Haute Lumière

Commerce · VI.09 · MMXXVI · daylight

La Bourse  /  Volume VI  /  Nº VI.09  /  Ten concept briefs

TEN CONCEPT BRIEFS · Chapter VI.09 — Conflict, Repair, and the Long Table

One page each. A reader who reads only these ten pages has the chapter.

---

BRIEF 1 — The Forum That Exists Before the Dispute

The idea. A conflict system is not a procedure. It is a place, a time, and a group of people who are known to be there. The expensive part of handling a dispute is almost never the hearing; it is getting a hearing to exist.

The Tribunal de las Aguas in Valencia has met at the Apostles' Door of the cathedral, outdoors and in public, on Thursdays at noon, for approximately a thousand years — about fifty-two sittings a year, on the order of 52,000 hearings. Eight judges, the síndicos, elected by the irrigators from among the irrigators. No lawyers, no written pleadings, no appeal. Decisions usually within minutes.

Worked example. Two members of a cooperative have been avoiding a dispute for eight months. In an organisation with no standing forum, raising it requires one of them to ask for a meeting to be convened, which is an act of escalation and therefore an act of courage. In an organisation with a first Thursday, it requires them to put an item on a list. The difference in the cost of the first step is the difference between the dispute being handled and the dispute being carried.

Why it matters. A forum that convenes on demand requires bravery, and bravery is not a renewable input. A forum that meets on a schedule requires a calendar. In the chapter's worked house the standing cost of that calendar is £17.50 per member per year — the whole of Ostrom's sixth principle, funded.

You already know this because you have watched a problem go unraised for a year and then get solved in eleven minutes once somebody finally put it on an agenda, and you have thought: all that was missing was the agenda.

---

BRIEF 2 — Ostrom's Fourth and Sixth Principles

The idea. Of Elinor Ostrom's eight design principles for enduring commons, two are about conflict and they are the two most often skipped when people copy the list.

• Principle four — monitoring. Monitors who actively audit conditions and behaviour are accountable to the appropriators, or are the appropriators themselves.
• Principle five — graduated sanctions. Violations meet sanctions that escalate, assessed by other appropriators or by officials accountable to them.
• Principle six — conflict-resolution mechanisms. Appropriators have rapid access to low-cost local arenas to resolve conflicts among themselves and with officials.

Read six closely. Every word is a specification: rapid, low-cost, local, arena, and with officials as well as among members.

The check. Cox, Arnold and Villamayor-Tomás (2010) re-analysed 91 studies of community-based resource management against the principles and found them broadly supported, reformulating the eight into 11 for precision. Monitoring and graduated sanctions were among the best supported.

Why it matters. Most organisations copying Ostrom take principles one, two and three — boundaries, congruent rules, collective choice — and stop, because those are the ones that read like constitution-writing. Four, five and six read like administration. They are where the system either holds or does not.

You already know this because you have belonged to a group with beautiful rules that fell apart the first time somebody broke one and nobody knew who was supposed to say so.

---

BRIEF 3 — What Litigation Actually Costs and Takes

The idea. Most organisations have exactly one written dispute mechanism and it is a court. Price it from the courts' own statistics rather than from anyone's impression.

The thirty months is the Administrative Office of the U.S. Courts' median time from filing to trial in civil cases. The cost medians are from the Federal Judicial Center's National Case-Based Civil Rules Survey (Lee and Willging, 2010) — and the crucial denominator is that they are medians across closed cases, most of which settled without trial. That is the cost of the route.

The system figure. RAND (Kakalik and Pace, 1986) found that of every dollar the American tort system spent, about 46 cents reached the injured party. Fifty-four cents of every dollar is the cost of finding out who was right.

Why it matters. The ratios are the argument. 913× on duration against a mediation, 43.5× against a three-week internal panel. Nobody rejects a conflict forum after seeing those numbers; they reject it before seeing them.

You already know this because you have watched a dispute about something small consume a year of somebody's attention, and known the whole time that the underlying facts could have been established in an afternoon.

---

BRIEF 4 — Arbitration Is Faster and It Is Not Neutral

The idea. Speed is real, and it is not free, and it is not distributed evenly.

