Where the standard advice comes from — and where it stops being true. “Pay off your debt before you retire” is close to universal, and for most households it is correct. The reasoning is sound: debt payments are a fixed obligation, a fixed obligation raises the income you must generate, and a higher required income means a higher withdrawal rate, more taxable draws, and less room to cut when markets fall. Kill the debt and every one of those improves. The trouble is that the argument is really about expensive, mandatory debt, and it gets applied to all debt. A 22% credit card and a 3% mortgage held against a seven-figure taxable account are not the same object, and no rule that treats them identically can be right about both.
Three tiers of debt. The distinction that makes the rest of this chapter usable:
| Tier | What it looks like | What to do |
|---|---|---|
| Oppressive | Credit cards, payday and title loans, high-rate auto and personal loans. Rates well above any return you can reliably earn; the balance grows faster than your assets. | Eliminate first, always. Nothing in this chapter applies until this is gone. See F1. |
| Working | The mortgage on the house you live in, a student loan, a business loan. Debt you took on to hold an asset you need, with a payment you must make. | Manage it — rate, term, and whether to prepay. That’s the question in F2. |
| Enhancing | Low-rate, collateralized, optional borrowing you could repay tomorrow if you chose, because you hold roughly 3–6× the balance in liquid assets. A pledged-asset line you draw on in a bad year; a cheap fixed-rate mortgage you chose to keep. | The subject of this chapter. Useful precisely because you don’t need it. |
The test for enhancing debt is not the rate alone. It is: could you write a check tomorrow and make this go away? If the honest answer is no, the debt is working or oppressive, whatever the interest rate says.
The mechanism: order only matters because you sell. This is the point that connects the whole chapter to Chapter 8. If you never add or remove a dollar, the order of your returns is irrelevant — multiplying the same set of growth factors in any sequence gives the same answer. Sequence risk is not created by bad markets; it is created by withdrawing from bad markets. Selling into a drawdown converts a paper loss into a permanent one and destroys the shares that would have carried the recovery. Every classic mitigation — a cash buffer, spending flexibility, guardrails, working longer — is a way to be less of a forced seller. Borrowing is simply the most direct one: instead of funding a bad year by selling shares at the bottom, you fund it with someone else’s money and repay when prices recover.
Why it is also a tax lever. Loan proceeds are not income. That single fact does more work than it first appears to in a retirement plan, because so many retirement costs key off a single number — your reported income:
- No capital gain is realized, so a low-basis position stays intact and untaxed.
- No addition to AGI, so it doesn’t drag more of your Social Security into taxability or blow past a bracket (Chapter 11).
- No IRMAA exposure, so a one-off large expense — a roof, a car, a wedding — doesn’t buy you two years of Medicare surcharges (Chapter 12).
- It creates room in a low-income year for the thing you actually want to do with that room: a Roth conversion at a low bracket.
Note the honest limit: investment interest on a portfolio loan is deductible only against investment income, and only if you itemize — and interest used for personal spending generally isn’t deductible at all. Treat the deduction as a bonus you probably won’t get, not part of the math.
And an estate lever. An appreciated asset held until death receives a step-up in basis: your heirs’ cost basis resets to the date-of-death value and the entire embedded gain escapes income tax (Chapter 28, and the placement logic in Appendix G). Sell that asset in your seventies to fund spending and you pay the gain; borrow against it and let it pass, and the loan is settled from the estate while the gain is forgiven. This is the whole of “buy, borrow, die,” and it is not a loophole in any exotic sense — it is the ordinary interaction of two ordinary rules. It is also the piece most exposed to policy change: the step-up has been a live reform target for years, and a plan that only works if it survives another two decades is a plan with a legislative assumption baked into it.
What you can borrow against, and what it costs in 2026. The instrument matters more than the idea, because the instruments differ on the one dimension that decides whether this helps or hurts — whether the lender can pull the money back when you most need it.
| Source | 2026 cost | Rate | Can it be called or cut? |
|---|---|---|---|
| Pledged asset line / SBLOC (taxable accounts only) | ~6.1–8.1% APR | Variable — SOFR + 2.40–4.40%, resets weekly; larger lines price better | Yes. A demand loan; maintenance call typically gives you 2–3 days |
| Margin loan | varies by broker | Variable, benchmark + spread | Yes — and requirements can be tightened mid-crisis |
| HELOC | prime + spread | Variable | Yes — lines can be frozen or reduced in a housing downturn |
| An existing fixed-rate mortgage you keep | your locked rate | Fixed | No — the most reliable cheap leverage most retirees will ever have |
| HECM standby line of credit (62+, see Ch 17) | rate + FHA premium | Variable; the unused line grows | No — cannot be cancelled or reduced, and non-recourse |
Read that last column before the cost column. The version of this strategy that actually reduces risk needs money that cannot be withdrawn at the bottom — which is why an already-locked fixed-rate mortgage and a HECM standby line are the two cleanest instruments here, and why a securities-backed line, the one people reach for first, is the one that needs the most discipline. Note too what 2026 pricing does to the arithmetic: at 6–8% variable against a long-run expected return that is not much higher, the leverage case is thin. The liquidity and tax cases survive at those rates. The “borrow to invest and pocket the spread” case largely does not.
