Haute Lumière

Commerce · V · MMXXVI · daylight

La Bourse  /  Volume V  /  Nº V.E1

For the Practitioner of Luminous Economics

Volume V — Labour, Value, Flourishing · Extension I of III Nine movements, one payroll.

---

THE PLATE

---

THE LETTER

The eleven chapters you have just read were written for an organisation with a remuneration committee, a claim run, a plan document and somewhere between five and fifty thousand employees. Two of them require more people than most countries' largest employers have. One of them — Chapter V.10's — asks you to field a panel of 879 people and you may have nine.

This chapter is Volume V re-cut for a practice.

It is not a simplification. The findings are the same findings: V.01's five prices · V.02's hours curve · V.03's concentration ceiling · V.04's employer share · V.05's payback month · V.06's cost-of-equity penalty · V.07's floor · V.08's meaning differential · V.09's capture ratio · V.10's protocol · V.11's paired series. What changes is which of them you can act on, which you can measure, and which are simply unavailable to you at any price — and the last category is real, it has three members, and they are named near the front rather than buried, because a practitioner's edition that pretends otherwise is worse than no edition.

Three of these instruments work better at your size than at the size they were written for, and one of them — Chapter V.07's floor — has a decision number that does not move at all when the headcount falls from two thousand to ten. That is not encouragement. It is arithmetic, and it is computed below.

And one of them is about you rather than about your staff. Volume V's hardest finding is Chapter V.08's, that the coefficient which dignifies care work is the coefficient that underpays it. At ten people you are on both sides of that coefficient at once, and the sum does not cancel. It compounds. That is the cut this chapter is built on, and we get to it in the first paragraph of the Arithmetic rather than the last.

— The Editors

---

DISCOVERY

What is already working, at exactly your size

Begin where the record is strongest, and notice what these have in common: every one of them is a unit small enough to hold in one head, and in every one the instrument was a rule about a relationship rather than a system about a population.

Buurtzorg's team of twelve. Chapter V.08 read it as evidence about judgement: neighbourhood nursing teams capped at twelve, holding their own caseload and rota, with a very thin office and no supervisory layer, delivering their outcomes on roughly 40 per cent fewer hours of care per client — 1.67 times the client outcome per hour of nursing. Read it again as evidence about size. The replicating unit was never the organisation. It was a team you could fit round a table, and what travelled was a cap and a caseload.

Adam Grant's five minutes. Callers who spent five minutes with a student whose scholarship their work had funded raised 142 per cent more weekly pledges and 171 per cent more weekly revenue — multipliers of 2.42 and 2.71 on a line an organisation had been trying to move for years. Lifeguards who read accounts of real rescues worked 43 per cent more hours and helped 21 per cent more. Chapter V.08 used these as evidence that meaning behaves like an input. The practitioner's reading is narrower and more useful: these are the only interventions in Volume V that cost nothing, need no sample, and work on one person at a time. They are available to you this week and they are not available to a company of fifty thousand in any form that survives contact with a communications department.

The Marcora buyout, at its actual scale. Chapter V.06 reported Italy's fund, CFI: on the order of 560 enterprises and 25,000 jobs. Do the division the headline does not. That is 44.64 jobs per enterprise. The average recovered Italian firm is a small firm, and the state's own arithmetic on it is EUR 14,000 per job held against EUR 24,000 of two years' unemployment benefit — a ratio of 0.583. The instrument that works best in Volume VI's own record was designed around businesses of forty-five people.

The Swiss training firm, which is mostly small. Chapter V.05's fourth cost-benefit survey found gross costs of CHF 28,070 per apprentice-year against productive output of CHF 31,240 — a net benefit of CHF 3,170 before the firm has hired anybody, and around two-thirds of Swiss training firms finishing ahead. Hold that beside this chapter's own arithmetic on the same question, because they disagree, and the disagreement is the most instructive thing in the chapter.

