Haute Lumière
Commerce · VII.03 · MMXXVI · daylight
TEN CONCEPT BRIEFS · Chapter VII.03 — Biodiversity as Infrastructure
One page each. A reader who reads only these ten pages has the chapter.
BRIEF 1 — The Service Radius
The idea. Pollination is not delivered to a farm. It is delivered to a distance, and the distance is measurable.
Across twenty-three studies covering sixteen crops on five continents, Ricketts and colleagues found that pollinator visitation rate falls by half at a median of about 0.6 km from natural habitat, and pollinator richness at about 1.5 km. The service decays outward from a source like light from a lamp.
Worked example. A forty-six-hectare woodland is, treated as a disc, 0.383 km across the radius. The ring of crop within one kilometre of its edge is 554.6 hectares. That ring is the asset's catchment, and every hectare inside it is a hectare the woodland is working on.
Why it matters. A service with a decay function can be sited, sized, bought and sold. Before this result, habitat was an amenity; after it, habitat is plant with a throughput. Everything else in this chapter — the leverage ratio, the royalty, the credit paper — is downstream of one number with a unit of distance on it.
You already know this because you already know that a warehouse is worth more near the motorway, and nobody had to prove to you that the motorway was the reason.
BRIEF 2 — The Leverage Ratio, L
The idea. A hectare of habitat is worth however many hectares of crop it can reach. That ratio is L, and it is the single largest controllable term in the value of habitat.
L = hectares of crop inside the service radius
------------------------------------------
hectares of habitatWorked example, and it is the chapter's sharpest finding. At Finca Santa Fe, two forest fragments of 46 ha and 111 ha sat at the farm's edge. Geometrically, those 157 hectares could reach 1,242 hectares of coffee — a ceiling of L = 7.9. Running the published service value backwards gives what actually happened: L = 2.07. A gap of 3.8 times, and none of it is biology. The blocks overlapped each other and spilled off the property.
Why it matters. L enters the value formula linearly. Doubling L doubles the value of every hectare of habitat on the farm, and L is changed by placement, which is free. Four metres of flowering margin running through the middle of a block is worth several times the same area banked along a boundary.
The instruction this yields. Compute L before arguing about how much habitat to have. Quantity is worth a factor of one; configuration is worth a factor of four.
You already know this because you already know that one fire exit in the middle of a floor beats two in the same corner, and for exactly the same reason.
BRIEF 3 — The Marginal Hectare
The idea. One formula prices a hectare of pollinator habitat, and every term in it is measurable on a real farm.
V = L · Y₀ · Δ · P · φ
V value per hectare of habitat per year
L leverage — crop hectares served per habitat hectare
Y₀ yield without the service, t/ha
Δ proportional yield uplift
P farmgate price, $/t
φ attribution — share of Δ owed to this habitatWorked example. Costa Rican coffee: L = 4.0 (dispersed placement), Y₀ = 1.0 t/ha, Δ = 0.20 (measured), φ = 0.70.
P = $1,320/t (trough) V = $ 739 /ha/yr capitalised at 7% $10,560
P = $3,000/t (mid) V = $1,680 /ha/yr $24,000
P = $6,000/t (peak) V = $3,360 /ha/yr $48,000Why it matters. Against a pasture return near $150/ha/yr, habitat wins by five to twenty-two times at every price. Against a coffee margin near $800, it wins above $1,429/t — sixty-five cents a pound. That is a date-and-price answer, and a credit committee can act on it.
The uncomfortable half. V is proportional to P. The value of a hectare of habitat is a derivative on a commodity price, and a clearing decision taken in a trough is a permanent bet against mean reversion.
You already know this because you have watched somebody sell a machine in a bad year and spend three good years renting one back.
BRIEF 4 — Attribution, φ
The idea. Not all of a measured uplift belongs to the habitat you are pricing, and an instrument that claims all of it loses its first dispute.
φ is the share of the measured yield differential attributable to this parcel rather than to managed hives, other remnant vegetation, drainage, aspect, or the simple fact that the blocks nearest the forest happen to be the older ones.
