Haute Lumière

Commerce · I.08 · MMXXVI · daylight

La Bourse  /  Volume I  /  Nº I.08

When It Stalls

Volume I — Transition: From Here to the Living Economy

---

THE PLATE

---

THE LETTER

You did the ninety days. You found the positive core, you signed a baseline, you sized the pilot so it fitted inside one person's discretion, you got a verified number, and you gave it to the one person who controls the next allocation.

Then it stopped.

Not dramatically. Almost nothing stops dramatically. The sponsor moved into a different role in month seven and the successor had four priorities of her own. A capital freeze came down across the group and the facility went to the bottom of a list. The plant that was running it had a bad quarter and every non-core initiative was quietly not renewed. Nobody cancelled anything. There was never a meeting. It simply became true that nothing has moved since March.

This chapter is about that, and it is the most useful chapter in the volume, because a stall is the ordinary condition of a transition rather than the exception to it. Every case in this book that worked has stalled at least once. Interface stalled. Mondragon stalled. The German solar build stalled so hard that four fifths of it went away. Every one of them restarted, and what is interesting is not that they restarted but what was true of the restart — because in all four the thing that came back was not the thing that had stopped.

So this is not a chapter about recovering momentum. It is a chapter about the one measurement you can only take when something has stopped moving, what it is worth, how long you have to use it, and the exact arithmetic of when to pick a programme up and when to put it down honourably and start a cleaner one.

You will leave with three numbers and one inequality. The numbers are what your stall is costing you per month, how much of what you built you still hold, and the year in which holding it stops being worth anything. The inequality tells you which way to jump. All four are computable this week from figures you already have.

— The Editors

---

DISCOVERY

What is already working

Begin where the method always begins: stalls are not where transitions die. They are where transitions are edited, and there is a body of well-documented cases in which a stop was followed by a restart on a narrower and stronger base.

Interface, 2008 to 2010. The carpet manufacturer whose Mission Zero programme opens Chapter I.01 met the financial crisis with commercial interiors demand falling away underneath it. Net sales fell roughly a fifth between 2008 and 2009. Capacity closed and headcount came down. The sustainability programme did not stop, and the reason it did not stop is the whole of this chapter: it had never been a spend line. It lived inside yield, scrap and energy cost, which is to say it lived inside the numbers that a downturn makes people examine more closely rather than less. Interface declared Mission Zero met in 2019, seven years after Ray Anderson's death and nine after the trough. What survived the stall was the part that had been embedded in the cost line. The part that lived on advocacy went quiet and came back afterwards.

Mondragon and Fagor, 2013. Fagor Electrodomésticos, a founding cooperative of the Mondragon federation, entered insolvency in the autumn of 2013 after several years of losses in a collapsed Spanish appliance market. The business did not restart. The people did: the federation used its inter-cooperative mechanisms to relocate the large majority of the Basque worker-members into other cooperatives over roughly the following two years, and Errasti, Bretos and Nuñez set out both the relocation and its limits in the record. The lesson is exact. The layer that restarted was the layer written into the federation's rules. The layer that depended on one cooperative's trading position did not, and no amount of commitment was going to make it.

Germany's photovoltaic build, 2012 to 2023. Installations ran at roughly 7.6 GW in 2012. Tariff reform and a change in the support régime took the annual rate to about 1.9 GW in 2014 and about 1.5 GW in 2016 — a fifth of peak, held there for four years. Then it came back: about 3.9 GW in 2019, 7.2 GW in 2022, and roughly 14.3 GW in 2023, which is nearly twice the 2012 rate. Eleven years from peak to a higher peak. What came back was the installed base and the installer and EPC workforce, which were local, contracted and distributed. What did not come back was domestic cell and module manufacture — Q-Cells in 2012, Conergy in 2013, SolarWorld in 2017 — because that layer had been carried by the tariff rather than embedded in anything local. Nobody could have told you in 2011 which of the two layers was load-bearing. The stall told everybody.

Evergreen Cooperatives, Cleveland, 2008 to 2020. The anchor-procurement cooperative model launched in 2008 with a laundry, a solar installer and, from 2013, a hydroponic greenhouse. The greenhouse in particular had years that were very hard indeed, and the model's early theory — build new cooperatives from nothing, pointed at anchor demand — stalled against the ordinary difficulty of building new businesses. The restart, from 2020, is the Fund for Employee Ownership: instead of founding cooperatives, acquire established, already profitable local businesses whose owners are retiring, and convert them. Same purpose, same anchors, a different acquisition mechanism entirely — and the change is legible only because the first version stopped long enough for somebody to ask which part of it had been doing the work.