Dauphin, Manitoba. Chapter V.07's Mincome: a saturation guarantee in one town, hospitalisations 8.5 per cent lower against matched communities, with accidents, injuries and mental-health admissions carrying most of the fall. The design worked because the unit was a place small enough that everybody in it was covered. Saturation is the one experimental design a small organisation can actually run, and it is the design that produced the strongest health result in the volume.

And Richer Sounds, in the detail nobody quotes. Chapter V.03 recorded the transfer of 60 per cent into trust in 2019, and separately that Julian Richer paid every employee below director level GBP 1,000 for each year worked — about GBP 4 million across some 500 people. Divide: an implied average of 8.0 years of service. The payment was the story; the divisor is the finding. A firm where the average person has been there eight years is a firm where every one of Chapter V.09's tenure arithmetic runs differently, and we run it below.

One pattern, and the whole chapter is built on it. The effects that are large at small scale act on an individual — five minutes, a cap of twelve, a lease clause, a town. The effects that need scale are measured as averages across a population. So the practitioner's strategy is not a smaller version of the corporation's. It is the opposite one: act where one person is the unit of action, and measure only where counting is exact.

---

THE ARITHMETIC

The founder's own coefficient, three thresholds, and where three instruments fail outright

First, the cut, and it is the reason this chapter exists.

Chapter V.08 wrote the worker's decision and the employer's saving as one equation. Let W_c be the wage available for the same skill elsewhere, δ the fraction of it forgone, m what the worker privately values the meaning at:

  the worker accepts   iff   m  >=  delta x W_c
  the employer saves                delta x W_c

They are the same term. In a company of fifty thousand that identity is a distributional fact about two parties. In a practice it is a fact about one person, because the person collecting the discount and the person paying it are the same person — and the arithmetic does not net to zero. It compounds, because the money does not stay in a wage. It stays in the firm.

Take a practitioner who draws £34,000 against a skill-matched external benchmark of £52,000:

  delta = 1 - 34,000 / 52,000                     34.6 %
  V.08's conservative delta                        5.5 %
  V.08's aggressive delta                         17.0 %
  the founder's delta, against the aggressive     2.04 x
  the differential, per year                    £18,000
  over eight years                             £144,000

Chapter V.08 said that eliminating every single separation would not repay a teacher-sized differential of 17.0 per cent. The founder's is 2.04 times that, and there is no turnover on the other side of it at all, because she is not going to leave.

Now price what that £144,000 actually became. It did not go into a pension. It went into the firm, which is also her job, which is also — on the definition Chapter V.03 used — very nearly the whole of her financial wealth. So price it exactly as Chapter V.03 priced an employee's stake, with a micro-firm's volatility in place of a listed employer's:

  micro-firm equity volatility                     60.0 %
  diversified index volatility                     17.0 %
  total variance  0.60^2                           0.3600
  covariance term (1.00 x 0.17)^2                  0.0289
  idiosyncratic variance                           0.3311
  idiosyncratic volatility                         57.54 %
    — V.03's participant carried 0.1311 and 36.2 %

Under the same lognormal, constant-relative-risk-aversion machinery, at a risk aversion of 2.0 and w = 1.00:

  exp(-gamma w^2 sigma_e^2 T / 2),  T = 10      0.036480
  exp(-gamma w^2 sigma_e^2 T / 2),  T =  3      0.370353

  the £144,000, held ten years                    £5,253
  the £144,000, held three years                 £53,331
  what the concentration costs, at ten years    £138,747

Three and a sixty-five hundredths pence in the pound. And the hurdle that follows — the excess return the practice must earn over an index merely to justify the position — is lambda*(w) = gamma w sigma_e^2 / 2:

  the founder                                33.11 points a year
  V.03's median participant at w = 0.634      8.31 points a year
  ratio                                              3.98 x

Chapter V.03 said of its participant that nothing in the productivity literature supports 8.31 points a year. Invert the same expression, w* = 2 lambda / (gamma sigma_e^2), and the ceiling falls out:

  at an ownership premium of 1.0 points a year          3.02 %
  at 1.5 points                                         4.53 %
  at 2.0 points                                         6.04 %

  the practitioner holds w = 1.00, which is            22.07 x

The practitioner is the most concentrated investor in the economy and nobody has ever sent her a statement. Chapter V.03's three rules were: never buy the stake with wages, cap the concentration, fund the exit from the first day. Her below-market drawings break the first rule every month, in cash, and the breach is invisible because it is recorded as not having been paid rather than as having been invested.

