent higher in 2020 and Interest Payments 0.4 percent higher in 2021 than it is in the extended Extended Baseline -1.6 -3.0 Extended Alternative Fiscal Scenario -3.6 -6.1 baseline projections. In addition, the Federal Reserve Payable-Benefits Scenario -1.6 -0.9 would respond, in CBO’s view, by raising interest rates to restrain the boost in overall demand and prevent infla- Total Deficit (-) or Surplus tion from rising above the central bank’s goal. As a result, Extended Baseline -4.5 -8.7 the interest rate on 10-year Treasury securities would be Extended Alternative Fiscal Scenario -7.0 -15.5 Payable-Benefits Scenario -4.5 -4.9 0.2 percentage points higher in 2020 and 2021 than it is in the extended baseline projections, CBO estimates (see Federal Debt Held by the Public Table 2-3 on page 43). Extended Baseline 92 144 Extended Alternative Fiscal Scenario 105 219 The economic and budgetary effects of the policies under- Payable-Benefits Scenario 92 106 lying the extended alternative fiscal scenario are highly Source: Congressional Budget Office. uncertain, as are the effects of the extended baseline. That uncertainty arises mainly from two sources: uncertainty The extended baseline projections generally reflect current law, following CBO’s 10-year baseline budget projections through 2029 and about future economic conditions and demographic then extending most of the concepts underlying those projections for the trends, and uncertainty about the macroeconomic effects rest of the long-term projection period. of policy changes. If future economic and demographic Under the extended alternative fiscal scenario, substantial tax increases conditions and their responses to policy changes differed and discretionary spending cuts would not take place as scheduled; from CBO’s projections, budgetary and economic out- instead, current law would be changed to maintain certain major policies comes would differ from those the agency estimates under that are now in place. Under the payable-benefits scenario, spending the extended alternative fiscal scenario. for Social Security would be significantly lower than it is in the extended baseline projections. For example, if federal borrowing rates were 0.1 percent- The estimates of deficits, surpluses, and debt include macroeconomic feedback. age point higher (or lower) than they are in the extended baseline projections, debt in the extended alternative fiscal scenario would be 225 percent of GDP (or 212 per- securities in 2049 would be 0.4 percentage points higher cent of GDP) rather than 219 percent in 2049. If total than the rate in CBO’s extended baseline projections. factor productivity growth was 0.1 percentage point higher (or lower), debt would be 209 percent of GDP (or In addition to the effects on output and interest rates 228 percent of GDP). Those estimated effects are roughly reported here, other effects would occur under the scalable for moderate changes in the economic variables. extended alternative fiscal scenario. In particular, the In particular, if interest rates were more than 0.5 percent- significant increase in federal borrowing would elevate age points higher than they are in the extended baseline projections or total factor productivity growth was more CHAPTER 2: THE LONG-TERM BUDGET OUTLOOK UNDER ALTERNATIVE SCENARIOS THE 2019 LONG-TERM BUDGET OUTLOOK 41 Table 2-2 . declined to zero and current revenues were insufficient Long-Term Economic Effects Under Two Scenarios to pay benefits specified in law, the Social Security Administration would no longer be permitted to pay ben- Relative to CBO’s Extended Baseline Projections eficiaries the full amounts to which they were entitled.7 2029 2049 CBO analyzed a payable-benefits scenario in which Social Real GDP (Percent) Security benefits would be limited to the amounts pay- Extended Alternative Fiscal Scenario -0.1 -2.5 able from dedicated funding sources beginning in 2033. Payable-Benefits Scenario n.a. 1.7 Interest Rates on 10-Year Although it is unclear how much payments for specific Treasury Securities beneficiaries would be reduced if total benefits were (Percentage points) limited to the amounts payable from dedicated fund- Extended Alternative Fiscal Scenario 0.1 0.4 ing, CBO estimated the amount of the total reduction Payable- Benefits Scenario n.a. -0.2 in annual benefits that would be necessary for the trust funds’ outlays to match revenues in each year after the Source: Congressional Budget Office. funds were exhausted. The required reduction would The extended baseline projections generally reflect current law, amount to 24 percent in 2033 and rise gradually to following CBO’s 10-year baseline budget projections through 2029 and 29 percent in 2049 (relative to the amounts in CBO’s then extending most of the concepts underlying those projections for the rest of the long-term projection period. extended baseline projections). Under the extended alternative fiscal scenario, substantial tax increases and discretionary spending cuts would not take place as scheduled; In CBO’s assessment, if benefits paid out were limited instead, current law would be changed to maintain certain major policies to revenues received by the Social Security trust funds, that are now in place. Under the payable-benefits scenario, spending federal deficits would decrease by 1.5 percent of GDP in for Social Security would be significantly lower than it is in the extended 2033 and by 3.8 percent in 2049 (relative to the amounts baseline projections. in CBO’s extended baseline projections). The cut in n.a. = not applicable. benefits would not be announced until 2033 and would therefore be unexpected (which matters for the projection of macroeconomic effects). That abrupt cut in benefits than 0.3 percentage points lower, projected debt as a in 2033 would cause a substantial drop in consumer percentage of GDP under the extended alternative fiscal spending and a corresponding increase in saving. It would scenario would grow to levels well outside of U.S. his- also probably induce some older workers to work more torical experience, which provides the empirical basis for hours or to delay retirement and save more. In addition, CBO’s models. some Social Security beneficiaries might return to work to supplement their income. Budgetary and Economic Effects of a Payable-Benefits Scenario Under the payable-benefits scenario, changes in over- Without legislative action, the combined trust funds all demand would lower GDP in the first few years for Social Security (known as Old-Age, Survivors, and following the reduction in benefits. In the long run, Disability Insurance, or OASDI) are projected to be however, increases in the labor supply and investment exhausted in calendar year 2032. Beyond that point, trust stemming from smaller budget deficits would boost fund balances would no longer be available to make up output and reduce interest rates. Those changes, which the gap between benefits specified in current law and are measured relative to amounts in CBO’s extended annual trust fund receipts. CBO’s extended baseline baseline projections, would generally decrease income projections reflect the assumption that the Social Security and wealth for older generations and increase them for Administration will pay benefits as scheduled under current law regardless of the status of the program’s trust 7. The balances of the trust funds represent the total amount that funds.6 However, if the trust funds’ combined balance the government is legally authorized to spend. For more details about the legal issues related to exhaustion of a trust fund, see 6. That approach is consistent with the requirement that CBO’s William R. Morton and Barry F. Huston, Social Security: What 10-year baseline projections incorporate the assumption that Would Happen If the Trust Funds Ran Out? Report for Congress funding for such programs is adequate to make all payments RL33514 (Congressional Research Service, June 11, 2018), required by law. https://go.usa.gov/xEtaw. 42 THE 2019 LONG-TERM BUDGET OUTLOOK june 2019 Figure 2-1 . Output per Person and Debt Under Three Scenarios Real Gross National Product per Person Thousands of 2019 Dollars, by Calendar Year 100 Actual Projected 97 Payable-Benefits Scenario 94 Extended Baseline 90 91 Extended Alternative Fiscal Scenario 80 70 60 50 40 30 20 10 0 2004 2009 2014 2019 2024 2029 2034 2039 2044 2049 Federal Debt Held by the Public Percentage of Gross Domestic Product, by Fiscal Year 220 Actual Projected 219 Extended Alternative Fiscal Scenario 200 180 160 140 144 Extended Baseline 120 106 Payable-Benefits Scenario 100 80 60 40 20 0 2004 2009 2014 2019 2024 2029 2034 2039 2044 2049 Source: Congressional Budget Office. The extended baseline projections generally reflect current law, following CBO’s 10-year baseline budget projections through 2029 and then extending most of the concepts underlying those projections for the rest of the long-term projection period. Under the extended alternative fiscal scenario, substantial tax increases and discretionary spending cuts would not take place as scheduled; instead, current law would be changed to maintain certain major policies that are now in place. Under the payable-benefits scenario, spending for Social Security would be significantly lower than it is in the extended baseline projections. Gross national product differs from gross domestic product, the more common measure of the output of the economy, by including the income that U.S. residents earn abroad and excluding the income that nonresidents earn in this country. The estimates of deficits, surpluses, and debt include macroeconomic feedback. CHAPTER 2: THE LONG-TERM BUDGET OUTLOOK UNDER ALTERNATIVE SCENARIOS THE 2019 LONG-TERM BUDGET OUTLOOK 43 Table 2-3 . Short-Term Economic Effects Under Two Scenarios Relative to the Extended Baseline Projections 2020 2021 2033 2034 Real GDP (Percent) Extended Alternative Fiscal Scenario 0.7 0.4 n.a. n.a. Payable-Benefits Scenario n.a. n.a. -0.8 -0.1 Interest Rates on 10-Year Treasury Securities (Percentage points) Extended Alternative Fiscal Scenario 0.2 0.2 n.a. n.a. Payable-Benefits Scenario n.a. n.a. -0.7 -0.4 Source: Congressional Budget Office. The extended baseline projections generally reflect current law, following CBO’s 10-year baseline budget projections through 2029 and then extending most of the concepts underlying those projections for the rest of the long-term projection period. Under the extended alternative fiscal scenario, substantial tax increases and discretionary spending cuts would not take place as scheduled; instead, current law would be changed to maintain certain major policies that are now in place. Under the payable-benefits scenario, spending for Social Security would be significantly lower than it is in the extended baseline projections. n.a. = not applicable. younger ones. That shift would stem not only from the is in the extended baseline projections in 2033 and direct effects of a drop in benefits, but also from macro- 2034. economic effects that would raise wages in the long run. Incorporating those macroeconomic effects into its •• The benefit cuts would cause some people to work analysis, CBO projects that the debt-to-GDP ratio would more and some to remain in the labor force longer stand at 106 percent in 2049, 38 percentage points below than they would have otherwise. Both of those factors the extended baseline projection—but still well above the would expand the supply of labor and thus the current level. economy’s output in the long term. How CBO Analyzed Outcomes Under the •• In CBO’s assessment, some workers who have not Payable-Benefits Scenario yet retired would respond to the prospect of smaller As with the extended alternative fiscal scenario, this benefit payments by boosting their saving and scenario represents significant changes to the fiscal policy reducing their spending.8 Those changes would lessen projected under current law. Because benefit cuts would the effect that smaller future benefit payments would be unexpected, workers would not adjust their saving and have on households’ future income and spending. hours worked beforehand. Hence, projections under this The resulting increases in saving and the labor supply scenario do not differ from those in the extended baseline would boost the capital stock and GDP. until 2033, when those cuts would begin. Thereafter, people would expect benefits to be reduced permanently. •• Federal debt would be lower than it is in the extended As a result, changes in investment and the labor supply baseline projections—increasing the amount of money would lead in the long term to greater output and lower available for (or “crowding in”) private investment interest rates than in CBO’s extended baseline projec- in capital goods and services, boosting the stock of tions. Although CBO has not analyzed every way in private capital, and making output greater than it which those changes would affect the economy in the would be otherwise. long term, the agency analyzed four of those effects for this report. •• The reduction in benefits would decrease retirees’ 8. In this analysis, CBO did not address the potential effects of income, pushing down the overall demand for goods moving households’ savings into or out of tax-deferred or taxable and services and causing output to be lower than it savings accounts. 44 THE 2019 LONG-TERM BUDGET OUTLOOK june 2019 Budgetary and Economic Outcomes Under the the central bank’s longer-term goal. In addition, the Payable-Benefits Scenario increase in the saving rate—and other factors—would In 2049, primary deficits under the payable-­ enefits b further reduce interest rates. Taken together, those effects scenario would be smaller than they are in CBO’s would cause the interest rate on 10-year Treasury secu- extended baseline projections—0.9 percent of GDP rities to be 0.7 percentage points lower in 2033 and instead of 3.0 percent of GDP. Adding debt-service costs 0.4 percentage points lower in 2034 than it is in the raises those amounts to 4.9 percent of GDP under the extended baseline projections, in CBO’s estimation (see payable-­ enefits scenario and to 8.7 percent of GDP in b Table 2-3 on page 43). the extended baseline projections (see Table 2-1 on page 40). The economic and budgetary effects of the policies under- lying the payable-benefits scenario are highly uncertain, For the payable-benefits scenario, CBO projects the as are the effects of the extended baseline. That uncer- following outcomes in 2049 (compared with outcomes in tainty arises mainly from two sources: uncertainty about the extended baseline): future economic conditions and demographic trends, and uncertainty about how reductions in Social Security •• Net interest costs would be 4.0 percent of GDP benefits would affect the economy and the budget. If (rather than 5.7 percent). future economic and demographic conditions and the macroeconomic effects of reduced Social Security benefits •• Total spending excluding net interest costs would be differed from CBO’s projections, budgetary and eco- 20.5 percent of GDP (rather than 22.5 percent). nomic outcomes would differ from those the agency esti- mates under the payable-benefits scenario. For example, •• Revenues would be 19.6 percent of GDP (rather than if interest rates on federal debt were 0.1 percentage point 19.5 percent). higher (or lower) than they are in the extended baseline projections, debt in the payable-benefits scenario would •• Debt would be 106 percent of GDP (rather than be 109 percent of GDP (or 102 percent of GDP) rather 144 percent). than 106 percent in 2049. If total factor productivity growth was 0.1 percentage point higher (or lower), debt In CBO’s assessment, the crowding in of private invest- would be 100 percent of GDP (or 111 percent of GDP). ment and the increase in the supply of labor and the capital Those estimated effects are roughly scalable for moderate stock would cause output to be higher and interest rates changes in the economic variables. to be lower in the long term under the payable-benefits scenario than they are in the extended baseline projec- The Size and Timing of Policy Changes tions. Specifically, real GDP would be 1.7 percent higher Needed to Meet Various Goals for in 2049, CBO estimates (see Table 2-2 on page 41). Deficit Reduction In addition, real GNP would be 2.3 percent higher CBO estimated the size of changes in spending or in 2049, and real GNP per person would be about revenues that would be needed if lawmakers wanted to $2,200 higher in that year (see Figure 2-1 on page 42). achieve some specific targets for federal debt held by the In contrast, the interest rate on 10-year Treasury securities public. The agency also assessed the extent to which the would be 0.2 percentage points lower under this scenario size of policy adjustments would change if such deficit than it is in CBO’s extended baseline projections. reduction occurred later, and it examined how waiting to resolve the long-term fiscal imbalance would affect differ- The policies underlying the payable-benefits scenario ent generations of the U.S. population. would have short-term effects as well. In CBO’s assess- ment, people would respond to smaller benefit payments The Size of Policy Changes Needed to Meet Various