The Repair Economy

Volume IV — Production and Regeneration

THE PLATE

A woman seen from behind, reading on a cushioned bench that looks out over a lotus pond toward a garden pavilion. — Plate IV.08The Bench at Four O'Clock.Repair is the only branch of production where the labour is visible, which is exactly why it is the only branch of production that is priced honestly.

THE LETTER

There is a sentence people say about repair that is almost true, and the gap between almost and true is where this chapter lives. The sentence is: it would be cheaper to fix things than to replace them, if only manufacturers let us.

The second half is a real fact about the world. Parts are withheld, schematics are proprietary, fasteners are exotic, and software refuses to recognise a component it did not personally introduce. All of that is documented, and it is the subject of the largest wave of consumer legislation in forty years.

The first half is not a fact. It is a question, and it has an arithmetic answer that changes by product, by part price, by how long the repaired thing will last — and, above all, by the wage of the person doing the work. For a smartphone display the answer is an emphatic yes and has been for years. For a mid-market washing machine it is a near-run thing that turns on about twenty pounds an hour. For a twenty-five euro kettle the answer is no, and no policy should try to make it otherwise.

So this chapter does not argue for repair. It hands you the function that decides it, computes it for three real products, and then follows the function where it goes — which is somewhere uncomfortable, because the term that dominates it is labour, and labour is the input that is expensive precisely where the environmental argument is loudest.

You will leave with one instrument you can run on any product in your portfolio in about ten minutes, the real policy record of the three jurisdictions that have tried this, and a clear-eyed answer to the question a finance director will ask first: does a repair line make money, and if so, whose money.

— The Editors

DISCOVERY

What is already working

Repair is not a proposal. It is one of the largest employers in the industrial world, and it has been profitable for a century in the categories where the arithmetic works. Begin with what already stands.

One occupation, in one country. Automotive service technicians and mechanics in the United States number 825,800, at a median of $50,620 a year, or $24.34 an hour, with employment projected to rise 5.0 percent over the decade to 2035. At the median that is a payroll of $41.80 billion a year paid to people whose entire job is keeping existing objects in service. No subsidy built it, no directive compelled it, and no one describes it as a circular economy initiative. It is simply the largest repair industry on earth and it has been operating at scale since before anybody had a word for it.

Renault, since 1949. Renault has remanufactured mechanical parts since the Choisy-le-Roi plant opened, and it sells those parts at 30 to 50 percent below new. The operation moved into the Re-Factory at Flins — 11,000 square metres, 180 vehicles a day, which at 250 working days is 45,000 a year. The circular business was then spun into a named subsidiary, The Future Is NEUTRAL, whose turnover has passed €1,000 million and which targets turnover above €2,300 million at an operating margin above 10.0 percent by 2030 — an implied operating profit of €230.0 million. Note what kind of document that is. It is not an environmental commitment. It is a margin guidance, published by a listed manufacturer, for a business made entirely of things other people had finished with.

Back Market. The European refurbished-electronics marketplace transacted $3,500.0 million of gross merchandise value in 2025, against $2,800.0 million the year before, on revenue of $495.0 million against $415.0 million — revenue growth of 19.28 percent at a take rate of 14.14 percent. It reached global EBITDA break-even for the first time. Its mature markets are better than that: France runs at a 35.0 percent EBITDA margin, Spain at 20.0 percent on €450.0 million of GMV and €55.0 million of revenue, a take rate of 12.22 percent. Second-hand is not a charity sector. It is a marketplace business with marketplace economics.

The community layer, measured. The Open Repair Alliance publishes what happens at volunteer repair events, and the data set is now real: 208,491 recorded repair attempts by 1,158 groups across 31 countries at 19,986 events — 10.43 attempts an event — with a fix rate of 53.0 percent, so 110,500 things that were going to be thrown away were not. The following release carried 305,649 records, 46.6 percent more in a year. There are over 2,500 Repair Cafés worldwide. This matters less for the tonnage than for the evidence: it is the only public data set in existence that records what actually breaks, by brand and by component, and it was built by volunteers with clipboards.

