Bond ETF Mechanics: Duration, Yield and Ladders

The constant-maturity machine

An index labeled "7–10 Year Treasuries" holds bonds until they age past the 7-year floor, then sells them to buyers of shorter paper and buys newly qualifying 10-year issues: a conveyor. Consequences:

• Duration is a policy, not a countdown. The fund sits near duration ~7.5 permanently. Your rate sensitivity on day 3,000 equals day 1: the exposure you chose is the exposure you keep, which is precisely what an allocation wants (Chapter 19) and precisely what a liability on a date doesn't (solved below).
• Roll-down is part of the engine. In a normal upward-sloping yield curve, a bond's yield falls (price rises) as it ages toward shorter maturities: the conveyor systematically harvests this "roll-down return," one reason bond-fund returns aren't just coupons.
• The yield you see is the yield of the machine's current cargo. The factsheet's YTM (strictly, usually yield-to-worst, the conservative version that assumes callable bonds get called at the worst moment for you) reprices continuously as the conveyor replaces cargo: the honest forward-looking number Chapter 22 crowned, now with its full name.

The convergence result: why horizon beats forecasting

The most useful, least-known arithmetic in fixed income: for a constant-duration fund, if rates rise, prices drop now but every subsequent coupon and replacement bond reinvests higher; and the two effects approximately cancel at a horizon around the duration (more precisely, near duration minus a small adjustment; academics formalize it as the "duration targeting" result: over horizons of roughly twice-duration-minus-one, starting yield dominates realized return almost regardless of the rate path).

Practical translations, blunt on purpose:

• Your expected return is the purchase-day YTM minus TER, provided your horizon comfortably exceeds the duration. Rate forecasts add almost nothing over that horizon: which is why Chapter 22 dismissed rate timing and why this book buys bond funds by matching duration to horizon, then stops thinking.
• 2022 revisited, final form: a duration-6.5 aggregate fund fell ~13%, and its YTM jumped from ~1% to ~4%: holders with 7+ year horizons were, in expectation, roughly restored by year 6–7 and better off after: the drawdown was the price of the raise. Holders who needed the money in 2023 had violated the matching rule, not been betrayed by the product.
• Corollary for direction-worriers: for a long-horizon rebalancing investor, rising yields are a gift with bad manners: they arrive as a drawdown and pay as income.

The second duration: credit spreads

Rate duration isn't the only sensitivity. Corporate and high-yield funds carry spread duration: sensitivity to credit-spread changes: which is why a corporate fund can fall while Treasuries rally (spreads widening in a risk-off, 2020's March exactly). Related machine-behaviors worth knowing: downgrade mechanics (a bond cut below investment grade exits IG indices, forcing sales into a distressed moment: the "fallen angel" effect, studied enough that fallen-angel indices exist to buy what IG funds must sell: a real structural niche), and evaluated pricing (Chapter 33's stale-NAV story: bond NAVs are built from pricing-service estimates, the ETF price often being the fresher fact in stress).

Fund vs ladder, settled properly

A ladder (individual bonds maturing each year, proceeds reinvested at the long rung) is a constant-maturity machine you run manually: which reframes the whole debate. Honest scoring:

• The fund wins on: diversification per euro (thousands of issuers vs your dozen), institutional trading costs (retail corporate-bond markups are brutal: often 0.5–2% per trade against a fund's basis points), automatic everything, and honesty (the fund's price marks to market daily; the ladder's "I haven't lost anything" is the same mark, unlooked-at).
• The ladder wins on: date-certain cash (a bond maturing in June 2031 pays par in June 2031: no price risk at the horizon: the one thing the conveyor structurally can't promise), psychological par-anchoring for those it genuinely helps, and, in some jurisdictions, tax-lot control.
• The verdict by use: allocations and stabilizer sleeves → funds, decisively. Known liabilities on dates (tuition 2030, house 2029, a retiree's next-five-years spending) → date-certain instruments.

And the instrument that collapses the trade-off: target-maturity ETFs (iBonds / BulletShares families): funds holding only bonds maturing in one named year, distributing along the way, then terminating at par-ish in December of that year. A ladder of these gives you both columns: fund diversification and costs, single-bond date-certainty: the correct modern answer for liability matching, and this book's recommended tool for it.

A worked example

Ana, 58, needs €30,000/year for the five years before her pension starts at 63, plus a lasting stabilizer sleeve.

• The dated liabilities: five target-maturity IG ETFs, 2027–2031, ~€30k each: each dissolves into cash on schedule; interim price wiggles are irrelevant by construction. Cost: fund-level fees ~0.10–0.18%, versus the 0.5–1.5% markups a retail five-bond-a-rung self-built ladder would have paid per position.
• The permanent sleeve: her Chapter 22 aggregate fund, duration 6.5, horizon open-ended: the conveyor, correctly matched.
• The 2022-style stress test, pre-written into her policy: if rates spike, the 2027–2031 funds barely move (short, shortening) and pay out par regardless; the aggregate sleeve draws down and re-yields: the plan happening, per Chapter 26's sentence.

One investor, three shapes from this chapter's diagram, each doing the only job it's built for: which is what "understanding the mechanics" purchases.