Top

Policy Options for Reducing the Federal Debt: Spring, 2024

Note: This document was updated on 5/21/2024 to include a conventional distributional analysis and a dynamic distributional analysis of the three policy bundles.
Summary: We analyze the budgetary and economic effects of three very different illustrative policy bundles that reduce federal budget deficits over time without shrinking the economy relative to current law with rising debt. The results also demonstrate how federal debt falls short of measuring the true fiscal burden being shifted to future generations.

Key Points

  • We consider three different bundles, with each bundle consisting of multiple fiscal policy options and its own distinct theme reflecting our discussions with experts across the policy spectrum: 1) raising taxes on corporations and high-income households; 2) broad-based changes to Social Security, Medicare and the employer deductibility of health care; and 3) a mixture of broad-based new tax revenues, including a carbon tax and a value-added tax, along with discretionary spending cuts.

  • As a constraint, all options were required to at least not reduce macroeconomic performance relative to current law, which has an increasing debt-GDP ratio. Over time, GDP and wages remain flat under Bundle 1 because the reduction in debt is offset with higher distortions. GDP climbs 9.8 percent by 2054 in Bundle 2 and by 5.9 percent with Bundle 3.

  • On a conventional basis, each policy bundle reduces deficits between $3.4 trillion (Bundle 2) and $6 trillion (Bundle 3) over the 10-year budget window. After 10 years, the bundles differ in their debt reduction, mainly due to differences in phase-in periods. None of the policy bundles fully stabilize the debt-GDP ratio over time, indicating that larger reforms are still needed.

  • Despite Bundle 3 reducing deficits the most over time, Bundle 2 produces more economic growth. The results also demonstrate how federal debt, as reported by the U.S. Treasury, falls short of measuring the true fiscal burden being shifted to future generations.

  • We present estimates of conventional distributional analysis using metrics commonly reported by scoring agencies. Conventional distributional analysis fails to account for the age of the taxpayer and various economic effects of a policy change. We show how dynamic distributional analysis corrects these shortcomings and often reverses the appearance of progressivity or regressivity that is indicated by conventional analysis.


Policy Options for Reducing the Federal Debt: Spring, 2024

Introduction

The Penn Wharton Budget Model projects that U.S. federal government debt held by the public will grow to 192 percent of the size of the economy (gross domestic product) by 2050 and to 217 percent by 2054. Our previous analysis has indicated that financial markets would likely demand large enough returns on federal debt to make continuation of current law unsustainable at around 190 percent of GDP.1 For the current analysis, we optimistically ignore this particular feedback channel to compare larger reform packages.2

We consider three different bundles, with each bundle consisting of multiple fiscal policy options and its own distinct theme: 1) raising taxes on corporations and high-income households; 2) broad-based changes to Social Security, Medicare, the employer deductibility of health care and non-defense discretionary spending; and 3) a mixture of broad-based new tax revenue and discretionary spending cuts. The policy bundles reflect our discussions across the political spectrum with leading economists, however, no single bundle represents the view of a single group or individual. Instead, we assembled broadly consistent views within each part of the political spectrum with the goal of deficit reduction without reducing economic activity relative to current policy with sharply rising debt.

We have also not tried to force an equal debt reduction across the bundles, either on a static or dynamic basis, at any specified time horizon. Doing so would essentially force the policy changes under Bundle 1 to expand enough that they would produce economic losses unless additional broad-based tax increases or spending cuts were introduced that impacted households without high income. In fact, none of the policy bundles---despite their apparent size---are sufficiently large enough to completely stabilize the debt-GDP ratio over time, which is quite informative of the scope of the problem.

Of course, policy bundles cannot be fully compared purely based on debt reduction and macroeconomic performance. We will update this document soon with more information about distributional effects.

Our Model

Our analysis of each policy bundle starts with our microsimulation model, which is estimated across a range of datasets to represent hundreds of thousands of different types of households, differentiated across 60 demographic and economic attributes. The PWBM microsimulation model interacts with various fiscal tax and spending modules. The PWBM tax module estimates individual income taxes, payroll taxes, corporate taxes, and estate taxes. It also simulates behavioral responses to changes in tax policy, calculating conventional estimates of the budgetary effects of tax policies and effective tax rates for different demographic groups. The various conventional elasticities applied include income classification, business formation classification, income timing and other material changes. The PWBM Social Security module estimates old age, dependent, survivor, and other auxiliary benefits as well as other rules. All major elements of benefits policy are parameterized, allowing for detailed analysis of reform proposals that highlights how structural shifts in demographic and economic forces affect Social Security’s finances.

This conventional modeling informs our core heterogeneous-agent overlapping-generations dynamic model (OLG). The OLG model allows for labor supply changes, capital intensity changes, and productivity changes in response to various tax (Bundle 1) and spending programs (Bundles 2 and 3). For example, a reduction in carbon from a carbon tax (Bundle 3) produces some productivity gains over time, which is consistent with data-driven estimates.

Our OLG model also fully merges the private and public healthcare sectors into the rest of the economy. As such, U.S. private healthcare (employer) and public healthcare (Medicare, Affordable Care Act, Medicaid, and other programs) are modeled in detail.3 When making labor supply and savings decisions, households simultaneously make decisions about whether to buy health insurance, whether to pay out of pocket for a healthcare shock, and even how to alter labor supply and savings to qualify for ACA premium support or Medicaid. Employers make competitive decisions about offering compensation and tax-preferred group healthcare in the presence of competition for talent in labor markets. Workers with low enough productivity--- where the cost of their health care coverage plus a minimum wage level of income is less than their marginal product---are not offered health coverage through their employer, consistent with current evidence. Healthcare premiums are computed in general equilibrium along with factor prices (wages and interest rates). Unlike factor prices, health prices are subject to various current-law rules regarding information restrictions and risk ratings, the net effect of which can lead to dynamic adverse selection over time and moral hazard.4 Investments in healthcare by the government or by an individual improve individual productivity (and wages) by health state, longevity, population size, and government revenue and spending, even in non-health programs. At the same time, non-health related changes in wages can impact elective decisions for healthcare and alter labor and savings decisions to qualify for Medicaid.