Four cases, four decades of evidence, and one pattern that holds across all of them: in every restart, what came back was the layer that had been embedded in something that ran without attention — a cost line, a federation rule, a local contract book, an anchor's purchase order. Nothing came back on the strength of enthusiasm, and nothing needed to.

Which means a stall is not simply a gap in the record. It is a reading.

---

THE ARITHMETIC

What a stall costs, what it is worth, and where the line sits

Here is the cut this chapter turns on, and it is the only one.

A stall is an ablation study you did not have to pay for. While you are pushing a programme you cannot tell what is being carried from what is being pushed — every part of it appears to be working, because your attention is underwriting all of it at once. Stop pushing and the question answers itself. Of everything that was running when you stopped, whatever is still running is, by definition, no longer dependent on you. That is the only clean measurement of institutionalisation there is, and a moving programme is structurally incapable of producing it.

Call it the survivor ratio, σ: the share of the programme's activity still running, with nobody pushing, at the end of the pause. It is the number the stall hands you, and we will see at the end of this movement that it is worth a great deal of money.

First, what a pause costs per year.

A paused programme is not one asset. It is five, and they decay at different rates.

Only the last of those is measured rather than assumed: team know-how decays at your own voluntary turnover rate, which at 15 percent a year gives 85 percent remaining after twelve months. The other four are stated as assumptions for you to replace with your own figures, and the model is written so that replacing them changes every number below.

Weight and sum them and you get the warm-start fraction — how much of a cold start you still hold after t years of pause:

        W(t) = Σ wᵢ · kᵢ^t

  t (yr)     W(t)    lost
     0.5    0.906    9.4%
       1    0.823   17.8%
       2    0.681   31.9%
       3    0.567   43.3%
       5    0.400   60.0%
      10    0.182   81.8%

About a fifth of what a programme cost to build stops being an asset in the first twelve months it is not being pushed. The instantaneous rate at the moment of the pause is 19.9 percent a year. That is the figure that belongs on the page when somebody says let us park it for now — not as an objection, but because parking has a price and it has never been quoted.

Second, the carrying cost, in money.

Take Chapter I.01's worked pilot: £180,000 to build, £62,000 a year of verified saving. Pausing it costs both the decay and the foregone return.

  year    decay    foregone    year total    cumulative
     1   31,950      62,000        93,950        93,950
     2   25,475      62,000        87,474       181,424
     3   20,460      62,000        82,460       263,885

A programme held still for twenty-four months has cost, in decay and foregone return, about exactly what it cost to build. That is the sentence for the paper. It is not an argument that stalling is wrong — sometimes a freeze is the correct call and the money is genuinely not there. It is an argument that a stall is a spending decision, and it should be taken by somebody who has seen the figure.

Third — and this is the one that decides — when restarting stops being worth it.

Restarting is not free even where W is high, because the second ask is priced differently from the first. Call the friction premium f: the share of a cold start you pay again purely because the programme is known to have stopped — re-litigated objections, a spent option, and a documented precedent that "we tried this." Restarting beats beginning again clean while

        W(t)  >  1 + f − 1/(1 − σ)

and with no survivors at all — σ = 0 — that reduces to the memorable form: restart while the warm fraction exceeds the friction premium, W(t) > f.

The window depends almost entirely on who is in the room:

The clock is not set by the programme. It is set by who will be in the room when you come back — which is why the second owner in Chapter I.01 was never a nicety.

And now the survivor ratio pays. Restarting at σ — resuming what still runs and rebuilding only the rest — rescopes both the rebuild and the ask by (1 − σ):

A survivor ratio of a quarter takes the hostile-room window from under three years to over seven. The measurement the stall gave you for nothing is what buys the restart back.

And here is where this fails, honestly.

Two conditions, and the first is a hard one.

The baseline cliff. W assumes the baseline decays. It does not always. If the underlying process was re-engineered during the pause — a new ERP, a line rebuild, a reorganisation that redrew the cost centres — the baseline does not lose comparability gradually. It goes to zero in a single step, and W falls by up to a quarter at once. A cliff at year two puts W at 0.55 with no survivors, which is barely above a new sponsor's friction and well below a hostile one. Under that condition this method loses, and the correct move is to declare the programme ended in writing and begin the successor clean.