This is the cut, and it is not an argument to pay yourself more than the business can afford. It is an argument about sequencing, and the sequencing is worth a factor you can compute. Match every pound of differential with a pound into anything diversified before the next pound of differential accrues, and w tends to 0.50 rather than 1.00:

  value of £1 at w = 0.50, over ten years          0.437032
  value of £1 at w = 1.00, over ten years          0.036480
  the sequencing is worth                             11.98 x

Nothing in that costs money. It reorders two payments.

Second: the three thresholds a practice can act on.

The floor (V.07), which is scale-free and is the gift of this chapter. Chapter V.07 priced a pay-floor guarantee on a workforce of two thousand. Run the identical arithmetic on ten. Median net monthly pay £1,850, monthly pay volatility 12.0 per cent — a sigma of £222.00 — with the floor set at 90 per cent of the trailing three-month median, £1,665.00:

  z = (1,665.00 - 1,850) / 222.00               -0.833333
  Phi(z)                                         0.202328
  phi(z)                                         0.281912
  E[(a - X)+] = sigma (z Phi + phi)                £25.15   per person per month
  a year                                          £301.84
  ten people                                    £3,018.44
  loaded replacement cost, each                    £3,000
  break-even fall in turnover                 10.06 points

Ten point nought six points — exactly the figure Chapter V.07 computed for two thousand people. The headcount appears in the numerator and the denominator and cancels. This is the only instrument in Volume V whose decision number is invariant to the size of the organisation, and it means the smallest employer in the country can sign the same paper as the largest and defend it with the same line.

The capture ratio (V.09), which is better for you than for a corporation. Chapter V.09's cut was that an employer rents health rather than owning it, and the lease is set by tenure. Model separations as exponential, hazard = ln2 / tenure, and the employer's share of a long-run benefit stream is r / (r + hazard) at a social discount rate of 3 per cent:

  median firm tenure 3.90 yr   hazard 0.177730   capture 14.44 %   hurdle 6.92 x
  a practice at 8.00 yr        hazard 0.086643   capture 25.72 %   hurdle 3.89 x
  the practice keeps                                       1.781 x as much

A practice with Richer Sounds' 8.0 years of average service keeps 25.72 per cent of the health it creates against a median firm's 14.44 per cent, and its hurdle on a long-dated health investment is 3.89 times rather than 6.92. Chapter V.09's whole diagnosis — that the wrong party was holding the asset — is less wrong about you than about anybody else, and it is the one place in this volume where being small is a straightforward advantage on the finance.

Your own week (V.02). Chapter V.02's calibration of the productivity-of-hours curve gives, from an elasticity of one at 49.0 hours and 0.35 at 56.0:

  2c = (0.35 - 1.00) / (ln 56 - ln 49)          -4.867769
  H* = 49 exp(-1 / -4.8678)                  60.17 h/week
  gap between the two maxima                   11.2 h/week

A practitioner working sixty hours is sitting on the total-output peak and 11.2 hours above the output-per-hour peak, which is a legitimate place to be and is exactly the place Chapter V.02's Kivimäki figures describe: at fifty-five hours a week and above, coronary heart disease risk 1.13 times and stroke risk 1.33 times a thirty-five-to-forty-hour week. Chapter V.02's structural answer was to make the limit a licence condition so that no supervisor has to be the one who declines the hours. You have no supervisor. The honest small-firm version is a written self-exception with a named control and a date, which sounds absurd until you notice that every version of this that works anywhere is a piece of paper somebody signed before the day it applied to.