The household prize. The PIRG Education Fund costed the consumer side: American households spend $1,767.00 a year buying new electronics, a 21.6 percent reduction is achievable through repair, which is $381.67 a household — and across 129,000,000 households, $49.24 billion a year. That is not a saving to the economy. It is a transfer, from manufacturers to households and to the people who do the repairing, which is why it is contested and why it took legislation.

Five things already working, and notice the pattern: not one of them is an argument about waste. Each is an argument about labour, margin, data or price. The environmental case for repair is genuine and it is not what has built any of these. What built them is that somebody found a category where the arithmetic already worked, and then did the arithmetic.

THE ARITHMETIC

The threshold, and where it fails

Repair-versus-replace has a closed form, and once you have it you will never again argue about it in the abstract.

The setup. Compare cost per year of service delivered. A repair costs the labour, the part and the fixed overhead of the job, and it buys some residual life. A replacement costs the new price and buys the life of a new unit.

  w    charge-out labour rate        currency / hour
  h    repair hours, incl. diagnosis  hours
  p    part cost                      currency
  k    fixed cost of the job          currency  (callout, logistics, bench)
  q    first-time fix probability     dimensionless
  L_r  residual life after repair     years
  L_n  expected life of a new unit    years
  C_n  replacement price              currency

q is in there because failed attempts are paid for too. Then:

                (w·h + p + k) · L_n
        R  =  -------------------------          repair when  R < 1
                   q · L_r · C_n

And rearranged for the variable that actually moves — the wage:

          q · L_r · C_n / L_n  −  p  −  k
   w*  =  --------------------------------
                        h

w* is the threshold labour rate. Below it, repair is cheaper per year of service. Above it, replacement is. It is the only number in this chapter you need to carry.

Case one — a drain pump in a UK washing machine. A mid-market machine costs £400.00 new and lasts 11 years; the repair buys 5 years of residual life; the pump is £45.00; the callout is £60.00; the job takes 1.20 hours; the trade fixes it first time 90 percent of the time.

  A  = q·L_r·C_n / L_n  =  0.90 × 5 × 400 / 11   =  £163.64
  w* = (163.64 − 45 − 60) / 1.20                 =  £48.86 / hour

At £45.00 an hour, R = 0.9717 and repair wins. At £70.00 an hour, R = 1.1550 and replacement wins. The whole decision sits inside a twenty-five-pound band of hourly rate, and nothing about the machine changed.

Case two — a smartphone display. New phone $799.00, life 4 years, repair buys 2.5, display module $120.00, $15.00 of handling, 0.75 hours, fixed first time 95 percent of the time.

  A  = 0.95 × 2.5 × 799 / 4  =  $474.41
  w* = (474.41 − 120 − 15) / 0.75  =  $452.54 / hour

Four hundred and fifty dollars an hour. At a $132.00 bench rate, R = 0.4932 — repair costs less than half. This is why phone repair is a real industry with real shops and no subsidy, and it is worth seeing that the reason is not sentiment about e-waste. It is that C_n is high, h is short, and the threshold is therefore nowhere near any wage anyone pays.

The honest negative, first form. Run the same instrument on a €25.00 kettle: life 5 years, repair buys 2, part €6.00, half an hour, fixed 80 percent of the time.

  A  = 0.80 × 2 × 25 / 5  =  €8.00
  w* = (8.00 − 6.00) / 0.5  =  €4.00 / hour

Four euros an hour. No legal wage in any member state clears it. Repair loses here, it should lose here, and a subsidy that makes it win is buying half an hour of a skilled person's life to save six euros of moulded plastic. A repair economy that will not say this about the kettle cannot be believed about the washing machine.