The appreciative bias against ending. This is the honest negative about the method itself. Discovery orients you to what is working, so a stalled programme will always present a survivor set, and a survivor set will always look like a reason to restart. The literature on escalating commitment — Staw's study of the big muddy, and Arkes and Blumer on sunk cost — is precisely about people who kept going because they had already gone. An appreciative method applied to a stall, without the inequality above, is an extremely efficient machine for justifying a restart that should not happen. The arithmetic is not decoration here. It is the guardrail that makes the appreciation safe.

---

DREAM

What becomes ordinary

In the organisation that has absorbed this, a programme has three states and not two, and everybody knows the names of all three.

Running. Dormant. Ended. Dormant is a real status with a real budget line, a trigger that put it there, a named custodian, and a sunset date. It is not a euphemism and nobody uses it as one, because ending is available and ending is respectable.

The reporting pack carries dormant programmes in their own short block. Each line shows the date it went dormant, its survivor ratio at the last reading, its warm fraction today, and the year the warm fraction crosses the friction premium. Four figures, one line each. Nobody presents them; they are simply there, the way the covenant headroom is simply there, and once a quarter somebody's eye stops on one and the programme comes back.

Measurement series do not stop when programmes do. That is the single change with the largest effect, and it costs almost nothing, because a report that has been automated is cheaper to leave running than to switch off and rebuild. The finance function knows this and treats an unbroken series as what it is: the audit evidence that a dormant asset is not impaired.

Ending is a rite rather than an absence. A programme that ends gets a written close — what it cost, what it returned while it ran, its final survivor ratio, and one paragraph on which layer proved load-bearing. Those closes are read. They are the most-cited internal documents in the organisation, because they are the only place where anybody wrote down what actually holds.

And a person who has run something that stalled is not marked by it. They are the person who has the reading. The next programme is given to them, and it is given to them because of the stall, which is exactly the right way round and which almost no organisation currently manages.

---

DESIGN

The structure that gets there

Four components. Each is small, and each is written into a template once rather than argued for each time.

1. The dormancy clause, in the facility.

Every regeneration facility carries a clause defining what happens when it stops drawing. The clause names the triggers — sponsor departure, capital freeze, covenant action, a measured survivor ratio below a floor — and states that on a trigger the facility converts to dormant rather than terminating. Dormant means four obligations survive, and only four: the measurement series continues, the verifier stays on retainer, the instrument documentation is reviewed once a year, and the second owner keeps a named fraction of their time on it.

It costs nothing to include and it is unavailable to include later. Put it in the template now, for every facility, whether or not you expect to use it.

2. The survivor register, taken at month three.

At the end of the third month of a stall — not the first, because the first month is still coasting, and not the twelfth, because by then people have tidied — somebody walks the programme and lists every activity that was running at the pause. Against each: is it still happening, with nobody pushing it?

That list is the survivor register and its bottom line is σ. Take it once. Taking it twice teaches people to perform for it, which is the ordinary fate of any measure that becomes a target.

3. The restart gate.

A restart proposal is one page and opens with four figures in this order: σ, the current W, the estimated f with the room named, and the crossing year. Then the scope — and the scope is the survivor set plus the smallest rebuild that makes it whole. Not the original programme. Restarting at the original scope is the commonest way a good restart fails, because it asks for the full facility again from a room that has already watched the full facility stop.

Estimating f is a judgement and should be written as one. It is the fraction of the original approval path you must walk again: count the signatures, mark which of those people are new, and mark which of them have the stall in their file. Three new signatures with the stall on the record is a high-f room, and a high-f room means a small scope or a clean successor, not a louder argument.

4. The closing rite.

When W falls below the threshold — or when the baseline cliff hits — the programme is ended in writing, publicly, with the close described in Dream. This is the component people leave out and it is the one that makes the other three honest. Dormancy is only a real status if ending is also a real status. Where nothing can be ended, everything is dormant, the register fills with fiction, and within two years nobody reads the block.

The sequence, for the person standing in a stall right now. Declare the pause and date it. Take σ at month three. Keep the series running whatever else you do. Re-cost W. Name the room and estimate f. Then either restart at σ or close it properly — and do one of the two inside ninety days, because the carrying cost is running the whole time and it is roughly £8,000 a month on a £180,000 programme.

---

DESTINY

How it holds when nobody is pushing

The pleasing part of this chapter is that its own mechanism is built to survive the thing it is about. Three reasons it holds.

The clause is in the template. A thing that has to be argued for each time depends on somebody being awake; a thing in the standard facility document depends on nobody deleting it. The second is a far lower bar and it is the only kind of durability that is free.

The measurement series is in the standing pack. Chapter I.01's rule was that anything reviewed monthly persists and anything reviewed by exception does not. Applied here it means the highest-decay component — baseline comparability, at 28 percent a year — is held at roughly 3 percent a year by a report that was already automated. The cheapest single intervention in this volume is not switching off a report.

