A refinance replaces one loan with another. It is a transaction with real costs — origination, appraisal, title, recording — and those costs are only recovered by keeping the new loan long enough. Everything else in the decision is secondary to that arithmetic.
Run the break-even
The decision tree
- Will you keep the home past break-even?If not, stop. No rate is low enough to overcome an unrecovered cost.
- Are you resetting the term?Refinancing into a fresh 30-year loan after several years of payments restarts front-loaded interest. Compare against a term matching your remaining years.
- Is mortgage insurance in play?If your equity has crossed 20%, removing PMI can be worth more than the rate change itself.
- Is this a cash-out?Different question entirely. You are borrowing against the house — judge it on what the money buys, not the rate.
- Has your credit improved?A score improvement since origination may unlock better pricing than the market rate alone suggests. See the 90-day blueprint.
The term-reset trapA borrower eight years into a 30-year loan who refinances into a new 30-year loan has extended the debt to 38 years. The payment falls and lifetime interest can still rise. If the goal is a lower payment, that trade may be acceptable — but make it knowingly.
What closing costs actually contain
| Line item | Typical range | Negotiable? |
|---|---|---|
| Origination / underwriting | 0.5–1.0% of loan | Yes — the main lever |
| Appraisal | $400–$800 | Sometimes waived |
| Title insurance and search | $700–$2,500 | Yes — you may shop it |
| Recording and transfer | $50–$500 | No — statutory |
| Discount points | 1% per point | Optional by design |
| Prepaid escrow | Varies | Not a cost — your money |
Note the last row. Escrow prepayments inflate the headline "cash to close" but are not a cost — that money funds your own tax and insurance account, and your existing escrow balance is refunded. Excluding it from the break-even calculation is the correct treatment.
Points: prepaid interest, nothing more
One point costs 1% of the loan and typically lowers the rate by 0.125–0.25%. It is a second break-even calculation nested inside the first. On a $340,000 loan, one point costs $3,400 and might save $50 monthly — a 68-month payback. Only sensible if you are certain about staying well past that.
Negative pointsThe reverse also exists: accept a slightly higher rate and receive a lender credit against closing costs. For anyone likely to move or refinance again within five years, this is frequently the better trade.
Cash-out refinancing: a different question
Cash-out is not a refinance decision, it is a borrowing decision. You are converting home equity into spendable money and securing that debt with the property.
Defensible uses
- Consolidating high-rate debt with the behaviour change that prevents rebuilding it
- Value-adding home improvements
- Funding a documented, high-return business need
Rarely defensible
- Discretionary consumption or travel
- Investing the proceeds in volatile assets
- Consolidating cards you will run up again
Shopping the loan properly
- Gather all quotes within a two-week window so the credit inquiries score as one event.
- Compare the standardised loan estimate form, not verbal quotes — section by section.
- Confirm identical lock periods; a 30-day lock is priced differently from a 60-day lock.
- Ask each lender to beat the best competing estimate. Have the document ready to send.
- Include a credit union and a mortgage broker alongside your existing bank.
Limits, errors and borders of the standard advice
Critical reviewRefinance advice is written almost entirely by parties paid per closed loan. That does not make it wrong, but it explains which rules of thumb survived: the ones that produce transactions.
"Refinance whenever you can drop the rate by 1%."
The claimA one-point rate reduction reliably justifies a refinance.
Where it breaksThe threshold ignores the three variables that actually determine the answer. On a small remaining balance, a full point saves less than the closing costs; on a large one, a quarter point can pay back in months. It ignores remaining term — refinancing into a fresh long loan late in an existing one can raise total interest even at a lower rate, because you restart the amortisation. And it ignores how long you will stay, which sets whether the payback period is ever reached. The rule is durable because it is simple and because it generates loans.
The borderIgnore the rate change and compute the break-even: total closing costs divided by monthly saving, compared against how long you will realistically hold the loan. Then check total interest over the new term, not just the payment.
