For fourteen years, the answer to “pay off mortgage or invest” was the same: invest. Mortgage rates sat near 3%. The S&P 500 compounded at a mid-teens annualized total return from 2010 through 2021. The math wasn’t even close — keep the cheap leverage, ride the index. That gospel calcified into a rule across the Boglehead community, JL Collins’s Simple Path to Wealth, and almost every personal-finance subreddit thread. It’s still the answer most advisors give today. But the inputs have shifted, and we ran the 2026 numbers. The math has flipped.

The claim: never pay off a low-rate mortgage early

The Boglehead position is precise. Paying down a fixed-rate loan is mathematically equivalent to buying a risk-free bond with duration matched to the loan’s remaining term. So the question reduces to: would you rather hold a guaranteed bond at your mortgage rate, or take equity risk for an expected higher return? For most of the post-2008 period, the answer was obvious. A 3% mortgage bond versus a 7-9% nominal S&P expectation left a 400-600 basis-point premium for taking equity risk — and historically, that premium showed up.

JL Collins codified this in two sentences. Pay off any debt above 5% as fast as possible. Below 4%, never pay it off early — it’s the cheapest money you’ll ever get. The 4-5% band was the gray zone where personal psychology mattered. That framework worked because two assumptions held: most people’s mortgages WERE below 5%, and forward equity returns WERE meaningfully higher than mortgage rates. The Boglehead consensus was right for fourteen years because both halves of that calculation agreed with it.

The math: should you pay off mortgage or invest at 6.5%?

Two things changed at once. Mortgage rates roughly doubled, and forward equity return expectations came down hard. Both halves of the calculation flipped.

Current 30-year fixed rate: Freddie Mac’s weekly mortgage rate survey showed 6.51% as of May 21, 2026, and 6.36% a week earlier. Call it “6.4-6.5%.”

Tax treatment for most homeowners: Post-TCJA, only about 9% of US filers itemize their deductions. The other 91% take the standard deduction — which the IRS just raised to $32,200 for married filing jointly in 2026 (up from $31,500 in 2025). For non-itemizers, mortgage interest provides exactly zero tax savings. Their after-tax cost of the mortgage equals the nominal rate. A 6.5% mortgage is a 6.5% guaranteed return on principal paid down.

Expected forward equity returns: The two most widely cited 2026 capital market assumptions bracket the consensus at the low end and the high end:

• Vanguard VCMM (October 31, 2025 run): 3.5-5.5% annualized nominal US equity return over the next decade. The firm cites elevated valuations and stretched profit margins as the constraints.
• JPMorgan 2026 LTCMA: 6.7% nominal US large-cap, holding steady from 2025.

Midpoint of the consensus: roughly 5-6% nominal. After a 15% long-term capital gains tax (where most MFJ households between $98,900 and $613,700 fall in the 2026 brackets), the after-tax expected return drops to roughly 4.3-5.7% in a taxable account. Midpoint ~5%.

Here’s what the spread looks like across mortgage rates, assuming a 4.7% expected after-tax stock return:

At today’s 6.5% rate, the expected after-tax stock return is running 180 basis points below the guaranteed return of paying down principal. And crucially, the paydown carries no variance — it’s a known 6.5% every year. The 4.7% stock expectation is a probabilistic midpoint with a wide distribution; bad sequences can push the realized return well below the paydown rate. The risk-adjusted gap is wider than the headline number suggests.

Worked example: Take a household with a $400,000 mortgage balance at 6.5% on a 30-year fixed, deciding where to send an extra $1,000 per month. Applied to principal, after ten years the mortgage balance drops to roughly $171,000 — about $168,000 less than the scheduled-payment path. Keep those extra payments going to retire the loan, and the mortgage is gone in roughly fifteen years instead of thirty, saving about $288,000 in lifetime interest. Alternatively, that same $1,000 a month invested in a taxable S&P index fund at the 4.7% expected after-tax return grows to roughly $153,000 over ten years on the central estimate, with a historical 25th-to-75th percentile range of about $130,000 to $185,000. Net household wealth at year 10: the paydown path is ahead by roughly $16,000 on the central expectation, and by more in adverse market scenarios — and unlike the investment path, paydown doesn’t expose the household to a tail where a bad decade leaves them behind.