by reducing their spending, which would decrease the Goals for Deficit Reduction overall demand for goods and services. As a result, real If lawmakers wanted debt in 2049 to match its current GDP would be 0.8 percent lower in 2033 and 0.1 per- level of 78 percent of GDP, they could cut noninterest cent lower in 2034 than it is in the extended baseline spending or raise revenues (or do both) in each year projections, CBO estimates. In CBO’s view, the Federal beginning in 2020 by amounts totaling 1.8 percent of Reserve would respond by lowering interest rates to boost GDP (see Figure 2-2). In 2020, 1.8 percent of GDP overall demand and prevent inflation from falling below would be about $400 billion, or $1,200 per person. If CHAPTER 2: THE LONG-TERM BUDGET OUTLOOK UNDER ALTERNATIVE SCENARIOS THE 2019 LONG-TERM BUDGET OUTLOOK 45 Figure 2-2 . The Size of Policy Changes Needed to Make Federal Debt Meet Two Possible Goals in 2049 If lawmakers aimed for debt in 2049 to equal: 42% of GDP (Its 50-year average) 78% of GDP (Its current level) Each year, they would need to reduce deficits as a share of GDP by: increase in 16% 2.9% 1.8% increase in revenues of GDP 11% revenues or of GDP or decrease in 15% 10% decrease in spending spending In 2020, that would amount to a reduction of: $630 billion $400 billion = $20 billion = $20 billion If the changes were equal percentage increases in all types of revenues, taxes per household in 2020 would be higher than they would be under current law by: +$2,100 +$1,400 Values are for households in the middle fifth of the income distribution. $12,500 $14,600 $12,500 $13,900 Current Average Tax New Average Tax Current Average Tax New Average Tax If the changes were equal percentage cuts in all types of noninterest spending, initial Social Security benefits in 2020 would be lower than they would be under current law by: -$2,800 -$1,900 Values are averages for people in the middle fifth of the lifetime earnings distribution who were born in the 1950s and who would claim benefits at age 65. $16,600 $19,400 $17,600 $19,400 New Average Benefits Current Average Benefits New Average Benefits Current Average Benefits Source: Congressional Budget Office In this figure, the indicated sizes of policy changes are relative to CBO’s extended baseline projections, which generally reflect current law, following CBO’s 10-year baseline budget projections through 2029 and then extending most of the concepts underlying those projections for the rest of the long-term projection period. The projected effects of the policy changes on debt include the direct effects of the policy changes and the feedback to the federal budget from faster economic growth. The effects on growth and the feedback to the federal budget reflect the positive economic effects of lowering the debt but do not reflect any assumptions about the specific details of the policy changes. GDP = gross domestic product. 46 THE 2019 LONG-TERM BUDGET OUTLOOK june 2019 such an adjustment was made in each year, the budget however, and the results do not reflect any particular would show a primary surplus of 0.2 percent of GDP assumptions about specific changes. Any policy change in 2030 and a primary deficit of 0.7 percent of GDP could alter productivity growth and people’s incentives to by 2049. If the changes came entirely from revenues or work and save, which would in turn affect overall eco- spending, they would amount to an 11 percent increase nomic output and feed back to the federal budget. in revenues or a 10 percent cut in noninterest spend- ing (relative to amounts in CBO’s extended baseline The Timing of Policy Changes Needed to Meet Various projections). Goals for Deficit Reduction The size of the policy changes needed to achieve a partic- Increases in revenues or cuts in noninterest spending ular goal for federal debt would depend, in part, on how would need to be larger than 1.8 percent of GDP to quickly that goal was expected to be reached. Regardless reduce debt to levels recorded in recent decades. If law- of the chosen goal for federal debt, lawmakers would makers wanted to decrease debt to 42 percent of GDP face trade-offs in deciding how quickly to implement (its average over the past 50 years) by 2049, they could policies designed to reduce or stabilize debt as a percent- increase revenues or cut noninterest spending (in relation age of GDP. The benefits of reducing the deficit sooner to amounts under current law) or adopt some combina- would include a smaller accumulated debt, smaller policy tion of those two actions beginning in 2020 by amounts changes required to achieve long-term outcomes, and less totaling 2.9 percent of GDP each year. In 2020, 2.9 per- uncertainty about the policies lawmakers would adopt. cent of GDP would be about $630 billion, or $1,900 per If lawmakers cut spending or increased taxes abruptly, person. people might have insufficient time to plan for or to adjust to the new system. To lower debt to its average over the past 50 years solely by increasing revenues or cutting noninterest spending, Over the first several years following their adoption, such lawmakers could make the following changes: policy changes would dampen overall demand for goods and services, thus decreasing output and employment •• If collections of the various types of revenues were below amounts projected under current law. That damp- increased proportionally, total revenues would need to ening effect is expected to be temporary, however, because be about 16 percent higher each year over the 2020– of how prices and interest rates would respond to the 2049 period. On average, that adjustment would reductions in demand and to the resulting actions by the result in federal taxes that were about $2,100 higher Federal Reserve. than they are under current law for households in the middle fifth of the income distribution in 2020. By contrast, if policymakers waited longer to reduce federal spending or increase taxes, more debt would •• If all types of noninterest spending were cut by an accumulate, which would slow the growth of output and equal percentage, spending overall would need to be income. Delaying implementation would thus mean that about 15 percent lower in each of the next 30 years. reaching any chosen target for debt would require larger For example, such cuts would lower initial annual changes. Nonetheless, if policymakers waited longer to Social Security benefits by about $2,800, on average, enact deficit-reduction policies, the economy probably for people in the middle fifth of the lifetime earnings would be affected less over the short term than it would distribution who were born in the 1950s and who first be if changes were made immediately. claimed benefits at age 65. Faster or slower implementation of policies to reduce In those examples, the projected effects on debt include budget deficits would tend to impose different bur- both the direct effects of the policy changes and the feed- dens on different generations. Reducing deficits sooner back to the federal budget that would result from faster would probably require older workers and retirees to economic growth. In general, reducing the federal debt sacrifice more but would benefit younger workers and increases the amount of money available for (or crowds future generations. Reducing deficits later would require in) private investment in capital goods and services, smaller sacrifices from older people but greater ones from which increases the stock of private capital and economic younger workers and future generations. output. The policy changes examined here are illustrative, CHAPTER 2: THE LONG-TERM BUDGET OUTLOOK UNDER ALTERNATIVE SCENARIOS THE 2019 LONG-TERM BUDGET OUTLOOK 47 Figure 2-3 . How Timing Affects the Size of Policy Changes Needed to Make Federal Debt Meet Two Possible Goals in 2049 Percentage of GDP The reduction in each year’s primary deficit needed to Starting Year make federal debt held by the public in 2049 equal . . . 1.8 . . . its current share of GDP (78 percent) 2020 2.9 . . . its 50-year average (42 percent) 2.2 2025 3.5 2.7 2030 4.4 0 1 2 3 4 5 Source: Congressional Budget Office. GDP = gross domestic product. CBO has analyzed those trade-offs in two ways. First, it such as across-the-board benefit cuts or tax rate increases estimated the extent to which the size of policy adjust- for all adults, that analysis suggests that the average ments would change if deficit reduction was delayed by income of generations born after the earlier implementa- five or 10 years. (CBO did not make any assumptions tion date would be lower under the policy with a 10-year about the specific policy changes that might be used to delay.9 In contrast, people born more than 25 years before reduce the deficit.) For example, if lawmakers sought to the earlier implementation date would have a higher aver- reduce debt as a share of GDP to its historical 50-year age income if action was delayed—mainly because they average of 42 percent in 2049 and if the necessary pol- would partly or entirely avoid the policy changes needed icy changes did not take effect until 2025, the annual to stabilize the debt. Generations born between those reduction in the primary deficit would need to amount to 3.5 percent of GDP rather than the 2.9 percent that would accomplish the same goal if the changes were made starting in 2020 (see Figure 2-3). If lawmakers chose to wait another five years to implement the policies (having them take effect in 2030), even larger changes would be 9. Those results are preliminary conclusions from an update of work that CBO published in 2010. See Congressional Budget Office, necessary; in that case, the required annual reduction in Economic Impacts of Waiting to Resolve the Long-Term Budget the primary deficit would amount to 4.4 percent of GDP. Imbalance (December 2010), www.cbo.gov/publication/21959. That analysis was based on a projection of slower growth in Second, CBO studied the effects on the average per capita debt than CBO now projects, so the estimated effects of a income of various generations from waiting to resolve the similar policy today would be close, but not identical, to the long-term fiscal imbalance. CBO compared economic effects estimated in that analysis. For a different approach to analyzing the costs of debt reduction for different generations, outcomes under two types of policies. One would stabi- see Shinichi Nishiyama and Felix Reichling, The Costs to Different lize the debt-to-GDP ratio starting in a particular year, Generations of Policies That Close the Fiscal Gap, Working Paper and the other would wait 10 years to do so. For policies 2015-10 (Congressional Budget Office, December 2015), www.cbo.gov/publication/51097. 48 THE 2019 LONG-TERM BUDGET OUTLOOK june 2019 two groups could either gain or lose from delayed action, changes would be relatively small compared with their depending on the specific details of the policy changes.10 lifetime earnings potential because, on average, future generations are expected to have much higher income CBO’s analysis indicates that delaying policy changes than current generations. would reduce the well-being of younger generations com- pared with a situation in which policy changes occurred Even if lawmakers waited to implement policy changes earlier. Moreover, the further in the future that a policy to reduce deficits in the long term, deciding about those change occurred, the more the well-being of older genera- changes sooner would offer two main advantages. First, tions would be improved and that of younger generations people would have more time to prepare by changing the would be worsened. However, the additional burden number of hours they work, the age at which they plan on younger generations resulting from delaying policy to retire, and the amount they choose to save. Second, policy changes that would reduce the debt over the long term would hold down longer-term interest rates and 10. Those conclusions do not incorporate the negative effects that could lessen uncertainty—thus enhancing businesses’ and would arise from a fiscal crisis if one occurred or effects that might arise from the government’s reduced flexibility to respond to consumers’ confidence. Those factors would boost output unexpected challenges. and employment in the near term. APPENDI X A Appendix A CBO’s Projections of Demographic and Economic Trends The Congressional Budget Office develops its assessment year in the first decade of the projection, 0.5 percent in of the long-term outlook for the federal budget on the the second decade, and 0.4 percent in the third. Over the basis of its projections of demographic and economic entire 30-year period, the U.S. population is projected to trends over the next three decades. Through 2029, the grow at an average annualized rate of 0.5 percent (com- economic and demographic projections presented in pared with a rate of 0.9 percent over the past 30 years): this report are the same as those that CBO published in Births account for an average annual increase of 1.2 per- January.1 For 2030 through 2049—the remaining years centage points, immigration adds 0.3 percentage points, of CBO’s extended baseline—the agency’s projections and mortality subtracts 1.0 percentage point. generally reflect historical trends and anticipated demo- graphic changes.2 (A set of annual projections is included The population is projected not only to grow more in this report’s supplemental data, available online at slowly but also to become older, on average, than in the www.cbo.gov/publication/55331.) past. In the agency’s projections, over the 30-year period, the share of the population that is 65 or older grows, Demographic Factors whereas the share that is of working age (defined as Both the size and composition of the U.S. population people ages 20 to 64) shrinks. As a result, CBO projects, influence the overall growth of the economy and affect a growing portion of the population will receive benefits federal tax revenues and spending. Rates of fertility, net from the Social Security and Medicare programs while immigration, and mortality determine the population a shrinking portion will be working and paying into the and thus the size of the labor force and the number of trust funds that support those programs. people receiving benefits from federal programs such as Social Security and Medicare. Because of changes to Fertility those rates, CBO projects the population to be smaller CBO projects a gradual rise in the total fertility rate over and to grow at a slower pace in the future than it pro- the next few years, increasing from a rate of 1.8 children jected last year. per woman in 2018 to a rate of 1.9 children per woman from 2022 through 2049.3 Fertility rates tend to be Population procyclical, meaning they often decline during recessions In CBO’s projections, the total population increases and rebound during recoveries. However, the U.S. fer- from 333 million at the beginning of 2019 to 388 mil- tility rate did not recover after the 2007–2009 recession; lion in 2049, and population growth slows from a rate of the rate (which averaged 2.0 children per woman in the 0.6 percent per year to 0.4 percent per year by the end of 20 years prior to the recession) peaked at 2.1 in 2007. the projection period (see Table A-1). The slowdown in growth is particularly pronounced for the population age 3. The total fertility rate represents the average number of children 16 or older, which grows on average by 0.8 percent per that a woman would have in her lifetime and is calculated as the sum of fertility rates for all ages between 15 and 49 in a given year. The total fertility rate can also be defined as the average 1. See Congressional Budget Office, The Budget and Economic number of children that a woman would have if, in each year Outlook: 2019 to 2029 (January 2019), www.cbo.gov/ of her life, she experienced the birth rates observed or assumed publication/54918. for that year and if she survived her entire childbearing period. 2. The extended baseline generally reflects current law, following In CBO’s long-term model, the likelihood that a particular CBO’s 10-year baseline projections through 2029 and then woman will have a child depends on such factors as that woman’s extending most of the concepts underlying those projections education, marital status, immigration status, and childbearing through the rest of the long-term projection period. history. 