The honest negative, second form — and this is the one that governs. Go back to the washing machine and ask what wage w* = £48.86 implies for the person holding the screwdriver. A charge-out rate is not a wage: it carries premises, insurance, the van, the parts inventory, the diagnostic subscription and the unbilled hours. In the United States, where both figures are published, the average shop labour rate is $132.00 an hour against a median technician wage of $24.34 — a multiple of 5.4232. Apply it:

  implied technician wage  =  48.86 / 5.4232  =  £9.01 / hour
  UK National Living Wage, April 2026        =  £12.71 / hour
  implied wage as a share of the legal floor =  70.9 %

The repair of a four-hundred-pound washing machine, done properly, cannot legally be a job in the United Kingdom. The minimum viable charge-out rate at the living wage is £68.93 an hour; the threshold is £48.86; the shop is short £20.06 an hour, or £24.08 on the job.

That is the sentence this chapter exists to make sayable. It is not an argument against repair. It is the statement of the constraint, and the constraint is distributional: repair is labour-intensive, and it is therefore structurally strongest exactly where labour is cheapest. The repair economy is not evenly available. It is easiest to build in low-wage economies, hardest in high-wage ones, and the countries legislating most energetically for it are the countries where the arithmetic is tightest. Any account of repair that does not say this is selling something.

The turn. Here is the figure that should change your mind about who the repair economy pays.

Back Market and Swappie sell the same thing — a refurbished phone. Back Market is a marketplace: it takes roughly ten percent of GMV, employs no technicians, touches no device, and reached EBITDA break-even globally with a 35.0 percent EBITDA margin in France. Swappie is a refurbisher: it buys the phones, employs the people who open them, and in 2024 turned €248.0 million of revenue into a net loss of €21.0 million — a net margin of −8.47 percent, reported as −8.5 percent, improved from −10.6 percent on a €22.0 million loss the year before.

The one profitable company in the refurbished-phone economy is the one that never touches a phone. The gap between the two margins is 43.47 points, and the whole of it is the labour term in the equation above. This is not a criticism of either firm. It is the instrument, confirmed in public accounts: value in the repair economy accrues to whoever is not doing the repairing, unless somebody deliberately arranges otherwise. Arranging otherwise is what the rest of this chapter is about.

One more number, for scale. Employment intensity per 1,000 tonnes of material: re-use and repair around 70.00 jobs, recycling 0.92, landfill 0.28 — 76.1× and 250.0× respectively. Treat that as directional and not as like-for-like: the re-use figure counts tonnes collected by social enterprises including their retail staff, the disposal figures count direct FTE against tonnes processed. The boundaries genuinely differ and the honest version of the claim is the weak one: repair is the most labour-intensive treatment of a broken object by one to two orders of magnitude. That is the same sentence as repair is the most expensive treatment where labour is expensive. Both are true and they are the same fact.

DREAM

What becomes ordinary

In the economy where this has been worked out, the repair decision is made with a number rather than a mood.

A product page carries the threshold rate alongside the price. Not a repairability score out of ten — the actual hourly rate below which repair is the cheaper year of service, computed from the manufacturer's own published part prices and service times. Two machines at the same price have different thresholds, and the buyer can see it, because the threshold is the thing they will actually experience in eight years' time when the pump goes.

Residual life is a published figure. Manufacturers state L_r for each major repair the way they state energy consumption, and they are held to it, because it turns out to be the single most powerful term in the equation: on the washing machine, lifting residual life from 5 to 8 years moves the threshold from £48.86 to £130.68 an hour. Nothing else comes close. Halving the hours gets £97.73. Abolishing the callout gets £98.86. Halving the price of the part — the thing every campaign asks for — gets £67.61. Durability is a labour subsidy, and it is the largest one available.

Repair work is a trade with a ladder. An apprentice diagnoses, a journeyman fixes, a master rebuilds, and the pay reflects the fact that a person who can find a fault in nine minutes is worth several who cannot. The trade has a register, portable qualifications and a body that defends its rates, in the way that electrical work does and for the same reason: it is skilled, it is consequential, and it cannot be offshored.