The analysis presented below, therefore, considers potentially complicated interactions of policies within each policy bundle. For Bundle 2, reducing the deductibility of employer healthcare increases other after-tax compensation through competitive labor markets, which is now taxable. But it also reduces private insurance coverage with more people electing ACA or Medicaid, at a cost to the government, or going without coverage, which can reduce their productivity. At the same time, Bundle 2 changes to Social Security and Medicare encourage more household saving and labor supply, that can materially impact non-payroll federal tax revenue and mitigate some of the reduction in health care coverage. A carbon tax (Bundle 3) can directly increase productivity over time while also raising immediate revenue.

In general, failing to account for interactions between fiscal policies can substantially understate or overstate the impact of a given policy change on total revenue, total costs, and the economy. Indeed, the distinct modeling of Social Security, Medicare, and the rest of federal government without capturing interactions can lead to first-order biases.

The Policy Bundles
Bundle 1: Increase Taxes on High Income Households and Corporations

This bundle includes the following revenue raisers and spending provisions:

  • Raise the top rate on ordinary income from 37 percent to 45 percent: Under current law, the top marginal income tax rate, which applies to any taxable income above $578,125 for single filers, is 37 percent and is scheduled to rise to 39.6 percent starting in 2026. This policy change would increase the top rate to 45 percent starting in 2025.

  • Introduce a third AMT income bracket, taxing at 45 percent income above $1 million: Under current law, high-income taxpayers must file taxes under the Alternative Minimum Tax (AMT) system in addition to the normal income tax system and pay taxes according to the greater of the two. This policy change would adjust the AMT so that it has a third bracket that applies a marginal rate of 45 percent on incomes above $1 million for married filers. This policy would take effect in 2025.

  • Tax capital gains and dividends at ordinary rates: Under current law, preferential tax rates are applied to capital gains held for a minimum period and qualifying dividends. Moreover, the basis of unrealized capital gains is stepped-up when assets are transferred at the time of death. This policy change would tax all income from capital gains and dividends using the ordinary income tax rate structure and would also subject capital gains without stepped-up basis to those rates at the time of the holder’s death, starting in 2025.

  • Double the Social Security taxable maximum earnings threshold: Under current law, only taxable earnings up to $168,600 are subject to OASDI payroll taxation. This policy change would raise that threshold to $337,200 starting in 2025.

  • Raise the corporate income tax rate from 21 percent to 28 percent: Under current law, corporations face a tax rate of 21 percent. This policy change would increase that rate to 28 percent, starting in 2025.

  • Implement paid family leave: Under current law, employees eligible for FMLA may take up to 12 weeks (about 3 months) of unpaid leave during a 12-month period for qualifying life events such as the care of a newborn or adopted child, to care for immediate family members, or to take medical leave. The policy change would provide up to 12 weeks of partial wage replacement, starting in 2025. PWBM assumes that over time the take-up rate of eligible leave takers increases from currently about 20 percent to 67 percent in 2060. This policy would take effect in 2025.

At first glance, including family leave as a new expenditure seems inconsistent with debt reduction. However, in our discussions with this part of the political spectrum, this expenditure seemed uniformly supported and part of a minimal viable package required for their support.

If enacted in 2025, Table 1 shows that Bundle 1 would generate net revenues (deficit reduction) of $3.7 trillion over the 10-year budget window and $16 trillion over the 2025-2054 period on a conventional basis, that is, before dynamic feedback effects.

Table 2 shows that the macroeconomic performance would differ very little over the next 30 years from current law, as GDP, the capital stock, hours worked, and average wages in 2054 are similar to the current-law baseline. However, debt-to-GDP declines from 217 percent under our baseline to 186 percent under this policy, although still growing and unsustainable. Intuitively, the economic effects are mainly a wash because the gains from debt reduction are mostly offset by the increase in economic distortions caused by the new tax rates.

Attempting further debt reduction with this approach would lead to economic decline relative to the baseline economy with growing debt. Distortions to economic activity increase by the square of the tax rates, which tends to favor broader tax bases over narrower ones for macroeconomic performance before considerations of distribution. Raising more revenue in Bundle 1 would generally require broader-based participation.

Table 1: Budget Effects of Bundle 1, Increase Taxes on High Income Households and Corporations

Conventional budgetary savings (+) or cost (-), billions of nominal dollars, excluding the effects of debt service

DOWNLOAD DATA
Policy 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2025-2034 2025-2054
Raise the top rate on ordinary income from 37 percent to 45 percent 69 83 74 78 80 83 86 90 94 98 834 3,885
Introduce a third AMT income bracket, taxing at 45 percent above $1 million 0 95 116 119 125 127 133 139 145 151 1,150 5,829
Tax capital gains and dividends at ordinary rates 37 59 44 46 43 42 44 45 47 48 455 1,799
Double the Social Security taxable maximum earnings threshold 77 107 109 114 118 122 126 131 134 138 1,176 5,052
Raise the corporate tax rate from 21 percent to 28 percent 68 98 103 106 107 108 111 114 116 119 1,050 4,322
Implement paid family leave -16 -26 -32 -36 -40 -44 -46 -50 -53 -55 -397 -2,224
Interaction effects 0 -70 -51 -53 -56 -58 -60 -63 -66 -69 -546 -2,683
Total 235 347 363 373 377 380 394 405 417 430 3,722 15,980

Table 2: Economic Effects of Bundle 1, Increase Taxes on High Income Households and Corporations

Percent Change from Baseline unless otherwise indicated

DOWNLOAD DATA
2030 2035 2040 2045 2050 2054
Gross domestic product -0.3 -0.4 -0.5 -0.5 -0.3 -0.1
Capital stock -0.3 -0.7 -0.9 -0.9 -0.7 -0.3
Hours worked 0.2 0.3 0.3 0.3 0.4 0.6
Average wage 0.2 0.0 -0.1 -0.1 -0.1 0.1
Debt-to-GDP ratio, baseline 111 127 145 167 192 217
Debt-to-GDP ratio, policy bundle 105 116 129 146 166 186
Bundle 2: Broad-based changes to Social Security, Medicare, the Healthcare Deduction, and Non-Military Discretionary Spending

This bundle includes the following provisions that mainly impact Social Security and Medicare:

  • Raise the full-benefit Social Security retirement age from 67 to 70 over time: Under current law, the Normal Retirement Age for Social Security, or the age at which beneficiaries can claim the full benefit amount that is calculated based on their career average monthly earnings, is 67 for individuals born in 1960 or later. This policy change would gradually continue to increase the normal retirement age (“NRA”) by two months per year until it reaches 70. The policy change would affect households who are 50 years old in 2025, who would face an NRA of 67 plus two months when they turn 62 in 2037. By delaying the time until the policy change takes full effect until 2037, this policy change creates a “flat region” for the increase in the NRA. The Appendix shows the impact of extending the NRA along its roughly current path that would impact individuals older than age 50 in 2025. A key insight shown in the Appendix is that the speed of adjustment has a first-order impact on potential macroeconomic gains.

