Pensions & LDI — the retirement balance sheet inside private markets

Built 2026-06-14. Structured data + edges: macro-pensions-ldi.json. Companion to macro-uncovered-risk-pools, macro-private-credit-marks, and self_marked_value.

Pension funds have migrated ~a third of the two largest US funds into manager-marked illiquids and use leveraged liability-driven investing (LDI) that already broke once — the UK September 2022 gilt doom-loop.

1. The UK 2022 LDI doom loop — the live precedent

• The 23 Sep 2022 mini-budget spiked 30yr gilt yields ~120bp+; leveraged LDI funds posted gilts as collateral, so falling prices triggered margin calls → forced gilt selling → higher yields → more calls (the BoE's "vicious spiral").
• UK DB liabilities hedged via LDI grew ~£400bn (2011) → ~£1.5T (2020), ~two-thirds of GDP. The BoE pledged up to £5bn/day (the "£65bn" headline = the 13-day ceiling) but actually bought ~£19.3bn — mostly an announcement effect.
• Margin calls >£70bn; forced LDI gilt sales >£36bn (23 Sep-14 Oct). Reform: leveraged LDI must hold a ≥250bp yield-resilience buffer + operational buffer. A 2025 gilt sell-off re-tested it — wider buffers held.

2. US public pensions — ~a third in valuation-priced illiquids

• Funded ratio ~80% (2024) → ~82.5% (2025); unfunded ~$1.27T. FY2025 returns averaged ~9.5% vs an assumed ~6.9% — the improvement is return-driven, so it reverses in a drawdown.
• Alts rose ~11% (2006) → ~26% (2016) → ~34% (2022) as public equity fell below 50%.
• CalPERS (~$556B, Jun 2025): PE ~17.7%, real assets ~13.1%, private debt ~3.8% (new private-credit target ~8%) → ~34% illiquid. CalSTRS (~$370B): PE 15.1%, real estate 12.8% → ~28% illiquid.

3. The denominator / self-marking problem

• Private-market values are manager/valuation-priced, not market-priced — the self-marked-value defect on the retirement balance sheet (self_marked_value). Stale marks lag public drawdowns, so >50% of US plans were over-allocated to PE in 2025 (S&P Global).
• H1-2025 secondaries: LP portfolios traded ~90% of NAV, tail-end (>10yr) below 75%, distressed sellers 25%+ discounts (Jefferies) — direct evidence the marks overstate realizable value.

4. Corporate DB (opposite sign) & the drawdown

• Milliman 100 funded ratio ~108% (end-2025) — corporate plans are overfunded, driving LDI de-risking and pension risk transfer (annuity buyouts that relocate liabilities onto life insurers, spec-insurance-bermuda). Do not conflate with underfunded public plans.
• Public plans run net cash outflow ~1.7% of assets (2024); mature plans pay benefits by selling assets — and you can't sell privates at NAV, so stress hits the liquid sleeve (or secondaries at a discount).

Synthesis

Pensions are where leveraged rate-hedging (LDI, proven fragile in 2022) and manager-marked illiquids meet the household's retirement claim. The same ~91% common factor (macro-cross-sectional-analysis) that prices the bank HTM hole and private-credit NAVs also prices pension alts and gilt/Treasury LDI collateral — and the demographic outflow removes the buy-and-hold cushion.

What is NOT asserted

• No claim US public pensions are insolvent (~82% funded) — the point is leverage + self-marked illiquids + net outflow.
• "£65bn" is the BoE ceiling; actual purchases ~£19.3bn.
• Corporate DB (~108%) and public DB (~82%) cut opposite directions — not conflated.
• Forced-selling-in-stress is graded weak (analytical), not a realized loss.
• Overlay edges are excluded from the proofs.

Sources: Bank of England — gilt market operation; SUERF — UK LDI lessons; TPR — LDI guidance; Equable — State of Pensions 2025; Pew — pension investment risk; CalPERS facts; CalSTRS portfolio; Milliman PFI Jan 2026; PBGC FY2024; S&P — pensions over PE target; Jefferies — H1-2025 secondaries.

← Research index · structured data: macro-pensions-ldi.json · macro-pensions-ldi.md