La Bourse / Volume I / Nº I.08 / Quiz, reflection, essays

A man standing at the head of a long boardroom table, speaking, the room lit evenly through a wall of glass. — Plate I.08 · Quiz, reflection, essaysThe Instrument Left Running.Nothing here has been used for a long time. One thing here has not stopped, and that is the only thing in the room that knows what happened.

ASSESSMENT · Chapter I.08 — When It Stalls

Three instruments: a ten-point quiz, eight reflection questions, five essay prompts. The quiz checks comprehension rather than recall. The reflections are private and first-person. The essays are arguable from more than one side.

THE QUIZ — ten points

Four on recall.

1. Define the survivor ratio σ, state where the numerator comes from, and say when it is taken and how often.

σ is the share of the programme's activity still running with nobody pushing, over the activity that was running at the pause. The numerator is taken from the activity list, not the calendar — a meeting that still happens is not an activity that still runs. Taken at month three: month one is still coasting on the last push, month twelve has been tidied. Taken once — repeated, it becomes a target and is trivially performed for. One mark for the ratio, one for either the activity-list discipline or the once-only rule.

2. Name the five components of the warm-start fraction and say which one is measured rather than assumed.

Discovery findings (25%), baseline comparability (25%), instrument documentation (20%), counterparty relationship (15%), team-specific know-how (15%). Only team know-how is measured — it decays at the organisation's own voluntary turnover rate. The other four decay rates are declared as assumptions in lib/verify/I_08.py and are meant to be replaced.

3. State the restart inequality in both forms, and say what f is.

Restart while W(t) > 1 + f − 1/(1 − σ); with no survivors this reduces to W(t) > f. The friction premium f is 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."

4. Name the three states a programme can be in under this design, and the two clauses that make dormancy usable.

Running, dormant, ended. Repayment must pause rather than accelerate on dormancy — a facility that accelerates will never be allowed to go dormant, so the programme is killed instead. And the restart terms must be pre-agreed, because that is what removes the signatures that generate f.

Four on application.

5. A director says: "Let's just park it until the capital position improves, and pick it up when things settle." What is the one thing missing from that sentence, and how would you supply it?

A price. Parking is a spending decision and has never been quoted as one. Supply it with two figures: the decay rate — roughly a fifth of the build cost in the first twelve months, 19.9 percent a year at the moment of the pause — and the foregone verified saving. The stronger answer notes the framing: this is not an objection to the pause. Sometimes the capital genuinely is not there. It is an argument that the pause should be decided rather than defaulted into, and most stalls are not decisions at all — they are the absence of a renewal nobody had to defend.

6. During a nine-month pause, the plant your programme ran in migrated to a new ERP and the cost centres were redrawn. What has happened to the arithmetic, and what do you do?

The baseline cliff. Comparability does not decay — it goes to zero in one step, removing up to 0.25 of W at once. At that point, with no survivors, the inequality is very unlikely to hold. Declare the programme ended in writing and begin the successor clean. Full marks require the consolation: discovery findings and instrument documentation survive a cliff untouched, so the successor starts well ahead of where the first one started.

7. Your survivor register comes back at σ = 0.05 after four months — almost nothing survived. Someone reads this as proof the idea was wrong. What is the better reading?

σ measures institutionalisation, not merit. A low σ says the programme was being carried entirely by one person's attention and was never embedded in a cost line, a rule, a contract or a routine. That is a finding about the design, and it is directly actionable: the successor is built to be embedded from day one. Credit any answer noting that the German photovoltaic case is exactly this — the manufacturing layer had a σ near zero and the installer layer did not, and nobody could have said which was which in 2011.

8. Why does restarting at the original scope so often fail, even where the arithmetic says restart?

Because you are asking a room that watched the full facility stop to approve the full facility again. Restarting at σ — the survivor set plus the smallest rebuild that makes it whole — is a different request: formalise four things that are demonstrably still running and fund the small remainder. Different meeting, different outcome. It also rescopes the ask by (1 − σ), which is what relaxes the inequality.

Two that require the arithmetic to be done.

9. A programme cost £240,000 to build and was returning £48,000 a year of verified savings. It has been paused for eighteen months. Using the chapter's model, what has the pause cost so far? Show your working.