  • Flattening Social Security benefits over time: Under current law, Social Security benefits are calculated by taking a beneficiary’s lifetime average monthly earnings and applying a set of fixed percentages to each segment of those earnings. This policy change would introduce a new minimum benefit in 2037 equal to the federal poverty line. The minimum benefit, therefore, increases benefits for some retirees, currently age 50 or younger in 2025, who have low income throughout their working ages. A new maximum benefit would also be phased in over time: a set of fixed percentages that apply to each segment of lifetime earnings would gradually decrease over a 20-year period, starting from the current-law values of 90 percent, 32 percent, and proceeding to 15 percent to 90 percent, 4 percent, and 0 percent, over 20 years. While the minimum benefit is fully implemented in 2037, only the first year of this 20-year period transition starts in 2037 so that it only applies to individuals 50 years of age and younger in 2025. Specifically, individuals who are 50 years old in 2025 can claim early benefits at age 62 in exchange for a benefit reduction or they can claim full benefits at the NRA. This individual would qualify for the full minimum benefit at the NRA (or a discounted value as early as age 62) but the maximum benefit would be reduced by around 1/20th of the full reduction in the maximum benefit noted above. The new maximum benefit would be fully phased in by 2056, that is, for individuals aged 30 or younger in 2025. When combined with the minimum benefit, this policy, therefore, mostly flattens Social Security benefits independently of earnings history. In essence, for the cohort born in 1994 or later, Social Security would be fully transformed from a pension into a retiree poverty-relief program, thereby requiring additional consumption to be financed using private savings. The Appendix considers the effect of applying this 20-year benefit adjustment at a faster rate.

  • Raise the Medicare eligibility age from 65 to 67: Under current law, most people qualify for Medicare coverage upon reaching age 65 and some individuals become eligible earlier if they have certain medical conditions. This policy change would increase the eligibility age by two months every calendar year to a maximum age of 67, while leaving unchanged the separate qualifying parameters for individuals with certain disabilities. This policy starts in 2025, with an eligibility age of 65 plus two months, so that people retiring in 2036 would be the first cohort to face the eligibility age of 67.

  • Convert Medicare to premium support: This policy would convert Medicare to a premium support system that would allow beneficiaries to choose from competing insurance plans while the federal government would share the cost of premiums. Previous proposals have differed in many respects, particularly how the federal contribution would be set. The policy change we model would allow private health insurance companies to submit bids to provide Medicare Part B services to beneficiaries.5 The average bid amount, including that from Medicare fee-for-service, would become the benchmark. The federal government would pay insurers the benchmark rate for a region minus a standard premium paid by enrollees. Enrollees that choose to enroll into an insurance plan that places a bid at the benchmark rate would pay a standard insurance premium directly to the insurer. That premium would be the same everywhere and would cover about 25 percent of total insured expenses under Medicare Part B. Beneficiaries that enroll into a plan with an above (below) benchmark bid would pay the standard premium and (less) the difference between the bid and the benchmark. This policy would take effect in 2025.

  • Limit the tax exclusion of employer-sponsored health insurance: Under current law, employees who enroll in employer-sponsored health insurance plans may exclude the full amount of their health insurance premiums from their taxable income. This policy would limit the amount of employer-provided health insurance premiums that can be excluded to a value equal to the median health insurance premium. This policy would take effect in 2025.6

  • Cut annual non-defense discretionary spending by 5 percent: Under current law, about 70 percent of the federal budget is for mandatory programs, for which spending is set by predetermined formulas, and interest payments on the national debt. About 30 percent of the budget is for “discretionary” programs, which are the programs for which spending is controlled by annual appropriations acts. About half of discretionary spending is defense spending, while the other half is comprised of education, transportation, housing, and more. This policy change would reduce the non-defense portion of discretionary spending by 5 percent each year relative to currently projected levels. This policy would take effect in 2025.

Table 3 shows that Bundle 2 would decrease net outlays by $3.4 trillion over the 10-year budget window and $25 trillion over the 2025-2054 period on a conventional basis. Notice that while the debt reduction during the first decade is a bit smaller than for Bundle 1, the debt reduction in later years is larger. The main reason is that the policies in Bundle 1 start in 2025. In contrast, the start of several of the larger spending reductions in Social Security and Medicare in Bundle 2 are delayed by over a decade---to give households time to respond with additional saving---and, even then are phased in over the subsequent 20 years.

As Table 4 shows, GDP increases by 9.8 percent by 2054. The debt-to-GDP ratio decreases significantly over the next 30 years, from 217 percent under the current-law baseline to 152 percent under Bundle 2 by 2054.

Most of the increase in GDP is driven by a larger capital stock, which increases by 19.8 percent by 2054. Households must replace some Social Security income and Medicare benefits with additional private saving. This “crowd in” of capital is especially effective for households facing fewer borrowing constraints, including higher-income households that are relatively more impacted by these changes due to progressive tilting of benefits in Social Security.

Labor supply (“Hours Worked”) also increases over time but by less than capital. Labor supply is impacted by several competing factors. First, labor supply increases to replace some lost retirement income. Second, labor supply responds to the larger wage rate, which increases by 7.5 percent over time. In equilibrium, wages would have increased a bit more without the increase in labor supply itself. Third, the move to a flat benefit completely disconnects payroll taxes paid from benefits received at the margin. Under current law, Social Security’s effective tax rate---which determines the labor supply distortion---is less than the statutory rate due to the payroll wage (and, hence, tax) and benefit linkage. That linkage gets removed with a flat tax, which, therefore, increases the effective tax rate to be equal to the statutory rate.

