Repair or replace: the arithmetic, and the three places it breaks

The rule everybody quotes is: if the repair costs more than half of a replacement, replace it. It is a decent rule. It is also a rule of thumb that was never meant to survive contact with a specific machine, and there are three common situations where it points the wrong way.

Where the rule comes from

It is a crude way of accounting for remaining life. A repair buys you the rest of the machine’s service life; a replacement buys you a whole one. If the appliance is halfway through its expected life and the repair costs half of new, you are paying half price for half the value, and it is a wash — so anything worse than that is a bad trade.

That reasoning is sound, which is why the rule persists. The problem is that it hides the two variables that actually matter — how old the machine is, and which part failed — inside a single ratio.

Break one: the rule ignores age entirely

A $300 repair on a $1,200 dishwasher is 25% and the rule says repair. But if that dishwasher is fourteen years old, you are spending $300 on a machine whose next failure is not far away, and the one after that is closer still. Conversely a $500 repair on a $1,000 washer is 50% — right on the line — and if the machine is three years old it is an obvious repair, because you are buying seven or eight more years.

Rough service lives worth holding in your head: refrigerators and ranges 13 to 15 years, washers and dryers 10 to 13, dishwashers 9 to 12, microwaves 8 to 10. Past those, every repair is a bet on a machine that has already outlived its design.

Break two: not all failures predict the next one

This is the one the ratio cannot see at all. A failure tells you something about the rest of the machine, and what it tells you depends entirely on what failed.

• A wear part that failed on schedule — door seal, drain pump, belt, igniter, heating element — tells you almost nothing bad. It is a consumable doing what consumables do. Repair it and the rest of the machine is unaffected.
• A control board tells you the machine has been running hot or has taken a supply surge, and boards are expensive enough that the ratio usually decides it anyway.
• A sealed-system or compressor failure on a refrigerator is close to terminal economically. It is the most expensive repair available, it needs certification to perform, and a machine that has lost its charge once is a machine whose circuit has been compromised.
• Rust or a cracked tub or cabinet is structural, and structural faults do not get better. Nothing downstream of that is worth spending on.

Break three: the replacement price is not the replacement cost

People compare the repair quote against the sticker price of a new unit, and the sticker price is the smallest part of it. Delivery, haul-away of the old machine, and installation are real money. A built-in or panel-ready unit is worse: cabinetry is cut to a specific depth and a replacement that is an inch different turns an appliance purchase into a carpentry job. For a built-in, the threshold for repair should be far higher than 50%, and sometimes a repair that looks absurd on the ratio is the cheapest thing in the room.

Our own calculator uses the 50% rule with an age adjustment, and it will give you a bad answer for a built-in or panel-ready appliance, because it does not know about your cabinets. That is a genuine limitation, not a disclaimer — if the unit is integrated, weigh the result accordingly.

What we would actually tell you

Run the numbers, then ask one question the numbers cannot answer: is this machine one you would choose again? A ten-year-old appliance that has never given trouble and does what you want is worth repairing past the point where the ratio says stop. One that has been repaired twice, is noisy, and was never quite right is worth replacing before the ratio says to.

The repair-or-replace calculator will do the arithmetic and show you every number behind the recommendation, including the ones that go against it. And if you need a real repair figure to put into it rather than a guess, that is what a $65 diagnostic is for — you get the number, and if you decide to replace instead, you have paid $65 to make a good decision rather than several hundred to make a bad one.