Key Findings

• We construct a counterfactual 2004 natural baseline, using rules and budget classification conventions in place in 2004, which predicts $2.50 trillion in mandatory spending by 2025, after adjusting for demographic, economic and medical and non-medical inflation. In contrast, actual mandatory spending was $4.17 trillion, a policy-driven excess of $1.66 trillion, or 40 percent of all mandatory outlays in 2025.

• Even excluding the COVID peak years (FY2020–2021, when the excess reached $2.5–2.7 trillion), the policy-driven excess has been rising: from $104 billion (7.4 percent of mandatory) in FY2006, to $763 billion (27.9 percent) in FY2019, to $1.66 trillion (40 percent) in FY2025.

• This shift from discretionary to mandatory classification reduces Congress’s annual oversight over federal spending. Discretionary spending has been compressed from an estimated 50.5 percent of total outlays under the 2004 natural baseline to 26.7 percent in FY2025. This $1.66 trillion displacement in discretionary capacity is larger than the entire military budget.

Introduction

The OBBBA’s DHS and DoD mandatory appropriations — and the pending Reconciliation 2.0 extension for ICE and CBP — represent the latest step in a shift toward mandatory classification that bypasses annual appropriations oversight and creates multi-year commitments harder to reverse.

But the trend toward greater mandatory spending has been ongoing for two decades. Some increase was inevitable: the share of the population aged 65 and over grew by 44 percent between FY2004 and FY2025, and healthcare costs continued to outpace general inflation. How much of the increase in mandatory spending is attributable to population aging, medical inflation, and other economic factors? How much to new legislation?

The 2004 Natural Baseline

This analysis constructs a 2004 natural baseline — what mandatory spending would have been under FY2004 programmatic rules, adjusted only for demographics, economic scale, and excess cost growth — and compares it to actual outlays. The resulting difference is the policy-driven excess.

FY2004 serves as the baseline year because it precedes both the ACA (enacted 2010, Medicaid expansion beginning 2014) and Medicare Part D (enacted 2003, costs beginning FY2006), and because the budget had returned to near-normal composition after post-9/11 increases. In FY2004, mandatory spending represented 54.0 percent of total outlays and discretionary represented 39.0 percent.

The 2004 natural baseline is built component by component. Social Security grows with the 65-and-over population share and a GDP-per-capita factor. Medicare and Medicaid add a 1.5 percent annual excess healthcare cost growth term, consistent with CMS National Health Expenditure trends and CBO projections. Other means-tested programs (SNAP, SSI, EITC, CHIP) include a cyclical adjustment tied to the unemployment rate, consistent with CBO and PWBM estimates of how these programs respond to recessions. Other mandatory spending (farm programs, federal retirement) grows modestly with GDP per capita at a lower elasticity. The policy-driven excess is the difference between actual mandatory outlays and the 2004 natural baseline in each year.

These findings are somewhat sensitive to the baseline year. FY2004 captures a pre-ACA, pre-Part D steady state; a different baseline would shift the magnitude of the policy-driven excess. We use an average 1.5 percent annual excess healthcare cost growth assumption, consistent with past experience, instead of applying different annual adjustments that are subject to more estimation error; a higher rate would increase the 2004 natural baseline somewhat (reducing the estimated excess) while a lower rate would do the opposite. We classify all COVID-era mandatory spending above the cyclically adjusted 2004 natural baseline as policy-driven, which is appropriate for measuring classification drift but could conflate emergency necessity with structural expansion. For structural deficit analysis, the pre-COVID FY2019 gap ($763 billion, or 27.9 percent of mandatory spending) and post-COVID gap may be more stable reference points.

Results

By FY2025, the 2004 natural baseline predicts $2.50 trillion, nearly double the FY2004 level of $1.24 trillion. This near-doubling reflects the 65-and-over population share rising from 12.4 percent to 17.8 percent (a 44 percent increase in the beneficiary-to-population ratio), real GDP per capita roughly doubling, and excess healthcare costs compounding over 21 years. Combined, these natural forces account for approximately 60 percent of FY2025 mandatory outlays. The remaining 40 percent is attributable to discrete policy choices and reclassifications.

What Has Driven the Policy-Driven Excess?

The major policy-driven additions to mandatory spending since FY2004 fall into five categories, each visible as a step-change in Figure 2.

Medicare Part D (from FY2006). The Medicare Modernization Act of 2003 added a prescription drug benefit as mandatory spending, accounting for roughly $50–70 billion per year in the early years of the policy-driven excess. Medicare spending increased 18.7 percent in 2006 alone, partly due to Part D.

Financial crisis and ARRA (FY2009–2013). The American Recovery and Reinvestment Act of 2009 and subsequent extensions added enhanced SNAP benefits, extended unemployment insurance, Medicaid FMAP increases, and refundable tax credits. A portion of this is captured by the model’s cyclical adjustment (elevated unemployment), but roughly $200–450 billion per year during 2009–2013 falls in the policy-driven excess.

ACA expansion and marketplace subsidies (from FY2014). The Affordable Care Act’s mandatory provisions represent the largest peacetime expansion of mandatory spending since Medicare and Medicaid were created in 1965. By FY2019, the Medicaid expansion added an estimated $310 billion above the 2004 natural baseline and marketplace subsidies added roughly $180 billion, together accounting for approximately 65 percent of the pre-COVID policy-driven excess of $763 billion.¹

COVID-19 emergency spending (FY2020–2022). The CARES Act, American Rescue Plan Act, and related legislation produced the sharpest spike: $2.49 trillion above the 2004 natural baseline in FY2020 and $2.73 trillion in FY2021. Enhanced unemployment benefits, economic impact payments, SNAP increases, and PPP/EIDL transfers were classified as mandatory direct spending under emergency authority rather than as supplemental appropriations.^2,3,4

IRA and OBBBA mandatory appropriations (from FY2022). Two recent laws extended the practice of using mandatory classification for programs traditionally funded through discretionary accounts. The Inflation Reduction Act (2022) included mandatory IRS funding and open-ended clean energy tax credits. The One Big Beautiful Bill Act (2025) created $191 billion in mandatory multi-year DHS funding and $156 billion for DoD. The pending Reconciliation 2.0 process (Senate HSGAC committee voted 8–5 on May 19, 2026) would add $72 billion in mandatory ICE/CBP funding.^5,6,7

Kent Smetters is the faculty director of the Penn Wharton Budget Model.

Footnotes

1. Estimating the Budgetary Effects of the Affordable Care Act — CBO ↩

2. The Long-Run Fiscal and Economic Effects of the CARES Act — Penn Wharton Budget Model ↩

3. Recovery Rebates in the American Rescue Plan Act — Penn Wharton Budget Model ↩

4. The Cost of the Employee Retention Tax Credit — Penn Wharton Budget Model ↩

5. Chairman Paul, HSGAC Republicans Advance ICE and CBP Funding — Senate HSGAC ↩

6. Senate-Passed Inflation Reduction Act — Penn Wharton Budget Model ↩

7. President Trump Signed Reconciliation Bill — Penn Wharton Budget Model ↩