Alexander Colvin's study of American Arbitration Association employment cases (2011) found:

Arbitration is about 18.1 months faster. The employee is about 1.70× more likely to win at trial.

Worked example. A firm writes a mandatory arbitration clause into its contracts and presents it internally as an efficiency: faster, cheaper, less adversarial. All three are true. It is also, on this evidence, a transfer of expected value from the employee to the firm, and the honest version of the board paper says so.

Why it matters. A conflict system designed entirely by the party who will usually be the respondent will drift into exactly this shape, and it will do so without anybody intending it. The design answer in this chapter is the both-ways invocation: either party may reach the terminal rung. It costs nothing and it is the term that makes a fallback legitimate rather than managerial.

You already know this because you have read a clause that was obviously written by one side and noticed that every one of its efficiencies happened to point the same direction.

---

BRIEF 5 — The Restorative Effect Size, Converted Into a Decision

The idea. Restorative conferencing works, the effect is small, and small is a number you can act on once you convert it.

The Campbell systematic review (Strang, Sherman, Mayo-Wilson, Woods and Ariel, 2013) pooled ten randomised controlled trials and reported a standardised mean difference on repeat offending of about d = 0.155.

The conversion, in full.

  d                                      0.155
  logit factor  π / √3                   1.8138
  odds ratio = exp(0.155 × 1.8138)       1.3246
  base two-year reconviction rate        50.0%
  control odds = 0.50 / 0.50             1.0000
  treated odds = 1.0000 / 1.3246         0.7549
  treated rate                           43.0%
  absolute reduction                     7.0 points
  number needed to treat = 1 / 0.070     14.3 conferences

Run about fourteen conferences and one person who would have reoffended does not. That is what d = 0.155 means in the world.

Why it matters. "A small but statistically significant effect" is not actionable and invites both dismissal and overclaiming. A number needed to treat of 14.3 is a budget line.

You already know this because you have seen a study reported as a breakthrough and the same study reported as a disappointment, and suspected that neither reporter had converted the effect into anything a person could do.

---

BRIEF 6 — Where Restorative Processes Lose

The idea. The evidence for restorative justice is credible precisely because it includes the arms where it did nothing and the arms where it went the wrong way.

The population it covers. Every randomised trial in the literature randomised cases where the offender had already admitted responsibility and both parties consented to a conference. That is not a methodological footnote. It means the evidence speaks about people willing to be accountable, and is silent about the person a conflict system most needs to handle: the one who denies everything and will not attend.

The arms that failed. In the Canberra RISE experiments, the drink-driving arm showed no reduction and a small increase in reoffending in the conference group. In one violent-crime arm, offending among Aboriginal offenders rose. Sherman and Strang have reported these as consistently as the positives.

The shape of the evidence. Stronger for violent crime than for property crime. Stronger for victim outcomes than for offender outcomes. Small on average, and in named populations, negative.

Why it matters. A design that assumes willingness has a hole exactly the size of unwillingness. This is why the chapter's ladder has a sixth rung and a fallback, and why it prices them rather than debating them.

You already know this because you have been in a mediation that worked beautifully and one that was a performance by somebody who had decided in advance to concede nothing, and you knew which was which in the first ten minutes.

---

BRIEF 7 — The Compliance Gap and the Enforcement Break-Even

The idea. People keep agreements they made more often than they keep orders handed to them — and the size of that difference decides whether you need enforcement at all.

McEwen and Maiman (1984) followed Maine small claims cases and counted whether the money actually moved:

Consent roughly doubles voluntary payment. That is the quoted half.

The unquoted half. Non-compliance with a mediated agreement is 29.4%, and a mediated agreement has nothing behind it. A court order leaves 66.2% outstanding — but that residual can be enforced. So there is a break-even in enforcement effectiveness:

  e*  =  (0.706 − 0.338) / (1 − 0.338)  =  0.368 / 0.662  =  55.6%

Consent alone beats a court order unless enforcement recovers more than about 56% of the unpaid balance.

The honest caveat, which the authors state themselves. These are cases that chose mediation. Selection is in the number.

You already know this because you have kept a promise you made and ignored an instruction you were given, on the same day, about the same sort of thing.

