Biodiversity as Infrastructure

Volume VII — Planetary and Cosmic

THE PLATE

A watercolour of hills in bands of gold, rust and green, wildflowers in the foreground. — Plate VII.03The Strip Through the Middle.The forest at the edge of the farm was worth two hectares of coffee for every hectare of itself. The same trees, cut into strips and put through the middle, could have been worth eight. Nobody moved them, because nobody had computed the multiplier.

THE LETTER

There is a version of this chapter that asks you to care about biodiversity, and it is not this one. You already care, or you would not have opened a volume called Planetary and Cosmic, and being asked again is a poor use of your morning.

What you probably do not have is a number you could take to a credit committee. That is the gap this chapter closes. Biodiversity is treated here the way a plant engineer treats a compressor: as a productive input with a measurable marginal product, a maintenance cost, a failure curve, a service radius, and a price at which it stops being worth running. Not because the moral case is weak — because the moral case has been made for sixty years, by better writers than us, and it has not moved the capital.

The arithmetic has started to move the capital. In the last two decades the field has produced something the first generation of conservation economics never had: measured yield responses on real farms, at real prices, with control blocks. Not willingness-to-pay surveys. Not global aggregates with error bars the width of the estimate. Field experiments with a treatment and a counterfactual, in coffee, blueberry, oilseed rape, almond, canola and watermelon, on five continents.

That literature lets us do something specific, and it is what this chapter is for. We are going to compute the marginal value of one hectare of pollinator-supporting habitat, on a real farm, at three real commodity prices, and then tell you the price at which the answer flips. Along the way we will find that the single largest lever on that number is not how much habitat there is but where it is, which is free, and that the compliance markets now buying this asset in England are pricing it off a variable that has nothing to do with it.

Then we will do the honest part: the delivery record of biodiversity offset markets, which is poor, with the audit numbers; and the substitutability assumption underneath them, which is false for a large class of systems, with the measurement. And then the design change that fixes it, which already exists, in statute, in another jurisdiction, since 2008.

— The Editors

DISCOVERY

What is already working

Start where the evidence is strongest, which is in the accounts of people who went and counted.

Finca Santa Fe, Valle General, Costa Rica. Taylor Ricketts and colleagues took an eleven-hundred-hectare coffee farm with two remnant forest fragments on it — forty-six hectares and one hundred and eleven hectares — and measured coffee yield as a function of distance from those fragments. Within one kilometre of forest, yields were twenty per cent higher, and the frequency of peaberries, the small malformed single-seed fruit that downgrades a lot, was twenty-seven per cent lower. They valued the service at roughly sixty thousand dollars a year, which across one hundred and fifty-seven hectares of forest is three hundred and eighty-two dollars per hectare per year — more, at the time, than the fragments would have returned as pasture. Two pieces of forest that the farm's accounts treated as nothing were earning a land rent.

Six hundred fields, forty-one crop systems, five continents. Lucas Garibaldi and a very large collaboration asked whether managed honey bees substitute for wild insects. They do not. Across all forty-one crop systems, visitation by wild insects raised fruit set; honey bee visitation raised it in a small minority. At equal visitation rates, the wild-insect effect was about twice the honey-bee effect. The finding matters commercially because a hive is a purchasable input and a landscape is not, and the industry had assumed the purchasable one was the substitute.

Three hundred and forty-four fields, small farms. The same group later found that on fields under two hectares — which is most of the world's farms — raising flower-visitor density closed the yield gap by a median of twenty-four per cent. That is not a marginal improvement. On a smallholding, that is the difference between a year that clears and a year that does not, and it is available without a purchased input.

Twenty-three studies, sixteen crops, on the shape of the service. Ricketts and colleagues then did the thing that turns all of this into an asset class: they measured how the service decays with distance. Visitation rate falls by half at a median of about six hundred metres from natural habitat; pollinator richness at about one and a half kilometres. A service with a measured decay function is a service you can site, size, and price. That single result is what makes the rest of this chapter arithmetic rather than advocacy.

And the enemies. John Losey and Mace Vaughan costed four services provided by native insects in the United States: pest control by native natural enemies at four and a half billion dollars a year, pollination by native insects at three point zero seven billion, dung burial at three hundred and eighty million, and recreation and wildlife nutrition at forty-nine point nine six billion — fifty-eight billion dollars a year in total. We are going to be hard on one of those numbers in a moment. It is still the case that somebody went and costed it, line by line, with a stated method, which is more than the field had in 1990.