The same five years, three ways
A retiree starts with $1,000,000 and spends $50,000 a year. The market delivers the same five returns either way — +25%, +18%, +12%, −15%, −25% — but in the bad-first ordering the two losing years come first. With no withdrawals at all, both orderings end at exactly $1,053,150. The withdrawals are what separate them.
| Scenario | Portfolio at yr 5 | Loan owed | Net |
|---|---|---|---|
| Good years first — sell each year | $855,949 | — | $855,949 |
| Bad years first — sell each year | $714,090 | — | $714,090 |
| Bad years first — borrow yrs 1–2 at 7% | $866,900 | $126,792 | $740,108 |
| Bad years first — borrow yrs 1–2 at 3% | $866,900 | $110,912 | $755,988 |
Order alone costs the unlucky retiree $141,859 — identical returns, identical average, worse sequence. Borrowing the first two years’ spending rather than selling into the crash recovers $26,018 of it at a 7% rate, and $41,898 at 3%. Two things are worth noticing. The move works, and the borrowing rate eats most of the benefit — which is exactly why 2026 pricing makes this a tactical instrument rather than a strategy. Also notice the sizing: at the worst moment, year 2, the debt is $103,500 against a $637,500 portfolio — about 16%, far below any advance rate, so no call is possible. Had the same retiree drawn all five years of spending on day one, that ratio would have been over 40% at the trough, and one more bad quarter would have forced the liquidation the loan existed to avoid.
The takeaway: borrowing helps by breaking the forced-sale link, and only while the loan is small enough that nobody can make you unwind it.
How much debt — and measured against what. Anderson’s recommendation is a debt-to-asset ratio of 15–35% for people within twenty years of retirement, on the reasoning that a household with no debt has given up a cheap, permanent option on its own liquidity. Two amendments make that safer to use. First, measure the ratio against liquid assets, not total assets: home equity and a pledged retirement account cannot answer a maintenance call. Second, treat the lower half of the range as the working zone. The number that actually protects you is not the ratio but the headroom — how far markets can fall before your drawn balance approaches the lender’s advance rate. Equities are typically advanced at 50–65%; if a 50% market decline would still leave your balance comfortably under that line, the loan is a buffer. If it wouldn’t, it is a second, correlated risk wearing a buffer’s clothes.
Does it really support a higher withdrawal rate? Cautiously, and much less than the strongest claims suggest. Anderson presents leveraged portfolios sustaining withdrawal rates far above the classic 4% (Chapter 7), which is where reviewers push back hardest. The most careful public test — Karsten Jeschke’s historical work at Early Retirement Now — found that a modest supplement (roughly a 3% portfolio withdrawal topped up by 1% from a margin line) did raise the sustainable rate to around 5.3% in most cohorts. It also found the failure mode plainly: for the 1929 and 1965 retirees, margin calls forced liquidation at the bottom and wiped the accounts out. Leverage did not hedge sequence risk in those cohorts; it amplified it. That is the honest shape of the evidence — a real but modest gain in most histories, and a catastrophic tail in the ones that matter most.
Who should not do this. The most common criticism of the strategy is not that the math is wrong but that the wrong people adopt it: the disciplined savers who could use it safely refuse on principle, and the people comfortable with debt — who shouldn’t — embrace it. Do not use enhancing debt if you carry any oppressive debt; if the borrowed money would fund spending you couldn’t otherwise afford; if you couldn’t repay the full balance from liquid assets on short notice; if most of your wealth sits in retirement accounts, which cannot be pledged; if a large single position is doing the collateralizing, so your collateral and your risk are the same bet; or if carrying debt in a falling market would cost you sleep — because the strategy only pays off if you can hold it through exactly that.
What it changes about the mortgage question. If a paid-off house is partly a sequence-risk defense — and it is, because it removes a mandatory payment — then a credit line you can draw on is a substitute for part of that defense, and the case for rushing to arrive mortgage-free softens for a specific kind of household: a low fixed rate, ample liquid assets outside retirement accounts, and the discipline not to spend the difference. For everyone else the standard advice still wins, and F2 now says so in those terms.