Haute Lumière
Commerce · VII.04 · MMXXVI · daylight
TEN CONCEPT BRIEFS · Chapter VII.04 — The Ocean and Atmospheric Commons
One page each. A reader who reads only these ten pages has the chapter.
BRIEF 1 — A Commons With No Edge
The idea. Ostrom's first design principle is clearly defined boundaries: who may draw, and from where. The atmosphere and the high seas cannot satisfy it. There is no line to walk, no neighbour to notice you, and no village that could be handed the rule.
This is not a difference of degree from a village fishery. It is a difference of kind, and it is why Chapter VI.04's successful commons do not scale up. Every one of them ran on a boundary plus mutual observation. Remove both and the whole institutional apparatus that makes a commons work has nothing to attach to.
The figure. The high seas run to 231 million square kilometres — 64 per cent of the world ocean. About 3,620 vessels fished it in 2016. That is one vessel per 64,000 square kilometres, each one invisible from every other.
Worked example. Imagine the Törbel alps with no fence, no register of households, and grazing rights held by anyone who owns a vehicle. Every rule Törbel wrote in the fifteenth century would still be correct and none of it would be operable. That is the high seas, and it has been that way since Grotius.
Why it matters. It tells you where not to spend effort. A regime for a commons with no edge cannot be built out of neighbourliness, monitoring by appropriators, or graduated local sanctions. It has to be built somewhere else entirely — and Brief 8 says where.
You already know this because you have worked in an organisation where a shared resource — meeting rooms, a budget line, an engineer's time — had no owner and no queue, and you watched perfectly reasonable people empty it without any of them doing anything wrong.
BRIEF 2 — Γ and Λ: Measuring the Gap
The idea. Everyone says fisheries bodies ignore their science. Almost nobody states the size of the gap, which means nobody can tell a body that is improving from one that is not. Two ratios fix that.
Γ = quota adopted / quota advised politics over science
Λ = catch taken / quota adopted behaviour over politics
Γ · Λ = catch taken / advice the total gapWorked example. ICCAT's scientists advised 8,500 to 15,000 tonnes of eastern Atlantic bluefin for 2009. The Commission adopted 22,000.
Γ (upper advice) 22,000 / 15,000 = 1.47 ×
Γ (lower advice) 22,000 / 8,500 = 2.59 ×
Λ (2007) 61,000 / 29,500 = 2.07 ×
total gap 61,000 / 15,000 = 4.07 ×The figure that makes it an instrument, not an accusation. CCAMLR's Γ for Antarctic krill is 0.111 — it operates a trigger level of 620,000 tonnes against its own precautionary limit of 5.61 million. Two bodies, one treaty family, a thirteen- to twenty-three-fold spread.
Why it matters. A ratio can be published annually, compared across bodies, and put beside a vote. An accusation cannot. The thing that has ever moved a quota back toward its science is the advice being visible next to the decision.
You already know this because you have seen a forecast that everyone knew was optimistic get approved anyway, and you have noticed how much harder that is to do when last year's version of the same forecast is printed on the facing page.
BRIEF 3 — The Deterrence Condition
The idea. Whether an operator complies is not a question about character. It is one line of arithmetic, from Gary Becker in 1968 and applied to fisheries by Sutinen and Andersen in 1985:
comply iff p · F ≥ G
p probability of being detected AND sanctioned
F the penalty if you are
G the net gain from violatingWorked example. A distant-water tuna longliner: 250 t a year at US$6,000 a tonne, so US$1.5 million gross. It under-reports 20 per cent — 50 tonnes, US$300,000 gross, US$180,000 net of the cost of taking it. Observer coverage on WCPFC longliners is 5 per cent; say half of what is reported gets sanctioned, so p = 0.025. A strong administrative fine is US$100,000.
expected penalty = 0.025 × 100,000 = $2,500
G / (p · F) = 180,000 / 2,500 = 72 ×The figure. Seventy-two. The violation pays seventy-two times what it is expected to cost.
Why it matters. It relocates the problem. Nobody in that wheelhouse is making a moral decision; they are reading a price. And a price can be changed by somebody who is not on the boat.
You already know this because you have parked illegally somewhere the penalty was smaller than the cost of the car park, and you did the sum without noticing you were doing it.
BRIEF 4 — The Deterrence Frontier, and the Probability Above One
The idea. p · F = G is a hyperbola, not a point. It tells you every combination of surveillance and penalty that deters the same violation — and, more usefully, which combinations are impossible.