The closes are read. They earn their own persistence, because they are the only internal document that answers the question everybody actually has.

Now the honest part: here is how this fails.

It fails when dormant becomes the polite word for dead, the register fills with programmes nobody intends to restart, and the block stops being read. The fix is in the clause: a hard sunset date on every dormancy, after which the status converts to ended automatically unless somebody signs to extend.

It fails when σ becomes a target and starts measuring attendance rather than output — a meeting that still happens is not an activity that still runs. Take the register once, from the activity list, not from the calendar.

It fails when the reserve is swept in a cost round, which is what happens to anything that lives in opex. Keep it inside the facility, where it is part of a committed structure rather than a discretionary line.

And it fails, most often, the same way Chapter I.01's method fails: somebody tries to win the restart philosophically before winning it numerically. In a stall the order matters even more than it did at the start, because the room has already heard the philosophy once.

---

DELIGHT

What it feels like

There is a particular quiet pleasure in coming back to something after a long gap and finding that one part of it never stopped. The report that kept arriving in an empty room. The supplier who kept sending the monthly figure because nobody told them not to. The junior analyst who kept the tab open.

It is a better feeling than the original launch, and it is a different feeling — launches are effortful and slightly anxious, and this is neither. It is the feeling of discovering that something you made has a small amount of life of its own. You did not have to be there. It was not being pushed and it moved anyway, which is the only definition of institution that has ever been worth anything.

And there is a second pleasure, in the ending. A programme closed properly, with its number written down and its lesson stated in one paragraph, is a clean thing to have made. It sits well. Most people carry a few unresolved projects around for years as a faint background debt; a written close discharges that debt entirely, and the relief of it is out of all proportion to the length of the document.

---

OPERATIONALIZE THIS

At the level of finance

The instrument: a dormancy covenant with a warm-start reserve, attached to the regeneration facility of Chapter I.01.

You are not proposing a new facility. You are proposing a clause and a small ring-fenced reserve inside one you already have, and the argument for it is an impairment argument, which is a language your finance function already speaks.

The mechanics.

• Trigger. Named events, written in advance: sponsor departure, group capital freeze, covenant action, or a survivor ratio below a stated floor. Written in advance is the point — a trigger negotiated during a stall is negotiated by people who are already arguing.
• Effect. The facility converts from drawing to dormant. It does not terminate, the baseline stands, and the repayment schedule pauses rather than accelerating. Say the last part explicitly; a facility that accelerates on dormancy will never be allowed to go dormant, and the programme will be killed instead.
• The four surviving obligations. Measurement series continues; verifier on retainer; documentation reviewed annually; second owner retains a named fraction of time. Nothing else.
• Reserve size. Enough to fund those four. On the £180,000 worked pilot that is on the order of £7,500 a year. Carry it inside the facility, never in opex.
• Sunset. A hard date — three years is a reasonable default — after which dormancy converts to ended unless extended by signature.
• Restart terms. Pre-agreed: restart at the survivor set plus the minimum rebuild, at the original economics, without re-approval, provided W exceeds f. Pre-agreeing the restart is worth more than the reserve, because it removes the signatures that generate f in the first place.

What the reserve is worth. Keeping the series running and the verifier retained changes baseline comparability from 72 percent a year to about 97, and the counterparty relationship from 80 to about 95. On the worked pilot:

  t = 1 yr    W plain 0.823    W dormant 0.908    preserved  £15,300
  t = 3 yr    W plain 0.567    W dormant 0.754    preserved  £33,600

  reserve, three years at £7,500          £22,500
  asset preserved                         £33,600      1.5x

A one-and-a-half-times return on an insurance product is respectable on its own. It is not the main argument.

The balance-sheet treatment, which is the main argument. Where the programme capitalised an asset, dormancy raises an impairment question at the next reporting date, and impairment is a real charge against a real number. An unbroken measurement series is the evidence that the asset's service potential is intact — you can show the cash generating unit's performance through the dormancy rather than asserting it. A reserve of £7,500 a year that protects a capitalised balance against a write-down is not an insurance product at all. It is the cheapest audit evidence available. Take that conversation to your auditors early; it is a discussion about service potential and useful economic life, which they have every year.

The counterparty. Internal treasury, exactly as in I.01 — the dormancy clause is a variation to an internal facility and can be documented in an afternoon. Where an external lender is involved, the clause is easier than you expect: a lender's alternative to dormancy is a defaulted facility with no savings stream, and dormancy preserves their security rather than diluting it.

