Guide

Liability-driven investing explained

Harbor Capital's closed defined-benefit pension plan had $4.2 billion in assets against $5.1 billion in actuarial liabilities — an 82% funding ratio that looked manageable on a spreadsheet. Then rates fell 120 basis points in nine months. Liabilities, discounted at a lower corporate bond yield, jumped $680 million. Equities rallied, but not enough: surplus volatility hit 14% annualized because assets were 55% equities and 45% bonds with duration 6.2, while liability duration was 14.8. Trustees asked for “more bonds,” but adding random long Treasuries without matching the liability profile widened tracking error on the wrong axis. The refactor adopted liability-driven investing (LDI): measure obligations in present value, match duration and key-rate exposures, layer interest-rate swaps where physical bonds were illiquid, and glide equity down as funding improved. Surplus volatility fell to 6.1% within two years; the plan reached 94% funded without chasing equity beta.

LDI is not a return-maximization strategy — it is a risk-alignment framework for any investor with dated cash obligations: pensions, insurers, endowments with spending rules, or individuals funding retirement withdrawals. This guide covers liability present value and discount curves, duration and cashflow immunization, funding ratios, derisking glidepaths, swap overlays, the Harbor Capital pension refactor, a technique decision table vs tactical asset allocation and bond ladders, pitfalls, and a production checklist.

What liability-driven investing is

Liability-driven investing (LDI) builds the fixed-income portfolio around the timing, size, and sensitivity of future obligations rather than around a market benchmark like the Bloomberg Aggregate. The core question is not “which bond index beats last quarter?” but “when rates move, do assets and liabilities move together?”

Traditional asset-only management optimizes return per unit of tracking error to an index. LDI optimizes surplus stability — the gap between asset value and liability present value. A pension with underfunded status can still pursue return-seeking assets, but the hedge portfolio (long bonds, swaps, futures) should neutralize parallel and non-parallel rate shifts that inflate obligations faster than assets respond.

LDI applies beyond pensions: life insurers match reserves to policy payouts; corporations hedge defined-benefit plans; universities align bond sleeves to scholarship commitments; retirees with known withdrawal schedules can LDI a portion of wealth while keeping growth assets for longevity risk.

Measuring liabilities

Present value and discount curves

Actuaries project benefit payments (or insurance claims) by year. Each payment is discounted to today using a discount curve — often high-quality corporate AA yields for U.S. pensions, government curves for public plans, or a bespoke curve for insurers. The sum is liability present value (PV).

Critical detail: liabilities are long-duration, convex instruments. A 1% drop in discount rates can raise PV by more than 10% for a mature plan. That sensitivity is why duration matching matters more than yield chasing.

Funding ratio

Funding ratio = market value of assets / liability PV. Above 100% is surplus; below is deficit. Trustees watch smoothed vs market funding (accounting vs economic truth). LDI decisions should use market-based liabilities — the number that moves when yield curves shift.

Inflation-linked obligations

COLA-adjusted pensions and CPI-linked annuities require hedging real rates, not just nominal duration. Pair nominal LDI with TIPS or inflation swaps; breakeven analysis from inflation breakevens guides sizing.

Duration matching and immunization

Immunization (Redington, 1952) sets asset duration equal to liability duration so that small parallel rate shifts change asset and liability PV by roughly equal amounts. For a single lump-sum liability in n years, hold a zero-coupon bond maturing in n years — perfect cashflow match.

Cashflow matching vs duration matching

Approach	Mechanism	Best for	Limitation
Cashflow matching (dedicated portfolio)	Buy bonds whose coupons and principal pay each liability date	Static schedules, insurers with predictable runoff	Illiquid for long ladders; reinvestment risk if benefits change
Duration matching	Match PV-weighted duration (and convexity where possible)	Pensions with evolving benefit profiles	Non-parallel curve shifts break the hedge
Key-rate duration	Match sensitivity at 2y, 5y, 10y, 30y pillars	Large plans with curve twist risk	More instruments to rebalance; model risk
Derivatives overlay	Swaps and futures adjust duration without selling equities	Underfunded plans gliding into bonds	Collateral, CSA terms, counterparty limits

Real pensions rebalance quarterly: liability duration drifts as members age and as discount rates move. Harbor Capital recomputes liability duration monthly and allows asset duration to deviate only within ±0.3 years before forced rebalancing.

Convexity and term premium

Duration matching is first-order. Large rate moves need convexity alignment too — liabilities often behave like long puts on rates. When term premium regimes shift independently of Fed expectations, static immunization drifts. Sophisticated LDI desks overlay macro views on premium without abandoning the hedge core.

Derisking glidepaths

Most underfunded plans cannot jump to 100% LDI overnight — they need equity return to close the gap. A glidepath rules how fast to shift from return-seeking assets (equities, alternatives) to hedging assets (long corporates, govies, swaps) as funding improves.