50 THE 2019 LONG-TERM BUDGET OUTLOOK JUNE 2019 Table A-1 . Average Annual Values for Demographic Variables That Underlie CBO’s Extended Baseline Projections Overall, 1989–2018 2019–2029 2030–2039 2040–2049 2019–2049 Growth of Population (Percent) 0.9 0.6 0.5 0.4 0.5 Contribution to Population Growth (Percentage points): Births 1.4 1.2 1.2 1.2 1.2 Net immigration 0.4 0.3 0.3 0.3 0.3 Deaths -0.9 -0.9 -1.0 -1.1 -1.0 Growth of Civilian Noninstitutionalized Population (Percent) a 1.1 0.8 0.5 0.4 0.6 Memorandum: Fertility Rate (Children per woman) 2.0 1.9 1.9 1.9 1.9 Life Expectancy at Birth, End of Period (Years) b 78.9 80.3 81.4 82.5 82.5 Life Expectancy at Age 65, End of Period (Years) b 19.4 20.1 20.8 21.5 21.5 Immigration Rate (Per 1,000 people in the U.S. population) 3.7 3.1 3.1 3.1 3.1 Source: Congressional Budget Office. The extended baseline projections generally reflect current law, following CBO’s 10-year baseline budget projections through 2029 and then extending most of the concepts underlying those projections for the rest of the long-term projection period. a The civilian noninstitutionalized population includes individuals age 16 or older who are not inmates of institutions or on active duty in the armed forces. b.Life expectancy as used here is period life expectancy, which is the amount of time that a person in a given year would expect to survive beyond his or her current age on the basis of that year’s mortality rates for various ages. Since then, the fertility rate has steadily declined, reach- Immigration ing 1.9 children per woman in 2010 and 1.8 children per Under current law, CBO projects, annual net immigra- woman in 2017 (the most recent year for which data are tion to the United States (a measure that accounts for available). all people who either enter or leave the United States in any year) would rise from 0.9 million people in 2019 CBO projects that total fertility rates will remain below to 1.2 million people in 2049. In an environment of the replacement rate—the fertility rate required for a relatively low birth rates, net immigration flows become generation to exactly replace itself—of 2.1 children a more important part of overall U.S. population growth. per woman for the next three decades.4 Over the next Between 2019 and 2029, projected net inflows account 30 years, that relatively low rate of fertility will contrib- for approximately half of overall population growth, but ute to slower population growth. CBO’s projection is by the last decade of the projection period that share is consistent with the recommendation made to the Social about four-fifths of all growth. Security Advisory Board by its 2015 Technical Panel on Assumptions and Methods.5 The new immigrants would largely consist of legal perma- nent residents (LPRs). Over the next two decades, average annual net flows of LPRs are projected to increase from approximately 860,000 LPRs per year in the first decade to approximately 890,000 over the second decade. In addition, the number of legal temporary residents is pro- 4. See Joyce A. Martin and others, Births: Final Data for 2017, National Vital Statistics Reports, vol. 67, no. 8 (National Center jected to increase steadily by approximately 80,000 per for Health Statistics, November 2018), www.cdc.gov/nchs/data/ year over the next 20 years. nvsr/nvsr67/nvsr67_08-508.pdf (988 KB). 5. See 2015 Technical Panel on Assumptions and Methods, Report CBO’s projections of annual net flows of foreign-born to the Social Security Advisory Board (September 2015), p. 9, people without legal status, which are informed by the https://go.usa.gov/cJYR5 (PDF, 3.4 MB). agency’s economic projections and by recent demographic APPENDIX A: CBO’S PROJECTIONS OF DEMOGRAPHIC AND ECONOMIC TRENDS THE 2019 LONG-TERM BUDGET OUTLOOK 51 trends, increase over that period.6 Growth in the U.S. agency projects that net immigration for all categories economy is an important factor because, in CBO’s would grow at a rate equal to overall population growth estimation, periods of faster growth over the past two in the prior year; that rate averages 0.4 percent annually decades have been associated with higher net flows of through 2049. foreign-born people without legal status. CBO expects that relationship to boost such immigration. However, Mortality estimates indicate that the number of foreign-born The mortality rate, which is the number of deaths per people without legal status in 2016 was the lowest since thousand people, has generally declined in the United 2004 despite relatively strong economic conditions in the States since the early 20th century, although the rate of United States, which implies that other factors have con- those improvements has slowed over time. For the most strained such immigration in recent years. CBO expects part, the mortality rate has decreased more quickly for those other factors to continue to hold down such younger people than for older people during that period. immigration in the near term.7 Nevertheless, over time, However, mortality rates rose in 2015 and 2016, the the agency expects economic growth to again become an most recent years for which data were available at the important factor for immigration. time this analysis was completed. The result was that life expectancy at birth declined in both years, marking the On the basis of recent data, CBO anticipates that net first decreases in this metric since 1993. Those declines flows of foreign-born people without legal status will are primarily driven by increases in mortality from be smaller in the near term than the long-term rela- Alzheimer’s disease, suicide, chronic liver disease, septi- tionship between immigration and economic growth cemia, and unintentional drug overdoses (in particular, would suggest; the agency projects zero net flows in 2019 opioids).8 (meaning that immigration is offset by emigration in this category). The agency expects annual net flows of for- CBO projects mortality rates for every five-year age eign-born people without legal status to increase signifi- group to decline at the same average pace each group cantly between 2020 and 2024, reaching approximately experienced from 1950 through 2015. After projecting 170,000 by 2024, and then remain roughly unchanged average mortality rates for men and women in each age through 2039, reflecting both economic growth and group, CBO incorporates differences in those rates for those other constraining factors. people 30 years of age and older on the basis of marital status, education, disability insurance status, and lifetime For projections beyond the next 20 years, CBO employs household earnings (for people under 30, the mortality a simplified approach: After 2039, under current law, the projections account for age and sex only). CBO projects lower mortality rates and thus longer life expectancies for people who are married, have more education, do not 6. CBO uses this term to refer to foreign-born people other than receive benefits through the Social Security Disability LPRs, refugees, asylees, temporary residents, and visitors. Most Insurance (DI) program, or are in higher-income foreign-born people without legal status either unlawfully entered the United States without inspection or lawfully entered groups.9 the United States in a temporary status and then unlawfully remained in the country after that temporary status expired. Some foreign-born people without legal status are beneficiaries 8. For an account of how factors affecting mortality and mortality under Temporary Protected Status or under policies whereby the improvement rates have changed over time, see National Center executive branch does not seek their immediate removal from for Health Statistics, Health, United States, 2017: With Special the United States (for example, Deferred Action for Childhood Feature on Mortality, www.cdc.gov/nchs/data/hus/hus17.pdf Arrivals); others are allowed into the United States while they (10.5 MB). await their removal proceedings in immigration courts. Many 9. For more information about mortality differences among groups of those foreign-born people without legal status are authorized with different earnings, see Tiffany Bosley, Michael Morris, to work in the United States, in which case they may apply for a and Karen Glenn, Mortality by Career-Average Earnings Level, Social Security number and must pay applicable federal taxes. Actuarial Study 124 (Social Security Administration, April 2018), 7. For the most recent estimates, see Jeffrey S. Passel and D’Vera https://tinyurl.com/yct5qdew (PDF, 301KB); Congressional Cohn, U.S. Unauthorized Immigrant Total Dips to Lowest Budget Office, Growing Disparities in Life Expectancy Level in a Decade (Pew Research Center, November 2018), (April 2008), www.cbo.gov/publication/41681; and Julian P. https://tinyurl.com/y9tmol2g. Official data on foreign-born Cristia, The Empirical Relationship Between Lifetime Earnings people without legal status are limited, so historical estimates are and Mortality, Working Paper 2007–11 (Congressional Budget very uncertain. Office, August 2007), www.cbo.gov/publication/19096. 52 THE 2019 LONG-TERM BUDGET OUTLOOK JUNE 2019 CBO’s projections result in an average life expectancy at projected a total fertility rate of 1.9 children per woman birth of 82.5 years in 2049, compared with 79.1 years in in each year for 2018 through 2048. 2019.10 Similarly, life expectancy at age 65 is projected to be 21.5 years in 2049, or 2.1 years longer than life expec- The lower fertility rate in the first three years of the tancy at age 65 in 2019.11 projection period eventually results in fewer births and a smaller working-age population throughout the entire Changes in Demographic Projections Since Last Year period than CBO projected last year. There are approx- CBO’s estimates of the U.S. population change as the imately 547,000 fewer births over the first half of the result of changes in rates of fertility, immigration, and projection period, but the effects of fewer births become mortality. most evident in the second half of the projection period, as the 2019–2021 birth cohort ages into its working and Population. In CBO’s projections, the population childbearing years. Between 2035 and 2049, the popu- increases from approximately 333 million in 2019 to lation age 16 and older contains about 3.7 million fewer 355 million in 2029—an average annualized growth people per year, on average, than CBO projected last rate of about 0.6 percent. Last year, CBO projected the year. Moreover, the combination of lower fertility rates population would grow slightly faster, increasing from and a smaller population of childbearing adults results 335 million to 358 million over that same period, an in roughly 60,000 fewer births per year, on average, than average annualized rate of 0.7 percent. Those revisions the agency projected last year. reflect changes to underlying data—specifically, unex- pectedly high mortality rates for 2015 and 2016 and Net Immigration. CBO’s projection of net immigration unexpectedly low fertility rates for 2016 and 2017—as is also lower than its projection last year. Between 2019 well as changes to the way the agency projects fertility and 2049, the agency projects the average net immi- rates and net immigration. gration rate to be 3.1 immigrants per thousand people, compared with an average rate of 3.2 over the same In the two decades following 2029, the population is period in last year’s report. Those revisions are attrib- projected to grow at an average annual rate of 0.4 per- utable to more recent data and adjustments to the way cent (revised down from 0.5 percent in last year’s report) CBO projects net immigration in the extended baseline. to 388 million by 2049 (5 million, or 1.4 percent, fewer people than projected last year). In total, over the next decade, CBO projects approxi- mately 352,000 (or 2.9 percent) fewer immigrants, on Fertility. The total fertility rate is projected to be net, than the agency projected last year. That change lower through 2021 than CBO projected last year. is driven primarily by smaller projected net flows of Total fertility rates have been persistently low since the LPRs and foreign-born people without legal status. In 2007–2009 recession. In recognition of that trend, CBO particular, CBO has revised its near-term projection of expects total fertility rates to remain low for the next few net flows of foreign-born people without legal status years, gradually rising from a rate of 1.8 children per to be lower on the basis of recent data that suggest net woman in 2019 to 1.9 children per woman by 2022, and flows for this category are likely to be smaller over the then remaining at that rate. By contrast, CBO last year next few years than previously projected. Indeed, the agency’s projection of zero net flows for 2019 reflects a downward revision of 171,000. Nevertheless, in CBO’s 10. Life expectancy as used here is period life expectancy, which is the amount of time that a person in a given year would expect to assessment, domestic economic conditions will return as survive beyond his or her current age on the basis of that year’s an important driver of immigration flows by the middle mortality rates for various ages. of the coming decade and, under current law, net flows 11. CBO projects life expectancy in 2090 to be 86.4 years at birth of foreign-born people without legal status will again be and 24.1 years at age 65. CBO’s projections of life expectancies positive. are longer than those of the Social Security trustees (85.7 and 23.5 years, respectively) but shorter than the projections (88.3 Last year, for the final 20 years of the extended baseline and 25.3 years, respectively) recommended in 2015 Technical projection, CBO based its projections of net immigra- Panel on Assumptions and Methods, Report to the Social Security Advisory Board (September 2015), pp. 13–20, https://go.usa.gov/ tion flows for all categories on the average growth in cJYR5 (PDF, 3.4 MB). net immigration published by the Census Bureau—a APPENDIX A: CBO’S PROJECTIONS OF DEMOGRAPHIC AND ECONOMIC TRENDS THE 2019 LONG-TERM BUDGET OUTLOOK 53 constant rate of 0.6 percent. After reassessing that Gross Domestic Product approach, the agency now projects that the same eco- CBO expects total output, or GDP, in the economy nomic forces driving immigration trends in the first to grow by an average of 3.9 percent per year over the decade of its projection will persist through the second 2019–2049 period (see Table A-2). In the agency’s pro- decade. Because of that change, net immigration is pro- jections, real (inflation-adjusted) GDP growth over that jected to grow at an average annual rate of 0.3 percent— period averages 1.9 percent per year, about what CBO half the rate of growth projected last year—resulting in projected last year for the 2018–2048 period. That rate a total of 400,000 (or 3.4 percent) fewer immigrants, on is less than the average growth of 2.5 percent for the past net, between 2030 and 2039. three decades. CBO expects that growth in real GDP per person will average 1.3 percent over the next three Beyond 2039, because of the significant uncertainty decades, less than the 1.6 percent growth of the past surrounding the mix of immigrants in the long run, three decades. CBO projects net immigration flows based on overall population growth in the prior year. As a result of that Projections of GDP. CBO projects that over the next change, net immigration is projected to grow at an aver- five years, GDP and employment will initially exceed age annual rate of 0.4 percent between 2040 and 2049, and then return to their long-run relationships with compared with the annual rate of 0.6 percent CBO their maximum sustainable levels. After five years, real projected last year. In total, CBO projects approximately GDP is then projected to grow at a pace that reflects 740,000 (or 6.0 percent) fewer net immigrants over that the increases in the supply of labor, capital services, and period than the agency projected last year. productivity described below. That projected pace also takes into consideration the influences of the marginal Mortality. Recent data show higher mortality rates than tax rates and increases in federal debt that CBO projects CBO expected last year for all age groups, but partic- in its extended baseline.12 ularly for people under 45 years of age. Those data led CBO to increase its projection of mortality rates for all Over the long term, total GDP is projected to be age groups in the near term and to reduce their rates of one-half of one percent below its potential (maximum mortality improvement over the next three decades. As sustainable) amount, as it has roughly been, on average, a result, CBO now projects approximately 970,000 (or over past decades. Those projected outcomes reflect 0.9 percent) more deaths over the next three decades CBO’s assessment that, during and after economic than the agency projected last year. downturns, actual output has fallen short of potential output to a greater extent and for longer periods than CBO’s new projections of mortality rates and mortality actual output has exceeded potential output during eco- improvement also affect the agency’s projections of life nomic booms.13 expectancies, which it now expects to be lower than it reported last year. Life expectancy at birth is projected to Projected real GDP growth over the next three decades be 82.4 years in 2048, 0.4 years shorter than CBO pro- is slower than the average annual rate of 2.5 percent jected last year, and life expectancy at age 65 is projected recorded over the past three decades because the labor to be 21.5 years, 0.2 years shorter than in last year’s force is projected to grow more slowly. On average, projection. CBO projects that real GDP will grow at an annual rate of 1.8 percent from 2019 to 2029. In the decade Economic Factors after 2029, average growth is projected to remain at The federal government’s revenues, spending, and debt 1.8 percent before rising to 1.9 percent over the 2040– depend on key economic factors such as the growth of 2049 period. The pattern of projected GDP growth gross domestic product (GDP), the size and composi- tion of the labor force, the number of hours worked, the distribution of earnings among workers, capital accumu- 12. The marginal tax rate is the percentage of an additional dollar of lation, productivity, inflation, and interest rates. CBO’s income from labor or capital that is paid in taxes. projections of those factors reflect the agency’s assessment 13. See Congressional Budget Office, Why CBO Projects That of various economic and demographic developments as Actual Output Will Be Below Potential Output on Average well as the effects of fiscal policy on economic activity. (February 2015), www.cbo.gov/publication/49890. 