---

BRIEF 8 — The Graduated Ladder, Costed

The idea. A sanction system with one setting is used for nothing, because the only available response is disproportionate to almost every case.

The six rungs. A word · a recorded word · a facilitated conversation · a panel hearing with a finding · suspension of a specific right · exclusion. The movement rule: a rung is used twice before the next becomes available, and nobody skips except for conduct defined in the membership agreement.

The cost, for a house of 200 members.

  STANDING (whether or not anything happens)
    training, 3 panellists × 2 days × £250        £1,500 / yr
    reserved panel time, 3 × 12 h × £30           £1,080 / yr
    convening and the record                        £920 / yr
    F                                             £3,500 / yr

  PER DISPUTE
    3 panellists × 3 hours × £30                    £270
    notice, minute, follow-up                        £60
    c                                               £330

  AT λ = 0.05 disputes per member per year (10 a year)
    F / N                                          £17.50 /member/yr
    λ · c                                          £16.50 /member/yr
    TOTAL                                          £34.00 /member/yr
                                                  (£2.83 /member/month)

Thirty-four pounds a member a year. With the arbitration backstop added it is £43.50, or £3.62 a month.

Why it matters. Almost everybody guesses this number an order of magnitude high, and the guess — not the cost — is why the clause gets written as the parties shall seek to resolve the matter amicably.

You already know this because you have been in an organisation where the only two responses to a problem were ignore it and fire somebody, and you watched the first one get used forty times.

---

BRIEF 9 — The Dispute-Rate Threshold

The idea. There is a rate above which running a repair ladder is cheaper than throwing people out, and it is computable before you write a single rule.

The cost of one exclusion. Replacement cost at 21% of a £30,000 salary (Boushey and Glynn, 2012) is £6,300; a 15% chance of a claim costing £8,500 to defend adds £1,275 expected. X = £7,575 — about 23.0× the cost of one internal hearing.

The threshold. Let q be the share of disputes ending in exit with no forum, s the share still reaching exclusion with one.

             F / N                    £17.50
  λ*  =  ------------------  =  -------------------  =  0.01119 /member/yr
          (q − s)·X  −  c       0.25×7,575 − 330
                                    = £1,563.75

      =  1.12 disputes per hundred members per year
      =  2.24 disputes per year in a house of two hundred

The cut. Because F is fixed, the threshold in whole disputes does not move with size. Fifty members or five thousand: 2.24 disputes a year, one every five and a half months.

The floor. The ladder must divert at least c / X = 4.4% of disputes from exclusion or no threshold exists — a forum that hears everything and changes nothing costs more than the exclusions it fails to prevent.

Why it matters. The observed rate in the worked house is 4.5× the threshold, and that is a quiet house. The decision was never whether to afford a conflict system.

You already know this because you can name, without effort, more than two people who left an organisation you were in over something that was never heard.

---

BRIEF 10 — The Shadow of the Last Rung

The idea. The terminal rung's job is not to be used. It is to set the price of every rung beneath it.

Robert Mnookin and Lewis Kornhauser named this in 1979: parties bargaining privately settle against their estimate of what would happen if they did not. Change what happens if they do not, and every private conversation changes, including the ones where nobody mentions it.

The cost of the last rung. A named external arbitrator on a standing retainer: £1,200 a year, plus £3,500 per hearing actually held, with 2% of ten annual disputes reaching it — 0.20 hearings a year. Expected annual cost £1,900, or £9.50 per member per year: a 27.9% uplift on the £34.00 system.

What the £9.50 buys. Not the 2%. The 98%. Non-compliance without a fallback runs at 29.4%; ten disputes a year produces 2.9 broken agreements, worth £22,271 a year if each becomes an exclusion — £111.35 per member per year. The fallback pays for itself if it prevents a quarter of one exclusion a year.

The term that makes it legitimate. Both-ways invocation: any member may escalate to the arbitrator, not only the organisation. Without it the fallback is a management instrument and the ladder beneath it is theatre. With it, the refuser who is correct — the minority member, the whistleblower, the person with a complaint against the panel's own constituency — has somewhere to go.

You already know this because you have negotiated harder with someone who had no alternative than with someone who did, and you did not have to be told to.