Bond ladders (and the ladder of linkers)

There are two ways to own bonds. This book has mostly talked about the first one, the constant-duration fund (Bonds in retirement: types, duration, the exact job). That is the permanent sleeve, the one that absorbs shocks and gets rebalanced. The second is the ladder. You own a series of bonds, each held to its own maturity, one "rung" per year of spending to fund: 2027, 2028, 2029, and on down the line.

A ladder is not a technical variation on the fund. It changes the whole frame. You stop running a portfolio against a market and start matching dated liabilities with contractual cash flows. Retirement becomes an actuary's problem rather than an investor's (Annuities and safety first: buying a floor, the same safety-first school, in bond form and reversible). Used well, the ladder is the most powerful tool in this book for three precise jobs: the bridge to your pension, the floor of the early years, and big dated expenses. Used badly, it is a pile of illiquid holdings that rebuilds, in nominal terms, the false security inflation eats away.

This chapter gives the principle, and the one property that sets a ladder apart: why "held to maturity" really does change the nature of the risk here, when it was an illusion for the permanent sleeve. Then the three uses and how each one gets built, the indexed version (the ladder of linkers, the closest thing to a guaranteed pension that exists), how to buy the rungs at retail, and the traps.

🔬 Where the ladder comes from: three generations of actuaries

The idea was not born on a FIRE forum. It comes from life insurers and pension funds, which have always faced the same problem: dated promises to keep with assets that move. The British actuary Frank Redington gave it its name in 1952, in "Review of the Principles of Life-Office Valuations" (Journal of the Institute of Actuaries): a fund is immunized when its value stops moving, to first order, under a general shift in interest rates. His recipe is already ours, matching the mean term of the assets to the mean term of the liabilities. Lawrence Fisher and Roman Weil put it to the data in 1971, in "Coping with the Risk of Interest-Rate Fluctuations" (Journal of Business): a matching strategy removes almost all of the interest-rate risk of a high-grade bond portfolio held to a definite horizon. Martin Leibowitz finally turned it into a manager's craft, in the two 1986 Financial Analysts Journal installments of "The Dedicated Bond Portfolio in Pension Funds". They draw a sharp line between cash-flow matching, which is exactly the ladder of this chapter, and duration immunization, which manages a horizon without pinning every single maturity.

Liability matching is therefore neither a forum trick nor a recent find. It is what pension funds do for a living, shrunk to household size. And that lineage also says what a ladder will never do, since no pension fund funds its most distant promises, or its benefit increases, this way.

The three jobs a ladder does§

Job 1: the bridge to your pension. The FIRE use case above all others, already met along the way (VPW, the Bogleheads' variable percentage withdrawal, Horizon, life expectancy, and 50-year retirements). Between the day you stop working and the day your pension or Social Security starts, there is an uncovered phase of 10 to 20 years in which the portfolio pays for everything. The floor part of those years is a dated liability: known, non-negotiable, the perfect candidate for matching. Building it is simple. One rung a year, sized to the uncovered floor, from year one to the year the pensions start. The rest of the money, relieved of the floor, is free to carry the comfort layer and the long term (The stock and bond allocation in retirement: covering the floor frees you toward the top of the allocation plateau). Where that last year falls is a claiming decision, not a market one (US health coverage and Social Security: the two ends of the gap).

Job 2: the floor of the early years (the fragile window). The same move, shorter. You match 5 to 8 years of floor on the day you retire, exactly the years when sequence risk peaks (Sequence of returns risk: the retiree's real enemy). It is the second bucket of the bucket strategy, made contractual and unambiguous (The bucket strategy: the promise, and the case against it). And unlike a cash buffer (The cash buffer: size, cost, and what it is really for), rungs at 3 to 8 years earn the market rate. The short ladder is a buffer that gives up almost nothing in opportunity cost.

Job 3: dated expenses. Putting the kids through college (2031 to 2036), paying off the last of a loan, a roof you know you will replace. Any liability with a date and a number can be matched to a rung, lifted out of the risky portfolio, and it stops distorting how you size the plan (How much you need).

What a ladder does not do. It does not replace the permanent bond sleeve, since the rebalanced shock absorber is the fund's job. And it does not cover the open-ended very long term. You do not ladder 40 years: the distant rungs are expensive against uncertain inflation, and longevity stays open at the end. Past the ladder, the right tools are the portfolio and, late in life, an annuity (Annuities and safety first: buying a floor).

Nominal or indexed: the question that decides everything§

A nominal rung of $40,000 maturing in 2038 will fund $40,000 of 2038 money, which is about $30,000 of today's at 2.5% inflation. Past 5 to 7 years, a nominal "guarantee" is a guarantee of shrinking purchasing power, the classic false security (Inflation and the withdrawal rate: the exact link). Three answers, cleanest first.

