An investor closing on a rental property this month faces a decision that has nothing to do with the price they just negotiated. It's the loan structure sitting on the term sheet: a 30-year fixed rate around 6.7%, or an adjustable-rate mortgage quoting three-quarters of a point lower with a catch buried in the fine print. That catch is the reset — the moment, five or seven years out, when the rate stops being fixed and starts tracking an index nobody in the room can predict today. With the Federal Reserve now cutting after eighteen months of holding steady, ARMs are back in more investor conversations than they've been since 2022, and the pitch sounds simple: lock in the discount now, refinance or sell before the adjustment hits. Whether that math actually works depends on numbers most buyers never run.
How a 5/1, 7/1, or 10/1 ARM Is Actually Built
The number before the slash is the fixed period in years; the number after it is how often the rate adjusts once that period ends. A 5/1 ARM holds its starting rate for five years, then recalculates annually. A 7/1 stretches the fixed window to seven years, a 10/1 to ten — and each additional year of fixed protection typically costs a borrower a bit of the rate discount, since the lender is carrying rate risk longer before it can reprice. After the fixed period, the new rate is set by adding a margin — a fixed percentage the lender bakes in at origination, commonly 2.5 to 3 percentage points on conventional ARMs — to whatever the index is doing that month. Since LIBOR was retired in 2023, that index is almost universally 30-day average SOFR, the Secured Overnight Financing Rate published by the New York Fed.
Caps are what keep an ARM from becoming unbounded risk, and they come in a three-number sequence like 2/2/5 or 5/2/5. Read left to right: the first number caps how much the rate can jump at the very first adjustment, the second caps every adjustment after that, and the third caps the total move over the life of the loan relative to the starting rate. A 2/2/5 ARM starting at 5.75% can't adjust past 7.75% at year six no matter what SOFR does, and it can never exceed 10.75% over the full 30-year term. A 5/2/5 structure — more common on 7/1 and 10/1 products — allows a bigger first jump but the same steady-state protection after that. There's also a rate floor, usually set at the margin itself, which matters more than borrowers expect: even if the index goes negative, the rate won't fall below that floor, so an ARM is not a symmetric bet on rates falling as much as it feels like one.
The adjustment itself isn't instant, either. Most ARMs use a look-back period — typically 30 to 45 days before the reset date — to average the index and set the new rate, which smooths out short-term SOFR volatility but also means the number a borrower gets isn't necessarily what the index is doing on the day the loan resets. Margins run wider on non-QM and DSCR investor loans than on owner-occupied conventional ARMs — often 3 to 4 percentage points instead of 2.5 to 3 — because the lender is pricing in the extra risk of a loan qualified on rental income rather than a W-2. An investor comparing a conventional 5/1 ARM against a DSCR 5/1 ARM on the same property needs to compare the margin, not just the headline start rate, since a lower teaser on the DSCR product can still land at a higher fully indexed rate once the fixed period ends.
Why the Trade Looks Different With the Fed Cutting
None of this works if the spread disappears.
For most of 2024 and 2025, ARMs were a hard sell. The Fed was holding at restrictive levels, the yield curve was inverted or flat, and short-term rates weren't meaningfully cheaper than the long end — so the discount on the initial ARM rate barely covered the reset risk. That relationship has shifted. With cuts now underway and SOFR easing alongside the fed funds rate, the curve has room to normalize, and lenders are pricing 5/1 and 7/1 ARMs at a real spread below the 30-year fixed again — often 75 to 125 basis points on investment-property loans, wider than the 25-to-50-point gaps typical of a flat-curve environment. That spread is the entire reason to consider an ARM. If it collapses to nothing, so does the case. Here's what the pitch conveniently leaves out: nobody, including the Fed, is committing to where SOFR sits in year six of a 5/1 ARM. The cutting cycle could continue, stall, or reverse if inflation surprises to the upside — and a borrower who signed up for the discount is exposed to whichever of those happens to be true when the fixed period ends.
An Illustrative Breakeven, Not a Promise
Run the numbers on a $500,000 loan to see where the discount actually goes. At a 30-year fixed rate of 6.75%, the principal-and-interest payment is roughly $3,243 a month. The same loan as a 5/1 ARM priced at 5.75% — a full point lower, plausible in the current spread environment — runs about $2,918, a difference of roughly $325 a month, or close to $19,500 saved over the five-year fixed period. That's real money, and it's the number every ARM pitch leads with. What it doesn't lead with is the other side: if SOFR plus margin puts the fully indexed rate at, say, 7.05% when the loan resets, the new payment jumps to roughly $3,343 — higher than the fixed-rate payment would have been the entire time, wiping out a chunk of the earlier savings in year six alone if the investor is still holding the loan. This is one illustrative scenario built on realistic mechanics, not a market forecast — actual rates, spreads, and payments will vary by lender, credit profile, and loan amount.
The Rule Lenders Don't Get to Skip: Qualifying at the Fully Indexed Rate
Under the Ability-to-Repay and Qualified Mortgage rules that came out of Dodd-Frank (12 CFR 1026.43), a lender cannot qualify an ARM borrower using the low introductory rate alone. The borrower has to be able to afford the payment at the fully indexed rate — the margin plus a reasonable index projection — or the intro rate, whichever produces the higher payment, calculated on a fully amortizing basis. That single rule is what separates 2026's ARM market from the one that blew up in 2007, when borrowers were routinely qualified at teaser rates that had no relationship to what they'd actually owe. It also means the underwriting discount most borrowers hope for from a lower initial rate mostly doesn't show up in their debt-to-income calculation — the lender is already assuming something closer to the reset-rate payment when deciding how much house, or how many rental units, the borrower can carry.
Who This Actually Fits — and Who It Doesn't
Take the 5/1 or 7/1 ARM if you have a defined exit inside the fixed window and the discipline to stick to it. That's the honest recommendation, and it applies to a narrower group of buyers than the rate sheet suggests:
- BRRRR investors who plan to refinance into permanent debt once the renovation is complete and the property is stabilized — often well inside 24 months, far short of a five-year reset
- Fix-and-flip operators using the ARM as bridge financing on a property they expect to sell within 12 to 18 months
- Investors on a documented three-to-five-year hold who are comfortable underwriting the exit at a higher exit rate, not just a favorable one, and building that into the deal from day one
- Anyone planning to carry the loan past its fixed period on the theory that rates will simply be lower by then — that's a bet, not a plan, and it belongs nowhere near an underwriting model
Skip the ARM if the plan is buy-and-hold for a decade or longer. A 30-year fixed at 6.7% is not exciting, but it removes an entire category of risk from a long hold — no reset to time, no refinance window to hit, no dependence on rates cooperating five years from now. The math above shows the discount is real for as long as it lasts. It just doesn't last as long as most buy-and-hold investors plan to.
None of this is personalized lending or tax advice — loan terms, margins, and caps vary by lender and by borrower profile, and the numbers here are built to show mechanics, not to promise an outcome. Anyone weighing an ARM against a fixed-rate loan on a specific property should run the fully indexed payment through their own underwriting, not the teaser rate on the term sheet, before signing anything.