Worked example. The chapter uses φ = 0.70. At L = 4.0, Y₀ = 1.0, Δ = 0.20 and P = $3,000/t, moving φ from 0.70 to 1.00 moves V from $1,680 to $2,400 a hectare — a forty-three per cent swing produced by an assumption nobody measured.
How to fix it rather than argue it. Match the paired blocks properly: cultivar, planting year, management regime, aspect. Then state φ in the contract as a term, with a review at year five. A stated assumption is auditable. An implied one is a future dispute.
Why it matters. Every credible number in ecosystem-service finance has an attribution term, and most published headline figures quietly set it to one. The discount is what makes the rest of the arithmetic survivable.
You already know this because you have seen a marketing team claim an entire quarter's growth, and you know exactly what the sales director said about it.
BRIEF 5 — Tonnage Is Not Value
The idea. Pollinator dependency is a cash-crop fact, not a calorie fact, and the two arguments have very different strengths.
The numbers. Of the 115 leading global food crops, 87 — 75.7 per cent — depend on animal pollination to some degree; 28 do not. Those eighty-seven are about 35 per cent of global production by volume. But Gallai and colleagues valued the pollination contribution at €153 billion, which is 9.5 per cent of the value of world crop production for human food — implying a denominator near €1,611 billion. IPBES puts the market value of crops directly affected by pollinators at $235–577 billion.
The tension, and the reading. Thirty-five per cent of the tonnage; nine and a half per cent of the value. Pollinator-dependent crops sell for roughly five times as much per tonne. The staples that feed people — wheat, rice, maize — are wind- or self-pollinated. The dependency sits in the expensive half of the diet: fruit, nuts, coffee, cocoa, oilseeds.
Why it matters. Argued as food security, the claim is weak and a hostile reader will dismantle it in one paragraph. Argued as cash-crop revenue, it is strong, specific, and lands with the people who own the land.
You already know this because you know a restaurant's margin is not in the bread.
BRIEF 6 — Why Pest Regulation Does Not Divide
The idea. A national aggregate that sounds enormous can be worthless at the scale where the decision is taken, and this is the cleanest example in the field.
Worked example. Losey and Vaughan costed pest control by native natural enemies in the United States at $4.5 billion a year. Divide by 160 million hectares of US cropland:
$4,500,000,000 / 160,000,000 ha = $28.12 per hectare per yearTwenty-eight dollars — below the cost of a single insecticide pass.
And it does not generalise. Bianchi and colleagues found natural enemy populations higher in complex landscapes in 74 per cent of studies, but pest pressure actually reduced in only 45 per cent — a 29-point gap between having the predators and having less damage. Karp and a large team, across 132 studies and 6,759 sites, found landscape composition explained little of the variation and the direction of effect was close to a coin flip. Dainese and colleagues, across 89 studies and 1,475 fields, did recover a signal — but it runs through the richness of service-providing organisms, not the quantity of non-crop habitat.
Why it matters. Pollination generalises and can be underwritten. Pest regulation does not, not yet, not at field scale. Carry the first as revenue and the second as an option with no strike price, and you will never be caught.
You already know this because you have seen a national average used to justify a local decision, and you know how that meeting ended.
BRIEF 7 — The Saturation Curve Is in the Denominator
The idea. The biodiversity–function curve saturates hard, and it saturates because of what was measured, not because of what is true.
The curve. Fit Y = a + b·ln(S). The marginal value of the next species is b/S, halving with every doubling:
S = 2 next species worth 0.5000 b
S = 4 0.2500 b
S = 8 0.1250 b
S = 16 0.0625 bThe sixteenth species is worth one eighth of the second. And only 12 per cent of polycultures in Cardinale's 574 experiments beat their own single most productive species. Read alone, that kills the case for high richness.
The denominator. Every one of those curves was fitted to one function, in one year, in one place. Isbell and colleagues widened it and found 84 per cent of 147 species promoted ecosystem functioning at least once — different species, different years, different functions.
Worked example — the two results reconciled. Let each function-year-place draw about twelve species from a pool of 147 and count the union:
species required = 147 · (1 − (1 − 12/147)^c)
c = 1 context 12.0 species 8.2 %
c = 10 84.3 57.3 %
c = 22 124.4 84.6 %
c = 30 135.6 92.2 %Twenty-two contexts reproduce the observed 84 per cent from a curve that looked flat after three species.