And the experiments underneath all of it. Five hundred and seventy-four controlled biodiversity experiments, synthesised by Bradley Cardinale and colleagues: on average a polyculture produces one point seven times the biomass of the average monoculture of its constituent species. Across the broader synthesis, species loss in the range of twenty-one to forty per cent reduces primary production by five to ten per cent. These are not correlations from landscapes. They are randomised plantings with replication, run for decades at Cedar Creek and a hundred other sites.

Six bodies of evidence, one pattern: the service is real, it is measurable, it has a shape in space, and it is already being delivered free to people who have not noticed. That is the positive core. Now we cost it.

THE ARITHMETIC

What works, what does not, and where the line sits

First, the dependency, stated exactly.

Of the one hundred and fifteen leading global food crops, eighty-seven — seventy-five point seven per cent — depend on animal pollination to some degree; twenty-eight do not. Those eighty-seven represent about thirty-five per cent of global production by volume. Nicolas Gallai and colleagues valued the insect-pollination contribution to world agriculture at one hundred and fifty-three billion euros in 2005, or nine point five per cent of the value of world crop production for human food — which implies a food-crop denominator of about one thousand six hundred and eleven billion euros. IPBES later put the annual market value of crops directly affected by pollinators at two hundred and thirty-five to five hundred and seventy-seven billion dollars.

Hold two of those numbers against each other, because the tension is the whole point. Thirty-five per cent of the tonnage. Nine point five per cent of the value. The difference is price: pollinator-dependent crops sell for roughly five times as much per tonne as pollinator-independent ones. The staples — wheat, rice, maize — are wind- or self-pollinated and are most of the calories. The dependency sits in the expensive half of the diet: the fruit, the nuts, the coffee, the cocoa, the oilseeds. This is not a food-security argument. It is a cash-crop argument, and it should be made as one, because that is where it is true and where the money is.

Second, the instrument. The marginal value of a pollinator-supporting hectare.

        V   =   L  ·  Y₀  ·  Δ  ·  P  ·  φ

  V    value per hectare of habitat, per year
  L    leverage — hectares of crop within the service radius,
       per hectare of habitat
  Y₀   yield without the service, tonnes per hectare
  Δ    proportional yield uplift from the service
  P    farmgate price per tonne
  φ    attribution — the share of Δ owed to this habitat

Take Finca Santa Fe and rebuild it term by term. Δ = 0.20, measured. Y₀ = 1.0 tonnes per hectare of green coffee, unserved. φ = 0.70, assumed — some of the uplift belongs to managed hives and other remnant vegetation, and an instrument that claims all of it will not survive its first audit.

That leaves L, and L is where the chapter turns.

Treat each fragment as a disc. Forty-six hectares is a radius of 0.383 kilometres; one hundred and eleven hectares is 0.594 kilometres. The ring of coffee within one kilometre of the first is 554.6 hectares; of the second, 687.6 hectares. Total reachable crop: 1,242 hectares, against 157 hectares of habitat.

  L, geometric ceiling   =   1,242 / 157   =   7.9

Now run it backwards from the published value. Three hundred and eighty-two dollars per hectare per year, at the coffee price of the 2001–2004 trough — about one thousand three hundred and twenty dollars a tonne, sixty cents a pound — implies:

  L, as measured   =   382 / (1.0 × 0.20 × 1,320 × 0.70)   =   2.07

The geometry says one hectare of habitat can serve nearly eight. The measurement says it served two. The gap is a factor of 3.8, and it is not biology. It is placement: two blocks at the edge of the farm, their service areas overlapping each other and spilling off the property, instead of strips running through the middle. L enters the formula linearly. Configuration is the multiplier, it costs nothing, and nobody had computed it.

Take L = 4.0 — a dispersed layout, between the measured two and the geometric eight — and run the instrument at three prices arabica has actually traded at in the last twenty-five years.

  P = $1,320/t   (trough,  ~$0.60/lb)   V = $  739 /ha/yr   capitalised $10,560
  P = $3,000/t   (mid,     ~$1.36/lb)   V = $1,680 /ha/yr   capitalised $24,000
  P = $6,000/t   (peak,    ~$2.72/lb)   V = $3,360 /ha/yr   capitalised $48,000

Capitalised at seven per cent. Against a net return to pasture of about one hundred and fifty dollars per hectare per year and a net margin on coffee of about eight hundred, the reading is immediate: the habitat hectare beats pasture by five to twenty-two times at every price in the range, and beats coffee above

  P*   =   800 / (4.0 × 1.0 × 0.20 × 0.70)   =   $1,429 /t   =   $0.65 /lb

Sixty-five cents a pound. Arabica has been below that in exactly one window since 1975: the coffee crisis of 2001 to 2004. Which means that for every year since, a dispersed hectare of pollinator habitat has out-earned the coffee that would replace it — and that the clearing decisions taken during the crisis were economically correct on the day they were made, and have been economically wrong every year since, and are irreversible on a thirty-year horizon while the price that justified them mean-reverted in three.