Table 3: Budget Effects of Bundle 2, Broad-based changes to Social Security, Medicare, the Healthcare Deduction, and Non-Military Discretionary Spending

Conventional budgetary savings (+) or cost (-), billions of nominal dollars, excluding the effects of debt service

DOWNLOAD DATA
Policy 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2025-2034 2025-2054
Raise the full-benefit Social Security retirement age from 67 to 70 0 0 0 0 0 0 0 0 0 0 0 1,610
Flat monthly Social Security benefit set at elderly poverty threshold and indexed to wage growth 0 0 0 0 0 0 0 0 0 0 0 2,532
Raise the Medicare eligibility age from 65 to 67 0 0 50 52 55 57 60 63 66 69 472 5,014
Convert Medicare to premium support 80 83 87 91 95 99 104 108 114 119 981 4,811
Limit exclusion of employer-sponsored health insurance 100 116 126 136 147 159 172 185 199 213 1,552 9,669
Cut annual non-defense discretionary spending by 5 percent 0 47 49 51 53 55 57 59 61 63 495 2,242
Interaction effects -1 -1 -6 -6 -7 -6 -7 -7 -9 -9 -59 -917
Total 179 245 306 324 344 364 385 408 431 455 3,441 24,961

Table 4: Economic Effects of Bundle 2, Broad-based changes to Social Security, Medicare, the Healthcare Deduction, and Non-Military Discretionary Spending

Percent Change from Baseline unless otherwise indicated

DOWNLOAD DATA
2030 2035 2040 2045 2050 2054
Gross domestic product 0.8 1.8 3.1 4.9 7.3 9.8
Capital stock 1.5 3.3 5.8 9.4 14.5 19.8
Hours worked -0.7 -0.5 -0.2 0.2 0.7 1.4
Average wage 0.6 1.2 2.2 3.6 5.5 7.5
Debt-to-GDP ratio, baseline 111 127 145 167 192 217
Debt-to-GDP ratio, policy bundle 106 114 123 134 144 152

The Appendix presents alternative versions of Bundle 2 with faster implementation dates for the retirement policies. The results demonstrate the first-order importance of the timing of the policies; faster implementation leads to significantly larger macroeconomic gains over time.

Bundle 3: A Mixture of Broad-based Tax Increases and Spending Cuts

This bundle includes the following broad-based tax increases and spending cuts:

  • Enact a 1 percent Value-Added Tax (VAT): Under current law, while it is very common for state governments to levy sales taxes, the federal government does not impose a broad-based consumption tax. This policy change would implement a value-added tax (VAT), or a tax on the difference between revenue and the cost of intermediate inputs, of 1 percent on a broadly defined base of consumption, with exemptions for certain categories such as health expenditures. This policy would take effect in 2025.

  • Disallow all itemized deductions except for charitable deductions: Under current law, when individuals calculate their income tax liability, they can take either the standard deduction or a set of itemized deductions, which includes deductions for mortgage interest, state and local taxes, and charitable contributions, among others. Starting in 2025, this policy change would disallow all itemized deductions, and all taxpayers would claim the standard deduction.

  • Expand the base of employment taxes to cover all pass-through income: Under current law, certain business earnings of limited partners and S corporation shareholders are not subject to self-employment payroll taxes. This policy would expand the self-employment payroll tax base to include all pass-through income, regardless of organizational form, and start in 2025.

  • Implement a carbon tax: Under current law, the U.S. does not have a carbon tax. This policy change would levy a tax on coal, oil products, and natural gas in proportion to their carbon content, that would be collected from fuel suppliers. This policy would enact a tax of $50 per ton of carbon produced and would reduce greenhouse gases by 7 percent in the first year (2025) and about 16 percent in 2054. This policy goes into effect in 2025.

  • Slow Social Security expenditure growth by indexing benefits to chained CPI: Under current law, Social Security benefits are adjusted for inflation each year using the Consumer Price Index (CPI). This policy change would instead use a different version of CPI – chained CPI, which is the measure used to adjust most parameters of the tax system – to adjust benefit payments each year for inflation. The chained CPI adjustments would be about 0.2 percentage points lower per year than using the CPI. This policy goes into effect in 2025.

  • Increase Medicare premiums: Under current law, premiums for Medicare parts B and D are income dependent. Beneficiaries with less than or equal to $103,000 of modified adjusted income pay $174.70 per month while households between $106,000 and 129,000 pay $244.60 per month. Households with higher MAGI pay higher monthly premiums. This policy change would increase Medicare premiums for the bottom income group to the level of income premiums in the second group (i.e. from 25 percent of Medicare cost to 35 percent). Under this policy, Medicare premiums would increase by an average of 27 percent. This policy would start in 2025.

If enacted in 2025, Table 5 shows that Bundle 3 would generate net revenues of $6.0 trillion over the 10-year budget window and $29 trillion over the 2025-2054 period on a conventional basis. As Table 6 shows, the capital stock and hours worked would be 12 percent and 1.5 percent higher, respectively, by 2054, which would increase GDP by roughly 6 percent. Debt-to-GDP would decrease from 217 percent under our current-law baseline to 146 percent under this policy.