You already know this because you know a team looks over-staffed in any one week and exactly right across a year.
BRIEF 8 — Two Markets in One Asset
The idea. The same hectare has two prices, set by unrelated variables, and confusing them is the most common error in this field.
Price one — the service. The measured agronomic value: $739 to $3,360 per hectare per year, depending on the commodity price.
Price two — the permit. England's biodiversity net gain regime. A hectare of medium-distinctiveness habitat in good condition scores 4 × 3 × 1.0 = 12 biodiversity units at baseline; created, the metric applies a difficulty multiplier of 0.67 and a temporal multiplier near 0.50, leaving 4.02 units per hectare. At a reported market price of £25,000 a unit that is £100,500 a hectare, one-off, against a thirty-year obligation — £6,538 per hectare per year at a five per cent annuity (factor 15.3725). At the lowest statutory credit tier of £42,000 a unit, £10,983 a year. The statutory schedule runs to £650,000 a unit.
The comparison. At $1.27 to the pound, the compliance price is $8,303 per hectare per year — 11.2× the measured service at trough prices, 4.94× at mid-cycle, 2.47× at peak. Against an English arable margin near £500, it is 13.1× the farming alternative.
Why it matters. The compliance price is set by the cost of the developer's alternative, which is planning refusal. It is not an ecological signal and must never be read as one. A landowner selling into it is selling a permit, and the agronomic value is a separate asset they have not sold.
You already know this because you know what a parking space costs in a city and what the tarmac cost.
BRIEF 9 — Substitutability and the Offset Ratio
The idea. Every offset market rests on one assumption — that a created habitat substitutes for a destroyed one — and for slow-forming systems that assumption is measurably false.
The measurement. Moreno-Mateos and colleagues synthesised 621 restored and created wetland sites against reference wetlands. Biological structure came in 26 per cent below reference; biogeochemical function 23 per cent below. The deficit was still measurable in sites restored a century earlier.
The arithmetic that follows. A one-to-one offset against a 77 per cent function delivery rate writes a 23 per cent loss into the instrument. To hold function constant:
1 / 0.77 = 1.30 hectares created per hectare lostAlmost no scheme requires it. Curran and colleagues showed that for old-growth forest, peatland and mature sclerophyll, measured recovery does not reach equivalence within any studied timeframe.
The honest statement. A created wetland is not a wetland with a delay. For specialist taxa with long establishment times, it is a different system that shares a name. Substitutability is a design choice, not a fact, and where it is chosen it should be priced with a ratio above one.
You already know this because you know a replacement hire is not the person who left, and that the org chart says otherwise.
BRIEF 10 — Performance-Linked Credit Release
The idea. The fix for offset delivery failure is not more monitoring. It is refusing to let a credit exist before the habitat does.
The record it answers. The National Research Council found in 2001 that no-net-loss was not being met, that much required mitigation was never built, and that the agency could not say which. The Government Accountability Office reported in 2005 that the Army Corps had no reliable means of knowing whether required mitigation was occurring at all.
The design change, with the precedent. The United States 2008 Compensatory Mitigation Rule rebuilt the instrument on four mechanisms:
| Mechanism | What it does |
|---|---|
| Performance-linked release | ~15% of credits on approval and site protection; the other 85% against ecological milestones |
| Bonded financial assurance | A bond sized to finish the work if the sponsor fails — converts a promise into a claim |
| Stewardship endowment | Funds maintenance beyond the sponsor's balance sheet |
| Advance mitigation preference | Habitat established before the impact, which removes the time lag |
Where England stands. The biodiversity net gain regime has the thirty-year securement and a public register — real, and more than most jurisdictions. It does not require financial assurance, does not tie unit release to measured performance, and permits units to be transacted before the habitat exists. Those three are statutory instruments, not primary legislation, and each has two decades of operating history behind it.
The number to watch. Not the price of a unit — the proportion of secured sites visited by someone qualified in the last thirty-six months. If it is not published, assume it is low.
You already know this because you have never paid a builder the full amount on the day the scaffolding went up.
All figures in these briefs are computed in lib/verify/VII_03.py, with their inputs and intermediate terms printed, and sourced in the chapter's Works Cited.