54 THE 2019 LONG-TERM BUDGET OUTLOOK JUNE 2019 Table A-2 . Average Annual Values for Economic Variables That Underlie CBO’s Extended Baseline Projections Percent Overall, 1989–2018 2019–2029 2030–2039 2040–2049 2019–2049 Growth of GDP Real GDP a 2.5 1.8 1.8 1.9 1.9 Nominal GDP (Fiscal year) 4.7 4.0 3.8 3.9 3.9 Real GDP per Person 1.6 1.2 1.3 1.5 1.3 Growth of the Labor Force 1.0 0.5 0.3 0.4 0.4 Labor Force Participation Rate 65.5 62.0 60.3 59.7 60.7 Unemployment Unemployment rate 5.9 4.5 4.6 4.5 4.5 Natural rate of unemployment 5.1 4.5 4.4 4.2 4.4 Growth of Average Hours Worked -0.1 * * * * Growth of Total Hours Worked 0.9 0.4 0.3 0.4 0.4 Earnings as a Share of Compensation 81 81 81 81 81 Growth of Real Earnings per Worker 0.9 1.2 1.1 1.0 1.1 Share of Earnings Below the Taxable Maximum 85 82 81 80 81 Growth of Productivity Total factor productivity in the nonfarm business sector 1.2 1.1 1.1 1.1 1.1 Real GDP per hour worked a 1.6 1.4 1.5 1.5 1.5 Labor force productivityb 1.5 1.3 1.5 1.5 1.4 Inflation Growth of the CPI-U 2.5 2.4 2.4 2.4 2.4 Growth of the GDP price index 2.1 2.1 2.0 2.0 2.0 Interest Rates Real rates On 10-year Treasury notes and the OASDI trust funds 2.2 1.3 1.6 2.0 1.6 Nominal rates On 10-year Treasury notes and the OASDI trust funds 4.7 3.7 4.0 4.4 4.0 On all federal debt held by the public c 4.8 3.1 3.6 4.0 3.6 Source: Congressional Budget Office. The extended baseline projections generally reflect current law, following CBO’s 10-year baseline budget projections through 2029 and then extending most of the concepts underlying those projections for the rest of the long-term projection period. CPI-U = consumer price index for all urban consumers; GDP = gross domestic product; OASDI = Old-Age, Survivors, and Disability Insurance; * = between -0.05 and 0.05. a.Real values have been adjusted to remove the effects of changes in prices. b.The ratio of real GDP to the labor force. Elsewhere, CBO reports other measures of labor productivity, such as the ratio of real potential GDP to the potential labor force. c.The interest rate on all federal debt held by the public equals net interest payments in the current fiscal year divided by debt held by the public at the end of the previous fiscal year. APPENDIX A: CBO’S PROJECTIONS OF DEMOGRAPHIC AND ECONOMIC TRENDS THE 2019 LONG-TERM BUDGET OUTLOOK 55 follows the pattern of labor force growth over the next for nearly the entire decline, while the effects of other three decades. factors largely offset one another. Real GDP per person is expected to increase at a slower People over age 65 tend to participate in the labor force pace than it has in the past—at an average annual rate of at lower rates than younger people—as of 2018, the aver- 1.3 percent over the 2019–2049 period, compared with age participation rate for prime-age people (those ages 25 1.6 percent over the past 30 years. That occurs mainly to 54) was 82 percent, whereas that for people over age because the labor force is projected to grow more slowly 65 was about 20 percent. Therefore, the ongoing aging than the overall population. of the population is expected to dampen the overall rate of participation in the labor force over the next 30 years. Changes in Projections of GDP Since Last Year. In Among the civilian noninstitutionalized population age CBO’s current projections, the level of real GDP is 16 or older, the share of people over age 65 has increased slightly higher in 2028 than the agency projected last from 16 percent to 20 percent over the past decade and year. Over the subsequent two decades, the agency’s cur- is projected to rise to 27 percent by 2049. In the mean- rent projection of real GDP grows slightly more slowly time, the share of the prime-age population is expected than it did last year; by 2048, real GDP is 1.6 percent to decline from 49 percent in 2018 to 45 percent by less than it was last year. GDP growth is projected to 2049. Without the effects of further aging of the pop- grow more slowly in the second decade (2030 to 2039), ulation—that is, if the age composition of the popula- mainly because growth in the labor force is slower in tion remained the same as it was in 2018—the overall this year’s projection than it was in last year’s projection. labor force participation rate over the next 30 years In the third decade, GDP growth is similar to last year’s would be roughly constant (and slightly higher than its projection. 2018 level), in CBO’s assessment. The Rate of Labor Force Participation and Aside from the aging of the population, CBO expects Labor Force Growth the effects on labor force participation of other demo- The size of the labor force depends on the rates at which graphic trends, economic trends, and fiscal policies under people of different demographic groups participate in current law to largely offset one another over the coming the labor market. Since the mid-2000s, the overall labor decades. In particular, two long-term trends are expected force participation rate in the United States has declined to put downward pressure on the participation rate: substantially, driven predominantly by the aging of the population.14 CBO expects that downward trend to •• Members (particularly men) of each generation that continue over the coming decades before slowing down followed the baby boomers tend to participate in and eventually leveling off toward the end of the 30-year the labor force at lower rates than their predecessors projection period. As a result, the labor force is projected did at the same age.15 (One notable exception to grow more slowly than the population. CBO’s pro- in later generations is that women younger than jections of the overall participation rate and labor force 35 generally participate at higher rates than female growth are broadly similar to its previous projections. baby boomers did at the same ages. However, as However, the agency has made larger revisions to the those later generations of women have aged, their participation rates of specific demographic groups. participation rates have also fallen below those of their predecessors.) Projections of the Labor Force Participation Rate. In CBO’s projections, the rate of labor force participation •• The marriage rate is projected to continue to fall, declines from 62.8 percent in 2019 to 61.0 percent in especially among men, and unmarried men tend 2029 and to 59.8 percent in 2040, where it remains to participate in the labor force at lower rates than roughly constant for the rest of the projection period. In married men do. CBO’s assessment, the aging of the population accounts CBO expects those forces to be mostly offset by two trends that are expected to increase participation: 14. The labor force participation rate is the share of the civilian noninstitutionalized population age 16 or older that participates in the labor force. 15. Baby boomers are people born between 1946 and 1964. 56 THE 2019 LONG-TERM BUDGET OUTLOOK JUNE 2019 •• The population is becoming more educated, and Americans’ participation rates since the last recession people with more education tend to participate in the are more structural and less cyclical than previously labor force at higher rates than do people with less estimated. education. Conversely, the agency has raised its projection of the •• Increasing longevity is expected to lead people to participation rate of prime-age people throughout the continue working to increasingly older ages. 30-year projection period. CBO now expects the partic- ipation rate of that group to be higher, on average, over In addition to the effects of those demographic trends, the next three decades than its current rate. That con- budgetary effects and incentives under current tax law, trasts with CBO’s previous projection that the rate would combined with economic trends, would also affect the decline slightly over the next three decades. The agency labor force. For example, rising federal deficits are pro- revised its projection primarily because the participation jected to slow growth in the stock of private capital and rate of prime-age people has rebounded more strongly to limit the growth of after-tax wages, thereby reducing in the past year than expected, which suggests that the the supply of labor. Meanwhile, under current law, tax rate’s decline after the last recession was driven more by rates on individual income are set to rise in 2026 when cyclical factors and less by structural factors than previ- some provisions of the 2017 tax act expire. In addition, ously estimated. as people’s income rises faster than inflation, more of their income is pushed into higher tax brackets through a Projections of Labor Force Growth. Because a fall- process known as real bracket creep, raising their effec- ing participation rate means that less of the growth in tive tax rates. After 2025, those higher tax rates and real population translates into labor force growth, the labor bracket creep are projected to decrease participation in force is expected to increase even more slowly than the the labor force because people would see a lower return population from 2019 to 2049. Although the population on their labor. age 16 or older is expected to grow by 0.6 percent per year, on average, the labor force is projected to grow at Changes in Projections of the Labor Force Participation an average rate of 0.4 percent per year. That represents a Rate Since Last Year. CBO’s current projection of the significant slowdown in labor force growth from earlier overall labor force participation rate is slightly lower periods: For example, the average annual growth rate was than previously projected for 2019 to 2024 and slightly 1.2 percent during the 1990–2006 period. higher than previously projected for 2025 to 2048. The new projection incorporates the agency’s reassessment In CBO’s projections, growth in the labor force of recent trends in the participation rates of different declines from an average of 0.5 percent during the demographic groups. 2019–2029 period to 0.3 percent during the 2030– 2039 period, driven by a decline in population growth Compared with its previous projections, CBO has over the next two decades as well as a decline in the par- lowered its projection of the participation rate of the ticipation rate. Labor force growth rebounds slightly, to youngest group of workers (ages 16 to 24) throughout an average of 0.4 percent per year, in the third decade of the 30-year projection period. The agency now expects the 30-year projection period; the labor force participa- that group’s rate of participation in the labor force to fall, tion rate is expected to have stabilized by then and there- from about 55 percent in 2018 to 54 percent in 2048, fore would no longer subtract from labor force growth. instead of rising to 57 percent as previously projected. CBO’s revision mainly reflects the observation that the Changes in Projections of Labor Force Growth Since participation rate of the youngest workers has declined Last Year. CBO’s current projection of labor force substantially since the 2007–2009 recession and has growth is similar to last year’s projection through 2029 failed to recover meaningfully in recent years despite but lower in the second decade because of the downward the growth of the economy.16 That development sug- revision in population growth discussed in previous sec- gests that the factors that have pushed down younger tions. In the third decade, projected labor force growth 16. For a discussion of CBO’s methods for projecting labor force Participation Rates, Working Paper 2018-04 (Congressional participation, see Joshua Montes, CBO’s Projection of Labor Force Budget Office, March 2018), www.cbo.gov/publication/53616. APPENDIX A: CBO’S PROJECTIONS OF DEMOGRAPHIC AND ECONOMIC TRENDS THE 2019 LONG-TERM BUDGET OUTLOOK 57 is similar to last year’s projection because higher partici- the labor force as a whole is expected to decline slightly. pation rates offset the downward revisions to population By 2049, the average number of hours that people work growth. is expected to be about 0.9 percent less than it is today. Other Labor Market Outcomes Total Hours Worked. Based on projections of the size In addition to the rate of labor force participation and of the labor force, average hours worked, and unem- the size of the labor force, CBO’s long-term labor market ployment, total hours worked are estimated to increase outlook also includes its projections for the unemploy- at an average annual rate of 0.4 percent between 2019 ment rate, the average and total number of hours that and 2049. This is slower than the average annual rate of people work, and various measures of workers’ earnings 0.9 percent over the past three decades. The drop in the over the next 30 years. The agency regularly updates growth of total hours is mainly because the population those projections to incorporate updates to historical is expected to grow more slowly in the future than it has data, reassessments of trends, and changes to its analytic over the past 30 years. methods. Average growth in total hours worked falls from 0.4 per- Unemployment. In CBO’s projections, the unemploy- cent in the first decade of CBO’s projections (2019 to ment rate falls from 3.9 percent at the end of 2018 to 2029) to 0.3 percent in the second decade, rising to 3.5 percent in 2019, about 1.1 percentage points below 0.4 percent in the third decade. A drop in population the agency’s estimate of the natural rate of unemploy- growth between the first and second decades is the main ment (the rate that results from all sources other than cause of the projected decline in growth of total hours fluctuations in overall demand related to the business worked. Growth in total hours worked increases in the cycle). As economic growth slows after 2019, the third decade because the decline in the rate of labor force unemployment rate rises, surpassing the natural rate by participation ends. 2022. (The natural rate of unemployment is projected to fall from 4.6 percent in 2019 to 4.5 percent in 2029.) Earnings as a Share of Compensation. Workers’ total From 2023 onward, the unemployment rate is expected compensation consists of taxable earnings and non- to remain roughly one-quarter of one percentage point taxable benefits such as employers’ contributions to above the natural rate, a difference that is consistent both health insurance and pensions. Over the years, the with the historical average relationship between the two share of total compensation paid in the form of earn- measures and with the projected gap of one-half of one ings has declined—from about 90 percent in 1960 to percent between actual and potential GDP. about 81 percent in 2018—mainly because the cost of health insurance has risen more quickly than total After 2029, both the actual and the natural rates of compensation.17 unemployment are projected to decline gradually as the labor force ages and becomes increasingly educated. CBO expects that trend in health care costs to continue, (Older and more educated workers tend to have lower which would further decrease the proportion of compen- actual and natural rates of unemployment.) By 2049, the sation that workers receive as earnings. However, under natural rate of unemployment is projected to be about current law, an excise tax on certain employment-based 4.2 percent, and the actual rate is projected to be about health insurance plans that have premiums above spec- 4.4 percent. ified amounts is scheduled to take effect in 2022. Some employers and workers are expected to respond by shift- Average Hours Worked. Different subgroups of the labor ing to less expensive plans, thereby reducing the share of force work different numbers of hours, on average. Men compensation consisting of health insurance premiums tend to work more hours than women do, for example, and increasing the share that consists of earnings. In and people between the ages of 30 and 40 tend to work CBO’s projections, the effects of the tax on the mix of more hours than people between the ages of 50 and compensation roughly offset the effects of rising costs for 60 do. In CBO’s estimation, those differences among health care until the effects of rising costs outweigh those groups will remain stable. However, over the long term, of the excise tax late in the projection period. As a result, the composition of the labor force is projected to shift toward groups that tend to work less (such as older work- 17. For more details, see Congressional Budget Office, How CBO ers). As a result, the average number of hours worked by Projects Income (July 2013), www.cbo.gov/publication/44433. 