Research
The repair-or-replace threshold, costed
The trade applies one number to every machine at every age. Worked out properly, the threshold is a curve, and the 50% rule sits on it at exactly one point.
Written by HYRE HVAC Research Desk Primary-source research, data analysis and fact checking
The finding
Repair when the repair quote, as a share of a replacement quote, is below the equipment’s expected remaining life divided by a new unit’s expected life.
For a central AC or heat pump that is 80% at four years, 51% at twelve and 33% at twenty. The popular 50% rule is right only at about 12.2 years. Treat each as a ceiling for a unit that just failed.
What is wrong with the 50% rule?
It ignores how long the repaired machine will keep running. The 50% rule puts two numbers in the same sentence, the repair in front of you and the replacement you are putting off, and that is useful.
What it cannot do is tell you what you are buying with the repair, because it contains no term for how long the repaired machine will go on running.
A $1,500 repair on a four-year-old system and the same $1,500 on an eighteen-year-old one score identically under the rule and are obviously different purchases.
Everyone writing about this knows it. Nobody fixes it because the missing number, expected remaining life, has not been available.
Until this month the best service-life figure on this site was 15 years for central cooling equipment.
It came from NAHB / Bank of America Home Equity, Study of Life Expectancy of Home Components, a February 2007 report of a 2006 telephone survey of manufacturers and trade associations.
The report itself says to use it as a general guideline only.
HYRE analysis: A telephone survey of manufacturers gives one number. What the threshold needs is not a number but a function: expected remaining life conditional on the equipment having already reached the age it is.
That is a property of a survival function, and we now have fitted survival functions for the major equipment classes, estimated from the age distribution of 18,496 surveyed households. The rest of this page is arithmetic on top of them.
How do you calculate the repair-or-replace threshold?
Divide the equipment’s expected remaining life by the expected life of new equipment. Both options buy years of service, and they buy different numbers of years, so the comparison has to be made per year rather than in total.
Write Cr for the repair quote, Cn for the replacement quote, R(a) for expected remaining life at age a, and L for the expected life of new equipment.
Repairing buys R(a) further years for Cr. Replacing buys L years for Cn. Repair is the cheaper order when Cr / R(a) < Cn / L, which rearranges to the form a homeowner can actually use:
Cr / Cn < R(a) / L. In words: the repair is worth making when it is a smaller fraction of the replacement quote than the equipment’s remaining life is of a new machine’s whole life.
Money has a time cost, so the general form replaces each life with its annuity factor A(n, d) = (1 − (1+d)−n) / d, giving a threshold of A(R(a), d) / A(L, d).
At a discount rate of zero this reduces to R(a) / L.
Both inputs come straight out of the fitted Weibull functions. For a central air conditioner or heat pump, new-equipment life L is 18.6 years; remaining life at ten is 10.5 years, at twenty 6.2.
Remaining life does not fall year for year: a machine that has survived to twenty is expected to give another 6.2 years, which is why the threshold curve flattens rather than reaching zero.
HYRE analysis: the convention turns out to be a special case: The curve crosses 50% at 12.2 years for a central air conditioner or heat pump, 14.0 years for a central furnace, and 13.7 for a ducted heat pump.
Those are ordinary mid-life ages for the equipment in question. The 50% rule is not made up. It is the right threshold for a system in the middle of its life, applied to every system regardless of age.
When does the 50% rule give the wrong answer?
On young equipment and on old equipment, in opposite directions. The gap is not even, and the two mistakes cost different amounts.
On young equipment the rule is far too strict: At 6 years the derived threshold for a central air conditioner or heat pump is 71%, 21 percentage points above the convention.
A repair quoted at 60% of a replacement on a six-year-old system is, on this arithmetic, still the cheaper order, and the rule says replace. This is the expensive error: it converts a repair into a full changeout.
On old equipment the rule is too lenient: At 20 years the threshold for the same class is 33%, 17 points below the convention.
A repair at 45% of a replacement passes the rule and fails the arithmetic.
This is the error that leaves people spending real money on a machine with 6.2 years left in it, and then spending again soon after.
The furnace and the heat pump behave differently from each other. The furnace threshold stays above the convention until 14.0 years because furnaces are longer-lived: expected life of 22.7 years against 18.6 for central cooling and 16.1 for a ducted heat pump.
A single rule applied to both is applying a cooling-equipment threshold to a furnace, which is one reason the rule feels wrong to technicians more often than it feels wrong on paper.
