Theta Scribe
Global Systems·

Charted: $2.3T Tax Expenditures, OASDI Empty by 2034, and $7.8T of GSE Credit

Aug 20, 2026 · 7 min read

The real fiscal levers sit outside the discretionary fight: JCT tax expenditures hit $2.3T in FY2026, OASDI reserves deplete ~2034, and Fannie/Freddie MBS still steers ~$7.8T of housing credit off-budget.

Loading interactive charts…

Congress argues about appropriation titles. Markets and households live inside a different ledger. The interactive map above answers the theme’s core question — which budget lines, trust funds, and off-balance vehicles move real levers? — by putting three plumbing layers on one screen: the tax-code shadow budget, the trust-fund entitlement spine, and the guarantee books that steer housing, education, and deposit risk without annual floor fights.

The shareable frame is narrow. JCT projects about $2.3 trillion of tax expenditures in FY2026 — larger than Social Security outlays on a like-for-like scale comparison. Combined OASDI reserves are on a mid-2020s Trustees path toward depletion around 2034. Fannie Mae and Freddie Mac still back roughly $7.8 trillion of mortgage-backed securities. Those three facts are not the same accounting identity. They are the same political identity: the fights that actually reprice retirement, health, and housing credit rarely show up as a single “nondefense discretionary” amendment.

The discretionary mirage

Lever classRough FY scaleBudget visibilityWhat it actually steers
All discretionary (approx)~$1.7THigh — annual billsDefense + agency ops
JCT tax expenditures (FY2026)$2.3TMedium — tax score, not outlayESI, pensions, CG rates, credits
SS + Medicare outlays (approx)~$2.2THigh on outlay side; trust-fund rulesPayroll taxes, COLAs, HI/SMI design
Net interest (approx)~$0.97THigh, hard to cutDebt service / first claim
GSE MBS outstanding~$7.8T stockLow — guarantee, not appropriationMortgage rates & credit availability
Federal Direct Student Loans~$1.6T stockMedium — credit reform scoringTuition finance & borrower relief

Read the table as a hierarchy of leverage density, not a sum. Discretionary is loud because it is annual. Tax expenditures and trust funds are quieter because they are permanent law. Off-balance credit is quietest of all: the stock can dwarf a year’s outlays while the appropriation line stays a footnote.

Pair this map with our tax-expenditure catalog for the line-item ranking, and with the Social Security depletion path for the reserve clock. This research post is the system diagram those two pieces sit inside.

Tax expenditures are the shadow budget

JCT’s FY2026 aggregate — $2.3T, after $2.2T in FY2025 and $11.7T over 2025–2029 in the CRFB summary — is the cleanest public meter of “spending through the Code.” The top ten lines alone are about $1.43T: pensions and retirement exclusions ($355B), preferential capital-gains and dividend rates ($252B), the employer-sponsored health exclusion ($240B), child/dependent credits ($128B), ACA subsidies ($105B), then charity, §199A, step-up at death, EITC, and SALT.

Toggle the dashboard’s Tax-exp rank and By family panels between JCT and Treasury. The ranking is not interchangeable. Treasury’s published FY2026 headlines put the employer medical exclusion at $296B — a $56B packaging gap versus JCT’s $240B ESI line — and surface imputed rent ($157B) and DC employer plans ($156B) that JCT folds into broader concepts. That is methodology, not a spreadsheet error. Analysts who cite only one office can cherry-pick a preferred giant.

Two caveats travel with every tax-expenditure chart. Do not sum line items as the revenue from simultaneous repeal — interactions and behavior matter. And refundable credits blur revenue loss and outlays; the deficit impact is real either way. For industrial-policy credit growth on the tax side, see our subsidies and industrial-policy series.

Trust funds are the entitlement spine

Social Security and Medicare are not “discretionary plus a trust-fund sticker.” They are dedicated-revenue machines with legal claim on special-issue Treasuries until reserves hit zero. In the mid-2020s Trustees framing used here (rounded for the map):

  • OASIretirement and survivors — depletes around 2033
  • OASDI combinedthe headline meter markets quote — around 2034
  • Medicare HI (Part A)around 2036
  • SMI (Parts B+D)adequately financed via premiums and general revenue, not a classic exhaustion clock
  • DI alone remains solvent far longer; the political clock is OASI/OASDI

The dashboard’s Trust funds panel shows an illustrative reserve runoff: OASDI and HI balances declining toward those depletion years. At OASDI exhaustion, continuing payroll taxes still cover a large share of scheduled benefits — historically scored near four-fifths in recent Trustees prints — but the ability to redeem accumulated reserves ends. That is a benefit cut by formula, not a shutdown metaphor, unless Congress changes law.