58 THE 2019 LONG-TERM BUDGET OUTLOOK JUNE 2019 the share of compensation that workers receive as earn- CBO projected last year. CBO’s current projections ings is projected to remain close to 81 percent through of wages and salaries are slightly lower than last year’s, most of the 2019–2049 period. mainly because updates to historical wage and salary data indicate that their share of GDP was, on average, Growth of Real Earnings per Worker. Projections of lower over the past decade than previously reported. As wages and salaries, nonwage compensation (such as a result, CBO revised down its projection of wages and employment-based health insurance), average hours salaries over the next three decades. worked, labor productivity (discussed below), and prices imply that real earnings per worker would grow Also, CBO’s current projection of the unemployment by an average of 1.1 percent annually over the 2019– rate is higher during the 2019–2023 period but slightly 2049 period. That rate is higher than the average annual lower from 2024 onward. The upward revision in the growth of 0.9 percent over the past 30 years. near term largely reflects the agency’s assessment that recent trends in hiring, layoffs, and retirement that had Distribution of Earnings. Over the past several decades, put downward pressure on the unemployment rate earnings have grown faster for higher earners than will not last as long as CBO estimated earlier. For the for lower earners. In CBO’s projections, the unequal 2024–2048 period, in contrast, the downward revision growth in earnings continues for the next three decades, occurred because the agency lowered its estimate of although that disparity falls over time. The distribution the natural rate of unemployment after reassessing the of earnings affects revenues from income taxes and pay- effects of the composition of the potential labor force. In roll taxes, among other things. Income taxes are affected particular, because younger workers tend to have higher by the earnings distribution because of the progressive natural rates of unemployment (on average, more than rate structure of the individual income tax; people with 10 percent, compared with prime-age workers’ 4 percent lower income pay a smaller share of their earnings than from 1990 to 2018), revising down their share in the people with higher income do. potential labor force, as CBO did, leads to a reduction in the estimate of the economywide natural rate of Social Security payroll taxes are also affected by the earn- unemployment. ings distribution. Those taxes are levied only on earnings up to a certain annual amount ($132,900 in 2019). Capital Accumulation and Productivity Below that amount, earnings are taxed at a combined In addition to growth in the labor force and the number rate of 12.4 percent, split between the employer and of hours worked, two other important factors affect the employee (self-employed workers pay the full amount); growth in output. One is the accumulation of capital, no tax is paid on earnings above the cap. The taxable including physical structures, equipment, land, and maximum has remained a nearly constant proportion of inventories used in production, along with intangible the average wage since the mid-1980s, but because earn- capital such as computer software. The accumulated ings have grown more for higher earners than for others, stock contributes a stream of services to production. The the portion of covered earnings on which Social Security second is the growth of total factor productivity (TFP), payroll taxes are paid has fallen from 90 percent in 1983 which is the growth of real output per unit of combined to 84 percent in 2017.18 The portion of earnings subject labor and capital services—that is, the growth of output to Social Security taxes is projected to fall to an average that is not explained by the growth of labor and capi- of 82 percent between 2019 and 2029, 81 percent in tal. Combined, the growth rates projected for the labor the following decade, and 80 percent between 2040 and supply, the capital stock, and TFP result in a projection 2049. of the average growth of labor force productivity. Changes in Projections of Other Labor Market Capital Services. Over the longer term, growth in the Outcomes Since Last Year. Projections of most other nation’s stock of capital will be driven by private saving, labor market outcomes are generally similar to what federal borrowing, total factor productivity, the after- tax rate of return, and international flows of financial capital. Private saving and international capital flows 18. Covered earnings are those received by workers in jobs subject tend to move with the after-tax rate of return on invest- to Social Security payroll taxes. Most workers pay payroll taxes on their earnings, although a small number—mostly in state and ment, which measures the extent to which investment local government jobs or in the clergy—are exempt. in the stock of capital results in a flow of income. CBO’s APPENDIX A: CBO’S PROJECTIONS OF DEMOGRAPHIC AND ECONOMIC TRENDS THE 2019 LONG-TERM BUDGET OUTLOOK 59 projection of that rate is consistent with the agency’s Nevertheless, CBO anticipates slower growth in labor projection that the real interest rate on 10-year Treasury quality than in the past. notes will be 1.4 percent in 2029 and 2.2 percent in 2049 (see “Interest Rates” on page 60). Another factor that is projected to slow the growth of nonfarm business TFP is the projected reduction in Total Factor Productivity. The annual growth of TFP in spending for federal investment. Under the assumptions the nonfarm business sector is projected to increase from used for CBO’s baseline, the government’s nondefense about 0.7 percent in 2019 to about 1.1 percent in 2022 discretionary spending is projected to decline over the and then to remain at that rate through 2049, yielding next decade to a much smaller percentage of GDP than an average annual growth rate of roughly 1.1 percent it has averaged in the past. About half of nondefense from 2019 to 2049. That projected growth rate is about discretionary spending from the 1980s onward has con- 0.3 percentage points slower than the average annual rate sisted of federal investment in physical capital (such as of 1.4 percent since 1950 and slightly slower than the roads and other infrastructure), education and training, average rate recorded since 1989. and research and development—all of which, in CBO’s judgment, contributed to TFP growth. Consequently, The projected path for nonfarm business TFP reflects lower nondefense discretionary spending as a percentage several considerations that, in CBO’s assessment, suggest of GDP would mean less federal investment, causing slower growth in coming decades than the long-term TFP to grow more slowly. historical average. For example, with the exception of a period of rapid growth in the late 1990s and early 2000s, Labor Productivity. Taken together, the projections of productivity has tended to grow more slowly in recent the labor force, capital services, and TFP result in labor decades than it did during the 1950s and 1960s. That force productivity that is expected to grow on average by long-term trend suggests that projections for the next 1.4 percent annually over the 2019–2049 period. When few decades should place greater weight on more recent, projections of total hours worked are used instead, real slower growth than on the relatively rapid growth of GDP per hour worked is expected to grow by an annual the more distant past. Thus, although CBO projects an average of 1.5 percent over the 2019–2049 period. acceleration of nonfarm business TFP growth from its unusually slow recent rate, the agency anticipates that Changes in Projections of Capital Accumulation and growth will return to a rate that is slower than its long- Productivity Since Last Year. CBO has revised its term historical average. analytic methods to account more fully for economic growth outside the nonfarm business sector—that is, in A number of developments support projections of slower the farm, household, nonprofit, and government sectors. growth in nonfarm business TFP. One is the anticipated That revision, which affects only the projection beyond slowing of growth in labor quality, a measure of workers’ the 10-year budget window, yields a more comprehen- skills that accounts for educational attainment and work sive accounting of the growth of private-sector capital experience that, in CBO’s analysis, is implicitly a part of services. For example, capital services from more sectors TFP. Following a relatively rapid rise during the 1980s outside the nonfarm business sector are now explicitly and 1990s, growth in labor quality slowed after 2000. included in CBO’s measure of capital services, in order In CBO’s assessment, that change results both from a to assess their contributions to GDP, whereas last year, gradual slowdown in the increase in average educational those services only implicitly contributed to GDP. As a attainment and from the burgeoning retirement of a result, CBO’s measure of capital services accounts for a relatively large and skilled portion of the workforce—the greater share of overall production than was the case last baby-boom generation. In coming decades, however, the year, and TFP accounts for less. slowdown in the growth of labor quality is expected to be partly offset by the aging of those remaining in the labor In addition, changes in historical data regarding the force, especially as better health and longer life expec- national income and product accounts that the Bureau tancy lead people to stay in the workforce longer than of Economic Analysis reported in July 2018 led CBO to did members of previous generations. (An older work- increase its projection of the growth in capital services force generally has a larger proportion of more highly and to lower its projection of the growth in TFP in the educated workers because they tend to remain in the nonfarm business sector. (As a result, TFP growth in that labor force longer than do workers with less education.) sector is expected to be slightly slower than it was in last 60 THE 2019 LONG-TERM BUDGET OUTLOOK JUNE 2019 year’s projections.) Those revisions offset each other and 0.4 percentage points less than the annual increase in the have little net effect on projected labor productivity. consumer price index. The GDP price index grows more slowly than the consumer price index because it is based The updated data and CBO’s revised analytic methods on the prices of a different set of goods and services and are reflected in its projection of labor force productivity a different method of calculation. over the 30-year projection period, which is lower than last year’s projection. This year, CBO projects that the Changes in Projections of Inflation Since Last Year. average annual rate of growth in labor force productivity Inflation in both measures of consumer prices is pro- would be roughly 1.4 percent from 2019 to 2048; last jected to be the same as the rates CBO projected last year year, CBO projected that rate would be roughly 1.5 per- for the 2018–2048 period. cent from 2019 to 2048. Interest Rates Inflation CBO projects the interest rates, both real and nominal, CBO projects rates of inflation for two categories: prices that apply to federal borrowing, including the rate on of consumer goods and services and prices of final goods 10-year Treasury notes and special-issue Social Security and services.19 Those rates influence nominal levels of bonds. It also projects the average nominal interest rates interest rates and income (that is, the levels without on federal debt held by the public and on the bonds held adjustments to remove the effects of inflation) and in the Social Security trust funds. Those rates influence thereby influence tax revenues, various types of federal the cost of the government’s debt burden and the evolu- expenditures that are indexed for inflation, and interest tion of the trust funds. payments on federal debt. After considering a number of factors, including slower Prices of Consumer Goods and Services. One measure growth in the labor force, slower growth in TFP, and of consumer price inflation is the annual rate of change higher government debt, CBO expects real interest rates in the consumer price index for all urban consumers on federal borrowing to be lower in the future than they (CPI-U). Over the 2019–2049 period, inflation in that have been, on average, over the past few decades. The measure averages 2.4 percent in CBO’s projections. real interest rate on 10-year Treasury notes (calculated That long-term rate is slightly less than the average rate by subtracting the rate of increase in the consumer price of inflation of 2.5 percent per year since 1990. CBO index from the nominal yield on those notes) averaged projects that, under a chained measure of CPI-U infla- roughly 2.9 percent between 1990 and 2007.21 That rate tion, prices will grow at a rate 0.25 percentage points less has averaged 0.8 percent since 2009 and is projected than the annual increase in the consumer price index.20 to be 1.4 percent in 2029. In CBO’s projections, the rate continues to rise thereafter, reaching 2.2 percent Prices of Final Goods and Services. After 2019, the in 2049. That rate is 0.7 percentage points below the annual inflation rate for all final goods and services average real interest rate on 10-year Treasuries over the produced in the economy, as measured by the rate of 1990–2007 period. CBO’s current projections of interest increase in the GDP price index, is projected to average rates are lower than last year’s. 19. Final goods and services include not only those purchased Factors Affecting Interest Rates. Interest rates are deter- directly by consumers, but also by businesses (for investment) mined by a number of factors. CBO projects the rates by and governments, as well as net exports. comparing how the values of those factors are expected 20. The chained CPI-U tends to grow more slowly than the to differ in the long term relative to their average values standard CPI-U for two reasons. First, it uses a formula that in the past. However, conclusions from such analyses better accounts for households’ tendency to substitute similar goods and services for each other when relative prices change. 21. Between 1970 and 2007, the real interest rate on 10-year Second, unlike the CPI-U, the chained CPI-U is little affected by Treasury notes averaged 2.8 percent; the average from 1954 statistical bias related to the sample sizes that the Bureau of Labor to 2007 was 2.6 percent. Historical inflation rates are taken Statistics uses in computing each index. Historically, inflation as from the consumer price index, adjusted to account for measured by the chained CPI-U has been 0.25 percentage points changes over time in the way that the index measures inflation. lower, on average, than inflation as measured by the CPI-U. See Bureau of Labor Statistics, “CPI Research Series Using CBO’s projections reflect that average difference between the two Current Methods (CPI-U-RS)” (accessed on March 28, 2018), measures. www.bls.gov/cpi/research-series/home.htm. APPENDIX A: CBO’S PROJECTIONS OF DEMOGRAPHIC AND ECONOMIC TRENDS THE 2019 LONG-TERM BUDGET OUTLOOK 61 depend greatly on the period being considered, as some also reducing the return on government bonds and recent decades show: Real interest rates were low in the other investments.24 1970s because of an unexpected surge in inflation. In the 1980s, when inflation declined at an unexpectedly rapid •• The share of total income received by higher-income pace, real rates were high.22 Interest rates fell sharply households is expected to be larger in the future during the financial crisis and recession that began in than during the 1990–2007 period. Higher-income 2007. households tend to save a greater proportion of their income, so the difference in the distribution of To avoid using any of those possibly less representative income is projected to increase the total amount of periods, CBO considered average interest rates and saving available for investment, other things being their determinants over the 1990–2007 period and then equal. As a consequence, the amount of capital per judged how different those determinants might be over worker is projected to rise and interest rates are the long term.23 That period was chosen for comparison expected to be lower. because it featured fairly stable expectations of inflation and no severe economic downturns or significant finan- •• TFP is projected to grow more slowly in the future cial crises. than it did from 1990 to 2007. For a given rate of investment, lower productivity growth reduces the Some factors reduce interest rates; others increase them. return on capital and results in lower interest rates, all In CBO’s estimates for the 2019–2049 period, several else being equal. factors tend to reduce interest rates on government secu- rities relative to their 1990–2007 average: •• CBO expects investors’ preferences for Treasury securities relative to riskier assets to remain •• The labor force is projected to grow much more elevated compared with