When should you refinance your mortgage?
When four things are true at the same time: today's market rate is below yours by more than your personal threshold (0.77 to 1.85 percentage points, depending mostly on your balance), you'll stay in the home past the break-even point, you can actually qualify, and you're not quietly resetting a nearly-paid-off loan into a fresh 30 years. Miss any one of those and the answer is KEEP — no matter how good the headline rate looks. As of July 24, 2026 the average 30-year fixed is 6.58% (Freddie Mac survey).
So when exactly should you refinance?
When the rate gap exceeds your computed threshold, you'll stay past break-even, you qualify, and the new term doesn't erase the savings. All four, simultaneously.
Most advice answers this question with a single number — "when rates drop a point" — which is why most advice is wrong. Refinancing isn't one test. It's four, and they're independent. Clearing three of them and failing the fourth means you lose money on a deal that looked good.
- The rate gap beats your threshold. Not a rule of thumb — a number computed from your balance, closing costs, and how long you'll stay. For most people it lands between 0.77 and 1.85 points.
- You'll stay past break-even. Closing costs divided by monthly savings gives you a month count. If you might move before then, refinancing costs you money.
- You can qualify. Credit, income, equity, and the seasoning rules for your loan type. A threshold you can't act on is trivia.
- The new term doesn't undo the gain. Refinancing 22 years of remaining loan into a fresh 30 lowers your payment while potentially raising your lifetime interest.
The rest of this page is each condition in turn, with the numbers. If you'd rather just have the answer for your own mortgage, the calculator runs all four and returns one word.
How big does the rate drop need to be?
Between 0.77 and 1.85 percentage points, depending mainly on your loan balance. Bigger balances need smaller drops, because the savings scale but the closing costs mostly don't.
This is the condition everyone gets wrong, because the popular answer — the "1% rule" — isn't derived from anything. Even the mortgage industry says so:
"Some finance professionals like to cite rules like 'refinance when you can get a rate X percentage point lower,' but basing a decision on a general rule can be costly." Peter Miller, HSH.com, updated October 10, 2025
He's right that the rules are wrong, and like almost everyone who says so, he doesn't supply a replacement. The defensible number comes from a closed-form solution published by economists Sumit Agarwal, John Driscoll and David Laibson (NBER Working Paper 13487), which accounts for something rules of thumb ignore entirely: the option value of waiting. Refinancing is a one-shot move. Spending it on a small dip forfeits the bigger dip that might arrive next quarter.
| Remaining balance | Drop you need | What the 1% rule tells you |
|---|---|---|
| $100,000 | 1.70 points | Refinance far too early |
| $200,000 | 1.31 points | Too early |
| $300,000 | 1.16 points | Slightly too early |
| $400,000 | 1.08 points | About right |
| $500,000 | 1.03 points | About right |
| $750,000 | 0.97 points | Slightly too late |
| $1,000,000 | 0.93 points | Too late — you're leaving money on the table |
Is half a point enough to refinance?
Almost never. A 0.50-point drop clears the threshold only on very large balances with unusually low closing costs and a long expected stay.
On a $300,000 balance with typical costs, you need roughly 1.16 points. Half a point isn't close. The exception is a borrower with a $750,000+ balance, a lender credit covering most costs, and no intention of moving — there, the threshold can fall under 0.80 and a large-enough half-point deal starts to make sense. For the median homeowner it doesn't.
Is 1 percent lower enough?
For balances above roughly $450,000, usually yes. Below about $250,000, usually no — you'd need closer to 1.3 points.
This is exactly why the 1% rule persists and exactly why it misleads. It happens to be approximately correct for a mortgage around $400,000–$500,000, which is close enough to the national average that the rule feels right to the people writing about it. It is meaningfully wrong at both ends of the range.
How long do you have to stay in the home?
Past the break-even point: closing costs divided by monthly savings. Under a typical refinance that's usually somewhere between 18 and 40 months.
Break-even is simple arithmetic and it's the condition people are most likely to wave through. If your closing costs are $5,000 and you save $210 a month, you break even at 24 months. Sell at month 20 and the refinance cost you money, regardless of how good the rate was.
Expected time in the home also feeds directly into the threshold itself, not just break-even — a longer horizon lowers the bar. Going from a 5-year to a 30-year expected stay knocks roughly a quarter of a point off the required drop at every balance, because there's more time to collect the savings and less chance you exit the loan before the option pays off.
Be honest rather than optimistic here. "We'll probably stay forever" is what almost everyone says, and the median US homeowner moves considerably sooner. If you're genuinely unsure, use the shorter number — the cost of waiting one more quarter is small, and the cost of paying $5,000 in closing costs 14 months before you list the house is not.
How soon after buying can you refinance?
It depends entirely on loan type. Conventional rate-and-term can be as little as six payments; FHA and VA streamlines require 210 days plus six payments; USDA and cash-out generally require 12 months.
This is the condition that catches people out, because it has nothing to do with the math and everything to do with program rules. You can have a perfect threshold gap and still be ineligible for another four months.
| Loan type | Minimum seasoning |
|---|---|
| Conventional rate-and-term | Generally six monthly payments before the new note date |
| Conventional cash-out | Six months on title, and the loan being paid off at least 12 months old |
| FHA Streamline | 210 days from the first payment date and six completed monthly payments |
| FHA cash-out | 12 months of occupancy and payments; no more than one 30-day late in 12 months |
| VA IRRRL (streamline) | 210 days from the first payment due date and six consecutive monthly payments |
| USDA Streamline | 12 months since closing, with all payments made in the month due |
Sources differ slightly on whether the VA's two conditions are "and" or "whichever comes first." The stricter reading — both — is the safe one to plan around.