Worked example. Invert Brief 3 both ways.
the fine that deters at today's coverage
F* = G / p = 180,000 / 0.025 = $7,200,000
= 4.8 × the vessel's annual revenue
the coverage that deters at today's fine
p* = G / F = 180,000 / 100,000 = 1.80The figure. 1.80. A probability cannot exceed one. At a US$100,000 fine, no monitoring regime of any density deters this violation — not an observer on every vessel, not a camera on every hauler, not a satellite over every square kilometre.
The frontier, read as a menu:
p F required as a multiple of annual revenue
0.025 $7,200,000 4.8 ×
0.12 $1,500,000 1.0 ×
0.50 $360,000 0.24 ×Why it matters. This is the chapter's central reversal. The high seas is not ungoverned because it is too big to watch. It is ungoverned because the fine is too small — and the fine is set by flag states whose revenue comes from the fleet. Fix F first, then buy p. Doing it the other way round buys nothing at any price.
You already know this because you have watched an organisation respond to a compliance failure by adding another audit, when everyone involved knew that failing the audit had never cost anybody anything.
BRIEF 5 — Patrol Is Not Observation
The idea. Two completely different problems have been called "monitoring the ocean" for forty years, and confusing them made the cheap one look impossible.
Patrol means being physically present where a violation is happening. Observation means holding a record of where vessels were and what they did.
Worked example. A patrol vessel with a 25 km detection radius steaming at 12 knots sweeps 26,669 km² a day.
231,040,000 km² ÷ 26,669 km²/day = 8,663 vessel-days
= 23.7 vessel-years
at $30,000/day = $260 millionFor one sweep — which detects almost nothing, because the ship has moved. Continuous satellite AIS and VMS analysis over the same water runs on the order of $10 million a year: twenty-six times cheaper than a single sweep that does not work.
The figure. 26 ×, and the direction of it is the whole point.
Why it matters. The ocean's size was always an argument about patrolling. It was never an argument about observation, and observation is what evidence requires.
You already know this because you stopped trying to catch people misusing a system by walking around, and started keeping a log — and the log worked and the walking never had.
BRIEF 6 — The Four Montreal Conditions
The idea. The Montreal Protocol is the one unambiguous success in global environmental governance. It succeeded because four things were simultaneously true. Each is a design condition, not luck, which means each can be asked about any other problem.
- A substitute existed, and the incumbents owned it. CFCs → HCFCs and HFCs: a molecular swap into largely existing equipment, with the patents held by the firms being regulated. The phase-out moved rents rather than destroying them.
- Production was concentrated. About 17 firms made the world's CFCs; DuPont alone about a quarter. Seventeen points to regulate.
- The science was unarguable and answered inside a career. Farman published the hole in 1985, the Protocol was signed in 1987, and healing was measurable by about 2000. The signers saw the reward.
- The trade provisions had teeth. Article 4 banned trade in controlled substances with non-parties — self-enforcing, as Scott Barrett showed, because the sanction cost the non-party more than the party.
The figure. The Multilateral Fund disbursed more than $4 billion, and the Protocol was already avoiding 11 Gt CO₂-eq a year by 2010 — about $364 million of one-off transfer per gigatonne-per-year avoided. The cheapest climate instrument ever built, and it was not built as one.
Why it matters. Four conditions is a checklist, and a checklist can be run against your own problem in an afternoon.
You already know this because you have seen one change land easily and an identical change fail elsewhere, and the difference was never the argument.
BRIEF 7 — Why Montreal Does Not Transfer to Carbon
The idea. "If we did it for ozone we can do it for carbon" is the most repeated and least examined sentence in climate policy. Run the four conditions and it fails all four — precisely, not vaguely.
Condition 1 — the substitute. There is no molecular swap for combustion. The replacement is a different capital stock across every sector, and the incumbent owns the resource, not the substitute. Montreal moved rents. Carbon strands them.
Condition 2 — concentration. This is the arithmetic one.
regulated points, Montreal 17 producers
motor vehicles in use 1,500,000,000 engines
ratio 88,000,000 × ≈ 7.9 orders of magnitudeThe concentration of fossil producers is real and beside the point: the regulated act is combustion, and combustion is everywhere. Vehicles are only one class of point.
Condition 3 — the signal. Farman to Protocol was 2 years; Protocol to visible healing about 13. About 25 per cent of a pulse of CO₂ is still airborne after a thousand years. The carbon signer's reward is the absence of a worse counterfactual, which is not observable by anyone, ever.
Condition 4 — the border test. A customs officer could open a cylinder and run a binary test. Carbon embodied in a shipment is a fact about the object's history, not a property resident in the object — so it must be certified, not measured. That is what the EU's border adjustment is building, sector by sector, and why it covers six of them and not an economy.