inclinations over the slowly than it did from 1990 to 2007. That slower 1990–2007 period. Investors began to have less growth in the number of workers would tend to appetite for risk in the early 2000s, and the demand increase the amount of capital per worker in the long for low-risk assets was strengthened by the economic term, reducing the return on capital and therefore fallout from the financial crisis, the slow subsequent recovery, and financial institutions’ response to increased regulatory oversight. The rise in demand for Treasury securities from those factors contributed to lower returns (that is, to lower interest rates). CBO 22. CBO calculates real interest rates by subtracting expected rates expects preferences for Treasury securities relative to of inflation from nominal interest rates. In general, borrowers riskier assets to gradually decline over the next three and lenders agree to nominal interest rates after accounting decades but to remain above their average levels from for their expectations of what inflation will be. However, if 1990 to 2007. inflation ends up being higher than was expected when the rates were agreed to, real interest rates will turn out to be lower At the same time, in CBO’s estimates, several factors than anticipated. If inflation ends up lower than expected, the opposite will occur. CBO uses the actual consumer price index, tend to boost interest rates on government securities adjusted to account for changes over time in the way that the relative to their average over the 1990–2007 period: index measures inflation, as a proxy for both what expectations of inflation have been in the past and what they will be in the •• Under CBO’s extended baseline, federal debt is future. One drawback is that if inflation fluctuates rapidly projected to be much larger as a percentage of GDP over time, changes in expectations may lag behind changes in actual inflation. Although CBO’s approach could mismeasure than it was before 2007—reaching 93 percent by expectations of inflation and real interest rates in some years, the 2029 and 144 percent by 2049. The latter figure is way inflation has varied over time suggests that CBO’s approach more than three and a half times the average over is a useful proxy over long periods, on average. the 1990–2007 period. Greater federal borrowing 23. A Bank of England study identified a similar set of determinants that account for the decline in real interest rates over the past 24. For more information about the relationship between the 30 years. See Rachel Lukasz and Thomas D. Smith, Secular growth of the labor force and interest rates, see Congressional Drivers of the Global Real Interest Rate, Staff Working Paper Budget Office, How Slower Growth in the Labor Force Could 571 (Bank of England, December 2015), https://tinyurl.com/ Affect the Return on Capital (October 2009), www.cbo.gov/ y3mrtoyv (PDF, 1.8 MB). publication/41325. 62 THE 2019 LONG-TERM BUDGET OUTLOOK JUNE 2019 tends to crowd out private investment in the long observable, for example. And, although the distribution term, reducing the amount of capital per worker of income is observable, neither models nor empirical and increasing both interest rates and the return on estimates offer much guidance for quantifying its effect capital over time. on interest rates. •• The capital share of income—the percentage of total In light of those sources of uncertainty, CBO relies not income that is paid to owners of capital—has been on only on economic models and findings from the research an upward trend for the past few decades. That share literature but also on information from financial markets is projected to decline over the next decade from its to guide its assessments of the effects of various factors current, elevated level but remain higher than its on interest rates over the long term. The current rate average has been over recent decades. The factors that on 30-year Treasury bonds, for example, reflects mar- appear to have contributed to the rise in income for ket participants’ judgments about the path that interest owners of capital (such as technological change and rates on short-term securities will take 30 years into the globalization) are likely to persist, keeping it above future. That market forecast informs CBO’s assessment the historical average. In CBO’s estimation, a larger of market expectations for the risk premium—the pre- share of income accruing to owners of capital would mium paid to investors for the extra risk associated with directly boost the return on capital and, thus, interest holding longer-term bonds—and for investment oppor- rates. tunities in the United States and abroad, and it points to considerably lower interest rates well into the future than •• The retirement of members of the baby-boom those of recent decades. generation and slower growth of the labor force will reduce the number of workers in their prime saving Projections of Interest Rates. The nominal interest rate years relative to the number of older people who on 10-year Treasury notes is projected to average 4.0 per- are drawing down their savings, CBO projects. As a cent over the 2019–2049 period and to reach 4.6 percent result, in CBO’s estimates, the total amount of saving in 2049. The real interest rate on 10-year Treasury notes available for investment is less than it otherwise is projected to average about 1.6 percent and, at the end would be (all else being equal), which tends to reduce of the period, to be 2.2 percent. the amount of capital per worker and thereby push up interest rates. (CBO estimates that this effect will The average interest rate on all federal debt held by only partially offset the positive effect of increased the public tends to be lower than the rates on 10-year income inequality on saving, leaving a net increase in Treasury notes because interest rates are generally lower savings available for investment.) on shorter-term debt than on longer-term debt and because Treasury securities are expected to mature, •• CBO anticipates that emerging-market on average, over periods of less than 10 years.25 CBO economies will attract a greater share of foreign projects a 0.4 percentage-point difference between the investment in coming decades than they did in rate on 10-year Treasury notes and the effective rate on the 1990–2007 period. As economic and financial federal debt over the 2030–2049 period. That difference conditions in those economies continue to improve, is projected to average 0.6 percentage points over the they will become increasingly attractive destinations next decade. The difference is larger over the coming for foreign investment. CBO projects that decade than for later years because a significant portion development would put upward pressure on interest of federal debt that will be outstanding during the next rates in the United States. 10 years was issued at the very low interest rates pre- vailing in the aftermath of the 2007–2009 recession. Some factors mentioned above are easier than others to (The average interest rate on all federal debt changes quantify. For instance, the effect of labor force growth more slowly than the 10-year rate because only a por- and rising federal debt can be estimated from available tion of federal debt matures each year.) Thus, in CBO’s data, theoretical models, and estimates in the literature. The extent to which other factors will affect interest 25. In particular, over the next decade, CBO expects the difference rates is more difficult to estimate. A shift in preferences between the rate on 3-month Treasury bills and the rate on for low- rather than high-risk assets is not directly 10-year Treasury notes to average 0.8 percentage points. APPENDIX A: CBO’S PROJECTIONS OF DEMOGRAPHIC AND ECONOMIC TRENDS THE 2019 LONG-TERM BUDGET OUTLOOK 63 projections, the average nominal interest rate on all fed- CBO’s projections of interest rates are different from eral debt held by the public is about 3.6 percent for the last year’s mainly because they are now based on a more 2019–2049 period and reaches 4.2 percent in 2049. comprehensive assessment of how changes in private investment affect the capital stock and thus the return on The Social Security trust funds hold special-issue bonds capital. Changes in the return on capital are estimated that generally earn interest at rates that are higher than to drive changes in interest rates across the economy. the average rate on federal debt. In CBO’s projections, Previously, changes in CBO’s measure of capital services the nominal interest rate on bonds newly issued to the owing to changes in investment in essence incorporated trust funds is equal to the rate on 10-year Treasury notes effects only on nonfarm business capital. Now, changes and averages 4.0 percent over the 2019–2049 period in capital services from changes in investment incor- and reaches 4.6 percent in 2049. The corresponding real porate effects on a broader range of capital, including rates are 1.6 percent, on average, over the full period and owner-occupied residential housing. 2.2 percent in 2049. That modeling improvement results in a smaller esti- Because interest rates have been low for much of the mated effect on capital services from a change in invest- past decade, CBO projects that the average interest ment (because of larger deficits and more crowding rate earned by all bonds held by the Social Security out, for example). Because CBO now incorporates the trust funds (both new and previously issued) would effect of changes in a capital stock measure that is more be slightly lower than the interest rate on newly issued comprehensive this year than it was last year, any given bonds over the next decade. The average interest rate change in private investment results in a smaller percent- on all bonds, which CBO uses to calculate the pres- age change in the agency’s capital stock measure than it ent value of future streams of revenues and outlays for did last year. In addition, residential housing depreciates those funds, is projected to average 3.8 percent over the more slowly than most other forms of capital, so the 2019–2049 period.26 immediate effect of residential investment on residential capital is relatively small. The smaller percentage effect Changes in Projections of Interest Rates Since Last on capital results in a smaller change in the return on Year. CBO’s current projections of interest rates are capital and ultimately a smaller change in interest rates lower than last year’s. The real rates on 10-year Treasury resulting from a change in investment. notes and the Social Security bonds are projected to average 1.6 percent over the 2019–2049 period and to Because of that modeling improvement, changes in be 2.2 percent in 2048. Last year, CBO projected that deficits have a smaller effect on interest rates in this year’s both rates would average 1.7 percent over the 2018– extended baseline projection. That occurs even though 2048 period and would be 2.4 percent in 2048. CBO has not changed its assessment of how changes in deficits affect private investment. In addition, slower 26. A present value is a single number that expresses a flow of past growth in both the labor force and TFP imply slightly and future income or payments in terms of an equivalent lump lower returns on capital and, in turn, lower interest rates. sum received or paid at a specific time. The value depends on All told, the average projected interest rate on 10-year the rate of interest, known as the discount rate, that is used to Treasury notes over the 2019–2048 period is 0.1 per- translate past and future cash flows into current dollars at that centage point lower than CBO projected a year ago. time. APPENDI X B Appendix BB Appendix Changes in Long-Term Budget Projections Since June 2018 The 30-year extended baseline projections for federal are 0.3 percentage points and 0.2 percentage points spending and revenues presented in this report differ lower, respectively, than projected last year. from the projections that the Congressional Budget Office published in 2018 because of certain changes •• Debt held by the public is projected to grow more in law, the availability of more recent data, changes to slowly than projected last year, rising from 78 percent the agency’s projections of demographic and economic of GDP this year to 141 percent in 2048; last year, factors, and other changes in assumptions and meth- CBO projected that it would rise from 79 percent of ods.1 CBO has also revised its methods of analyzing GDP in 2019 to 152 percent in 2048. uncertainty and fiscal scenarios that are alternatives to the extended baseline projections. This appendix com- The revised projections of deficits and debt resulted pares CBO’s current projections with the previous ones. primarily from lower projected spending, which was par- Because most of last year’s projections ended in 2048, tially offset by a small reduction in projected revenues. the appendix generally makes comparisons only through that year. •• Projected discretionary spending throughout the 30-year projection period is lower than CBO Measured as a percentage of gross domestic product anticipated last year because appropriations for relief (GDP), budget deficits and federal debt held by the pub- and recovery efforts related to hurricanes and wildfires lic are now projected to be smaller over the next three were smaller in 2019 than in 2018.2 decades than CBO projected last year. •• Net spending for interest on debt over the 30-year •• In CBO’s extended baseline projections, deficits are period is lower in this year’s projections than it was projected to grow from 4.2 percent of GDP this year in last year’s because less debt is projected to be to 8.6 percent in 2048, which are 0.4 percentage accumulated and because CBO has revised downward points and 1.0 percentage point lower, respectively, its projections of the average interest rate on that debt than projected last year (see Figure B-1). (see Appendix A). •• Primary deficits—deficits excluding net spending for •• Projected outlays for Social Security (throughout interest—are projected to grow from 2.4 percent of the 30-year period) and major health care programs GDP this year to 3.0 percent of GDP in 2048, which (over the first 10 years) are slightly smaller than they were last year because the most recent data show reductions in the number of beneficiaries and in 1. The extended baseline projections generally reflect current spending, respectively. law, following CBO’s 10-year baseline budget projections and then extending most of the concepts underlying those projections for the rest of the long-term projection period. •• Revenues are projected to be slightly lower because of For the 2018 extended baseline projections, see Congressional new administrative and tax data. Budget Office, The 2018 Long-Term Budget Outlook (June 2018), www.cbo.gov/publication/53919. For the 10-year projections underlying the extended baseline projections in this report, 2. Projections of discretionary spending are based on the most see Updated Budget Projections: 2019 to 2029 (May 2019), recent appropriations for each discretionary program and are www.cbo.gov/publication/55151. For the changes in projections increased over time to account for inflation. (In addition, total of demographic and economic factors since 2018, see discretionary spending is subject to caps that are specified in law Appendix A of this report. and that limit discretionary outlays through 2021.) 66 THE 2019 LONG-TERM BUDGET OUTLOOK JUNE 2019 Figure B-1 . CBO’s 2018 and 2019 Extended Baseline Projections of Deficits and Federal Debt Held by the Public Percentage of GDP Deficits 5 Annual deficits over the next three decades are 0 smaller in this year’s Primary Deficit projections than they -3.0 2019 Projection were in last year’s. The -3.3 2018 Projection differences are larger −5 for total deficits than for Deficit primary deficits because projected net spending for interest has fallen. -8.6 2019 Projection −10 -9.5 2018 Projection 2018 2023 2028 2033 2038 2043 2048 Federal Debt Held by the Public 160 152 2018 Projection 140 141 2019 Projection Projected debt is 120 also lower now. The difference between the two sets of 100 projections grows over time, reaching about 11 percent of GDP. 80 0 2018 2023 2028 2033 2038 2043 2048 Source: Congressional Budget Office. The extended baseline projections generally reflect current law, following CBO’s 10-year baseline budget projections and then extending most of the concepts underlying those projections for the rest of the long-term projection period. The primary deficit is the deficit excluding net spending for interest. GDP = gross domestic product. Changes in Projected Spending Noninterest Spending In CBO’s extended baseline projections, spending as As a percentage of GDP, noninterest spending—that a percentage of GDP is lower than projected last year is, spending for Social Security, spending for the major because of reductions in both noninterest spending and federal health care programs, other mandatory spending, net spending for interest. and discretionary spending—is projected to be lower throughout the 30-year period than projected last year APPENDIX B: CHANGES IN LONG-TERM BUDGET PROJECTIONS SINCE JUNE 2018 THE 2019 LONG-TERM BUDGET OUTLOOK 67 Figure B-2 . CBO’s 2018 and 2019 Extended Baseline Projections of Outlays and Revenues Percentage of GDP Outlays 25 Noninterest spending and net Noninterest 23.1 2018 Projection spending for interest are both 22.5 2019 Projection lower in this year’s projections 20 than they were in last year’s. Noninterest spending is now 15 projected to be 0.3 percent of GDP lower in 2019, and 10 0.6 percent of GDP lower in 2048, than projected last year. 6.3 2018 Projection Although this year’s projection of 5 Net Interest net spending for interest is about 5.5 2019 Projection the same as last year’s for 2019, it is about 0.7 percent of GDP 0 lower for 2048. 