Is repairing really greener than replacing? (the carbon math)

Evidence grade: A2 — Institutional literature

Answer: Usually, by a wide margin — but not by magic. The carbon ledger has three big lines: making a new product (mining, factories, shipping — the "non-use" or embodied footprint), using it, and the repair itself. Published life-cycle studies put manufacturing a new washing machine, TV, or phone in the tens to hundreds of kilograms of CO₂e, while the repair's materials cost a small fraction of that — so repair usually wins. The honest exception: products whose footprint is dominated by electricity use, where keeping an inefficient old unit running can cost more carbon than building its efficient successor.

Effort: N/A — reference, no procedure

The three lines of the ledger

Every repair-vs-replace decision is really this comparison:

1. Non-use phase of the new product — raw material extraction, manufacturing, distribution, and end-of-life disposal. The EEB's Coolproducts analysis "considers the energy needed to produce and distribute new products and dispose of old ones — the so-called nonuse phases, distinguishing from when the product is in use or the use phase," noting this manufacturing footprint "is often overlooked in climate assessments" [1].
2. Use phase — the electricity (and water) the product consumes over its life. For some products this is a footnote; for others it's nearly the whole story.
3. The repair itself — the adhesive, spare part, sandpaper, and the trip to fetch them. This side of the ledger is qualitatively far below the replacement side — but the honest caveat belongs here: no published, consumer-facing LCA gives a carbon figure for a specific repair adhesive, so the repair side of this article stays qualitative rather than pretending at false precision. (Precise adhesive-level carbon figures: unverified — not found in published literature.)

The published numbers: repair usually wins big

The European Environmental Bureau modeled what happens if the EU simply keeps its appliances and electronics one year longer [1]:

Across just these four product groups, one extra year of life would save around 4 million tonnes of CO₂ annually — the equivalent of taking over 2 million cars off the roads [1]. And the study stress-tested the obvious objection, assuming "very optimistic annual efficiency improvements of 5%" in new models. The conclusion: "Put simply, in terms of GWP, it's always better to repair than to replace one of these products" [1]. In reality, the footnote adds, washing-machine and vacuum efficiency improves only ~1% per year, and "for smartphones and notebooks, new products actually tend to consume more energy than their predecessors" — so the use-phase argument barely applies to electronics at all [1].

WRAP's reuse case studies tell the same story in per-item terms: providing one tonne of washing machines for direct reuse yields a net saving of 0.5 tonnes CO₂-eq — "just over 30 kg CO₂-eq per machine" — while even the preparation-for-reuse route (collection, refurbishment, resale) nets 0.2 tonnes, about 12 kg per machine [4]. For TVs the savings are larger still: one tonne reused directly avoids 8 tonnes CO₂-eq, "just over 100 kg CO₂-eq per TV" [4]. Current UK reuse levels already avoid 2,600 tonnes CO₂e per year from washing machines and 156,000 tonnes from TVs [4].

When repair loses: the use-phase exception

Here is where honesty requires the other side. MIT researchers ran life-cycle assessments on remanufacturing refrigerators, dishwashers, and clothes washers versus buying new, and found that "despite savings achieved in production, appliance remanufacturing is a net energy-expending end-of-life alternative" [2]. The reason: "the use phase constitutes 88% to 95% of the total life cycle energy use of the appliances" [2]. When a new generation is dramatically more efficient, each extra year of the old unit's electricity swamps the manufacturing savings:

• Remanufacturing a 1992-model clothes washer instead of buying a 2003 model expended 44% more energy per cubic meter over the life cycle [2].
• Remanufacturing a 2008 refrigerator expended 32% more energy than buying new [2].
• Raw materials and manufacturing are only 5–12% of an appliance's life-cycle energy — so the production side is the small line, not the big one, for these products [2].

A University of Michigan life-cycle optimization study on refrigerator-freezers reached the same shape of answer: the energy-optimal replacement lifetime can be as short as 2–12 years, and "scrappage programs encouraging retirement of old, inefficient refrigerators can serve to reduce household energy consumption" [3].

There is also a subtler version of the exception on the repair side. WRAP's sensitivity analysis shows that if a washing machine needs more than 2 refurbishments to achieve its second life, "the avoided greenhouse gas emissions from preparation for reuse switch to become net emissions" [4]. A repair that keeps failing is not green — it is a series of carbon spends that buys no useful life. And WRAP cautions that its own tool's results depend on primary data it considers not yet robust, so product-level figures should be treated as order-of-magnitude, not gospel [4].

Run the math on your own repair: three questions

1. What's the embodied carbon of the replacement? For appliances and electronics, published LCAs put this in the tens to hundreds of kg CO₂e — the dominant line for anything that doesn't plug in, and a serious line even for things that do. That is the amount your repair has to beat.
2. How much of the old product's footprint is electricity? If it's a tool, a chair, a phone, a pipe fitting — the use phase is trivial and the repair wins almost by inspection. If it's an old fridge, dryer, heater, or air conditioner, compare the efficiency gap against the years you'd keep it: a 1990s refrigerator can use roughly twice the energy of a modern one, and at that gap the new unit's manufacturing debt can pay back [2][3].
3. How big is the repair's own footprint? A cartridge of epoxy and a spare part are small beside manufacturing a whole product (exact figures unverified). The red flag is repetition: heavy parts, multiple attempts, or shipping a bulky item back and forth can erode the advantage — WRAP's two-refurbishment line is the published version of "know when to stop" [4].