What is the threshold at each age and discount rate?
| Equipment | Age | Expected remaining life | Threshold, undiscounted | at 3% | at 5% | at 8% |
|---|---|---|---|---|---|---|
| Central furnace (all fuels) | 2 yr | 20.7 yr | 91% | 94% | 95% | 97% |
| Central furnace (all fuels) | 4 yr | 18.8 yr | 83% | 87% | 90% | 93% |
| Central furnace (all fuels) | 8 yr | 15.4 yr | 68% | 75% | 79% | 84% |
| Central furnace (all fuels) | 12 yr | 12.6 yr | 55% | 63% | 68% | 75% |
| Central furnace (all fuels) | 16 yr | 10.3 yr | 45% | 54% | 59% | 66% |
| Central furnace (all fuels) | 20 yr | 8.4 yr | 37% | 45% | 50% | 58% |
| Central furnace (all fuels) | 24 yr | 7.0 yr | 31% | 38% | 43% | 50% |
| Central air conditioner or central heat pump | 2 yr | 16.7 yr | 90% | 92% | 93% | 95% |
| Central air conditioner or central heat pump | 4 yr | 14.9 yr | 80% | 84% | 87% | 90% |
| Central air conditioner or central heat pump | 8 yr | 11.8 yr | 63% | 70% | 73% | 78% |
| Central air conditioner or central heat pump | 12 yr | 9.4 yr | 51% | 57% | 62% | 68% |
| Central air conditioner or central heat pump | 16 yr | 7.6 yr | 41% | 48% | 52% | 58% |
| Central air conditioner or central heat pump | 20 yr | 6.2 yr | 33% | 40% | 44% | 50% |
| Central air conditioner or central heat pump | 24 yr | 5.2 yr | 28% | 34% | 37% | 43% |
| Ducted central heat pump | 2 yr | 14.3 yr | 89% | 91% | 92% | 94% |
| Ducted central heat pump | 4 yr | 12.8 yr | 79% | 83% | 85% | 88% |
| Ducted central heat pump | 8 yr | 10.4 yr | 64% | 70% | 73% | 77% |
| Ducted central heat pump | 12 yr | 8.6 yr | 54% | 60% | 63% | 68% |
| Ducted central heat pump | 16 yr | 7.3 yr | 46% | 52% | 55% | 61% |
| Ducted central heat pump | 20 yr | 6.4 yr | 39% | 45% | 49% | 54% |
| Ducted central heat pump | 24 yr | 5.6 yr | 35% | 40% | 44% | 49% |
Maximum repair cost worth authorizing, as a percentage of a replacement quote, by equipment class, age and discount rate. Bold cells are within 2.5 points of the 50% convention. Expected remaining life is shown at the undiscounted basis; it does not change with the discount rate, only the annualization does.
The discount rate is the reader’s, not ours, and it moves the answer a long way.
At zero the 50% crossing for a central air conditioner or heat pump falls at 12.2 years; at 5% it is 16.9 years and at 8% it is 20.0.
A higher discount rate favors repair, because it discounts the far-off benefit of a long-lived new machine. Someone financing a replacement should use their borrowing rate.
Do energy savings from a new system change the answer?
Yes, and always toward replacement. The threshold above compares purchase costs only. It leaves out that the new machine may be cheaper to run.
The adjustment is simple. If replacing saves ΔC a year in energy, the replacement option’s annual cost falls by ΔC, and the threshold falls by ΔC × R(a) / Cn.
The size of that shift depends entirely on your replacement quote, which is why it cannot be tabulated in dollars here: HyreHVAC holds no equipment price dataset and does not publish installed prices.
It can be sized against a quote you already have. Take the $190-a-year figure computed on the working-system comparison: a 30% cut in cooling energy plus an 80%-to-95% furnace change, on national average use and EIA 2024 prices.
For a central air conditioner or heat pump at fifteen years old, with 8.0 years of expected remaining life, the threshold then falls by about 15 percentage points against a $10,000 replacement quote, 25 points against a $6,000 one, and 10 points against a $16,000 one.
Those quote figures are illustrative arithmetic on a number you would supply, not our estimate of what a system costs.
HYRE analysis: The shift is real but it is second-order against the age effect, which spans 52 percentage points across the equipment’s life.
And it only applies where the efficiency gain is real: an old label reading SEER cannot be divided into a new one reading SEER2, because they are different test procedures. The comparison page handles that trap in full.
How did we derive the threshold?
Two-parameter Weibull functions fitted to the EIA RECS 2020 equipment-age distribution at 1.5% assumed annual stock growth, published in full with their residuals and sensitivity on /research/equipment-lifespan-evidence/. This page introduces no new empirical input; it is arithmetic over that study’s output.
R(a) = ∫ from a to ∞ of S(t) dt, divided by S(a): the mean residual life of a unit that has already survived to age a. Integrated numerically to an 80-year horizon at 20,000 steps, which is stable to well under 0.05 years, an order of magnitude finer than anything printed.
Equivalent annual cost. Repair is the cheaper order when C_r / A(R(a), d) < C_n / A(L, d), where A(n, d) = (1 − (1+d)^−n) / d and A(n, 0) = n. The threshold reported is that inequality rearranged into C_r / C_n.
L is the fitted mean, not the median: 22.7 years for a central furnace, 18.6 for central cooling equipment, 16.1 for a ducted heat pump. The mean is the right quantity in an annualization because it is what the buyer receives in expectation.
Central furnaces, central air conditioners and heat pumps, and ducted heat pumps. Boilers, window units and storage water heaters are excluded: the first two fit the survival model badly, as the lifespan study states, and none of the three is the subject of the repair-or-replace quote this page is about.