Status-based: at 80% funded, 50% hedge; at 90%, 75% hedge; at 100%, 90% hedge.
Time-based: reduce equity 2% per year regardless of markets (predictable but can deepen deficits in bear markets).
Hybrid: Harbor uses funding ratio triggers with a floor — never derisk below 70% funded without sponsor contribution.

Liability-relative glidepaths differ from target-date funds: TDFs assume individual longevity; pensions assume cohort mortality and benefit law. Do not copy a 2040 retirement fund glidepath for a frozen pension with average retiree age 72.

Instruments in an LDI portfolio

Long corporate and government bonds — physical duration; prefer liquid issues for rebalancing.
Interest-rate swaps — receive fixed, pay floating to extend duration without deploying cash; see interest-rate swaps.
Treasury futures and bond ETFs — tactical duration adjustments; watch basis and roll in stressed markets.
Inflation swaps and TIPS — for real-liability hedges.
Buy-in / buy-out — transfer liability to an insurer (terminal LDI for some sponsors).

Collateral management on swaps is non-optional: when rates rise sharply, variation margin calls can strain liquidity if the plan holds illiquid alternatives. Harbor keeps 8% of assets in T-bills and repo for margin after a 2022 near-miss.

Harbor Capital pension sleeve refactor

Pre-refactor: 55/45 equity/bond, liability duration 14.8, asset duration 6.2, no swap overlay, quarterly actuarial reports only. Post-refactor rules:

Liability engine: monthly PV and duration using AA corporate curve; key-rate buckets at 5y, 10y, 20y, 30y.
Hedge sleeve target: match 85% of liability duration when funded ≥ 90%; scale linearly from 50% hedge at 75% funded.
Implementation: long investment-grade credit ladder plus receive-fixed swaps for the 20y+ bucket where bonds are scarce.
Equity cap: return-seeking sleeve capped at max(15%, (funding − 70%) × 1.5%) of total assets.
Contribution policy: sponsor tops up when market funding drops below 80%; derisking pauses until restored.

Outcome 2023–2025: surplus volatility halved; two rate-cut cycles produced modest surplus gains because liabilities rose less than hedged assets. Equity upside was forgone in 2024's rally — an explicit trade trustees accepted for predictability.

Technique decision table

Goal	LDI approach	Alternative	When alternative wins
Stabilize pension surplus vs rates	Duration-matched bond + swap hedge	More equities for return	Young open plans with long contribution horizon
Fund known retirement withdrawals	Cashflow-matched bond ladder	4% withdrawal rule on balanced fund	Small portfolios where ladder is impractical
Beat peers quarterly	Not the primary LDI objective	Tactical asset allocation	Endowments without rigid payout formulas
Hedge inflation on COLA benefits	TIPS + inflation swaps	Commodity futures	Imperfect CPI linkage; basis risk acceptable
Close deficit fast	Glidepath with return sleeve + growing hedge	Lump-sum sponsor contribution	Sponsor balance sheet can inject cash now
Retail income planning	Partial LDI near retirement	Target-date fund	Simple automation; no custom liability model

Common pitfalls

Using accounting liabilities for hedging — smoothed numbers lag markets; hedges drift.
Matching duration but not curve shape — twist moves hurt belly-heavy liabilities hedged only with 30-year bonds.
Ignoring credit spread risk — AA discount rates widen in crises; corporate bond assets may fall while liabilities rise.
Over-derisking into a deficit — locking in low funding with no return sleeve.
Swap collateral surprise — margin calls without liquidity buffer force fire sales.
Confusing LDI with “all bonds” — wrong duration is worse than a balanced fund.
Static hedge for dynamic benefits — workforce turnover and benefit law changes require rolling liability models.
Forgetting longevity risk — people living longer extends liabilities beyond any bond ladder without mortality updates.

Production checklist

Publish monthly liability PV, duration, and key-rate durations on a market curve.
Compute funding ratio (market) and surplus volatility trailing 36 months.
Set hedge ratio policy as a function of funding status and sponsor covenant.
Rebalance when asset minus liability duration exceeds ±0.3 years.
Stress-test parallel ±100 bps and 2s10s twist scenarios quarterly.
Maintain liquidity sleeve for derivative margin (target 5–10% of assets).
Document glidepath triggers and board pre-approval for derisking steps.
Separate nominal hedge from inflation-linked sleeve for COLA benefits.
Attribution: report surplus change split into rates, credit, equity, and contributions.
Annual review with actuary on mortality, benefit changes, and discount methodology.

Key takeaways

LDI aligns bond and derivative hedges to obligation PV and duration — not to a generic bond index.
Funding ratio drives how much return-seeking vs hedging assets you can hold.
Duration matching is necessary but not sufficient — key-rate and convexity gaps matter in real curve moves.
Swaps and futures extend duration efficiently; collateral and liquidity planning are part of the strategy.
Harbor Capital cut surplus volatility by half with a funding-triggered glidepath and monthly liability analytics.