Third: the honest negatives, and there are three.

Chapter V.10's instrument is unavailable to you at any budget, and its own gate will certify it as sound. Run the panel arithmetic on ten people. The standard deviation of a nought-to-ten life-satisfaction item is 1.9; the test-retest correlation is 0.60; the variance of the within-person change is 2 sigma^2 (1 - rho):

  variance of the change                            2.8880
  SD of the change                                1.699412 points
  (z_a + z_b), computed by bisection              2.801585
  minimum detectable effect at n = 10               1.5056 points
  against V.10's firm-scale target of 0.20            7.53 x too coarse

  pairs needed for 0.20 points, simple random          567
  design effect 1 + 11 x 0.05                         1.55
  pairs needed, clustered                              879
  against the practice's ten                          87.9 x

One and a half points on a ten-point scale is a life event, not a management decision. And now the part worth pausing on, because it is a lesson about instruments rather than about wellbeing. Chapter V.10's front-page gate is minimum detectable effect / largest documented nuisance effect >= 2.0, and the documented mode effect is 0.20 points. The practice scores 7.53. It sails through. A gate that protects against bias will certify an instrument whose binding constraint is variance, and the ten-person firm is the case that proves it — a series so imprecise that no housekeeping artefact could possibly dominate it, because nothing could.

Nor does patience rescue it. Keep a quarterly census for three years — eight quarters before a change and four after — and the team mean has a standard error of 1.9 / sqrt(10):

  standard error of a team of ten's mean          0.600833 points
  eight before, four after                        1.715614 sd of that mean
  in points                                         1.0308
  AR(1) rho                                           0.50
  sum term  S(k - s) rho^s                        6.007812
  variance inflation 1 + (2/k) S                  2.501953
  worth this many independent quarters              3.1975
  honest detection                                2.101645 sd  =  1.2627 points
  still too coarse by                                 6.31 x

Three years of disciplined quarterly measurement lands at 1.2627 points — worse than the naive reading, because the honest correction for month-to-month correlation takes power away rather than adding it. There is no budget that fixes this. What is available instead is in the Design movement and it is not a survey.

Chapter V.05's training payback does not land inside the contract, ever, and the reason is that your scarce input is your own billable hour. Take a three-year trainee on £22,000, £23,500 and £25,000, against a qualified person's annual output of £60,000, using Chapter V.05's published relative productive values of 0.26, 0.41 and 0.54 — the third of which reproduces as 0.81 x 0.84 x 0.80 = 0.5443. The principal teaches six hours a week in year one, four in year two, two in year three, over forty-six weeks, and her hour bills at £120:

  yr  share  output    wage   supervision      net    cumulative
  ---------------------------------------------------------------
   1   0.26  £15,600  £22,000     £33,120  -£39,520     -£39,520
   2   0.41  £24,600  £23,500     £22,080  -£20,980     -£60,500
   3   0.54  £32,400  £25,000     £11,040   -£3,640     -£64,140

It never crosses. And you can say precisely how far short it is, because the whole of the margin available for teaching is the three-year output less the three-year wage:

  three-year output                               £72,600
  three-year trainee wage                         £70,500
  the margin available for teaching                £2,100
  the principal's supervision hours                   552 h
  break-even rate for the principal's hour           £3.80 an hour
  against her billable rate of £120, short by        31.5 x

A practitioner can afford to spend her own time on training at three pounds eighty an hour. That single figure is the whole of the poaching problem for a small firm, and it has nothing to do with poaching. Chapter V.05's rule was train alone when P < T, pool when P > T; the practitioner's P is not merely larger than her T, it is larger than the contract. Even free supervision leaves £2,100 across three years, which is a rounding error on the decision.