That is the Balenciaga cut of this chapter, and it is worth saying plainly. The marginal value of a hectare of habitat is not an ecological quantity. It is a derivative on a commodity price. A land-use decision taken on a spot price is a permanent bet against mean reversion, and it is a bet the farmer did not know they were placing. Everything in the Design movement follows from neutralising it.

Third, what does not work. Pest regulation does not survive division.

Four and a half billion dollars a year sounds like an argument. Divide it by one hundred and sixty million hectares of United States cropland:

  $4.5bn / 160,000,000 ha   =   $28.12 per hectare per year

Twenty-eight dollars a hectare — below the cost of a single insecticide pass. The national aggregate does not underwrite a field-scale decision, and anybody presenting it as though it does will be caught.

It gets harder. Reviewing the landscape studies, Bianchi and colleagues found natural enemy populations higher in complex landscapes in seventy-four per cent of cases — and pest pressure actually reduced in only forty-five per cent. A twenty-nine point gap between having the predators and having less damage. Then Daniel Karp and a very large team assembled one hundred and thirty-two studies across six thousand seven hundred and fifty-nine sites and found that landscape composition explained little of the variation in pest or enemy abundance, and that the direction of the effect was close to a coin flip. Matteo Dainese and colleagues, across eighty-nine studies and one thousand four hundred and seventy-five fields, did recover a consistent signal — but it runs through richness of the service-providing organisms, not through the amount of non-crop habitat.

So: pollination generalises and can be underwritten. Pest regulation does not and cannot, not yet, not at field scale. An honest biodiversity balance sheet carries the first as revenue and the second as an option with no strike price.

Fourth, the saturation curve, and the thing it hides.

Fit the standard form to the experimental record, Y = a + b·ln(S). The marginal value of the next species is b/S, so it halves 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 b

The sixteenth species is worth one eighth of the second. And only twelve per cent of polycultures in Cardinale's five hundred and seventy-four experiments beat their own single most productive species. Read alone, that is devastating to the case for high richness: plant three species, bank most of the gain, and the rest is sentiment.

Except that every one of those curves was fitted to one function, in one year, in one place. Forest Isbell and colleagues widened the denominator — many functions, many years, many sites — and found that eighty-four per cent of the one hundred and forty-seven plant species in the pool promoted ecosystem functioning at least once. Different species, in different years, for different functions.

Those two results are usually presented as a controversy. They are not. Model it: let each function-year-place draw about twelve species from a pool of one hundred and forty-seven, and count the union.

  species required  =  147 · (1 − (1 − 12/147)^c)

  c =  1 context     12.0 species    8.2 % of the pool
  c =  5             51.0           34.7 %
  c = 10             84.3           57.3 %
  c = 22            124.4           84.6 %
  c = 30            135.6           92.2 %

Twenty-two contexts reproduce the observed eighty-four per cent from a curve that looked flat after three species. The saturation is in the denominator, not in the field. A three-species mix is sufficient for one function in one year, and that is exactly what the experiments measured and exactly what a farm does not experience.

Fifth, the Dasgupta Review, differentiated.

Partha Dasgupta's 2021 Review for HM Treasury is the most serious attempt to put the biosphere on a national balance sheet. Its central measurement: between 1992 and 2014 — twenty-two years — produced capital per head doubled, human capital per head rose thirteen per cent, and natural capital per head fell forty per cent. Annualised:

  produced capital   +3.20 % / yr
  human capital      +0.557 % / yr
  natural capital    −2.30 % / yr

  the spread, produced less natural      5.50 points per year
  the ratio shift over the span          3.33 ×

And the Impact Inequality: current living standards require about 1.6 Earths. Written in the ratio from Chapter I.01, S = D / (R · r), an overshoot of 1.6 closes — at flat demand — in 23.7 years if the biosphere regenerates at two per cent and 47.2 years at one per cent. r is the only term any of this can move, and it is the term the accounts do not contain.

Where its own authors flag weakness, and they do flag it. The Review states plainly that accounting prices for most natural capital are not known and cannot presently be estimated; that its biosphere model treats the whole living world as a single aggregated regenerating asset; and that the services easiest to price are provisioning services, which biases any account built from market data toward exactly the things that were never in danger. Clive Spash and Frédéric Hache have pressed the deeper objection: inclusive wealth accounting requires substitution between capital forms at the margin, and for several ecosystem types that substitution is not merely difficult to price but physically unavailable. The Review's framework is right about the direction and cannot yet supply the prices. Say both.