The counter: three honest exceptions

If we stopped here, this would read as a sermon. There are three places where the inversion verdict breaks down — and they matter more than the math itself for most readers.

1. Tax-advantaged accounts always come first. A 50-100% instant return from a 401(k) employer match beats any mortgage paydown. Same logic for Roth IRA contribution room ($7,500 in 2026; $8,600 if you’re 50+) — that room is use-it-or-lose-it each year. HSA contributions get triple tax advantage if you have a qualifying high-deductible plan. The funding order is: capture every dollar of employer match first, max the HSA if eligible, max the Roth IRA, then — and only then — compare extra mortgage principal against contributing to a taxable brokerage. The Verdict here is on the LAST decision in that chain, not the whole chain.

2. Rate lock-in changes the math entirely. Most US homeowners are still sitting on 3-4% mortgages from the 2020-2021 refi window. Their after-tax mortgage cost is the same 3-4%, but their expected after-tax stock return is also the new 4-5%. The spread is still positive for them — keep the cheap money, invest the spread, the old rule still applies. The verdict you’re reading only fires for: new home buyers at current rates, people refinancing today, ARMs that already reset, second mortgages, and HELOCs (which are running 7-8%+). For roughly 60% of US mortgage holders sitting on legacy sub-5% rates, this changes nothing.

3. Liquidity is asymmetric. Paying down principal locks the money in home equity, which is the most expensive asset to access in an emergency. A cash-out refinance today costs 7-8% plus 2-3% in fees. A HELOC carries variable rates that can rise quickly. A portfolio in a taxable brokerage settles in 2-3 business days at whatever the market gives you that day. If your emergency fund is under six months of expenses, build that first — the paydown verdict assumes you already have liquidity. Reasonable readers can also keep some allocation in taxable equities as a liquidity option even when the headline spread is mildly negative; the same logic that revived 60/40 applies to keeping some equity exposure now.

The verdict

If your mortgage is locked below 5%, invest the spread. The Boglehead rule still holds for you. If you’re shopping a 6.5%+ rate today and not itemizing, paying down beats investing on average — for the first time in over a decade.

Run yourself through four questions:

1. Are you getting your full 401(k) employer match? If no — do that first, before any extra principal goes anywhere.
2. Is your mortgage rate ≤ 5%? If yes — invest the spread. The old rule still applies to you.
3. Do you itemize? Mortgage interest plus state/local taxes plus charitable deductions must exceed $32,200 MFJ. If yes — the after-tax math shifts roughly 80-100 bps back toward investing.
4. Do you have six months of liquid emergency savings? If no — build that before paying extra principal.

If you answered Yes → No (rate above 5%) → No → Yes — meaning the match is captured, your rate is above 5%, you’re not itemizing, and your liquidity is solid — the math now favors paying down.

The “invest the spread” rule isn’t wrong. It’s just no longer guaranteed. For fourteen years it was — and that’s why it felt like a law. It was an artifact of ZIRP and a structural bull market in US equities. With mortgage rates near 6.5% and forward equity returns running 3.5% to 6.7% per Vanguard and JPM, the spread that made the rule a near-certainty has closed. We’re back in a regime where the answer depends on your specific rate, your tax situation, and your liquidity. That’s not a rule. That’s a decision.

One more honest beat. The mortgage paydown number is reliable in a way the stock number isn’t. A 6.5% guaranteed return becomes more attractive when expected stock returns drop, AND it becomes more attractive when the dispersion around those expected returns widens. Both happened in 2025-2026. The next decade of US equity returns could surprise to the upside — Vanguard’s range tops out at 5.5%, JPM’s central estimate is 6.7%, and there are credible analysts at higher numbers. But a 6.5% mortgage rate doesn’t care about that range. It pays exactly 6.5%, no asterisks. Anyone telling you the old rule still works — including retirement-math personalities still quoting 8% withdrawals — is fighting the last cycle.