The exit is Chapter V.05's fourth instrument and it is the Danish one. Denmark's Employers' Reimbursement Fund charges DKK 2,821 per full-time employee — on the survey's own implied rate of DKK 7.4630 to the euro, a ten-person firm pays DKK 28,210, or EUR 3,780 a year, and draws reimbursement only if it hosts. The free rider and the trainer pay the same and only the trainer is repaid. At your size that is not a policy preference. It is the only structure in which the arithmetic works.

And the third: Chapter V.09's medical-spending endpoint is beyond every employer in the volume, and by a factor that does not respond to money. The two-arm constant is 2 (z_a + z_b)^2 = 15.697759, so detecting a $300 change in annual spending against a standard deviation of $12,000 needs:

  n per arm                                         25,116
  both arms                                         50,233 employees
  against the practice's ten                         5,023 x

You are not under-resourced on this. You are five thousand times short, and so is almost everybody. The correct response is Chapter V.09's own: do not buy the claim, buy the endpoint you can read, and the endpoint a practice can read exactly is a count — who left, who was absent, who was hurt — because at ten people a count is not an estimate. It is a census with no sampling error at all, and one leaver is ten points.

---

DREAM

What becomes ordinary

Describe it in the present tense, because a dream in the future tense is a wish.

There is a figure on the front of the management accounts that was not there before, and it is her own δ. It is computed once a year against a skill-matched external benchmark rather than against what other people in her trade pay themselves, and it is 34.6 per cent, and it is not a source of shame because it is not a verdict. It is a financing fact: the business is funded, this year, by £18,000 from one household, and the accounts say so on the line where it is true.

Beside it there is a note. Not a figure of speech — a dated, indexed, subordinated loan from her to the company, which ranks ahead of her equity and which a buyer would have to settle. Eight years of it at three per cent is £160,062 against £144,000 unindexed; the indexation alone is £16,062 and it cost one clause. She has stopped subsidising the business and started lending to it, and the only thing that changed was that somebody wrote it down.

Every pound of differential is matched by a pound into something that is not this company before the next pound accrues. She does not experience this as discipline. It reads the way a seatbelt reads. And the position she is building is worth 43.70 pence in the pound rather than 3.65 — a factor of 11.98 bought entirely with the order of two payments.

The four-item record exists and nobody calls it a survey. Ten people, the same four questions in the same words in the same month, kept in a folder that will be ten years old one day, read as ten individual series rather than as one mean. Nobody computes an average. Nobody benchmarks it. When somebody's line moves, the person whose line it is is the one who says what moved it, which is the only reading a sample of one can support and is also, as it happens, the reading Chapter V.10 says survives the ordinal problem.

The floor arrived in the second year and nobody has mentioned it since, which is the sign it worked. It costs £3,018 a year across ten people and it pays for itself if turnover falls by 10.06 points, and turnover is a count so everybody knows.

And she is not training anybody alone. There are three practices in the arrangement, the apprentice is employed by the pool, and the rotation is what teaches. The month somebody qualifies and leaves, nobody in the building describes it as a loss, because the pool's arithmetic was never built on holding them.

---

DESIGN

The order, and the money, to the pound

The order the findings are actually used in. Not the order of the chapters — the order of dependency, because each step produces the input the next one needs.

Steps one to four are the first ninety days and cost almost nothing. Steps five to eight are the second ninety. Steps nine and ten run alongside and are never finished.

And the money. The whole apparatus, costed, under £6,000.

Under the ceiling of £6,000 by £580. Two things about that table are worth saying out loud.

The apparatus costs less than the guarantee, and the guarantee is not a cost. It is an expected top-up against a distribution, it returns itself at 10.06 points of turnover, and turnover at ten people is a number everybody in the room can verify without a system.

Nothing in it is a platform, a licence or a subscription to a dashboard. It is one benchmark, six of your own hours, two of somebody else's, a train fare, and a reserve — and the single most valuable line, your own δ, costs £360 and has never been computed by anybody in your position.