Table 5: Budget Effects of Bundle 3, A Mixture of Broad-based Tax Increases and Spending Cuts

Conventional budgetary savings (+) or cost (-), billions of nominal dollars, excluding the effects of debt service

DOWNLOAD DATA
Policy 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2025-2034 2025-2054
Enact a 1 percent Value-Added Tax (VAT) 68 94 97 100 104 108 112 115 119 123 1,039 4,468
Disallow all itemized deductions except charitable deduction 18 128 255 264 276 288 301 314 327 339 2,508 12,779
Expand the base of employment taxes to cover all pass-through income 49 66 68 71 73 76 79 82 85 88 738 3,305
Implement a carbon tax 85 80 77 75 72 75 80 83 86 89 803 3,482
Slow Social Security benefit growth by indexing benefits to chained CPI 3 5 8 12 15 19 22 26 30 33 173 1,636
Increase Medicare premiums 68 70 72 75 77 80 82 85 88 91 788 3,391
Interaction effects 0 0 0 0 0 0 0 0 0 0 -1 -5
Total 290 443 577 596 617 645 675 705 735 763 6,047 29,056

Table 6: Economic Effects of Bundle 3, A Mixture of Broad-based Tax Increases and Spending Cuts

Percent Change from Baseline unless otherwise indicated

DOWNLOAD DATA
2030 2035 2040 2045 2050 2054
Gross domestic product 1.1 1.9 2.7 3.6 4.7 5.9
Capital stock 2.2 3.7 5.4 7.3 9.7 12.2
Hours worked 0.3 0.4 0.6 0.8 1.1 1.5
Average wage 0.9 1.5 2.2 3.0 3.9 4.9
Debt-to-GDP ratio, baseline 111 127 145 167 192 217
Debt-to-GDP ratio, policy bundle 101 107 114 123 134 146

Federal debt is reduced by substantially more relative to Bundle 2. Moreover, Bundle 3 benefits from some productivity gains associated with lower carbon emissions. Finally, all policies in Bundle 3 commence in 2025 and without gradual phase-in upon commencement.

However, Bundle 3 improves economic performance by less than Bundle 2. The reason is that federal debt fails to properly capture transfers across generations. For example, the entire current federal debt held by the public, equal to around $27 trillion at start of 2024, is less than the $62.7 trillion the PWBM calculates in present value benefits paid (or projected to paid) by the Social Security and Medicare programs alone to past and current generations in excess of the present value of payroll taxes collected from them. Ultimately, it is the totality of inter-generational transfers---not the fraction that happens to be ran through the U.S. Treasury auctions---that matters for economic performance. Bundle 2, therefore, reduces inter-generational transfers by more than indicated by federal debt alone.

Distributional Analysis: Conventional Estimates

Conventional distributional analysis reports how household post-fisc income changes due to changes in taxes less transfers or benefits received. This information is represented as a cross section of the population for a given year. Conventional analysis often focuses on tax provisions alone, maybe net of various provisions like the EITC. However, we generally include the effects of tax and spending provisions---including Medicaid, Social Security benefits, and food stamps---except spending that is widely shared---such as roads and military---that would be too subjective in nature to further differentiate by income group.

Table 7 reports the conventional analysis by policy bundle type---Bundle 1, Bundle 2, and Bundle 3---and across three different policy years 2025, 2034 and 2054. The reader can use the pulldown options in Table 7 to select the between the policy bundle and year, for a total of nine different combinations. Different years are provided because the policy bundles phase-in over time and at different speeds. For example, most of the policies in Bundle 1 start in 2025 while many of the policies in Bundle 2 phase in over the next 20 years. With conventional distributional analysis, the results do not reflect changes in macroeconomic variables.

Notice that Bundle 1 (Table 7a) appears to be the most progressive change of the three bundles, with top incomes seeing their post-fisc income fall by over $1.8 million per year in 2025 while lowest incomes see their post-fisc income increase by $245. This difference grows even more across time, as the policies in Bundle 1 are fully phased in.

Bundle 2 (Table 7b) shows a very different pattern, with all income groups losing money on a post-fisc basis. The top 5 percent income groups appear to lose less money than some lower income groups. This seemingly odd behavior is driven by the shortcomings of conventional distributional analysis itself which does not distinguish by age. Relative to Bundle 1, Bundle 2 relies on both changes to working age post-fisc income---mostly the limit on the exclusion of employer-sponsored health insurance from wage income---along with changes to retirement benefits. However, households with incomes in the top 5 percent of the income distribution typically have sources of wage income whereas older households typically rely more on capital income (derived from past savings), Social Security benefits and other sources of non-wage income. As such, retirees tend to appear below the top 5 percent in the income distribution even for retirees previously inside the top 5 percent during their working ages. Hence, conventional distributional analysis does not classify these households based on lifetime (or “lifecycle”) incomes.

Bundle 3 (Table 7c) also reports a decline in post-fisc income for all income groups, although more dollars are lost at higher income. However, as with Bundle 2, the percent reduction in post-fisc income is smaller at the highest 0.1 percent income level relative to top 1 percent, again reflecting the failure of conventional analysis to capture lifecycle effects.

Table 7: Distributional Effects

(a) Bundle 1, Increase Taxes on High Income Households and Corporations

Income group 2025 2034 2054
Average income change, after taxes and transfers Percent change in income, after taxes and transfers Average income change, after taxes and transfers Percent change in income, after taxes and transfers Average income change, after taxes and transfers Percent change in income, after taxes and transfers
First quintile 245 4.8% 790 11.1% 1,705 15.2%
Second quintile 415 0.9% 1,340 2.3% 2,910 2.8%
Middle quintile 495 0.6% 1,785 1.8% 3,935 2.3%
Fourth quintile 395 0.3% 2,050 1.2% 4,620 1.6%
80-90% -470 -0.2% 1,180 0.5% 3,010 0.7%
90-95% -5,430 -1.7% -5,195 -1.2% -8,550 -1.1%
95-99% -19,545 -3.3% -23,720 -3.0% -45,430 -3.2%
99-99.9% -117,860 -6.1% -154,075 -6.3% -302,605 -7.1%
Top 0.1% -1,845,915 -12.0% -2,489,035 -12.5% -4,509,630 -12.8%

(b) Bundle 2, Broad-based changes to Social Security, Medicare, the Healthcare Deduction, and Non-Military Discretionary Spending

Income group 2025 2034 2054
Average income change, after taxes and transfers Percent change in income, after taxes and transfers Average income change, after taxes and transfers Percent change in income, after taxes and transfers Average income change, after taxes and transfers Percent change in income, after taxes and transfers
First quintile -60 -0.3% -390 -13.3% -1,370 -14.3%
Second quintile -335 -0.7% -900 -1.5% -4,190 -4.0%
Middle quintile -565 -0.7% -1,410 -1.4% -7,390 -4.3%
Fourth quintile -730 -0.6% -1,815 -1.1% -7,570 -2.7%
80-90% -810 -0.4% -2,040 -0.8% -6,890 -1.5%
90-95% -835 -0.3% -2,090 -0.5% -6,645 -1.0%
95-99% -740 -0.1% -1,945 -0.3% -6,075 -0.5%
99-99.9% -710 0.0% -1,900 -0.1% -6,120 -0.2%
Top 0.1% -640 0.0% -1,780 0.0% -5,530 0.0%