The manufacturer runs the repair line and is glad to. It sells the part at a margin, books the service revenue, keeps the telemetry on what fails and feeds it back into design — and it holds the customer relationship for the whole life of the object rather than for the ninety days of the warranty. The performance contract is normal: the firm sells clean laundry, not laundry machines, and therefore wants the machine to last, which is Walter Stahel's argument from 1976 arriving about fifty years late and entirely intact.

And the second-hand price of a well-made thing is high, publicly quoted, and part of how people decide. Durability shows up in the resale column the way fuel economy shows up in the running-cost column, which means it finally shows up in the purchase decision — which is the only place it has ever needed to show up.

DESIGN

The four levers, and the policy record of each

The equation gives you exactly four ways to make repair win, and they are not equally powerful. Here they are on the washing machine, each acting alone.

Lever	Change	`w*` becomes	Movement
Residual life `L_r`	5 → 8 years	£130.68	+£81.82
Fixed cost `k`	£60 → £0, drop-off	£98.86	+£50.00
Hours `h`	1.20 → 0.60	£97.73	+£48.86
Part price `p`	£45 → £22.50	£67.61	+£18.75

Now read the policy record against that table, which is the point of having it.

The European Union. Directive (EU) 2024/1799, adopted 13 June 2024, published 10 July 2024, in force 30 July 2024, applying from 31 July 2026. It obliges manufacturers to repair at a reasonable price and within a reasonable time, inside and outside the guarantee; obliges them to supply parts and tools to any repairer, professional or not; gives the consumer 12 months of extra legal guarantee when they choose repair over replacement; and builds a European repair platform, common interface by 31 July 2027 and fully operational by 1 January 2028. Light-means-of-transport batteries join on 18 February 2027.

Read honestly: the directive is strong on p — the weakest lever — and on the information asymmetry, and it is silent on L_r, h and k. Its scope is Annex II, which reaches only goods that already carry ecodesign repairability requirements: Annex II is a list, not a right. Vacuum cleaners are named in it and no repairability requirements exist for them, so the naming does nothing. The twelve-month guarantee extension is the exception and it is the cleverest thing in the instrument, because it acts on L_r from the consumer's side — choosing repair now buys a warranted year, which is exactly the term the arithmetic says is decisive.

France. The indice de réparabilité — scale 0 to 10, five criteria, five product categories, mandatory from 1 January 2021, and from 7 January 2025 replaced for televisions and washing machines by a durability index. The government evaluated its own instrument using more than 20,000,000 sales lines over three years from two retailers. The result: the effect on sales of more repairable products was positive and statistically significant online, and positive and not statistically significant in store. Awareness reached 55.0 percent of respondents, and 66.0 percent of consumers called it useful.

That is an honest and rather modest result for the most admired repair policy in the world, and the reason is visible in the table above: an index acts on information, and information is not one of the four levers.

What did move. The bonus réparation — a direct subsidy at the till, €25.00 for a washing machine at launch, doubled to €50.00 in 2024. Repairs funded went from 181,000 in 2023 to 715,227 in 2024, a factor of 3.952, with payouts rising from €4,000,000 to €24,300,000 — an average of €22.10 a repair in 2023 and €33.98 in 2024.

And now compute what size of subsidy the arithmetic says is needed. The shop must clear the minimum viable charge-out of £68.93; the threshold is £48.86; so the bonus B that closes the gap on this job is:

  B  =  68.93 × 1.20  +  45  +  60  −  163.64  =  £24.08

Twenty-four pounds and eight pence, against a launch bonus of twenty-five euros. Sterling and euro are not the same unit and the machine is UK-priced, so that is an agreement of magnitude and not an identity — but the magnitude is the finding. France did not guess. It picked a number about the size of the gap, and the volume moved by a factor of four when the number doubled.