What replaces the survey. Chapter V.10's own conclusion was that the protocol is expensive and the sample is cheap. At ten people invert it completely: the sample is a census and free, and the protocol is the entire product. Four items in the Office for National Statistics wording, one month, one wording, one position in the sequence, for ever. No mean is ever published. No benchmark is ever quoted. A then-test rides on the second wave and costs one question. What you are building is not an instrument. It is a record — ten biographies with a frozen question in them — and it is the only design in which n = 10 is a strength rather than a defeat, because the comparison it supports is within a person and therefore immune to the problem that breaks league tables.

Governance, in one line, because a practice of one still needs it the moment there are three. Two signatures on any change to the four items, and a named person other than you who holds the folder. Chapter V.10's finding was that most organisational wellbeing series are three years old at most because somebody changed something and nobody could say what. Yours will be ten years old, and that will be the most valuable thing in the building that money cannot buy.

---

DESTINY

How it holds when the practice is tired

A practice has one failure mode a firm does not: the person who built every one of these instruments is also the person who has to sell next month's work, and in a thin quarter all of it looks like overhead.

The note holds it. A differential recorded as a loan is a liability of the company, and a liability cannot be quietly forgotten the way a habit can. It has to be repaid, waived in writing, or carried — three outcomes, all of them visible, none of them the fourth outcome that currently happens, which is that it evaporates. This is the same fastening Chapter V.08 named for the Mission Differential Account and it works here for the same reason: irreversibility is not willpower, it is double-entry.

The matching rule holds it. A sequencing rule survives a bad month because it does not require a payment, only an order. When there is no differential there is nothing to match.

The count holds it. Chapter V.11's paired page — one flow, one stock — is a layout decision before it is anything else, and at ten people the stock measure is free: who left, who was off, who was hurt, how many hours were actually worked. Chapter V.11's rule is that a flow can always be raised for a period by drawing the stock down. In a practice the stock being drawn down is very often the principal, and the paired page is the only document that will say so.

Now the honest part. This fails when δ is computed against the trade rather than against the skill, because the answer comes back near zero and everybody relaxes — Chapter V.08's exact failure mode, and it bites harder here, since a practitioner's obvious comparison group is other practitioners, all of whom are underpaying themselves by the same 34.6 per cent.

It fails when the record is read as a mean. Ten numbers averaged is a statistic with a standard error of 0.600833 and no business being quoted; ten numbers read as ten lines is ten conversations.

It fails when the floor is offered and then withdrawn in a thin quarter, which converts an instrument into a memory of one, and the memory is worse than never having offered it.

And it fails, most often, when the pool is never formed — because forming it requires ringing two people who are notionally competitors and saying that your training payback is £3.80 an hour. That sentence is the hardest thing in this chapter to say out loud and it is the one that produces the arrangement.

---

DELIGHT

What it feels like

There is a specific relief in seeing your own number.

Not a pleasant number: 34.6 per cent, on a page, in your own handwriting. But the thing people describe, having done it, is not discouragement. It is that a private three-in-the-morning arithmetic — is this devotion or is it foolishness — stops being a question about character and becomes a question about money, with a figure attached, taken by an adult in daylight. Chapter V.08 said that people handed that number describe the feeling as being taken seriously. When you hand it to yourself the effect is the same and it arrives faster.

Then the second one, which is quieter and comes later. Somebody on the payroll notices that the floor arrived and did not have to be asked for. They do not thank you, because there is nothing to thank — it is simply how the month works now. The absence of the conversation is the whole of the gift, and it cost £25.15 a person a month against a distribution you measured yourself.

And there is the folder. Four questions, the same words, the same month, ten lines. Open it in the fifth year and you are reading something nobody in any corporation has: not a benchmark, not a score, not a quartile — a decade of one small group of people answering one honest question, with no incentive attached to the answer. A question that does not want anything is rare enough to be felt, and at ten people it is the only kind you can afford to ask.

---

OPERATIONALIZE THIS

At the level of finance

The instrument: a founder's differential note, indexed, subordinated, with a concentration covenant on the marginal pound.