The ladder of linkers. Every rung indexed to prices. The 2038 cash flow will be $40,000 real, by contract. This is the "guaranteed pension" object met earlier (Inflation-linked bonds: the only contract written in real terms: a 30-year TIPS ladder funding about 4.5% a year when real yields sit near 2%, about 3.9% at 1%). It is the canonical answer as soon as the ladder runs past 5 to 7 years. The window of real yields in which such a ladder beats the 4% rule, and how much the horizon you pick moves the answer, are on the plate in Inflation-linked bonds: the only contract written in real terms.
The inflated nominal ladder. Nominal rungs that grow (40,000 × 1.025^n). It covers expected inflation and stays naked against surprises. Fine for the short rungs, steadily more fragile after that.
The short rolling ladder. Nominal, 3 to 5 years only, rebuilt at the top every year. Inflation over a short horizon carries little uncertainty. The long-horizon risk stays in the portfolio, handled by its building blocks (Defensive assets: the map and the roles).

The doctrine that follows is simple. Nominal is fine when the ladder is short. When it is long, it is indexed or nothing. It is not a new doctrine either. Zvi Bodie and Michael Clowes had already made it the heart of Worry-Free Investing (2003), a whole book on funding a retirement out of inflation-indexed bonds rather than out of stocks.

Building it, rung by rung§

A US investor is unusually well served here: everything the last section described can be bought at retail, in an afternoon, with no gatekeeper to work around. The Treasury sells notes, bonds and TIPS at auction to anyone with $100, and the secondary market at any broker quotes every issue still outstanding, which is how you buy the exact maturity you need instead of whichever one is on the auction calendar this month.

Vehicle	Useful horizon	Indexed	Cost	Where
T-bills, money market fund	Years 1-2	No	~0	Any broker, TreasuryDirect
Treasury notes and bonds	Any maturity out to 30 years	No	Free at auction, a few dollars at a broker	TreasuryDirect, any broker
TIPS	5, 10 and 30 at auction, any live maturity on the secondary market	Yes	Same	TreasuryDirect, any broker
Defined-maturity Treasury and corporate ETFs	One fund per year into the 2030s	No	About 0.07 to 0.10%	Any broker
Defined-maturity TIPS ETFs	One fund per year into the 2030s	Yes	About 0.10%	Any broker
Brokered CDs	3 months to 10 years	No	Dealer spread	Any broker

Which rung for which job. Years one and two are not rungs at all: that is the cash buffer, in T-bills or a money market fund (The cash buffer: size, cost, and what it is really for). From year three out to about year seven, nominal notes do the job cheaply, with a maturity available for every year. Past that, buy TIPS, for the reason the last section gave: at ten years and beyond, a nominal rung guarantees the wrong thing.

Buying them. At auction you place a noncompetitive bid and take the yield the auction sets, in $100 increments, through TreasuryDirect or through your broker. Notes come in 2, 3, 5, 7 and 10-year terms, bonds in 20 and 30, TIPS in 5, 10 and 30. But a ladder really gets built on the secondary market, which quotes every maturity still alive on whatever day you choose. One practical detail: a security sitting in TreasuryDirect has to be transferred to a broker before it can be sold, so keep the whole ladder in the brokerage account. Terms checked on treasurydirect.gov in August 2026; the Treasury does change them, so read the site before you buy.

The hole in the TIPS calendar. Not every year has a TIPS maturing in it, and there is a well-known stretch of the 2030s with none at all. The standard fix is the shoulder years: buy extra of the maturities on either side of the gap and spread the proceeds across the missing ones. It costs a little precision, never the structure.

Defined-maturity ETFs, the ready-made rung. These funds hold a basket of bonds that all mature in the same year, then liquidate and pay out. They are the rung in a box, and they make a small ladder practical: one ticker a year instead of twenty settlements. The two big lineups, iShares iBonds and Invesco BulletShares, run one fund per year across Treasuries, corporates, municipals and, more to the point here, TIPS. Check the quality of the basket (Treasury or broad investment grade, never high yield, see the traps below) and the exact month the fund liquidates. That indexed line matters: a retiree who cannot hold individual bonds, inside a plan account for instance, still gets a real ladder.

Brokered CDs, the insured cousin. Same shape with a bank behind it instead of the Treasury, insured by the FDIC up to $250,000 per depositor per bank, sometimes at a slightly better yield: a fine rung, as long as you stay inside the insurance limit and accept that its interest, unlike a Treasury's, is taxed by your state.

Where to hold it. Treasury interest is exempt from state and local income tax, worth real money in a high-tax state. TIPS have the opposite problem, since the annual increase in principal is taxed in the year it accrues with no cash arriving to pay the bill: put the TIPS rungs in a tax-deferred or Roth account whenever you have the room (US accounts, and the order you drain them).