DREAM

What becomes ordinary

In the version of this that has already happened, a farm plan has two maps.

The first is the cropping map, which every farm has had for a century. The second is the service map: habitat in yellow, the one-kilometre service radius drawn around each parcel, and the hectares of crop inside each ring counted. The leverage ratio L sits at the bottom of the sheet like a soil pH, and it is managed. When a farm buys land or takes ground out, the first question is what it does to L, and that question takes four minutes because the map exists.

Habitat is planted in strips, because the arithmetic says so. The four-metre margin running through the middle of a block is worth four times the same area banked at the boundary, and that is now common knowledge on the same level as knowing which way to plough a slope.

The farm's lender has a line for it. Not a green loan with a rate concession and a reporting burden — an ordinary secured facility, advanced against a contracted service royalty, underwritten the way a lender underwrites a milk contract. The credit paper states the leverage ratio, the measured differential, the price floor, and the term. Nobody in that transaction uses the word biodiversity and the transaction is entirely about biodiversity.

The grower pays a share of the measured yield differential rather than a rent, so when the commodity falls the payment falls and the habitat is not the first thing cut. When the commodity rises the habitat owner participates. The instrument floats with the thing it is a derivative of, which is the only structure that survives a full price cycle, and it means that no window of low prices is ever again long enough to make the clearing look right.

Compliance markets exist alongside this and are understood for what they are. When a developer pays twenty-five thousand pounds for a biodiversity unit, nobody mistakes that price for the ecological value of the habitat; it is the price of the developer's alternative, which is planning refusal. Two markets in the same asset, priced off unrelated variables, and everyone in both of them knows it.

And the accounts carry r. Not as a disclosure, not as a narrative section — as a rate, beside the depreciation schedule, showing which stocks are regenerating faster than they are drawn and which are not. The auditors have a standard for it. It took about four years and it was not controversial in the end.

DESIGN

The structure that gets there

Layer one: measure the differential on your own ground.

Paired blocks. Same cultivar, same age, same management, one inside the service radius and one outside, minimum three seasons. Yield and quality both — at Santa Fe the peaberry reduction was worth attending to separately from the tonnage. This is not a research programme; it is four afternoons a year and a spreadsheet, and without it every number downstream is somebody else's.

Layer two: compute L before you plant anything.

Draw the service radius. Count the crop hectares inside it. Divide by the habitat hectares. If L is under two, the habitat is in the wrong place and moving it is the cheapest intervention available. Do this before arguing about how much habitat to have, because the placement decision is worth a factor of four and the quantity decision is worth a factor of one.

Layer three: contract the service as a royalty, not a rent.

A share of the measured differential, not a fixed payment per hectare. Sized in the Operationalize movement below. The royalty structure does three things a rent cannot: it is self-indexing against the commodity cycle, it is payable only on a measured result so it survives audit, and it makes the habitat owner and the grower parties to the same number rather than opposite sides of a price.

Layer four: if you enter a compliance market, enter it knowing what sets the price.

England's biodiversity net gain regime is the largest mandatory market of its kind. A hectare of medium-distinctiveness habitat in good condition scores 4 × 3 × 1.0 = 12 biodiversity units at baseline. Create that hectare and the metric applies a difficulty multiplier of 0.67 and a temporal multiplier around 0.50 for a twenty-year target, leaving 4.02 units per hectare. At a reported market price of twenty-five thousand pounds a unit, that is one hundred thousand five hundred pounds a hectare, one-off, for a thirty-year obligation — six thousand five hundred and thirty-eight pounds per hectare per year at a five per cent annuity. Against an English arable net margin near five hundred pounds, that is thirteen point one times the farming alternative. Priced at the lowest statutory credit tier of forty-two thousand pounds a unit — the government's deliberately unattractive last resort, in a schedule that runs to six hundred and fifty thousand pounds — the same hectare is ten thousand nine hundred and eighty-three pounds a year.

Now hold that against the measured agronomic service, which came out between seven hundred and thirty-nine and three thousand three hundred and sixty dollars a hectare. At a stated 1.27 dollars to the pound, the English compliance price is eight thousand three hundred and three dollars a hectare a year — eleven point two times the measured service at trough prices, four point nine times at mid-cycle, two point five times at peak.

That is not a contradiction and it is not a scandal. It is two different goods. One price is the marginal product of pollination in a crop. The other is the cost of the cheapest way a developer can obtain planning consent. They have no reason to converge and they do not. A landowner who sells into the compliance market is selling a permit, not a service, and should price it as one.

Layer five: govern for delivery, because the record says delivery is where this fails.