2018 2023 2028 2033 2038 2043 2048 Revenues 25 19.8 2018 Projection 20 19.4 2019 Projection 15 Projected revenues over the next three decades are also 10 slightly lower in this year’s projections than they were in last year’s—0.4 percent of GDP lower 5 by 2048. 0 2018 2023 2028 2033 2038 2043 2048 Source: Congressional Budget Office. The extended baseline projections generally reflect current law, following CBO’s 10-year baseline budget projections and then extending most of the concepts underlying those projections for the rest of the long-term projection period. GDP = gross domestic product. (see Figure B-2). Most of that change stems from lower year; and 5.0 percent of GDP in 2048, rather than the projections of discretionary spending. 5.5 percent projected last year. The reduction throughout the 30-year period occurred primarily because appro- Discretionary Spending. In CBO’s current projections, priations for 2019 that are designated as emergency outlays for discretionary spending as a percentage of requirements (generally to respond to wildfires and other GDP equal 6.3 percent of GDP in 2019, rather than major disasters) are substantially lower in 2019 than they the 6.4 percent projected last year; 5.0 percent of GDP were in 2018. So far, appropriations for 2019 amount to in 2029, rather than the 5.4 percent projected last $2 billion—a sharp reduction from the $108 billion that 68 THE 2019 LONG-TERM BUDGET OUTLOOK JUNE 2019 was appropriated in 2018, mostly for relief and recovery demographic and economic projections). For DI out- efforts related to Hurricanes Harvey, Irma, and Maria lays, the projected reduction after the first decade is and wildfires in western states. In accordance with the also driven by a reduction in the projected share of the statutes that govern its projections, CBO develops its population that would receive disability benefits. CBO projections for discretionary spending by starting with now projects a long-run age- and sex-adjusted rate of appropriations for the most recent year available and disability incidence—the share of workers who are adjusting those amounts for inflation over time. CBO’s awarded disability benefits in each year out of all workers projections last year were based on the amounts appro- who are insured under DI but not receiving benefits at priated for 2018; the current projections are based on the the start of the year—of 5.2 per 1,000.4 Last year, the much smaller amounts appropriated for 2019.3 projected rate was 5.4 per 1,000. CBO revised it because there have consistently been fewer new DI beneficiaries Spending for Social Security. In CBO’s current projec- than the agency expected in recent years. The revised rate tions, outlays for Social Security as a percentage of GDP is also closer to current longer-term historical averages. are slightly lower than the agency anticipated last year. Specifically, the average rate from 1990 through 2018, Although projected spending for Social Security in 2019 a time during which DI policy has remained fairly is about the same as projected last year (4.9 percent of steady—and also from 1990 through 2007, the period GDP), it is slightly lower throughout the next 10 years covering the last two full business cycles—was about and thereafter. In 2048, that spending is projected to 5.2 per 1,000. The downward revision reduces the total equal 6.2 percent of GDP, rather than the 6.3 percent of projected number of DI beneficiaries in 2048 by about GDP projected last year. 3 percent. Over the next decade, the revisions to outlays are due Spending for Major Health Care Programs. CBO’s to slight reductions in projected spending for both of current projection of federal spending for the major Social Security’s components—Old-Age and Survivors health care programs, measured as a percentage of GDP, Insurance (OASI) and Disability Insurance (DI). CBO’s is slightly lower over the next 10 years than it was in last current projections of the number of OASI beneficiaries year’s projections and about the same thereafter. The and DI beneficiaries over the next 10 years are lower change consists mainly of small revisions in projected than the previous projections by about 1 percent and outlays for Medicare and for subsidizing health insurance 5 percent, respectively. Those revisions reflect recent purchased through marketplaces and related spending. data showing that people claimed OASI benefits later than expected and that fewer people were awarded DI Medicare. Spending for Medicare net of offsetting benefits than expected. In addition, CBO has reduced receipts (which are mostly premiums paid by beneficia- its projections of population growth. The effect of fewer ries) is projected to be about 0.1 percent of GDP lower beneficiaries is slightly offset by a higher-than-anticipated in 2019 than anticipated last year and less than 0.1 per- cost-of-living adjustment that beneficiaries received in cent of GDP lower, on average, over the first decade of January 2019. the projection period. That revision was made mainly because the most recent data indicate that spend- After 2028, the slight reduction in OASI outlays is ing for Medicare’s Part D (which covers prescription driven by downward revisions to CBO’s projections of drugs) and Part A (Hospital Insurance) has been lower the population and to projections of wages and salaries than expected. After the first decade, net spending for (see Appendix A for a discussion of changes in CBO’s Medicare is projected to be about the same as projected last year. 3. To project discretionary spending, CBO assumes that such spending would generally adhere to the caps that are specified in Medicaid, CHIP, and Marketplace Subsidies. Throughout current law through 2021 and then increase gradually, to account the first decade, outlays for Medicaid and the Children’s for inflation, through 2029. Afterward, discretionary spending remains roughly constant as a percentage of GDP in CBO’s projections. (It is not precisely constant as a percentage of GDP 4. The adjustment accounts for changes since 2000 in the age and because CBO’s projection of GDP includes the macroeconomic sex makeup of the population that has worked long enough effects of the policies underlying the extended baseline and recently enough to satisfy work requirements for disability projections.) benefits but is not yet receiving those benefits. APPENDIX B: CHANGES IN LONG-TERM BUDGET PROJECTIONS SINCE JUNE 2018 THE 2019 LONG-TERM BUDGET OUTLOOK 69 Health Insurance Program (CHIP), together with spend- For Medicare, the average annual rate of excess cost ing to subsidize health insurance purchased through the growth implicit in CBO’s baseline projections is about marketplaces established under the Affordable Care Act 1.2 percent from 2020 through 2029, a slightly higher and related spending, are projected to be, on average, less rate than last year’s projection of 1.0 percent from 2019 than 0.1 percent of GDP lower than projected last year. through 2028. (The increase reflects slightly higher pro- That reduction is driven by lower projections of pre- jected spending per beneficiary because CBO revised its miums for insurance purchased through marketplaces, methods to incorporate updated data from the Centers reflecting updated data and technical revisions. After the for Medicare & Medicaid Services.) The projected rate first decade, projected spending is also less than 0.1 per- of excess cost growth for 2030 is 1.2 percent, the same cent of GDP lower than projected last year, reaching as last year’s estimate for 2029. Excess cost growth is 3.3 percent of GDP in 2048. projected to average 1.1 percent over the full projection period, a slightly higher rate than last year’s estimate for Methods Underlying Projections of Health Care Spending. the 2018–2048 period (1.0 percent) but the same as the To project long-term spending for the major health historical average of 1.1 percent from 1985 to 2017. care programs, CBO used the same method that it used last year. Namely, it combined estimates of the number For Medicaid, the average annual rate of excess cost of people who are projected to receive benefits from growth implicit in CBO’s baseline projections for the those programs with fairly mechanical estimates of the federal share of such spending is 1.8 percent from 2020 growth of spending per beneficiary (adjusted to account through 2029, up by about 0.2 percentage points from for demographic changes to the beneficiaries in each last year’s estimate for 2019 through 2028. The rate program). CBO has estimated the growth of spend- for 2030 is 1.8 percent, up by about 0.2 percentage ing per beneficiary by combining projected growth in points from last year’s estimate for 2029. Those changes potential nominal GDP per person with projected excess were the cumulative result of many updates that CBO cost growth for each program. (Potential GDP is the made to its baseline projections for legislative, eco- maximum sustainable output of the economy; excess nomic, and technical reasons. The rate of excess cost cost growth is the extent to which health care costs per growth is projected to average 1.6 percent over the person, after being adjusted for demographic changes, full projection period, which is about 0.3 percentage grow faster than potential GDP per person.) For both points higher than last year’s estimate for the 2018– the 10-year and the 30-year periods, potential GDP per 2048 period and 0.9 percentage points higher than the person is projected to grow at an average rate of about 1985–2017 average. 3.4 percent per year, about the same rate that CBO projected last year. For private health insurance premiums, which CBO uses as an input to its calculation of marketplace subsidies, Through 2029, CBO used the rate of excess cost growth the average annual rate of excess cost growth implicit for Medicare, Medicaid, and private health insurance in CBO’s baseline projections is 1.8 percent from 2020 premiums that is implicit in the agency’s 10-year baseline through 2029 (which is slightly lower than last year’s projections for each of those categories. For 2030, the estimate of 2.0 percent for the 2019–2028 period). The rate equals the average rate from the last 5 years of those rate for 2030 is 1.5 percent, which again is slightly lower projections (2025 to 2029), which is different for each than last year’s estimate of 1.6 percent for 2029. The rate category. After 2030, the rate for each category moves of excess cost growth is projected to average 1.4 percent linearly, by the same fraction of a percentage point each over the full projection period, which is about 0.1 per- year, from that category-specific rate to a rate of 1.0 per- centage point lower than last year’s estimate for the cent in 2049.5 2018–2048 period and 0.7 percentage points lower than the 1988–2017 average. Other Mandatory Spending. CBO’s projections for 5. For more information, see Congressional Budget Office, The other mandatory spending are slightly lower than they 2016 Long-Term Budget Outlook (July 2016), Chapter 3, were last year. (Other mandatory spending includes out- www.cbo.gov/publication/51580. In contrast to outlays for the larger health care programs, outlays for CHIP are projected to be lays for retirement programs for federal civilian and mili- a constant percentage of GDP after 2029. tary employees, certain programs for veterans, refundable 70 THE 2019 LONG-TERM BUDGET OUTLOOK JUNE 2019 tax credits, the Supplemental Nutrition Assistance have been levied on—particularly proprietors’ income Program, and all other mandatory programs aside from and monetary interest income, which is the share of per- Social Security and the health care programs described sonal interest income that does not come from marketed above.) On average over the 30-year projection period, goods and services. Average tax rates on those types of outlays for other mandatory spending as a percentage of income have therefore been lower than CBO previously GDP in CBO’s projections are less than 0.1 percent of estimated. As a result, CBO has lowered its projections GDP lower than projected last year. That small change of average tax rates in the future. Additional factors was the cumulative result of several updates that CBO contributing to the downward revisions include lower made for legislative, economic, and technical reasons. taxable distributions from pension plans than projected previously, changes in the relationship between earnings Net Spending for Interest and payroll tax receipts that have taken place in recent In CBO’s current projections, net spending for years and that are projected to persist, and a down- interest—that is, the government’s interest payments ward revision to CBO’s forecast of wages and salaries. on debt held by the public, offset by interest income Receipts from corporate income taxes are also projected that the government receives—is lower throughout the to be 0.1 percent of GDP lower over the 2019–2028 30-year projection period than it was in last year’s pro- period; that change results from new data from corporate jections (see Figure B-2 on page 67). That spending is income tax returns for 2016 and improvements in CBO’s lower because the agency’s projections of interest rates, modeling of the income of multinational corporations. deficits, and federal debt held by the public are likewise lower. (For a discussion of changes to CBO’s projections Those effects are partially offset by an increase in pro- of interest rates, see Appendix A.) jected revenues from customs duties that reflects new tariffs imposed by the Administration during 2018. In For the 2019–2028 period, net spending for interest 2019 and over the 2019–2028 period, those revenues is projected to average 2.5 percent of GDP; last year, are now projected to be 0.2 percent of GDP higher than the projected average was 2.7 percent. It is projected to CBO projected last year.6 equal 3.0 percent of GDP in 2029 (down 0.2 percent- age points from last year’s projections) and 5.5 percent Changes in Social Security’s of GDP in 2048 (down 0.7 percentage points from last Projected Finances year’s projections). Social Security’s 75-year actuarial deficit––a measure of the program’s budgetary shortfall over a 75-year period–– Changes in Projected Revenues is currently projected to be 1.5 percent of GDP (which In CBO’s extended baseline projections, revenues as a is about the same as estimated last year) or 4.6 percent percentage of GDP are slightly lower throughout the of taxable payroll (which is slightly higher than last year’s 30-year period than they were in last year’s projections. estimate of 4.4 percent).7 Although in 2019 they are projected to be about the same, by 2048 they are projected to be about 0.4 percent 6. The projections of revenues from customs duties over the of GDP lower than CBO projected last year. Most of next decade were most recently published in May 2019; see the revisions occur in the first decade of the projection Congressional Budget Office, Updated Budget Projections: 2019 to 2029 (May 2019), www.cbo.gov/publication/55151. CBO’s period. extended baseline projections incorporate the assumption that the new tariffs would continue throughout the projection The downward revisions to total revenues as a share of period at the rates in effect at the beginning of May. For more GDP result from CBO’s slightly lower projections of information about CBO’s approach to projecting revenues from individual income taxes, payroll taxes, and corporate customs duties, see Congressional Budget Office, The Budget and income taxes. Receipts from individual income taxes Economic Outlook: 2019 to 2029 (January 2019), www.cbo.gov/ publication/54918. and from payroll taxes are now each projected to be 0.1 percent of GDP lower over the 2019–2028 period 7. The actuarial deficit is computed as the sum of the present value than CBO projected last year. Those changes are mainly of projected tax revenues and the trust funds’ current balance minus the sum of the present value of projected outlays and a driven by new administrative and tax data that suggest year’s worth of benefits at the end of the period. The result is lower tax receipts than CBO had projected. Also, the negative, indicating that the program’s long-term cost is greater Bureau of Economic Analysis has revised upward its esti- than its income. A present value is a single number that expresses mates of some of the sources of income that those taxes a flow of current and future income (or payments) in terms of an equivalent lump sum received (or