The honest negative, third form. The fund is nonetheless underspent. The electrical and electronic envelope is €410.0 million over six years, of which €130.0 million was budgeted for 2022 to 2024 and €29.3 million actually paid — 22.5 percent uptake. A subsidy correctly sized and poorly reached is still a subsidy nobody used, and the binding constraint on reach is k: the consumer has to find an accredited repairer, carry the thing there, and claim. Which is the second-strongest lever in the table, unaddressed.

The United States. Consumer-electronics right-to-repair is now a working patchwork rather than a campaign. California's SB 244 and Minnesota's Digital Fair Repair Act came into force on 1 July 2024; California requires parts, tools and documentation for 3 years on goods priced $50.00 to $99.99 and 7 years at $100.00 and above. Washington and Colorado took effect on 1 January 2026, Connecticut on 1 July 2026, Texas on 1 September 2026. Oregon went furthest and banned parts pairing — the practice of using serialised software to refuse a component the manufacturer did not supply — for devices made after 1 January 2025.

The parts-pairing ban is the only measure in any of the three jurisdictions that acts on h, because a pairing check is not a part cost, it is an hour cost: it turns a twenty-minute swap into a diagnostic dead end. On the arithmetic, Oregon's is the most economically consequential right-to-repair law yet written, and it is the one that got the least attention, because banning a software behaviour photographs badly and a score out of ten photographs well.

DESTINY

How a repair line holds

A repair business sustains itself on three things, and fails on the absence of any one.

Volume density, not volume. Repair economics are dominated by k — the cost of getting the broken thing and the skilled person into the same room. A shop with four hundred jobs a year drawn from three postcodes is a business; the same four hundred jobs drawn from a county is a van with a person in it losing money. Density is the asset. It is why the strongest repair networks are urban, and why the mail-in model exists: posting a phone is cheap, posting a washing machine is not, and the categories sort themselves accordingly.

A parts position. The margin in every durable repair business is in parts, not hours, because hours are capped by the wage and parts are not. A repairer without a parts position is selling labour at cost and will not survive a recession. This is also why the directive's parts-access provisions matter more to the shop's survival than to the consumer's bill.

Diagnostic capital. The difference between a good technician and an average one is not manual skill, it is time-to-fault, and time-to-fault is h. A firm that records every fault it has ever seen, by model and symptom, compounds a genuine asset — which is precisely what the Open Repair Alliance's 208,491 records are, given away.

Now the failure modes, named.

It fails when the shop takes categories the arithmetic refuses — the kettles, the twenty-pound accessories — because the customers are pleasant and the jobs are quick, and each one loses money slowly.

It fails when a subsidy is treated as revenue rather than as a temporary correction to k and p. A repair line that clears only with the bonus is a policy-dependent business, and policy changes; the 22.5 percent uptake figure above is a reminder that the money moves in both directions.

It fails when the wage is set below the trade rate to make the arithmetic work. That is not a business model, it is a countdown: the good technicians leave, time-to-fault rises, h rises, and the threshold the shop was already failing to clear moves further away. Every repair business that solves its margin problem out of the wage is solving it by consuming the only stock it has.

And it fails, most quietly, when nobody ever computes w* at all — when repair is a conviction rather than a threshold, and the shop discovers which categories were losing money only in the year it closes.

DELIGHT

What it feels like

The pleasure of repair is not virtue. Virtue is a thin fuel and it burns out around the fourth identical fault of the day. The pleasure is diagnosis.

There is a particular quality of attention in the ten minutes between it does not work and I know why, and anyone who has had it wants it again. It is the same attention a doctor has, or a person tracing a fault in an argument: the object is telling you something and the whole task is to stop assuming and listen. When it arrives — when the symptom and the mechanism snap together — there is a small internal click that is one of the reliable joys of working life, and it costs nothing to provide.

Then there is the tray of eleven screws in the order they came out, and the machine closing up with nothing left over. There is the customer's face, which is not gratitude exactly but a kind of relief at having been told the truth about an object they own. And there is the specific satisfaction of a thing that was dead and is now working, which is available in no other branch of manufacturing, because manufacturing makes things that were never dead.