(c) Bundle 3, A Mixture of Broad-based Tax Increases and Spending Cuts

Income group 2025 2034 2054
Average income change, after taxes and transfers Percent change in income, after taxes and transfers Average income change, after taxes and transfers Percent change in income, after taxes and transfers Average income change, after taxes and transfers Percent change in income, after taxes and transfers
First quintile -35 -0.2% -50 -0.2% -260 -0.6%
Second quintile -125 -0.3% -320 -0.5% -1,065 -1.0%
Middle quintile -310 -0.4% -1,015 -1.0% -2,675 -1.5%
Fourth quintile -670 -0.5% -2,800 -1.6% -6,475 -2.2%
80-90% -1555 -0.7% -6,145 -2.3% -12,670 -2.7%
90-95% -3,125 -1.0% -10,095 -2.5% -19,135 -2.7%
95-99% -7,620 -1.3% -17,835 -2.3% -38,740 -2.9%
99-99.9% -26,405 -1.5% -74,900 -3.3% -145,345 -3.7%
Top 0.1% -152,640 -1.4% -502,490 -3.3% -966,405 -3.5%
Distributional Analysis: Dynamic Estimates

Dynamic distributional analysis considers households across the income and age distribution, including the unborn represented by a negative age index at the time of the reform. It asks how much each household in the (income, age) combination values the proposed policy change over their entire lifetime, represented as a one-time transfer at the time of the policy change. Dynamic distributional analysis is the standard in academic research, where conventional analysis is rarely used due to several key limitations that dynamic analysis addresses.

A positive equivalent variation means that the person would be better off under the policy reform; a negative equivalent variation means that the person would be worse off under the policy reform. For example, as shown in Figure 8, for Bundle 1, a person of age 40 at the time of the policy change and with gross income in the lowest 20th percent prior of the income distribution gains $27,100 of value from Bundle 1. Put differently, this household is indifferent between the adoption of Bundle 1 and receiving a one-time payment of $27,100 without Bundle 1. However, a household age 30 in the highest 20th percentile loses the equivalent of $26,100, as shown by the negative value. This household would be exactly indifferent between Bundle 1 and a one-time payment of $26,100 to avoid permanently adopting Bundle 1.

Notice that all Bundles tend to leave future generations (those unborn with negative index values). However, Bundle 2 produces the largest gains without almost any losses, except to older households in the highest 20th percentile who lose mostly due to macroeconomic capital deepening that reduces interest income earned on their assets.

The policies in Bundle 2 are slowly phased in over time. However, these changes are anticipated in advance by households who, unless they are borrowing constrained, respond with more savings and labor. These increases also reduce the “fiscal externality” caused by the existing tax system, as households are incented to undo some of their “freeriding” (reductions in labor and capital saving) to provide a given amount of government revenue. This reduction in freeriding can have economywide positive fiscal externalities.

Table 8: Equivalent Variations of Bundles

Willingness to pay for the policy change

(a) Bundle 1, Increase Taxes on High Income Households and Corporations

(b) Bundle 2, Broad-based changes to Social Security, Medicare, the Healthcare Deduction, and Non-Military Discretionary Spending

(c) Bundle 3, A Mixture of Broad-based Tax Increases and Spending Cuts


Appendix

In this appendix, we present alternative results for Bundle 2 by changing the age of households that are affected by the two retirement policies in 2025:

  • Raise the full-benefit Social Security retirement age from 67 to 70
  • Flat monthly Social Security benefits

In the main text, both options apply starting with individuals who are 50 years old in 2025. That gives them 12 years---until they turn 62 in 2037---to prepare for those policy changes to take effect. Here we present the budget and economic effects for three alternative assumptions where either one or both policies start for those who are 62 in 2025, thereby expediting the effect of the policy. Figure A1 shows how the full-benefit Social Security retirement ages change under the different scenarios based on beneficiaries’ age in 2025.

Figure A1: Change in Full Retirement Age

DOWNLOAD DATA
Appendix Option 1: Bundle 2, but Social Security retirement age increases for 62-year-olds in 2025.

Table A1 shows that if the retirement age increases for 62-year-olds in 2025, that policy change decreases net outlays by $3.6 trillion over the 10-year budget windows and by $28.9 trillion over the next 30 years. Compared to Bundle 2, this change in retirement eligibility lowers deficits by $134 billion over 10 years and $3.9 trillion over the next thirty years.

As Table A2 shows, the faster and bigger cuts to entitlements increase the capital stock by 23.2 percent, hours worked by 2 percent, and GDP by 11.4 percent. The debt-to-GDP ratio in 2054 is 151 percent, which is close to the 151 percent under Bundle 2. Compared to Bundle 2, the capital stock is 3.4 percentage points larger, and GDP is 1.7 percentage points larger.

Table A1: Budget Effects of Bundle 2, but Social Security retirement age increases for 62-year-olds in 2025

Conventional budgetary savings (+) or cost (-), billions of nominal dollars, excluding the effects of debt service

DOWNLOAD DATA
Policy 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2025-2034 2025-2054
Raise the full-benefit Social Security retirement age from 67 to 70 0 0 1 1 2 7 14 25 37 47 134 5,523
Flat monthly Social Security benefit set at elderly poverty threshold and indexed to wage growth 0 0 0 0 0 0 0 0 0 0 0 2,532
Raise the Medicare eligibility age from 65 to 67 0 0 50 52 55 57 60 63 66 69 472 5,014
Convert Medicare to premium support 80 83 87 91 95 99 104 108 114 119 981 4,811
Limit exclusion of employer-sponsored health insurance 100 116 126 136 147 159 172 185 199 213 1,552 9,669
Cut annual non-defense discretionary spending by 5 percent 0 47 49 51 53 55 57 59 61 63 495 2,242
Interaction effects -1 -1 -6 -6 -7 -6 -7 -7 -9 -9 -59 -917
Total 179 246 307 325 345 371 399 433 468 502 3,575 28,874

Table A2: Economic Effects of Bundle 2, but Social Security retirement age increases for 62-year-olds in 2025

Percent Change from Baseline unless otherwise indicated

DOWNLOAD DATA
2030 2035 2040 2045 2050 2054
Gross domestic product 1.5 2.9 4.5 6.5 8.9 11.4
Capital stock 2.5 5.1 8.4 12.6 17.8 23.2
Hours worked -0.2 0.2 0.5 0.9 1.4 2.0
Average wage 0.8 1.8 3.0 4.6 6.6 8.6
Debt-to-GDP ratio, baseline 111 127 145 167 192 217
Debt-to-GDP ratio, policy bundle 105 114 123 133 143 151
Appendix Option 2: Bundle 2, but the 20-year transition to a flat monthly benefit starts with 62-year-olds in 2025 rather than 50-year-olds.

