Clearing debt is a guaranteed return; investing is not
The claim
Paying down a debt returns its interest rate with certainty, which should be compared to expected investment returns on a risk-adjusted basis, not a raw one.
Why we rate it strong
The core comparison is arithmetic. The judgement about how large an expected-return premium justifies taking risk is where reasonable people differ.
Repaying a debt returns exactly its interest rate. There is no VarianceThe squared spread of returns around their mean, and the raw material of volatility. Higher variance around the same expected return is worse for anyone who must sell on a fixed date., no tax on the return, and no Sequence riskThe risk that returns arrive in an unlucky order. Two retirements with identical average returns can end very differently if one meets its bad years while withdrawals are shrinking the balance.Read the evidence on this →. Compare that to a DiversificationOwning enough different things that no single outcome decides your result. It is the one adjustment that reduces risk without a matching reduction in expected return.Read the evidence on this → portfolio's roughly 7% nominal Expected returnThe probability-weighted average of what could happen — not a prediction of what will. A strategy can have the higher expected return and still be the wrong choice for someone who cannot survive its bad tail. with a VolatilityHow much returns scatter around their average, usually measured as a standard deviation. It is a proxy for risk rather than risk itself: it treats an unexpected gain and an unexpected loss identically. near 17% and a real possibility of losing money over any given decade, and the high-rate case is not close.
The comparison people get wrong is treating an expected return and a guaranteed return as equivalent numbers. A certain 6% and an expected 7% are not one percentage point apart in any meaningful sense — the second carries a distribution wide enough that a full decade of underperformance is unremarkable. This is why our tool leans toward the debt whenever the gap is under roughly two points.
Two things outrank both. An Employer matchMoney your employer adds to your plan in proportion to what you contribute, typically up to a few percent of salary. Declining it is declining part of your stated compensation.Read the evidence on this → is an immediate return on the contributed amount that no debt rate matches, so it comes first. And a starter cash buffer comes before aggressive repayment, because without one, the next unexpected bill goes straight back onto the card and the effort resets.
Tax changes the arithmetic in both directions. Deductible interest lowers the effective debt rate; taxable gains lower the effective investment return. The calculator handles both.
Where this breaks down
- Liquidity is not symmetric. Money paid into a mortgage is hard to get back out; money in a taxable account is not. That option value is worth something.
- Debts with a fixed low rate become less burdensome as inflation erodes their real value — a real consideration for long-dated fixed-rate mortgages.
- The psychological return on being debt-free is real and shows up in the behavioural data, even where the arithmetic marginally favours investing.
Terms used on this page
The same definitions the underlined words open, written out so nothing on this page depends on a click.
- Employer match
- Money your employer adds to your plan in proportion to what you contribute, typically up to a few percent of salary. Declining it is declining part of your stated compensation. Evidence →
- Diversification
- Owning enough different things that no single outcome decides your result. It is the one adjustment that reduces risk without a matching reduction in expected return. Evidence →
- Volatility
- How much returns scatter around their average, usually measured as a standard deviation. It is a proxy for risk rather than risk itself: it treats an unexpected gain and an unexpected loss identically.
- Sequence risk
- The risk that returns arrive in an unlucky order. Two retirements with identical average returns can end very differently if one meets its bad years while withdrawals are shrinking the balance. Evidence →
- Variance
- The squared spread of returns around their mean, and the raw material of volatility. Higher variance around the same expected return is worse for anyone who must sell on a fixed date.
- Expected return
- The probability-weighted average of what could happen — not a prediction of what will. A strategy can have the higher expected return and still be the wrong choice for someone who cannot survive its bad tail.
Sources
Follow these rather than taking our word for the summary.
James J. Choi, David Laibson and Brigitte C. Madrian (2011). $100 Bills on the Sidewalk: Suboptimal Investment in 401(k) Plans
The Review of Economics and Statistics, 93(3), 748-763
Finding: A substantial share of employees forgo employer matching contributions they could take with no downside, leaving guaranteed money unclaimed.
Gene Amromin, Jennifer Huang and Clemens Sialm (2007). The tradeoff between mortgage prepayments and tax-deferred retirement savings
Journal of Public Economics, 91(10), 2014-2040
Finding: A large share of US households accelerated mortgage payments while under-contributing to tax-deferred accounts, an arbitrage that cost them money on the authors' estimates.