paid) at a specific time. APPENDIX B: CHANGES IN LONG-TERM BUDGET PROJECTIONS SINCE JUNE 2018 THE 2019 LONG-TERM BUDGET OUTLOOK 71 Those projections result from several factors. On the economic factors by using simulations from a multivar- one hand, CBO has lowered its projection of payroll iate statistical model.9 That approach accounts not only taxes, making the actuarial deficit larger. That reduction for the uncertainty of long-term trends of individual was a result of new administrative and tax data, reduced factors but also for the uncertainty of those factors’ long- projections of the labor force, and downward revisions term movement in relation to one another. to wages and salaries throughout the projection period. Incorporating into the analysis another year (2093) with CBO has also significantly increased the number of a relatively large difference between Social Security reve- factors that it varies when analyzing uncertainty. Last nues and outlays also increased the actuarial deficit. On year, CBO varied four key factors—the labor force the other hand, CBO has lowered projected outlays for participation rate, the growth rate of total factor pro- Social Security—making the actuarial deficit smaller— ductivity (TFP) in the nonfarm business sector, interest because the agency has reduced its projections of the rates on federal debt held by the public, and excess cost number of beneficiaries in the OASI and DI programs growth for Medicare and Medicaid spending.10 This year, and its projections of wages and salaries. (The reduction CBO varied seven key factors, the first three of which are to projections of the number of beneficiaries was made demographic, the next three of which are economic, and partly because of downward revisions to the long-run the last of which relates to health care: rate of disability incidence and population growth.) The small increase in the actuarial deficit when measured •• The total fertility rate, as a percentage of taxable payroll also reflects a slightly lower projection of the share of earnings that is subject to •• The rate of mortality improvement, Social Security payroll taxes over the next 30 years.8 •• The net immigration rate, CBO projects that if current law governing the program’s taxes and benefits did not change, the DI trust fund •• The growth rate of TFP in the nonfarm business would be exhausted in fiscal year 2028, the OASI trust sector, fund would be exhausted in calendar year 2032, and the combined trust funds would also be exhausted in cal- •• Interest rates on federal debt held by the public, endar year 2032. Last year, those exhaustion dates were three years earlier for the DI trust fund, the same year for •• The civilian unemployment rate, and the OASI trust fund, and one year earlier for the com- bined trust funds. The change in the date of exhaustion •• Excess cost growth for Medicare and Medicaid of the DI trust fund is due to lower projections of the spending. number of DI beneficiaries over the next 10 years. Introducing demographic factors into this year’s Changes in Analyzing Uncertainty analysis—specifically, varying the civilian unemployment CBO has changed the methods that it uses to analyze the rate together with the demographic factors, which affect uncertainty of its projections. To illustrate that uncer- tainty last year, the agency created two projections of fed- 9. A multivariate statistical model is one that describes the statistical eral debt—one in which key factors were varied in ways properties (such as mean, standard deviation, and correlations) of multiple variables. CBO’s simulations for the growth rate that would raise projected deficits in relation to CBO’s of total factor productivity in the nonfarm business sector, the extended baseline projections and another in which those interest rates on federal debt held by the public, and the civilian factors lowered projected deficits. The ranges of variation unemployment rate are based on a vector autoregressive (VAR) that the agency used for the economic factors were based multivariate model. CBO’s VAR model incorporates parameters on historical movements and potential future develop- that vary with time, allowing the variables to have time-varying ments of each individual factor. This year, CBO instead statistical properties. In particular, the model allows the variables to exhibit highly persistent (but not necessarily permanent) analyzed the long-term uncertainty surrounding the key deviations from their historical averages. CBO estimated the parameters of the VAR model using annual data from 1953 to 2018. 8. Beyond the 30-year projection period, the share of earnings subject to Social Security payroll taxes is held constant in CBO’s 10. Total factor productivity is the average real output per unit of projections. combined labor and capital services. 72 THE 2019 LONG-TERM BUDGET OUTLOOK JUNE 2019 the size of the working-age population—has allowed place––including the individual income tax provisions CBO to make a more comprehensive assessment of the of Public Law 115-97 (often called the 2017 tax act in uncertainty of the economy’s future amount of labor. CBO’s publications), which are scheduled to expire in In last year’s analysis, by contrast, CBO varied one 2026 under current law. In that scenario, projected defi- economic factor, the labor force participation rate, to cits were larger than in the extended baseline projections. quantify that uncertainty. In the second and third scenarios, projected deficits were larger still. CBO’s new method of quantifying uncertainty in its projections and the additional uncertainty stemming In this report, CBO has described two scenarios in addi- from the demographic factors included in this year’s tion to the extended baseline projections: an extended analysis result in noticeable differences from last year’s alternative fiscal scenario and another scenario, called ranges of budgetary outcomes. Last year, CBO estimated the payable-benefits scenario, which incorporates the that in 2039, federal debt under current law could be assumption that outlays for Social Security would be as much as 43 percent of GDP higher or 35 percent of reduced to equal the program’s total annual revenues GDP lower than it was in the agency’s extended baseline once the combined Social Security trust funds were projections. Also, CBO noted that those estimates did exhausted. not cover the full range of possibilities. And the agency did not quantify the degree of its certainty that actual The Extended Alternative Fiscal Scenario debt would equal a value between those estimates. Under Last year, CBO projected that debt held by the public the new method, CBO now estimates that if future tax in the extended alternative fiscal scenario would equal and spending policies did not vary from those specified about 210 percent of GDP in 2048, which was about in current law, there is a two-thirds chance that federal 60 percentage points more than in that year’s extended debt held by the public in 2039 could be as much as baseline projections. This year, CBO projects that it 62 percent of GDP higher or 42 percent of GDP lower would equal 211 percent in 2048, which is 70 percent- than it is in the agency’s extended baseline projections. age points more than in this year’s extended baseline projections. The larger difference this year results from CBO’s current analysis of uncertainty extends 20 years several modeling changes. into the future; last year, the analysis extended 30 years into the future. The likely range of uncertainty that Two of the changes make the difference larger. First, in CBO’s models produce for projections of debt is less this year’s extended alternative fiscal scenario, revenues informative after 20 years because the key parameters are lower than in the extended baseline projections by governing the economic effects of fiscal policy in the a larger amount than they were last year because CBO agency’s models are based on the nation’s historical expe- has modeled the long-term effects of the policy changes rience with federal borrowing. At the high end of a range in the extended alternative fiscal scenario in more detail 30 years in the future, projections of debt as a percentage than it did last year. Second, economic output is lower of GDP would grow to amounts well outside that histor- in the extended alternative fiscal scenario than it is in ical experience. the extended baseline projections, pushing down some kinds of noninterest spending—and in CBO’s long- Changes in Alternative Scenarios for term projections, such spending is now less sensitive to Fiscal Policy changes in economic output over the next decade than Last year, CBO published a report that described three it was previously. As a result, the reduction in economic fiscal scenarios that were alternatives to the extended output in relation to the extended baseline projections baseline projections.11 The first of the three scenarios, pushes down noninterest spending by a smaller amount called the extended alternative fiscal scenario, incor- in this year’s extended alternative fiscal scenario than in porated the assumption that current law was changed last year’s. to maintain certain major policies that are now in Those changes are partially offset by a modeling improve- ment that reduces the effects of deficits on the return on 11. Congressional Budget Office, The Long-Term Budget Outlook Under Alternative Scenarios for Fiscal Policy (August 2018), capital and interest rates. Those effects are now based www.cbo.gov/publication/54325. on a more comprehensive assessment of how changes in APPENDIX B: CHANGES IN LONG-TERM BUDGET PROJECTIONS SINCE JUNE 2018 THE 2019 LONG-TERM BUDGET OUTLOOK 73 private investment affect the capital stock and thus the in reducing benefits by 24 percent in calendar year 2033 return on capital—which is a key factor driving changes (the year after the program’s combined trust funds are in interest rates throughout the economy, in CBO’s view. projected to be exhausted) and by 29 percent in calendar This year, CBO has expanded its measure of the capital year 2049. In 2017, CBO estimated that the reduction stock that is affected by changes in private investment to in benefits would amount to 28 percent in calendar year include owner-occupied residential housing. Last year, 2031 (the year after the projected exhaustion of the com- that measure mainly included nonfarm business capital bined trust funds in that analysis) and greater percent- stock. Because the measure is now more comprehensive, ages in later years.12 any given change in private investment (for example, a change resulting from larger deficits, which crowd out This year, CBO analyzed the payable-benefits scenario private investment) now results in a smaller percentage with the same suite of dynamic macroeconomic models change in the measure. That smaller percentage effect on that the agency used in analyzing the extended alterna- the capital stock results in a smaller change in the return tive fiscal scenario. In the 2017 analysis, CBO used a on capital and ultimately in a smaller change in interest simpler set of models; assumed that people would not rates. Because of that modeling improvement, changes change their decisions about consumption, saving, or in deficits have a smaller effect on interest rates, and ulti- work in anticipation of receiving lower Social Security mately on federal debt held by the public, in this year’s benefits; and assumed that they would not change their extended alternative fiscal scenario. decisions about saving or work after receiving those lower benefits. The Payable-Benefits Scenario In CBO’s current payable-benefits scenario, debt equals Furthermore, the current analysis incorporates the ways 106 percent of GDP in 2049, which is 38 percentage in which those changed decisions about work and saving points below its level in the extended baseline projec- would affect the economy and feed back into the federal tions. CBO last examined a payable-benefits scenario in budget. It also incorporates the ways in which those 2017 and projected that debt would equal 111 percent decisions would affect the overall demand for goods and of GDP in 2047, the last year of the extended baseline services when benefits were unexpectedly cut. The previ- projections at the time, which was 39 percentage points ous analysis did not account for any of those effects. below its level in those extended baseline projections. Also, CBO now projects that limiting Social Security 12. See Congressional Budget Office, The 2017 Long-Term Budget benefits to amounts payable from revenues would result Outlook (March 2017), www.cbo.gov/publication/52480. List of Tables and Figures Tables 1-1. CBO’s Extended Baseline Projections 6 1-2. Assumptions About Outlays and Revenues Underlying CBO’s Extended Baseline Projections 19 1-3. Summary Financial Measures for the Social Security System 22 1-4. Effective Marginal Federal Tax Rates Underlying CBO’s Extended Baseline Projections 30 1-5. Middle Two-Thirds of the Ranges of Possible Outcomes for Key Factors Used to Quantify the Uncertainty in CBO’s Projections of Federal Debt Held by the Public 34 2-1. Budget Projections Under Three Scenarios 40 2-2. Long-Term Economic Effects Under Two Scenarios Relative to CBO’s Extended Baseline Projections 41 2-3. Short-Term Economic Effects Under Two Scenarios Relative to the Extended Baseline Projections 43 A-1. Average Annual Values for Demographic Variables That Underlie CBO’s Extended Baseline Projections 50 A-2. Average Annual Values for Economic Variables That Underlie CBO’s Extended Baseline Projections54 Figures 1-1. Federal Debt Held by the Public Since 1790 7 1-2. The Federal Budget in 2019 and 2049 8 1-3. CBO’s Budget Projections in Brief 10 1-4. Total Deficit, Primary Deficit, and Net Interest 12 1-5. Population, by Age Group 15 1-6. Average Annual Growth of Real Potential GDP 17 1-7. Outlays and Revenues in Selected Years 20 1-8. Composition of Federal Outlays 21 1-9. Federal Outlays for the Major Health Care Programs, by Category 24 1-10. Spending for Social Security and the Major Health Care Programs in 2019 and 2049 25 1-11. Average Annual Growth of Health Care Costs per Beneficiary 26 1-12. Other Federal Noninterest Spending 27 1-13. Increases in Federal Revenues 28 1-14. Shares of Income Taxed at Different Rates Under the Individual Income Tax System 29 1-15. Uncertainty in CBO’s Projections of Federal Debt Held by the Public 31 2-1. Output per Person and Debt Under Three Scenarios 42 2-2. The Size of Policy Changes Needed to Make Federal Debt Meet Two Possible Goals in 2049 45 2-3. How Timing Affects the Size of Policy Changes Needed to Make Federal Debt Meet Two Possible Goals in 2049 47 B-1. CBO’s 2018 and 2019 Extended Baseline Projections of Deficits and Federal Debt Held by the Public 66 B-2. CBO’s 2018 and 2019 Extended Baseline Projections of Outlays and Revenues 67 About This Document This volume is one of a series of reports on the state of the budget and the economy that the Congressional Budget Office issues each year. In keeping with CBO’s mandate to provide objective, impartial analysis, the report makes no recommendations. Overseen by Julie Topoleski and prepared with guidance from Devrim Demirel, Edward Harris, John Kitchen, John McClelland, David Weaver, and Jeff Werling, the report represents the work of many analysts at CBO. Xiaotong Niu prepared the visual summary. Ricci Reber wrote Chapter 1 with contributions from James Otterson. Jaeger Nelson and Kerk Phillips wrote Chapter 2. Aaron Betz and Ricci Reber wrote Appendix A with contributions from Gloria Chen, Edward Gamber, and Robert Shackleton. Marina Miller wrote Appendix B with contributions from Jaeger Nelson, James Otterson, and Kerk Phillips. Jessica Banthin, Barry Blom, Lori Housman, Jamease Kowalczyk, Noah Meyerson, Eamon Molloy, Sam Papenfuss, Lisa Ramirez- Branum, Dan Ready, Sarah Sajewski, Emily Stern, Robert Stewart, and Rebecca Yip contributed to the analysis. Charles Pineles-Mark developed the long-term budget simulations with assistance from Nathaniel Milhous, Marina Miller, Xiaotong Niu, and Ricci Reber. Aaron Betz, Junghoon Lee, Jaeger Nelson, James Otterson, Kerk Phillips, and Robert Shackleton prepared the macroeconomic simulations with contributions from Mark Lasky. Edward Harris coordinated the revenue simulations, which were prepared by Kathleen Burke, Paul Burnham, Bayard Meiser, Shannon Mok, and Kurt Seibert. Jimmy Chin, Nathaniel Milhous, and Adam Staveski fact-checked the report. The report builds on the 10-year projections of the budget and economy that CBO released earlier this year, which reflected the contributions of more than 100 people at the agency. Wendy Edelberg, Mark Hadley, Jeffrey Kling, and Robert Sunshine reviewed the report. Christine Bogusz, Bo Peery, Benjamin Plotinsky, and Elizabeth Schwinn edited it, and Robert Rebach prepared it for publication. Nathaniel Milhous, Marina Miller, and Charles Pineles- Mark prepared the supplemental data. The report is available on CBO’s website (www.cbo.gov/ publication/55331). CBO continually seeks feedback to make its work as useful as possible. Please send any comments to communications@cbo.gov. Phillip L. Swagel Director June 2019