The number that decides it. One inequality, on the front page:

                                    1
        W(t)   >   1 + f   −   ---------
                                 1 − σ

With no survivors that is W(t) > f. If it holds, restart at σ. If it does not, end it in writing and start the successor clean — and the close is worth writing well, because it is the document the successor will be approved on.

The first ninety days of a stall.

---

APPRECIATIVE QUESTIONS

Twelve, for a room

Discovery — what is already working

1. Think of something here that stopped being pushed and kept running anyway. What was it, and what was it embedded in that made that possible?
2. Which of our reports would still arrive next month if everyone who commissioned it left tomorrow — and what does that tell us about which reports we should be building?
3. Name a programme that stalled here and came back stronger. What was different about the version that came back?

Dream — what becomes possible

1. If a stall were an ordinary status with a budget line rather than a silence, what would somebody feel able to tell us that they currently do not?
2. Imagine our reporting pack carries a dormant block with four figures per line. Whose eye do we most want to stop on it, and what would make it stop?
3. If ending something well were as respectable here as launching something, what would we end this quarter — and what would that free?

Design — what we build

1. What is the smallest thing we could keep running through a pause, and what would it preserve?
2. Who would take the survivor register at month three, and what would they need to be able to write it honestly?
3. Which of our facility templates could carry a dormancy clause by the end of the month, and who signs that change?

Destiny — how it holds

1. What would have to be true for a dormant programme here to be picked up by somebody who had no part in building it?
2. Where have we already written a good close — a piece of work that ended and left a usable record — and what made that one get written?
3. If this practice were quietly dying, the register would fill with things nobody intends to restart. Who would notice that first, and what would they do about it?

---

WORKS CITED

Anderson, R. C. (1998). Mid-Course Correction: Toward a Sustainable Enterprise. Chelsea Green.

Anderson, R. C. and White, R. (2009). Confessions of a Radical Industrialist. St. Martin's Press.

Arkes, H. R. and Blumer, C. (1985). "The Psychology of Sunk Cost." Organizational Behavior and Human Decision Processes, 35(1), 124–140.

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

Crist|Kolder Associates. Volatility Report. Annual editions. (Reported median CFO and CEO tenure in the Fortune 500 and S&P 500.)

Efficiency Valuation Organization. International Performance Measurement and Verification Protocol (IPMVP), Core Concepts. Successive editions.

Errasti, A., Bretos, I. and Nuñez, A. (2017). "The Viability of Cooperatives: The Fall of the Mondragon Cooperative Fagor." Review of Radical Political Economics, 49(2), 181–197.

Hirschman, A. O. (1970). Exit, Voice, and Loyalty: Responses to Decline in Firms, Organizations, and States. Harvard University Press.

Kelly, M. and Howard, T. (2019). The Making of a Democratic Economy: Building Prosperity for the Many, Not Just the Few. Berrett-Koehler.

Meadows, D. H. (2008). Thinking in Systems: A Primer. Chelsea Green.

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

Pressman, J. L. and Wildavsky, A. (1973). Implementation: How Great Expectations in Washington Are Dashed in Oakland. University of California Press.

Staw, B. M. (1976). "Knee-deep in the Big Muddy: A Study of Escalating Commitment to a Chosen Course of Action." Organizational Behavior and Human Performance, 16(1), 27–44.

United States Bureau of Labor Statistics. Job Openings and Labor Turnover Survey (JOLTS). Monthly series. (Quits rate, used as the sanity check on the turnover input.)

Weick, K. E. (1984). "Small Wins: Redefining the Scale of Social Problems." American Psychologist, 39(1), 40–49.

Whyte, W. F. and Whyte, K. K. (1991). Making Mondragón: The Growth and Dynamics of the Worker Cooperative Complex, 2nd edn. ILR Press.

Wirth, H. Recent Facts about Photovoltaics in Germany. Fraunhofer Institute for Solar Energy Systems ISE. Annual editions. (German annual PV additions, rounded.)

Note on figures. The warm-start model, the carrying-cost schedule, the restart threshold, the dormancy reserve and the baseline cliff are all computed in lib/verify/I_08.py and are reproducible there. The component weights and four of the five decay rates are stated in that module as assumptions, to be replaced by the reader's own figures; the fifth, team know-how, is driven by the organisation's measured voluntary turnover. German photovoltaic additions are rounded from Fraunhofer ISE's annual series. Interface net sales for 2008 and 2009 are from the company's own reporting of those years. Fagor relocation figures follow Errasti, Bretos and Nuñez (2017).