One last accounting note: recycling is not repair. Recycling a broken product and buying new still pays the full manufacturing footprint of the replacement — the energy to melt, reprocess, and rebuild. The carbon win comes from keeping the product alive, not from keeping its materials in the loop [1].

Honest limits

• These are scenario studies, not your product's number. EU-wide models, UK case studies, and US appliance data give the shape of the answer — the order of magnitude — not a carbon figure for your specific repair.
• MIT's appliance study measures energy, not carbon. Energy ≠ CO₂e: where the grid is clean, the use-phase penalty of an old appliance shrinks; where it's coal-heavy, it grows. The direction holds; the payback point moves with your electricity mix.
• WRAP's own caveat applies. Its researchers wrote that the reuse-benefits tool needs better primary data "to make the conclusions more robust" — treat per-item figures as estimates, not measurements [4].
• A failed repair buys nothing. Repair wins only when it restores real, useful life. WRAP's finding that benefits turn negative past two refurbishments is the quantitative form of that rule [4].
• Adhesive-level carbon figures are unverified. No published LCA was found giving a consumer-facing CO₂e figure for a specific repair adhesive; the "repair side is small" claim in this article is qualitative, not measured.

Sources

1. European Environmental Bureau — "Coolproducts don't cost the earth: they don't make them like they used to" (briefing, 2019). Extending the lifetime of all washing machines, notebooks, vacuum cleaners, and smartphones in the EU by one year would save "around 4 million tonnes of carbon dioxide (CO₂) annually by 2030, the equivalent of taking over 2 million cars off the roads for a year"; manufacturing/distribution/disposal account for ~72% of a smartphone's total climate impact, ~52% of a notebook's, ~25% of a washing machine's, ~21% of a vacuum cleaner's; smartphone +1 yr = 2.1 Mt CO₂/yr, notebook +1 yr = 1.6 Mt, washing machine +1 yr = 0.25 Mt, vacuum +1 yr = 0.1 Mt; "Put simply, in terms of GWP, it's always better to repair than to replace one of these products"; tested against "very optimistic annual efficiency improvements of 5%"; in reality washer/vacuum efficiency improves ~1%/yr and "for smartphones and notebooks, new products actually tend to consume more energy than their predecessors." URL: https://eeb.org/wp-content/uploads/2019/09/Coolproducts-briefing.pdf
2. Boustani, A., Sahni, S., Graves, S.C. & Gutowski, T.G. — "Appliance Remanufacturing Life Cycle and Energy Economic Savings" (MIT, IEEE ISSST paper). "The results show that, despite savings achieved in production, appliance remanufacturing is a net energy-expending end-of-life alternative"; "the use phase constitutes 88% to 95% of the total life cycle energy use of the appliances"; remanufacturing a 1992 clothes washer vs a new 2003 unit meant "expending 44% more energy per cubic meter"; remanufacturing a 2008 refrigerator expended "32% more energy... compared to a new refrigerator"; "raw materials processing and manufacturing constitute 5 to 12% of appliance life cycle energy." URL: http://web.mit.edu/ebm/www/Publications/57_Paper.pdf
3. Horie, Y.A. (2004) — "Life Cycle Optimization of Household Refrigerator-Freezer Replacement." Center for Sustainable Systems, University of Michigan, CSS04-13. "The optimal lifetime for the energy objective ranges between 2 to 12 years"; "scrappage programs encouraging retirement of old, inefficient refrigerators can serve to reduce household energy consumption." URL: https://css.umich.edu/publications/research-publications/life-cycle-optimization-household-refrigerator-freezer
4. WRAP — "Benefits of Reuse: Case Study — Electrical Items" (Nov 2011). "Providing 1 tonne of washing machines for direct reuse... can result in a net GHG saving of 0.5 tonnes CO2-eq. This is just over 30kg CO2-eq per machine"; preparation-for-reuse route: 0.2 tonnes net, "about 12kg CO2-eq per washing machine"; TVs: 8 tonnes CO₂-eq per tonne for direct reuse ("just over 100kg CO2-eq per TV"); current UK reuse avoids 2,600 tonnes CO₂e/yr (washing machines) and 156,000 tonnes (TVs); "if more than 2 refurbishments are required, the avoided greenhouse gas emissions from preparation for reuse switch to become a net emissions"; the project "clearly indicated the need to improve the quality of the primary data used in the tool to make the conclusions more robust." URL: https://www.wrap.ngo/sites/default/files/2020-09/WRAP-Electricals%20reuse_final_0.pdf

Repair Lab · Release 1 · Evidence last checked September 30, 2026 · A project of AOJEL