It is one document, it needs no counterparty, and it converts the largest and least-recorded transaction in the small-enterprise economy — a household financing a business out of forgone wages — into a dated claim that survives the founder, the accountant and the sale.

The terms.

The worked case. A differential of £18,000 a year accrued for 8 years at an index of 3.0 per cent, annuity factor ((1 + i)^n - 1) / i = 8.892336:

  the note                                      £160,062
  the same differential unindexed               £144,000
  the index alone is worth                       £16,062

The balance-sheet treatment. A director's loan account, which every accountant in the country already prepares and no auditor will argue about. It is a liability of the company at amortised cost and an asset of the household. Two things make it real rather than decorative: it must be genuinely repayable — an irrevocably waived loan is a capital contribution and should be described as one — and the computation of the principal must be documented the way any related- party transaction is documented, because that documentation is what makes it survive a buyer's diligence. An undocumented founder's subsidy is a gift to whoever buys the company. A documented one is a claim they must settle, and Chapter V.03's and Chapter I.11's succession arithmetic both turn on which of those two it is.

The counterparty. Yourself, for the first three years, which is why this can be signed on Monday. Then, at a transaction: the buyer, the incoming partner, or the trust. And at any point in between, a lender — because a business whose owner has been documenting her own subvention is a business with a legible history of what it actually cost to run.

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

    household exposure to this firm
   ---------------------------------   <=   w*  =  2 lambda / (gamma sigma_e^2)
    household financial wealth

At an ownership premium of 1.5 points a year, this chapter's parameters put the right-hand side at 4.53 per cent. The practitioner is at 1.00, which is 22.07 times the ceiling, and no covenant can move her there in one year or in ten. So the covenant is written on the marginal pound rather than the stock, and that version is achievable this month:

  w tends to 1.00, unmatched                       0.036480    =  3.65 pence
  w tends to 0.50, matched pound for pound         0.437032    = 43.70 pence
  the covenant is worth                              11.98 x

Eleven point nine eight times, for a rule about the order of two payments. Put that line on the front of the note, because it is the sentence that gets the covenant signed — and unlike every other instrument in this volume, there is nobody to persuade.

The first ninety days on a page.

---

APPRECIATIVE QUESTIONS

Twelve, for a room

Discovery — what is already working

1. What is the one thing this practice does that a client has thanked you for twice, and has anybody ever costed the hour it takes?
2. Who here has stayed far longer than they had to, and when you ask them why, what do they actually say?
3. Think of a five-minute contact with the person your work was for — yours or somebody's. What changed in the week afterwards that you could have counted?

Dream — what becomes possible

1. If the wage you are forgoing were on the front of the accounts as a number rather than as a silence, what conversation would become possible that is not possible now?
2. Imagine the folder is ten years old and holds ten honest lines. What is the first question you would take to it?
3. If your household's exposure to this firm were halved without the firm losing a penny, what would you attempt next year that you will not attempt now?

Design — what we build

1. What is the skill-matched benchmark for your own hour, and who would you ask to find it this month?
2. What is the smallest guarantee you could offer the people here that would change one real decision for one of them — and what would it cost to find out?
3. Which two practices near you have rotations different enough from yours to teach something you cannot, and who introduces them to each other?

Destiny — how it holds

1. If you were unavailable for three months, which of these instruments would still be running, and what did you do that made that true?
2. What would be the first sign that the record had become a score rather than a question, and who would notice before you did?
3. When this practice is sold or handed on, which of the things you have paid for personally will the buyer have to pay for — and which will they simply receive?

---

WORKS CITED

Ainsworth, B. E., Haskell, W. L., Herrmann, S. D., et al. (2011). "2011 Compendium of Physical Activities." Medicine & Science in Sports & Exercise, 43(8), 1575–1581.

Arbejdsgivernes Uddannelsesbidrag (AUB), Denmark. Employers' Reimbursement Fund, established 1977; contribution rates as published by the fund.