Does refinancing restart your 30 years?
It does unless you deliberately choose a shorter term. That's how a lower rate can still cost you more interest over the life of the loan.
Say you're eight years into a 30-year mortgage. You have 22 years left. Refinance into a fresh 30-year and your payment drops — partly from the better rate, but also because you've stretched the remaining balance over 96 extra months. The monthly saving is real. The lifetime saving may not exist at all.
The honest comparison holds the payoff date constant: what does the remaining balance cost at the old rate over 22 years, versus the new rate over 22 years? That's why our calculator reports a matched-term interest figure, and why that number is often smaller than what other tools show. It isn't being pessimistic. It's refusing to count term extension as savings.
Two legitimate ways around this: refinance into a shorter term outright (with the 15-year averaging 5.96% this week against the 30-year's 6.58%, that spread does real work), or take the 30-year and keep making your old payment, which pays the loan down on roughly the original schedule while preserving the flexibility to drop back if money gets tight.
When should you NOT refinance?
When your rate already beats the market, when you're moving soon, when you can't qualify, or when the only gain is a lower payment bought with a longer term.
This deserves saying plainly because almost nobody publishing refinance advice has any incentive to say it: most American homeowners should not refinance right now. Roughly half of outstanding US mortgages carry a rate below 4%. Against a 6.58% market, those loans are an asset. A homeowner sitting at 3.25% on a $400,000 balance is saving on the order of $733 a month compared with borrowing the same money today. There is no refinance that improves on that.
The specific cases where the answer is KEEP:
- Your rate is below today's market. Obvious, and yet the most common reason people run the numbers at all is that they saw a headline about rates falling.
- The gap is real but under your threshold. This is the genuinely interesting case — refinancing would work, but waiting is mathematically better. Our calculator calls this GETTING CLOSE.
- You're likely to move inside the break-even window.
- You're near the end of the loan. Late in an amortization schedule most of your payment is already principal. There's little interest left to save, and the fee is the same.
- Your credit or equity has slipped since origination. The rate you're quoted may not resemble the rate in the headline.
Should you wait for rates to drop further?
If your gap is below the threshold, yes — that's precisely what the threshold means. If it's above, waiting to time the bottom usually costs more than it saves.
The threshold already contains the answer to this question. It isn't the point at which refinancing starts saving money — you save money the instant rates dip below yours. It's the point at which the savings are large enough that you should stop holding out for more. Below it, waiting is correct. Above it, you're speculating on rate direction with a several-thousand-dollar option, and forecasters are reliably bad at this.
The past eighteen months make the case. The 30-year ranged from roughly 5.98% to 6.95%, bottoming in February 2026. When it dipped, ICE counted nearly 5 million homeowners moving "in the money" on a swing of a few tenths of a point. Most did nothing, and the window shut. Anyone who refinanced on the first small dip of that run also gave up the better one that followed. Both errors cost money; only one of them is discussed.
Does the time of year matter?
Not in any way you can trade on. Mortgage rates track the 10-year Treasury and inflation expectations, not the calendar.
There's a mild seasonal pattern in purchase activity — spring is busier, and lenders are sometimes hungrier for refinance volume in slow winter months, which can show up as slightly better pricing. It is not large enough to wait for. What genuinely moves rates is macroeconomic: inflation prints, employment data, and the eight scheduled FOMC meetings a year. If you're watching anything, watch those.
What if you just want a lower payment?
That's a different goal from saving money, and it has cheaper solutions — a recast, or a longer term without the closing costs.
Worth separating, because the two get conflated constantly. If the actual problem is monthly cash flow rather than lifetime interest, refinancing is an expensive way to solve it. A recast — where you make a lump-sum principal payment and the lender re-amortizes your existing loan — lowers your payment while keeping your current rate, usually for a few hundred dollars rather than several thousand. For anyone holding a sub-4% mortgage, that distinction is worth thousands: refinancing would mean surrendering the best financial asset in the house.
What happens if you never check?
About 20% of households who should refinance never do, forgoing a median of roughly $11,500 each. The failure is attention, not arithmetic.
Benjamin Keys, Devin Pope and Jaren Pope studied a large sample of US mortgages (NBER Working Paper 20401) and found that roughly one in five households for whom refinancing was clearly optimal simply didn't do it. Median forgone savings: about $160 per month, or roughly $11,500 in present value — around $5.4 billion across their sample. A 2024 University of Chicago working paper models the cause directly as inattention. People aren't running the numbers and choosing wrong. They aren't running them.
Which is the whole reason this site exists. The four conditions above change slowly — but the first one changes every week.
How we work this out
Market rates come from the Freddie Mac Primary Mortgage Market Survey, published Thursdays, and the Optimal Blue Mortgage Market Indices, both via FRED®. Thresholds are the Agarwal–Driscoll–Laibson closed form, solved exactly with the Lambert W function rather than approximated — assumptions and full tables are on the threshold page. Break-even is closing costs divided by monthly savings, cross-checked against a matched-term interest comparison so that stretching your loan never counts as saving. Seasoning requirements were verified in July 2026 and reflect program minimums, not lender overlays.