Why it matters. A lesson misapplied is worse than a lesson not learned, because it spends the credibility of the case that worked.
You already know this because you have watched a playbook that worked brilliantly in one business unit get mandated across a company where none of its preconditions held.
BRIEF 8 — The Commodity Test, and the Chokepoint
The idea. Here is what the ocean can borrow, and it is condition 4 — the one carbon cannot have.
A tuna is a physical object that crosses a customs border in a box. Its legality can be attached to it with a document and tested in one place by one person. That is exactly what Article 4 needed, and it is why the ocean's version of Montreal is a paperwork regime, not a navy.
Then move the test off the water.
moving points at sea 3,620 vessels across 231,040,000 km²
fixed points on land 20 hub ports, which do not move
ratio 181 ×Worked example. CCAMLR's Catch Documentation Scheme for toothfish, in force since 1999, put a document on every consignment. Estimated illegal catch fell from tens of thousands of tonnes in the mid-1990s to a residual an order of magnitude smaller. The mechanism was a piece of paper stapled to a box.
The EU's IUU Regulation does the same with catch certificates and carding: Thailand was yellow-carded in April 2015 and cleared in January 2019 — 45 months, during which it fitted vessel monitoring to its commercial fleet and rebuilt its port controls. The FAO Port State Measures Agreement, in force since June 2016, is the same instrument as treaty law.
Why it matters. Three schemes exist and they are not one. A single interoperable catch documentation standard, recognised by the three markets that buy most traded seafood, is the ocean's Article 4 — and it is buildable without a new treaty.
You already know this because you have seen a control that nobody obeyed when it was a policy, and that everybody obeyed the moment it became a required field in the system that issues the invoice.
BRIEF 9 — The Compliance Assurance Bond
The idea. A fine against a foreign-flagged vessel is a judgment, and a judgment is worth its probability of collection. Collateral already posted is worth its face. That single substitution is the instrument.
The mechanics. Size the face on the deterrence frontier at the coverage the vessel actually carries, not by negotiation with the industry.
at 100% electronic monitoring, p = 1.00 × 0.50 = 0.50
F* = G / p = 180,000 / 0.50 = $360,000
bond face, rounded up = $400,000 (26.7% of revenue)
surety premium at 2% = $8,000 / yr
electronic monitoring = $9,400 / yr
total cost to the operator $17,400 / yr = 1.16% of gross revenueThe number that decides it. Expected penalty divided by illicit gain, which must exceed 1.00:
today (0.025 × $100,000) / $180,000 = 0.014
bonded (0.500 × $400,000) / $180,000 = 1.11 an 80 × improvementThe part that is not obvious. The counterparty is a surety, not a state. The underwriter now has money at risk against the vessel's conduct, so the underwriter reads the monitoring data — every year, including the years nobody is interested. You have created a paid, permanent, private monitor with a balance-sheet interest in accuracy, and you did not legislate one.
Balance-sheet treatment. A bond is a contingent liability with a fee. The premium is opex; posted cash is restricted cash, an asset. The P&L moves by $17,400 while the expected penalty moves eighty-fold.
You already know this because you have taken a deposit from somebody, and you know exactly how differently people treat money you are already holding.
BRIEF 10 — A Fine Is a Price
The idea. The honest limit of everything above. The deterrence model assumes a risk-neutral operator computing one expected value. Most compliance, in most fisheries, is not computed — it is normative, as Sutinen and Kuperan showed: people follow rules they regard as legitimate, and they do it far more cheaply than any penalty regime could buy.
Worked example. Uri Gneezy and Aldo Rustichini studied Israeli day-care centres that introduced a small fine for collecting children late. Late collections rose, and stayed high after the fine was withdrawn. The fine converted a moral obligation into a purchase, and once a thing has a price the question of whether to buy it is entirely reasonable.
Why it matters here. Set F at seven million dollars and you have told every honest skipper in the fleet what the going rate for the sea is. The frontier is a necessary condition for deterring a calculating violator. It is not a sufficient condition for a fleet, and a regime built only on it can destroy the free compliance it was quietly running on.
What to do with that, rather than be paralysed by it. Three things. Aim the penalty at the operators who are already calculating — which is a small, known, repeatedly-named set — rather than at the fleet. Make the bond a record the compliant operator accumulates, so the instrument rewards a good history rather than merely punishing a bad one. And publish Γ and Λ, because the thing norms respond to is visibility, not price.
You already know this because you have been trusted with something and found that the trust was a stronger constraint than any rule would have been — and you know how quickly that would have changed if somebody had put a tariff on it.