Cooperrider, D. L., Whitney, D. and Stavros, J. M. (2008). Appreciative Inquiry Handbook, 2nd edn. Crown Custom Publishing.

Dolan, P. and Kavetsos, G. (2016). "Happy Talk: Mode of Administration Effects on Subjective Well-Being." Journal of Happiness Studies, 17(3), 1273–1291.

England, P., Budig, M. and Folbre, N. (2002). "Wages of Virtue: The Relative Pay of Care Work." Social Problems, 49(4), 455–473.

Forget, E. L. (2011). "The Town with No Poverty: The Health Effects of a Canadian Guaranteed Annual Income Field Experiment." Canadian Public Policy, 37(3), 283–305.

Grant, A. M. (2008). "The Significance of Task Significance." Journal of Applied Psychology, 93(1), 108–124.

Grant, A. M., Campbell, E. M., Chen, G., Cottone, K., Lapedis, D. and Lee, K. (2007). "Impact and the Art of Motivation Maintenance." Organizational Behavior and Human Decision Processes, 103(1), 53–67.

Gray, B. H., Sarnak, D. O. and Burgers, J. S. (2015). Home Care by Self-Governing Nursing Teams: The Netherlands' Buurtzorg Model. The Commonwealth Fund.

Kish, L. (1965). Survey Sampling. Wiley.

Kivimäki, M., Jokela, M., Nyberg, S. T., et al. (2015). "Long working hours and risk of coronary heart disease and stroke." The Lancet, 386(10005), 1739–1746.

Krueger, A. B. and Schkade, D. A. (2008). "The reliability of subjective well-being measures." Journal of Public Economics, 92(8–9), 1833–1845.

Kruse, D. L. (2016). "Does employee ownership improve performance?" IZA World of Labor, 311.

Meulbroek, L. K. (2005). "Company Stock in Pension Plans: How Costly Is It?" Journal of Law and Economics, 48(2), 443–474.

Muehlemann, S. and Wolter, S. C. (2014). "Return on Investment of Apprenticeship Systems for Enterprises: Evidence from Cost-Benefit Analyses." IZA Journal of Labor Policy, 3:25.

Office for National Statistics. Personal Well-being in the UK. Annual statistical bulletins, 2012–, with accompanying user guidance.

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

Pencavel, J. (2015). "The Productivity of Working Hours." The Economic Journal, 125(589), 2052–2076.

Strupler, M. and Wolter, S. C. (2012). Die duale Lehre — eine Erfolgsgeschichte, auch für die Betriebe. Rüegger Verlag.

US Bureau of Labor Statistics. Employee Tenure Summary; Employer Costs for Employee Compensation; and Occupational Employment and Wage Statistics.

Vieta, M., Depedri, S. and Carrano, A. (2017). The Italian Road to Recuperating Enterprises and the Legge Marcora Framework. Euricse Research Report, Trento.

Volpp, K. G., Troxel, A. B., Pauly, M. V., et al. (2009). "A Randomized, Controlled Trial of Financial Incentives for Smoking Cessation." New England Journal of Medicine, 360(7), 699–709.

Note on figures. Every figure in this chapter is computed in lib/verify/V_E1.py and prints with its inputs, its units and its source; run python3 lib/verify.py V.E1. Figures carried from Chapters V.01 to V.11 are recomputed here rather than quoted, so that a reader checking this chapter need not open another module. The practitioner's drawings, the skill-matched benchmark, the micro-firm's equity volatility, the risk aversion, the two horizons, the trainee's wages, the principal's billable hour, the qualified person's output, the quarter-to-quarter correlation and the practice's median tenure are stated assumptions, printed as such, and every threshold moves when they are replaced. The concentration arithmetic inherits Chapter V.03's model wholesale, including its weakness: a compensating-differential estimate and a lognormal certainty equivalent are both ranges wearing a point estimate, and the ordering of the conclusions survives the range while the magnitudes do not.