The output is a ratio of the reader’s two quotes. HyreHVAC holds no equipment price dataset, publishes no installed prices, and the illustrative quote levels used to size the running-cost adjustment are labeled as the reader’s numbers, not ours.
Every threshold, crossing age and remaining-life figure on this page is computed at build time from the same module that writes the downloadable CSV. A number in the prose and the same number in the table cannot disagree.
What are the limits of this threshold?
R(a) is the expected remaining life of a randomly chosen unit that has reached age a. The machine you are being quoted on has just failed, which is evidence about it that the survival function does not contain.
Its true remaining life is shorter than R(a), so every threshold on this page is an upper bound.
How much shorter depends on what failed: a capacitor says almost nothing, a compressor or a cracked heat exchanger says a great deal.
The failure-type judgment lives on the repair or replace calculator; this page cannot supply it.
No unit is observed failing; the survival function is inferred from the age distribution of survivors; and the medians move by several years across the plausible range of the stock-growth assumption.
A threshold derived from an uncertain survival function is an uncertain threshold, and the crossing ages here should be read to the nearest year at best.
The 50% crossing for a central air conditioner or heat pump moves from 12.2 years at a zero discount rate to 20.0 at 8%. We publish four rates rather than choosing one, because choosing one would hide the sensitivity behind a decimal point.
The model treats a repaired unit as continuing on its original survival curve, which is the neutral assumption.
If a major component replacement genuinely extends life beyond that curve, the threshold is too low; if the repair is a patch on a failing system, it is too high. No public dataset resolves this for residential HVAC.
Comfort, an unsafe condition, a refrigerant that is no longer produced, the disruption of a failure in February, and a house-wide project that makes the work cheap to fold in are all real and none of them is in this arithmetic.
A cracked heat exchanger is a safety decision, not an economic one: if you smell gas or a carbon monoxide alarm sounds, leave the house and call your gas utility or 911 from outside.
The ratio is only meaningful if the replacement quote is for equipment that actually meets the load and the repair quote is for a repair that actually fixes the fault. A replacement quote for an undersized system and a repair quote that will be followed by another are not two sides of the same comparison.
Questions
Is the 50% repair-or-replace rule correct?
What is the correct repair-or-replace threshold for my system?
Why does the threshold not fall to zero on very old equipment?
Does the rule differ between a furnace and an air conditioner?
Should energy savings from a new system change the threshold?
What discount rate should I use?
Does this tell me what a repair or a replacement should cost?
Written and audited by
HYRE HVAC Research Desk
Primary-source research, data analysis and fact checking
We are a research desk, not a sales floor. We read the federal microdata file, the statute or the manufacturer data sheet ourselves, and we publish the figure with the document it came from and the date we retrieved it.
Where a number cannot be traced to a primary source, we publish the shorter page and say what we could not verify.
The counts below are generated from the published pages themselves, last counted September 28, 2026, and they are what we have actually published rather than what we intend to.
- 13
- studies published
- 12
- federal sources read and cited
- 8
- studies published with their full dataset as CSV
- 51
- jurisdictions reproduced against EIA’s own tables
How this desk works
- Primary sources only. Federal data comes from the agency that collects it, in the file that agency publishes. We do not cite an article that cites a source; we download the source and compute the figure ourselves.
- We validate against the agency before we publish. First, we use each federal microdata file to reproduce the agency’s own published tables. Our cooling research reproduces EIA’s state estimates and standard errors for all 51 jurisdictions. That check caught a variance formula that was off by a factor of four.
- Every estimate carries its uncertainty. These are survey figures, not counts. Standard errors are computed from the replicate weights the federal file supplies and printed beside the estimate. An estimate too imprecise to publish is reported as such rather than printed.
- Nothing is typed by hand. Prose, tables and charts all read from one dataset built by script, so a number in a sentence and the same number in the table below it cannot disagree.
- We publish the data, not just the conclusion. 8 of our 13 studies offer the full computed table as a CSV download on the page, so you can check the analysis or disagree with it. A study without a row-level dataset gets no download link and claims none in its structured data.
- We correct in public. Where we have published a figure wrongly we fix the figure, rewrite any analysis that rested on it rather than patching the number underneath it, and leave a dated correction note on the page.
- We do not install or sell HVAC equipment, and we take no payment for placement, ranking or a favorable mention. Nobody buys a position on this site.
Data as of Survival functions fitted to EIA RECS 2020 public use microdata v7. Authorship on this site is organizational: the analysis belongs to the desk rather than to a named individual, and we do not publish credentials we do not hold.
Our editorial policy sets out how we source, date and correct what we publish.
The data behind this page
Every figure on this page is computed from one file, and that file is published here so the analysis can be checked, disagreed with, or reused.
Sources & retrieval dates
Put your own two quotes through it
The threshold is a ratio, so it needs your repair quote, your replacement quote and the age of the machine. The calculator adds the failure-type judgment this analysis deliberately leaves out.