"A no-cost refinance is free."
The claimLender-paid closing costs mean the refinance costs nothing.
Where it breaksThe costs are paid, just not at the table. They are recovered through a higher rate, or added to the principal so you borrow them and pay interest on them for decades. On a long hold, a slightly higher rate is far more expensive than the fees it replaced. The label survives because "no cost" is a powerful marketing phrase, and because the alternative — a lower rate with fees — requires the borrower to do arithmetic at exactly the moment they are tired of paperwork.
The borderLender-credit structures are genuinely the right choice when you expect to move or refinance again within a few years, or when you lack cash for fees. Ask for the same loan quoted both ways and compare total cost over your expected holding period — the crossover is usually only a few years out.
"Never reset the clock — always refinance into a shorter term."
The claimRestarting a thirty-year term destroys the progress you have made, so shorten it.
Where it breaksTerm and payment are separable, which the rule treats as identical. You can take a long term for the payment flexibility and pay it down on a shorter schedule voluntarily, keeping the option to stop in a bad month — an option a shorter contractual term takes away permanently. Forcing a higher required payment also crowds out retirement contributions and buffer building, and the advice is often given to people whose mortgage rate is below what they could earn elsewhere, where accelerating repayment is the weaker use of the money.
The borderShorter terms are right when the rate discount is meaningful, the higher payment is comfortable, and you value certainty over flexibility. Otherwise take the longer term and overpay by choice — same amortisation, more optionality.
"Cash-out refinancing is cheap money — use it to consolidate debt."
The claimReplacing high-rate unsecured debt with low-rate mortgage debt is obviously beneficial.
Where it breaksThe rate falls and the risk profile changes fundamentally. Unsecured debt is converted into debt secured on your home, so a payment problem that previously damaged your credit can now cost you the house. The term also lengthens dramatically, and a lower rate over thirty years frequently costs more in total than a higher rate over four. Most importantly, the empirical failure is behavioural: consolidation without changing what caused the debt leads to the cards being reused, leaving both the mortgage and the balances.
The borderDefensible when the rate gap is large, the underlying spending problem is genuinely resolved, you keep the repayment period short rather than stretching it, and the equity position leaves real margin. Never as a first response to revolving debt — see the sequencing options first.
Our break-even method is only as good as the holding period you assume, and people are systematically bad at predicting how long they will stay in a home — the input that dominates the answer is the one you cannot know. We also quote no rates or fee levels, so the calculations must be completed with current quotes. And mortgage products, prepayment penalties, portability and the tax treatment of interest differ so much between countries that some of these trade-offs do not exist in your market while others we omit do.
Frequently asked questions
Is there a minimum rate drop worth refinancing for?
No fixed threshold. On a large balance, 0.4% can break even in under two years; on a small balance, even 1.5% may not. The calculator above answers it directly for your numbers.
What about a recast instead?
A recast re-amortises your existing loan after a lump-sum principal payment, lowering the payment for a fee of a few hundred dollars — no new loan, no closing costs, and the rate stays the same. If your rate is already good and you have cash, recasting is often superior.
Should I refinance to a 15-year term?
Lower rate and dramatically less lifetime interest, at the cost of a materially higher payment and less flexibility. A middle path: keep the 30-year term and pay it like a 15-year, retaining the option to fall back on the smaller required payment.
How long does the process take?
Typically 30–45 days. Respond to documentation requests within 24 hours — borrower delay is the most common cause of a blown rate lock.
The bottom line
Calculate the break-even honestly, exclude escrow prepayments from costs, match the new term to your remaining years rather than defaulting to 30, and gather every quote inside a fortnight. If the break-even is under two years and you are confident you will stay, refinancing is straightforwardly good. Past four years, the answer is usually no.
Revision log. Jul 2026: added break-even model and recast comparison. Rate-sensitive content reviewed quarterly.