W(1.5) = 0.25(0.90^1.5) + 0.25(0.72^1.5) + 0.20(0.85^1.5) + 0.15(0.80^1.5) + 0.15(0.85^1.5) = 0.2135 + 0.1527 + 0.1567 + 0.1073 + 0.1175 = 0.7478. Decay: (1 − 0.7478) × 240,000 = £60,528. Foregone: 1.5 × 48,000 = £72,000. Total ≈ £132,500 — a little over half what the programme cost to build, in eighteen months of nothing happening. Credit any answer within £5,000. The point of the question is that a stall has a number, and the number is large.

10. Same programme. The CFO who sponsored it has left; her successor has the stall in his file, so you estimate f = 0.45. The survivor register came back at σ = 0.30. Restart, or begin again clean?

Threshold: 1 + f − 1/(1 − σ) = 1 + 0.45 − 1/0.70 = 1.45 − 1.4286 = 0.0214. W(1.5) = 0.7478, which exceeds 0.0214 by a very wide margin. Restart — and not narrowly. The stronger answer states the scope: restart at σ, resuming the thirty percent still running and rebuilding the other seventy, not proposing the original facility again. And the strongest answer notices what the arithmetic just demonstrated — a survivor ratio of only three in ten was enough to make a hostile-room restart the cheaper move, which is the chapter's entire claim about what a stall hands you.

REFLECTION — eight questions, for one person and a pen

These are not for a room. Write the answers by hand if you can; the slowness is the point.

Think of something you started that stopped without anyone deciding to stop it. What was the last thing that happened, and what would have had to happen for it to continue?

Of everything you currently maintain by personal effort, what would still be running three months after you stopped? Answer honestly, and then ask what the difference is between the things on each side of that line.

What have you left dormant for years without ever writing a close for it? What would it cost you to write one — and what do you suspect that would free?

Recall a time you restarted something and it went badly. Was the problem the thing, or the scope you restarted it at?

Where have you kept going because you had already gone? Be specific about the moment you first suspected it and what you did with the suspicion.

Who in your working life would notice if a piece of your work quietly stopped? If the honest answer is nobody, what is the smallest thing you could embed it in?

What report, series, habit or record have you switched off that you now wish you had left running? What did that cost you when you came back?

What would it take for stopping something to feel as respectable to you as starting it? Write the sentence you would need to be able to say out loud.

ESSAY PROMPTS — five

Each is arguable from more than one side. Each requires at least one source the chapter cites and at least one it does not.

1. Is the survivor ratio a measurement or an artefact? The chapter claims σ is the only clean measurement of institutionalisation and that a moving programme cannot produce one. Argue either that this is a genuine natural experiment — an unplanned removal of the intervention — or that it is confounded beyond usefulness, since a stall is rarely random and the same conditions that caused it also select which activities survive. Engage the chapter's German photovoltaic case, and at least one source on natural experiments or quasi-experimental inference that the chapter does not cite.

2. Escalation, or persistence? Staw's "big muddy" and Arkes and Blumer's sunk-cost work describe continuing because one has already begun. The chapter's whole design is about continuing things that have stopped. Argue whether the restart inequality genuinely distinguishes persistence from escalation, or whether it supplies a quantitative costume for the same behaviour. Use Staw (1976) directly and at least one more recent empirical treatment of escalation of commitment.

3. The clause that nobody uses. The dormancy covenant costs nothing to include and pays out only when nobody is paying attention. Argue either that pre-committed procedural clauses of this kind reliably survive organisational neglect — Ostrom's design principles would suggest conditions under which they do — or that they are systematically defeated by the same forces that caused the stall. Use Governing the Commons, and one source on contract incompleteness, covenant design or governance mechanisms that the chapter does not cite.

4. Which layer is load-bearing? Germany's photovoltaic stall retained the installer base and lost domestic manufacture; Interface retained what lived in the cost line; Mondragon retained what was written into federation rules. Argue whether "embedded in something that runs without attention" is a real predictive criterion — one you could apply in advance — or a description available only in retrospect. Use Wirth's Fraunhofer ISE series and Errasti, Bretos and Nuñez (2017), plus one industrial policy or economic geography source the chapter does not cite.

5. The right to end things. The chapter argues that dormancy is only a real status where ending is also a real status, and that a person who has run something that stalled should be given the next programme because of the reading they now have. Argue the counter-case: that organisations which make ending easy end things that should have been carried through, and that a degree of irreversibility — Chapter I.01's own prescription — is what produces anything durable at all. Use Hirschman's Exit, Voice, and Loyalty, and one source on project termination, innovation portfolio management or organisational learning from failure that the chapter does not cite.