Haute Lumière

Commerce · II.03 · MMXXVI · daylight

La Bourse  /  Volume II  /  Nº II.03  /  Ten concept briefs

TEN CONCEPT BRIEFS · Chapter II.03 — Non-Dual Economics

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

---

BRIEF 1 — A Boundary Is an Election

The idea. The divisions that organise economics — producer and consumer, firm and environment, worker and owner, self and other — are not observations. They are elections made inside accounting standards, and the standards say so, in writing, with the alternatives listed beside them.

This is not a critique of accounting. It is what accounting is. A set of accounts is a statement about a reporting entity, and the reporting entity has to be defined before a single number can be written. The definition is the first act, it is a choice, and every figure downstream inherits it.

Worked example. The GHG Protocol offers three consolidation approaches — equity share, financial control, operational control. A firm with 500 kt of emissions from wholly owned operations and a fifty-fifty joint venture emitting 300 kt may report 500 kt, 650 kt or 800 kt. All three are compliant. The spread is 60 percent of the lowest figure, and a footnote records which was taken.

The test that keeps this honest. If a boundary were a fact, moving it would be an error. Moving these is a disclosure. That difference is the whole brief.

Why it matters. Once you see a boundary as an election, three questions become available that were not available before: where is it now, where else could it legitimately be, and who is paid by where it sits. The rest of the chapter is those three questions applied four times.

You already know this because you have watched a departmental budget argument in which nobody disagreed about any number — only about which costs belonged to which department. Everyone in the room understood, without saying so, that the line was the thing being negotiated.

---

BRIEF 2 — The Consolidation Boundary

The idea. A company's balance sheet contains what the company controls, and control is defined by a rule that has changed twice in living memory.

The rules, in order.

The arithmetic. One unit of outside at-risk equity holds 33.33 units of assets off balance sheet under a three percent rule, and 10.00 under a ten percent rule. The 2003 reform cut that leverage by 3.33×.

The worked case. The Powers Report (2002) found that Chewco's outside equity of $11.5 million included $6.6 million funded by loans cash-collateralised by Enron itself. The equity was never genuinely at risk, so the three percent test never held. Consolidating the entities retroactively reduced Enron's reported 1997–2000 net income by $586 million. No asset moved.

What to notice. The scandal is usually told as a story about lying. The accounting story is narrower and more useful: a rule that draws a boundary at a numerical threshold creates a market in the threshold.

Why it matters. Every ratio a lender covenants on — leverage, gearing, interest cover — is computed on the consolidated entity. Move the entity and you move every ratio without touching the business.

You already know this because you have seen a group reorganisation that changed nothing about the work and everything about the reported figures, and you have wondered which one of those was the point.

---

BRIEF 3 — The Boundary Yield

The idea. The sliver of outside equity that holds an entity off a balance sheet is paid a fee measured against the assets it moves, and risks only itself. So the boundary has a return, and the return can be quoted.

  boundary yield  =  fee on assets moved  /  outside equity fraction

The table.

The reading that changes things. FIN 46(R) did not abolish this trade. It repriced it, by a factor of 3.33. The most consequential accounting reform of its decade was a price change, and it should be read as one: it raised the cost of a boundary election rather than removing the election.

The general form. Wherever a standard sets a numerical threshold for where a line falls, the reciprocal of that threshold is the leverage available to whoever is willing to stand on the line, and any fee paid for standing there divided by the threshold is their return. This holds for thin-capitalisation rules, for beneficial-ownership tests, for materiality thresholds, and for safe-harbour percentages of every kind.

Why it matters. It tells you what a reform did. Reforms that change a threshold change a price. Reforms that replace a threshold with a judgement — IFRS 10 — change who has to defend the answer, which is a different and usually stronger move.

You already know this because you have watched a deadline or a limit become a target the moment it was published, and you have noticed that the people best positioned to sit exactly on it were the ones who read the rule most carefully.

---

BRIEF 4 — Scopes, and the Fence Around the Firm

The idea. The GHG Protocol splits emissions into scope 1 (own combustion), scope 2 (purchased energy) and scope 3 (the rest of the value chain, in fifteen categories). The split is a fence, and most companies report from inside it.

How much is inside. Two independent measurements:

Different data, different method, same neighbourhood: a firm reporting scopes 1 and 2 has drawn a fence around roughly one-fifteenth of what it causes.

