Running a Refinance Break-Even Correctly

Break-even math on a refinance is the single most confidently-wrong calculation I watch borrowers do, and I've watched a lot of them do it. Take closing costs, divide by monthly savings, get a number, feel good about it. The number is arithmetically fine. It's also financially misleading more often than not, because it treats a refinance as a rate swap when it's actually a new loan with its own clock, its own amortization curve, and its own set of assumptions the borrower never agreed to. ## Where The Napkin Math Comes From Here's the version everyone runs. Balance of $400,000, rate drops from 7.1% to 6.2%, closing costs land at $6,000, monthly savings come out to $220. Divide one by the other, get roughly 27 months, and that's the pitch. It's clean. It fits on the back of a business card, and loan officers like it for exactly that reason: it closes conversations fast. But the formula assumes exactly one thing changed: the rate. Nothing else in the file moved. That assumption is almost never true, and it's not true in a way that's easy to miss if you're not the one staring at amortization schedules for a living. ## The Reset Nobody Mentions Refinancing resets the loan to month one. Doesn't matter how far along the original note was; the new loan doesn't know or care that the borrower already survived the worst of the front-loaded interest. A borrower five years into a 30-year mortgage has already crossed the steepest part of the interest curve and started building real equity through principal paydown, the kind of paydown that compounds. Refinance into another fresh 30, and that progress just resets. The clock restarts, and the borrower is back at the top of the curve. So take that same $400,000 example. Say 25 years remained on the old loan. New loan gets written for a fresh 30. The borrower just tacked five years onto their payoff date without necessarily meaning to. Lower rate: good. Longer runway: worth questioning. Smaller payment: worth understanding why it's smaller. Three separate effects, and the standard break-even calculation bundles them into one number that only tells you about the first. More than once, borrowers have pushed back on this in person, insisting the math has to be wrong because the payment is lower and the rate is lower, so what's the catch. The catch is the term. It's always the term. The fix is tedious in practice, if not complicated in principle: run two amortization tables, not one payment comparison. Take the remaining schedule on the old loan, run it against the full schedule on the new one, and match the end dates if the borrower can swing it. Some lenders will write the new loan for 25 years instead of defaulting to 30, specifically to preserve the original timeline. That one decision, term-matching over a fresh 30, moves lifetime cost more than most rate improvements do on their own. ## The Numerator Gets Undercounted Too Closing costs are usually low-balled, and it's not always intentional. Origination fees, appraisal, title insurance, recording fees, prepaid interest, escrow funding due at closing. All of it counts. All of it is cash the borrower wouldn't have spent otherwise, and skipping any line item just moves the break-even point closer than it actually is. Points make this messier. Buying down the rate with a discount point is its own nested break-even problem, separate from the main one: does the monthly savings recoup the upfront cost before the borrower sells the house or refinances again. Points generally need something like seven to eight years to pay off. A borrower who moves every four or five years, which describes a large share of the market according to National Association of Realtors tenure data, may never see that money again. Lender credits work the other direction. No-closing-cost refinances roll the fees into a slightly higher rate, so upfront cost drops toward zero while the rate improvement shrinks along with it. Fine choice for someone who expects to move soon. This is a different break-even profile entirely, though, and it needs its own math rather than getting folded into the standard formula as though it were the same animal. ## Four Places The Payment Comparison Actively Misleads Cash-out refinances break the comparison on contact, because the borrower is now comparing two different loan balances, not two rates on the same one. A lower rate on a bigger balance can still produce a lower monthly payment. That looks like a win on the calculator, but the comparison is the wrong one to be running; at that point the borrower has made a decision to draw on home equity, and it should get measured against a HELOC or a home equity loan, not against the old mortgage payment. Term changes cut the opposite way. Refinance a 30 into a new 30 at a lower rate, and total time in debt usually goes up even while the payment goes down. Compare that against refinancing into a 15-year term instead: payment barely moves, sometimes it rises, and total interest owed falls hard because the curve compresses rather than stretches. The standard formula tells that second borrower there's no benefit, when the benefit is actually substantial. It just doesn't show up in the one number the formula is built to produce. ARM-to-fixed conversions are about certainty more than savings. A borrower approaching their first adjustment isn't asking whether the new fixed payment beats today's ARM payment; they're asking whether it beats what the ARM becomes once the index resets under current caps. That's a modeled comparison against a future state, not a static one against the present. Holding period undoes everything, given a short enough stay. A 27-month break-even means nothing to someone who sells the house at month fourteen. Freddie Mac's research on refinance behavior treats the break-even period as a reasonable heuristic only for borrowers who stay put and match loan terms. Outside that lane, the tool stops measuring what people assume it measures, and starts measuring something closer to nothing. ## What The Real Comparison Needs Three numbers, not two. Closing costs, monthly payment difference, and total interest paid across a matched horizon on both loans, old and new. The first two get you the napkin math everyone already knows. The third is the one that catches the reset problem and the term-extension problem, and there's no way around it, no shortcut: you run both amortization tables in full, or you're guessing. None of this makes the simple formula worthless. For a borrower staying long-term, matching their remaining term, skipping the cash-out entirely, it's a decent, honest tool. Outside that lane, the monthly savings number is answering a question nobody actually asked, and the real answer was sitting in the amortization table the whole time, not in the smaller number that shows up in the checking account.