The plumbing point is sharper than the scare headline. Payroll-tax rates, taxable-maximum design, COLA formulas, and HI financing move more retirement and hospital cash than most appropriation riders. If you only watch the 302(b) allocations, you are watching the wrong valve.

Off-balance vehicles steer credit without floor fights

Guarantees and federal credit programs are where fiscal capacity becomes private balance-sheet capacity. The dashboard’s Off-balance scatter plots editorial scores for budget visibility against policy leverage, with bubble size equal to stock:

  • GSE MBS (~$7.8T)high leverage, low visibility; mortgage credit availability is a federal guarantee story dressed as a secondary-market utility
  • FDIC-insured deposits (~$10.5T)the systemic backstop stock; almost never debated as “spending,” always present when banks wobble
  • FHA/VA insurance (~$2.1T)housing credit for specific borrower channels
  • Direct Student Loans (~$1.6T)more visible because credit-reform scoring and forgiveness fights put it on front pages
  • PBGCsmaller stock, real pension-guarantee optionality when plans fail

Filter to housing credit and the thesis concentrates: mortgage rates and underwriting standards are fiscal-plumbing outcomes even when the unified deficit print barely moves. Student-loan forgiveness debates are the rare moment an off-balance stock becomes a partisan appropriation-style fight — which is why they look anomalous next to the quiet GSE book.

What moves when Congress actually turns a valve

If the question is “what changes household cash and credit next year,” the ranked levers look nothing like a CBO discretionary baseline:

  1. ESI exclusion and ACA subsidy designhealth compensation and Marketplace enrollment
  2. Retirement exclusions and required distribution rulesprivate retirement saving volume
  3. Capital-gains / dividend rate structure and step-uprealization timing and estate planning
  4. OASDI payroll tax / benefit formulanear-retiree cash flows after depletion risk becomes political
  5. GSE guarantee fees and FHFA capital rulesmortgage spreads without a new housing bill
  6. Net interest paththe rising first claim that crowds the rest of the unified budget

Every item above can move hundreds of billions over a decade without a single popular “spending cut” headline — or can explode into one when the trust-fund date or a forgiveness score hits the news.

Caveats

  • Tax-expenditure line items must not be summed as a joint repeal score; packaging differs between JCT and Treasury
  • Trust-fund depletion years and reserve paths are rounded mid-2020s Trustees framing for the mapnot a live SSA score sheet; see the dedicated SS post for report-vintage detail
  • Plumbing-layer pie and visibility/leverage scores are editorial composition tools, useful for exploration, not official CBO tables
  • Off-balance stocks are public outstanding/exposure round numbers (GSE MBS, Direct Loans, insured deposits, FHA/VA, PBGC)definitions differ across FHFA, ED, FDIC, and PBGC releases
  • Budget yardsticks (SS, Medicaid, discretionary, net interest) are approximate FY-scale comparisons, not identical accounting concepts
  • Payroll-tax effects on ESI are excluded from the income-tax expenditure estimates shown in the rank panel

Methodology

JCT aggregates and top-10 FY2026 ranking follow the Committee for a Responsible Federal Budget summary of JCT estimates (February 2026). Treasury largest items are from the Office of Tax Analysis Tax Expenditures FAQ (FY2026 estimates). Trust-fund depletion years and reserves are rounded from recent SSA / Medicare Trustees intermediate assumptions for dashboard scale. Off-balance stocks use public FHFA / Education Department / FDIC / related release round numbers for outstanding guarantees and loan books. The pie of plumbing layers is an editorial map that juxtaposes annual flows with guarantee stocks — deliberately not a single official identity.

Unlike the tax-expenditure catalog alone, this piece is the system diagram: tax code + trust funds + off-balance credit as one fiscal-plumbing stack. Unlike the Social Security depletion post alone, it asks which other valves sit beside OASDI when households and markets actually reprice.

Related reading: US tax expenditure catalog 2026 and Social Security trust-fund depletion path.