If enacted in 2025, Table A3 show that this option would decrease deficits by $3.6 trillion over the 10-year budget window and by $34 trillion over the next 30 years. Compared to Bundle 2, the flat monthly benefits now decrease outlays by $167 billion over the 10-year budget window and by $9 trillion over the next 30 years.

Like Appendix Option 1, this option cuts entitlement spending faster and bigger compared to Bundle 2. As shown in Table A4, the result is that households save even more, increasing capital by 29 percent and hours worked by 2.2 percent in 2054. Combining the higher capital with an increase in labor generates GDP that is 29 percent larger than in the baseline economy, and 4.3 percentage points larger than under Bundle 2. The debt-to-GDP ratio under the policy is 141 percent in 2054, compared to the 217 percent under the current-law baseline.

Table A3: Budget Effects of Bundle 2, but the 20-year transition to a flat monthly benefit starts with 62-year-olds in 2025 rather than 50-year-olds

Conventional budgetary savings (+) or cost (-), billions of nominal dollars, excluding the effects of debt service

DOWNLOAD DATA
Policy 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2025-2034 2025-2054
Raise the full-benefit Social Security retirement age from 67 to 70 0 0 0 0 0 0 0 0 0 0 0 1,610
Flat monthly Social Security benefit set at elderly poverty threshold and indexed to wage growth 0 0 1 3 6 11 18 28 41 57 167 11,594
Raise the Medicare eligibility age from 65 to 67 0 0 50 52 55 57 60 63 66 69 472 5,014
Convert Medicare to premium support 80 83 87 91 95 99 104 108 114 119 981 4,811
Limit exclusion of employer-sponsored health insurance 100 116 126 136 147 159 172 185 199 213 1,552 9,669
Cut annual non-defense discretionary spending by 5 percent 0 47 49 51 53 55 57 59 61 63 495 2,242
Interaction effects -1 -1 -6 -6 -7 -6 -7 -7 -9 -9 -59 -917
Total 179 246 307 327 350 375 404 436 472 512 3,608 34,024

Table A4: Economic Effects of Bundle 2, but the 20-year transition to a flat monthly benefit starts with 62-year-olds in 2025 rather than 50-year-olds

Percent Change from Baseline unless otherwise indicated

DOWNLOAD DATA
2030 2035 2040 2045 2050 2054
Gross domestic product 1.7 3.4 5.5 8.1 11.1 14.1
Capital stock 3.0 6.4 10.6 16.0 22.6 29.1
Hours worked -0.9 -0.4 0.1 0.7 1.4 2.2
Average wage 1.1 2.4 3.9 5.9 8.4 10.7
Debt-to-GDP ratio, baseline 111 127 145 167 192 217
Debt-to-GDP ratio, policy bundle 105 113 121 130 136 141
Appendix Option 3: Bundle 2, but Social Security retirement age increases and flat monthly benefits start with 62-year-olds in 2025.

As shown in Table A5, this policy combines Appendix Options 1 and 2 and decreases deficit by $3.7 trillion over the 10-year-budget window and by $38 trillion over the next 30 years. Compared to Bundle 2, this policy decreases deficits by $301 billion over the 10-year budget window and by $13 trillion over the next 30 years. That is a 50 percent increase in deficit reduction compared to Bundle 2.

As a result, Table A6 shows that capital increases by 34 percent relative to the current-law baseline, working hours increase by 2.7 percent and GDP increases by 16 percent. The debt-to-GDP ratio under the policy is 129 percent in 2054, compared to the 217 percent under the current-law baseline. Compared to Bundle 2, this is a 22-percentage point decrease in the debt-to-GDP ratio.

Table A5: Budget Effects of Bundle 2, but Social Security retirement age increases and flat monthly benefits start with 62-year-olds in 2025

Conventional budgetary savings (+) or cost (-), billions of nominal dollars, excluding the effects of debt service

DOWNLOAD DATA
Policy 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2025-2034 2025-2054
Raise the full-benefit Social Security retirement age from 67 to 70 0 0 1 1 2 7 14 25 37 47 134 5,523
Flat monthly Social Security benefit set at elderly poverty threshold and indexed to wage growth 0 0 1 3 6 11 18 28 41 57 167 11,594
Raise the Medicare eligibility age from 65 to 67 0 0 50 52 55 57 60 63 66 69 472 5,014
Convert Medicare to premium support 80 83 87 91 95 99 104 108 114 119 981 4,811
Limit exclusion of employer-sponsored health insurance 100 116 126 136 147 159 172 185 199 213 1,552 9,669
Cut annual non-defense discretionary spending by 5 percent 0 47 49 51 53 55 57 59 61 63 495 2,242
Interaction effects -1 -1 -6 -6 -7 -6 -7 -7 -9 -9 -59 -917
Total 179 246 309 328 352 382 417 461 509 559 3,742 37,936

Table A6: Economic Effects of Bundle 2, but Social Security retirement age increases and flat monthly benefits start with 62-year-olds in 2025

Percent Change from Baseline unless otherwise indicated

DOWNLOAD DATA
2030 2035 2040 2045 2050 2054
Gross domestic product 2.3 4.4 6.9 9.8 13.2 16.4
Capital stock 3.9 8.1 13.3 19.6 27.1 34.3
Hours worked -0.5 0.0 0.5 1.1 1.9 2.7
Average wage 1.3 2.9 4.9 7.2 9.9 12.4
Debt-to-GDP ratio, baseline 111 127 145 167 192 217
Debt-to-GDP ratio, policy bundle 104 111 118 124 127 129



This analysis was produced by Kody Carmody, Jon Huntley, Ed Murphy, Brendan Novak, and Felix Reichling under the direction of Kent Smetters. Alexander Arnon, Sophie Shin, and Aidan O'Connell contributed to the analysis. Mariko Paulson prepared the brief for the website.