Worked example. PUMA's 2010 account found €145 million of environmental cost across the value chain, of which €8 million arose in operations PUMA owned. The remainder sat with tanneries, mills, farms and freight — all of them, in the conventional portrait, somebody else.

What scope 3 is and is not. It is deliberately double-counted across entities, so corporate inventories cannot be summed to a national total. That is a design feature and it is stated in the standard. It makes scope 3 an excellent tool for finding where to act and a poor one for adding up.

Why it matters. The abatement curve drawn across the whole chain is usually cheaper than the one drawn to the factory gate — which means the fence is not only a reporting choice, it is a cost choice. Companies that moved the fence outward frequently found the cheapest tonnes on the other side of it.

You already know this because you have been in a procurement conversation where the cheapest improvement available was in somebody else's plant, and the only obstacle was that it was somebody else's plant.

---

BRIEF 5 — The Outsourcing Move

The idea. Sell the plant, contract to buy the same output, and a reported emissions figure can fall to zero while the physical emissions do not change at all.

  reported, on a scope 1+2 boundary    100 kt  ->    0 kt   (-100 %)
  physically emitted                   100 kt  ->  100 kt   (   0 %)

Why the standard does not catch it automatically. The GHG Protocol requires base-year recalculation for structural change — acquisitions, divestments, outsourcing. But a company that does not report the relevant scope 3 category has nothing to recalculate. The requirement is real and it is conditional on a disclosure the company chooses whether to make.

The general form. Any target expressed over a perimeter that the target-setter also controls can be met by moving the perimeter. This is not confined to emissions. It applies to headcount targets met by contracting, to injury rates met by reclassifying contractors, to cost-per-unit met by moving a cost centre, and to leverage covenants met by deconsolidation.

The fix, in one clause. Freeze the perimeter in the document that sets the target, and require restatement of the baseline on any perimeter change regardless of what is reported. That clause is the whole of the Operationalize movement, and it is two sentences long.

Why it matters. Because a target without a frozen perimeter measures the skill of the person defining the perimeter, and measures nothing else.

You already know this because you have seen a team hit a number by changing what counted, and you have noticed that nobody could point to a rule they broke.

---

BRIEF 6 — Household Production and the Growth Wedge

The idea. GDP counts a meal cooked for money and not the identical meal cooked for love. The line between them is a line in a statistical standard, and it moves.

Pigou's case, 1920. A man marries his housekeeper and the national dividend falls. The following morning's output is unchanged; a £30,000 wage has simply left the accounts.

The measurement. The BEA's household production account (Bridgman et al., 2012) finds that including household production would have raised US nominal GDP by 39.0 percent in 1965 and 25.7 percent in 2010.

The arithmetic those two numbers make possible.

  ln(1.390 / 1.257) / 45 years  =  0.22 percentage points a year
  cumulative over 45 years      =  10.6 %

For forty-five years, measured growth overstated total production by about a fifth of a point annually, because work crossed from kitchens onto payrolls. Some of that crossing was genuine specialisation with genuine gains. Some was reclassification recorded as growth. The national accounts cannot tell you the split, by construction.

The independent check. The ONS valued UK household services at £1.24 trillion in 2016, 63.1 percent of GDP. Dividing gives an implied nominal GDP of £1.97 trillion, against £1.99 trillion in the national accounts — 1.2 percent apart. Two publications with different purposes describing one economy.

Why it matters. Any policy judged on GDP growth is partly judged on how much unpaid work has been marketised during the period. That is a real effect with a computable size, and it is roughly a fifth of a point a year.

You already know this because you have done work at home that would have cost money to buy, and you have never once seen it in a national figure.

---

BRIEF 7 — The Arm's-Length Principle

The idea. When two parts of the same company transact, there is no market price, because there is no market. The OECD requires the transaction to be priced as though the parties were independent — a boundary that does not exist, priced as if it did.

The mechanics. Five permitted methods (comparable uncontrolled price, resale price, cost plus, transactional net margin, profit split), each producing a range rather than a point. Documentation defends the choice; examination contests it.

Worked example. €1,000 million of in-country sales; a licence to the group's intellectual property holder priced at three percent under one method and six percent under another.

  taxable profit moved out of the market country   EUR 30.0 m
  tax moved at a 25 % rate                         EUR  7.5 m

No physical or commercial fact changed. One method was selected instead of another.