Repair is the only part of the economy where the good day and the profitable day are the same day. That is not a small structural fact. It is why the trade retains people at wages that the arithmetic says are too low, and it is also why it must not be asked to.

OPERATIONALIZE THIS

At the level of finance

The instrument: a repair line financed as a spare-parts working-capital facility, with a residual-life warranty as the security.

You are not asking for a sustainability budget. You are asking for inventory finance against a book of service revenue — a structure every asset-based lender already understands.

The mechanics.

Structure. A revolving working-capital facility secured on the spare-parts inventory, sized at the parts float plus ninety days of labour cost. Parts are good collateral: they are fungible, they have a published price, and they do not spoil.
The threshold gate. No category enters the line until w* has been computed for it at your own fully loaded charge-out rate, with q taken from your own first-time-fix data and not from the trade's. Categories where R ≥ 1 are declined and the decline is recorded. This is the covenant that keeps the line honest, and it is the one number the credit committee should see.
Balance-sheet treatment. Capitalise the diagnostic data set. A recorded fault library reduces h on every future job in that model family, has a measurable life, and is a genuine intangible asset. Your auditors will want evidence of the h reduction; the fault library produces it as a by-product. Parts inventory sits at the lower of cost and net realisable value as normal; reman cores are inventory, not scrap, from the moment a deposit is taken.
The core deposit. Charge a refundable deposit on the broken unit, refunded on return, exactly as Renault does for mechanical cores. It converts your supply of inputs from a hope into a receivable and it is the single most effective way to secure p.
Counterparty. The manufacturer first, as an authorised service partner: they carry the parts risk and the warranty liability, and after 31 July 2026 they carry a legal obligation to supply you. Insurers second — they are the largest buyer of repair in the world and they buy on cost per claim, which is R under another name. Consumers last, because they are the most expensive channel to acquire and the least predictable to schedule.
The residual-life warranty. Warrant L_r explicitly — twelve months, or the directive's extension, whichever is longer — and price it. A warranted residual life is the product. Without it you are selling an hour of someone's time; with it you are selling years of service, which is the unit the customer actually buys.

The number that decides it. One figure on the front page, per category:

                q · L_r · C_n / L_n  −  p  −  k
       w*  =  ----------------------------------   >   your loaded charge-out rate
                             h

If w* exceeds what you must charge to pay the living wage plus overhead, the category is a business. If it does not, the category is a subsidy, and you should say which one you are running.

The first ninety days.

Day	Action	Artifact
1–15	Pull `C_n`, `p`, `h` and warranty claims for your top ten categories	The input sheet
16–30	Compute `w*` and `R` for each; rank	The threshold table
31–45	Measure `q` on your own last hundred jobs, not the trade's average	First-time-fix baseline
46–60	Take the top two categories to a parts supplier; negotiate `p` and core deposits	Parts term sheet
61–75	Drop-off point live in the densest postcode, to attack `k`	The counter
76–90	First month's line P&L against the threshold table	The one page

APPRECIATIVE QUESTIONS

Twelve, for a room

Discovery — what is already working

Which thing in this building has been repaired rather than replaced more than once, and who decided that each time?
Where do we already sell service rather than product, and what did we learn about the customer from doing it?
Which of our people can find a fault fastest, and what do they do that the rest of us do not?

Dream — what becomes possible

If every product we sell published the hourly rate below which repairing it beats replacing it, which of ours would we be proudest to publish?
What would change in our design reviews if the residual life after the most common repair were a specified requirement rather than an outcome?
If our repair line were the most profitable thing we do in eight years' time, what would we have had to start recording this year?

Design — what we build

Which single category in our range has the highest threshold rate, and what is stopping us starting there next month?
Of the four levers — residual life, hours, fixed cost, part price — which one do we actually control most, and have we ever tried moving it deliberately?
Where is the densest cluster of our broken units geographically, and what would a counter there cost?