  1. That analysis from October 2023 also projected a debt-to-GDP ratio of 188 percent by 2050, which we have recently updated to 192 percent.  ↩

  2. For more technical readers, the October 2023 analysis reflected inputs from our dynamic “moonshot OLG” model that allows for aggregate uncertainty, thereby allowing the equity premium and risk-free rate to be endogenous, as determined by equilibrium relationships. The resulting state space is about 20,000 times larger than our “core” OLG model. Our “core” OLG model allows, like most OLG models, allows for uninsurable idiosyncratic risk but with deterministic future prices determined by equilibrium conditions. As a result, the “core” model can accommodate a very rich policy space but requires deciding whether (i) the time-varying government’s borrowing rate (risk-free rate) is endogenous along with a time-varying but fixed equity premium that is not influenced by policy or (ii) whether the government’s time-varying borrowing rate is fixed along with an endogenous time-varying endogenous equity premium that is influenced by policy. Choice (i) will generally lead to explosive and unsustainable debt paths much earlier than choice (ii). We use choice (ii) in the “core” OLG model, which means that debt only impacts the equity premium with no impact on the government’s borrowing rate. We instead forecast the government’s borrowing rate using a duration issuance model and market data by duration. Our first decade is very similar to CBO.  ↩

  3. Model documentation is being prepared and will be released over time.  ↩

  4. For example, group policies offered by employers cannot individually risk rate. Moreover, the ACA only allows for some pricing difference by age but at a rate increase that is below actuarially fair.  ↩

  5. CBO’s 2017 analysis of such a policy found that it would decrease Medicare outlays by 8 percent and reduce average premiums by 8 percent.  ↩

  6. This policy was described in greater detail in a 2022 CBO report.  ↩

,Current Law,62 year old in 2025,50 year old in 2025
100,65,65,65
99,65,65,65
98,65,65,65
97,65,65,65
96,65,65,65
95,65,65,65
94,65,65,65
93,65,65,65
92,65,65,65
91,65,65,65
90,65,65,65
89,65,65,65
88,65,65,65
87,65.16666667,65.16666667,65.16666667
86,65.33333333,65.33333333,65.33333333
85,65.5,65.5,65.5
84,65.66666667,65.66666667,65.66666667
83,65.83333333,65.83333333,65.83333333
82,66,66,66
81,66,66,66
80,66,66,66
79,66,66,66
78,66,66,66
77,66,66,66
76,66,66,66
75,66,66,66
74,66,66,66
73,66,66,66
72,66,66,66
71,66,66,66
70,66.16666667,66.16666667,66.16666667
69,66.33333333,66.33333333,66.33333333
68,66.5,66.5,66.5
67,66.66666667,66.66666667,66.66666667
66,66.83333333,66.83333333,66.83333333
65,67,67,67
64,67,67,67
63,67,67,67
62,67,67.16666667,67
61,67,67.33333333,67
60,67,67.5,67
59,67,67.66666667,67
58,67,67.83333333,67
57,67,68,67
56,67,68.16666667,67
55,67,68.33333333,67
54,67,68.5,67
53,67,68.66666667,67
52,67,68.83333333,67
51,67,69,67
50,67,69.16666667,67.16666667
49,67,69.33333333,67.33333333
48,67,69.5,67.5
47,67,69.66666667,67.66666667
46,67,69.83333333,67.83333333
45,67,70,68
44,67,70,68.16666667
43,67,70,68.33333333
42,67,70,68.5
41,67,70,68.66666667
40,67,70,68.83333333
39,67,70,69
38,67,70,69.16666667
37,67,70,69.33333333
36,67,70,69.5
35,67,70,69.66666667
34,67,70,69.83333333
33,67,70,70
32,67,70,70
31,67,70,70
30,67,70,70
29,67,70,70
28,67,70,70
27,67,70,70
26,67,70,70
25,67,70,70
24,67,70,70
23,67,70,70
22,67,70,70
21,67,70,70
Age	Lowest 20th	20 to 40	40 to 60	60 to 80	Highest 20th
-20	47000	62300	80800	95600	76400
-10	35700	39800	50400	47300	2400
0	24700	23400	28800	14100	-4800
10	16500	12500	14700	-1300	-8000
20	12400	8000	8900	-3100	-9300
30	19000	17800	21800	29000	-26100
40	27100	28500	34900	46000	42300
50	17400	29000	37200	66400	81800
60	6800	8700	16100	15900	-17500
70	-1300	-5500	-13200	-25800	-91400
80	0	-3600	-7200	-16900	-59300
Age	Lowest 20th	20 to 40	40 to 60	60 to 80	Highest 20th
-20	144300	262700	336400	488000	704300
-10	124700	210300	273200	376200	528600
0	95300	155800	206500	281400	379000
10	65900	111300	148300	211300	286900
20	42600	76900	111300	163600	235900
30	59800	67200	83800	138400	277000
40	81800	94600	129200	177800	402500
50	85600	152500	197000	321800	616800
60	70500	104600	166800	207300	350500
70	6100	9600	8700	4300	-29600
80	6200	9700	10100	6900	-8300  
Age	Lowest 20th	20 to 40	40 to 60	60 to 80	Highest 20th
-20	79500	132500	168600	219500	266000
-10	61000	90000	112700	132700	133800
0	37800	49900	61200	58800	24600
10	19100	17400	19500	3800	-4600
20	8300	-1400	-3100	-7200	-9000
30	-900	-3500	-6800	-19000	-68900
40	-9600	-15400	-12800	-15200	-41200
50	-15300	-26600	-27000	-37800	-49800
60	-40400	-78700	-125100	-151100	-222100
70	-12000	-107400	-183000	-311500	-352700
80	-5500	-85700	-121700	-185300	-188500