In a simulator. A ladder is modeled by equivalence. For a matched pension bridge, take the cost of the ladder out of the simulated capital, then enter the covered floor as income (Using the FIRE simulator: the side income and pension sliders carry the structure). The failure probability left on the screen is then the failure of the comfort layer alone. That is the whole point, and the example in Inflation-linked bonds: the only contract written in real terms put numbers on it.

⚠ The traps in building one

Four recurring traps. Yield chasing, first: swapping Treasuries for corporates to pick up a point. On an instrument whose entire purpose is the certainty of the cash flow, adding default risk defeats the object (Bonds in retirement: types, duration, the exact job: high yield, never). Then excessive granularity: fifteen separate holdings of $12,000, each with its own settlement and its own bid-ask spread. Defined-maturity funds solve that. Then the ladder as a prison, the whole portfolio laddered "for safety". A ladder has no growth, no room to spend more, and nothing to leave. It matches the floor, never the comfort (Annuities and safety first: buying a floor, the same rule as an annuity). And the forgotten rung, last. A ladder is alive: realized inflation moves and so does spending. Revisit it at the annual review (The annual review: the retiree's checklist), not only on the day you build it.

🧮 The bridge Claire and Idris built

Take the couple from Choosing your strategy: criteria, comparison, worked case again: a floor of $45,000, pensions in 13 years that will cover about 53% of it. The liability to match is 13 years × $45,000 of floor, less what the comfort portfolio can deliver in any weather. They decide to match 100% of the floor for years 1 to 6 (the fragile window) and 60% for years 7 to 13. The build reads line by line. Years 1 and 2, T-bills and a money market fund ($92k). Years 3 to 8, nominal Treasury rungs and defined-maturity Treasury funds maturing 2028 to 2033, each sized up 2.5% a year (about $220k). Years 9 to 13, TIPS bought for those exact maturities (about $120k). Total cost, about $430k out of $1.7M. The balance, $1.27M, carries the comfort layer on guardrails and then VPW, at an effective withdrawal rate of about 1.8%. Failure of the comfort layer becomes a footnote, failure of the floor is zero by contract until the pensions start. The price paid is the expected return the $430k gives up: measured, accepted, written down.

Claire and Idris's liability, year by year. The annual need does not move in today's dollars ($45k of floor, $58k of comfort); what changes is who pays it. The rungs cover 100% of the floor for the first six years and 60% after that, and the pensions take over in year 14. The comfort portfolio carries only the balance, $13k a year at first and $31k later, an average of $22.7k over the thirteen years: the 1.8% of the remaining $1.27M.

The essentials§

A ladder is one rung per year of liability, held to maturity. Matching the asset to the liability cancels interest-rate risk, where "hold it to maturity" was only an accounting illusion for a permanent sleeve. It is the one literal "guaranteed" in this book.
Its three jobs are the bridge to your pension, the floor of the fragile window (a buffer that earns its keep) and dated expenses. Never the permanent sleeve, which is the fund's job, and never the open-ended long term, which belongs to the portfolio and then to an annuity.
Nominal is fine when short. Long means indexed or nothing: a nominal rung at 12 years guarantees shrinking purchasing power. The ladder of linkers is the canonical answer.
Building one at retail is easy in the US: Treasuries and TIPS at auction from $100 or on the secondary market for the exact maturity, defined-maturity ETFs (indexed ones included) when the ladder is small or the account cannot hold individual bonds, brokered CDs as the insured cousin. Mind the TIPS gap in the 2030s, the quality of any fund basket, and put the TIPS rungs in a tax-advantaged account. And revisit the ladder every year.
The traps are yield chasing (the outright fallacy), too many tiny holdings, the ladder as a prison that matches the comfort too, and the forgotten rung. The ladder serves the floor; the portfolio serves the life.

Going further§

Allan Roth and the TIPS ladder tools (tipsladder.com): the finished version of the idea, priced on live quotes.
treasurydirect.gov: terms, auction calendar and how a noncompetitive bid works, from the issuer.
The defined-maturity ETF lineups (iShares iBonds, Invesco BulletShares): the product pages, for the actual build.
Wade Pfau, Safety-First Retirement Planning (2019): liability matching as a doctrine (Annuities and safety first: buying a floor).
Zvi Bodie and Michael Clowes, Worry-Free Investing (2003): the case for funding a retirement out of indexed bonds, written for a general reader. The three founding papers cited above are in The library: sites, papers, books, tools.
In this book: Inflation-linked bonds: the only contract written in real terms (the real rung, and what a guaranteed ladder delivers), Bonds in retirement: types, duration, the exact job (funds against individual bonds, the real debate), The bucket strategy: the promise, and the case against it (the second bucket made contractual), The cash buffer: size, cost, and what it is really for (rungs zero and one).