The aggregate. Tørsløv, Wier and Zucman (2023) put roughly 36 percent of multinational profits in havens — about $600 billion in 2015 and $200 billion of tax. The OECD's Action 11 work, on different data, put the annual loss at $100–240 billion, four to ten percent of global corporate income tax receipts. Two strategies disagreeing about method, agreeing about magnitude.

Why it matters. It is the clearest case in economics of a boundary that is known to be a fiction, is required to be priced, and moves a sum the size of a mid-table national budget every year.

You already know this because you have transferred something between two parts of one organisation and been asked what it was worth, and you have noticed that the answer depended entirely on who was asking and why.

---

BRIEF 8 — The Worker/Owner Line

The idea. A payment to a person above the line is an expense and reduces operating profit. The identical payment to the identical person below the line is a distribution and does not. Enterprise value is quoted as a multiple of the line.

Worked example. A firm with €10 million of EBITDA and €40 million of payroll moves ten percent of payroll into profit distribution to the same people.

  moved                                      EUR  4.0 m
  EBITDA after                               EUR 14.0 m   (+40 %)
  enterprise value at 8x, before / after   80.0m / 112.0m
  created by the reclassification            EUR 32.0 m

Identical cash, identical people, same month.

The honest negative, immediately. An equity analyst normalises that in one line. And the people who were creditors of the payroll are now residual claimants carrying risk they did not carry before. Moving this line without moving governance with it is a transfer of risk dressed as a transfer of ownership.

Where it is done properly. The National Center for Employee Ownership counts roughly 6,500 ESOPs in the United States with some fourteen million participants — about 2,150 each. The UK's Employee Ownership Trust, created by the Finance Act 2014, provides a conversion route with the tax treatment written around it. In both, the ownership carries voice, information rights and a claim on the residual, which is what makes the reclassification real.

Why it matters. The worker/owner boundary is the one where the accounting move and the human move can be made independently — and where making only the accounting move is the most common failure in the whole field.

You already know this because you have been paid a bonus and understood perfectly well that it was not the same thing as owning any part of what you built.

---

BRIEF 9 — The Double-Count, and Ostrom's First Principle

The idea. This is the brief that argues against the chapter, and it is the most important one.

Refuse to draw a boundary and you cannot count. Take a supply chain of n equal tiers, each truthfully reporting its full upstream footprint. Tier one's emissions are counted n times; the sum of inventories over physical emissions is (n+1)/2.

The same factor applies to claimed reductions. One tonne abated at the bottom of a ten-tier chain can be claimed, truthfully and simultaneously, by every firm above it.

Ostrom's first design principle. In Governing the Commons (1990), the first of the eight principles for a long-enduring commons is clearly defined boundaries — of the resource and of who holds rights in it. The empirical record behind that principle is the strongest in the field, and this chapter does not argue with it.

The distinction that survives. The claim is that a boundary is chosen, not that it can be dispensed with. An accounting with no boundary has no unit. A number with no unit cannot be audited, financed, or defended in a court.

Why it matters. It is what makes the rest of the chapter usable. A reader who took "boundaries are conventions" to mean "draw no boundaries" would produce an unauditable document and a legal entity their creditors could not read.

You already know this because you have been in a project where everyone was responsible for quality, and you have seen what that did to quality.

---

BRIEF 10 — The Boundary Register

The idea. One artifact makes the whole chapter operational, and it fits on a page: one row per boundary, five columns.

The fifth column is the one that does the work, and it is the one that will be resisted. It is not an accusation: often the answer is benign — the joint venture partner, whose covenants are set on the equity-share basis — and writing it down is what makes benign cases visible as benign.

The rule that makes it hold. Any change to an elected boundary requires the prior period to be restated on both bases and presented side by side. That single clause removes most of the value of moving a line quietly, which is most of the reason lines are moved quietly.

Sequence. Emissions perimeter first: clearest published alternatives, lowest internal charge, an external standard to point at. Intra-group pricing last, with tax counsel in the room, because there the register is discoverable.

What it is not. It is not a campaign to move every line outward. Some elections, once visible, should stay exactly where they are — and the register is how you find out which.

You already know this because you have kept a list of the assumptions behind a model, and you have noticed that the list was more useful to the next reader than the model was.

---

All figures in these briefs are computed in lib/verify/II_03.py and sourced in the chapter's Works Cited.