Destiny — how it holds

What would we have to pay a master technician for this to still be a trade people enter in ten years, and what would w* have to be to support it?
Which categories should we decline, and who here has the authority to decline one?
If the subsidy ended tomorrow, which parts of this line would still clear — and how would we know before tomorrow?

WORKS CITED

Directive (EU) 2024/1799 of the European Parliament and of the Council of 13 June 2024 on common rules promoting the repair of goods. Official Journal of the European Union, 10 July 2024.

European Commission. Directive on repair of goods — implementation and scope. Consumer protection law portal.

Regulation (EU) 2023/1670 of 16 June 2023 laying down ecodesign requirements for smartphones, mobile phones other than smartphones, cordless phones and slate tablets.

Loi n° 2020-105 du 10 février 2020 relative à la lutte contre le gaspillage et à l'économie circulaire (loi AGEC). République française.

Arrêté du 10 novembre 2023 portant diverses dispositions relatives aux fonds dédiés au financement de la réparation, du réemploi et de la réutilisation des produits relevant du principe de responsabilité élargie du producteur. Légifrance.

Commissariat général au développement durable and Direction interministérielle de la transformation publique. Évaluation de l'impact de l'indice de réparabilité sur les choix des consommateurs. Ministère de la Transition écologique, 2023.

Behavioural Insights Team. Leveraging Behavioural Insights to Design and Test the Repairability Index in France. BIT, 2023.

ADEME. Bonus réparation — bilan et barèmes. Agence de la transition écologique, successive years.

California Senate Bill 244 (2023), Right to Repair Act, Civil Code §1793.03.

Minnesota Digital Fair Repair Act, Minn. Stat. §325E.72 (2023).

Oregon Senate Bill 1596 (2024), Right to Repair.

Colorado House Bill 24-1121, Consumer Right to Repair Digital Electronic Equipment.

Perzanowski, A. (2022). The Right to Repair: Reclaiming the Things We Own. Cambridge University Press.

U.S. Bureau of Labor Statistics. Occupational Outlook Handbook: Automotive Service Technicians and Mechanics. U.S. Department of Labor.

PIRG Education Fund. Repair Saves Families Big. U.S. PIRG Education Fund, 2023.

Open Repair Alliance. The Rise of Community Repair. Open Repair Alliance and The Restart Project, 2024, with the accompanying open data release.

RREUSE. Job Creation in the Re-use Sector: Data Insights from Social Enterprises. RREUSE, 2021.

Department of Climate Change, Energy, the Environment and Water. Employment in Waste Management and Recycling. Australian Government.

Back Market. Back Market Enters New Phase of Growth. Company press release, 2025, and Back Market Clears $3.5 Billion in 2025 GMV.

Swappie Oy. Annual accounts for the financial year 2024, as reported by AIM Group, 2025.

Renault Group. The Future Is NEUTRAL: The Circular Economy Is Stepping Into a New Era, and the Re-Factory Flins press kit.

Stahel, W. R. and Reday-Mulvey, G. (1981). Jobs for Tomorrow: The Potential for Substituting Manpower for Energy. Vantage Press.

Stahel, W. R. (2019). The Circular Economy: A User's Guide. Routledge.

Ellen MacArthur Foundation (2013). Towards the Circular Economy: Economic and Business Rationale for an Accelerated Transition.

Ostrom, E. (1990). Governing the Commons: The Evolution of Institutions for Collective Action. Cambridge University Press.

Cooperrider, D. L. and Whitney, D. (2005). Appreciative Inquiry: A Positive Revolution in Change. Berrett-Koehler.

Note on figures. The repair ratio, the threshold rate, all three worked cases, the four-lever table, the closing subsidy, the French and US policy dates and every firm figure quoted above are computed in lib/verify/IV_08.py and reproducible with python3 lib/verify.py IV.08. The jobs-per-tonne comparison is reported as directional because the boundaries of its two sources differ, and the euro–sterling agreement between the computed subsidy and the French bonus is reported as an agreement of magnitude, not an identity.