Which is the next movement, and it is the honest one.

DESTINY

How it holds, and how it has failed

The delivery record of biodiversity offset markets is poor and it is documented.

The United States wetland record. The National Research Council's 2001 review of compensatory mitigation under the Clean Water Act found that the national goal of no net loss was not being met, that a substantial share of required mitigation was never constructed or failed its permit conditions, and — the finding that mattered most — that the permitting agency generally could not say which. Four years later the Government Accountability Office reported that the Army Corps of Engineers had no reliable means of knowing whether required compensatory mitigation was occurring at all; district files frequently lacked the documentation to tell.

And where delivery happened, equivalence did not. David Moreno-Mateos and colleagues synthesised six hundred and twenty-one restored and created wetland sites against reference wetlands. Biological structure came in twenty- six per cent below reference; biogeochemical function twenty-three per cent below — and the deficit was still measurable in sites restored a century earlier.

Run that through an offset ratio and the arithmetic is unforgiving. A one-to-one offset against a seventy-seven per cent function delivery rate is a twenty-three per cent loss written into the instrument. To hold function constant you need

  1 / 0.77   =   1.30 hectares created per hectare lost

and almost no scheme requires it. Thomas Curran and colleagues went further and showed that for slow-forming systems — old-growth forest, peatland, mature sclerophyll — measured recovery ratios do not reach equivalence within any studied timeframe. Sophus zu Ermgassen and colleagues, reviewing the global no-net-loss record, found that ecological outcome data exist for only a small fraction of offsets worldwide, and that among those with documented outcomes only a minority demonstrate no net loss.

The substitutability assumption is the load-bearing member and for many taxa it is false. 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. England's early-adopter studies point at the same class of problem from the governance side: much of the reported net gain has come from enhancement of existing habitat rather than creation, and monitoring and enforcement sit with planning authorities whose ecological capacity is thin.

Now the design change that fixes it, and the precedent is twenty years old.

The United States 2008 Compensatory Mitigation Rule rebuilt the instrument around four mechanisms, and England's regime currently has one of them.

Performance-linked credit release. Only about fifteen per cent of credits are released on instrument approval and site protection. The remaining eighty-five per cent release against ecological performance standards written into the instrument, milestone by milestone. A credit that does not exist cannot be sold.
Bonded financial assurance. A bond or letter of credit sized to complete the work if the sponsor fails. This converts a promise into a claim.
A long-term stewardship endowment plus permanent site protection, so the obligation outlives the sponsor's balance sheet.
A preference hierarchy favouring advance mitigation — habitat established before the impact — which eliminates the time lag that caused most of the documented failures in the first place.

England has the thirty-year securement and a public register, which is real and is more than most jurisdictions have. It does not require financial assurance, it does not tie unit release to measured performance, and it permits units to be transacted before the habitat exists. Those three are statutory instruments, not primary legislation, and each has a working precedent with two decades of operating history. That is the whole of the fix, and it is unusually cheap.

The failure mode after that is the ordinary one: an obligation that nobody is resourced to check. A register without an inspection budget is a list of intentions. The number to watch is not the price of a unit. It is the proportion of secured sites visited by someone qualified in the last thirty-six months, and if that number is not published, assume it is low.

DELIGHT

What it feels like

The pleasure in this work is the pleasure of finding that the thing you were prepared to argue for on principle turns out to pay, and pays more than you thought, and pays for a reason nobody had bothered to write down.

There is a specific version of it, and it happens at the map. You draw the service radius around the habitat you already have, count the crop hectares inside it, and find that the number is two when it could be eight — and that the fix is not money, or permission, or a policy, but where the next planting goes. That is a good afternoon. It has the quality of finding a room in a house you have lived in for years.

And then the field itself, which is the part no spreadsheet carries. A flowering strip through the middle of a block in the second week is loud. Not metaphorically — audibly, at eight in the morning, from ten metres away. People who have put one in describe going to stand near it for no reason, and then having to account to themselves for why they were standing there.

The arithmetic is how it gets approved. The noise is why it gets kept.

OPERATIONALIZE THIS

At the level of finance

The instrument: a habitat service royalty, with a lender advance against it.

Not a conservation easement, which pays for absence. Not a rental agreement, which breaks in a price trough. A royalty on a measured differential, which is the structure of a production payment and is recognisable to any agricultural lender.

The parties. The habitat owner grants a covenant maintaining the parcel as pollinator habitat for a stated term. The grower — or, better, the mill, co-operative or buyer, who has the harvest records — pays a percentage of the measured yield differential attributable to the service. Where habitat owner and grower are the same person, the counterparty is the lender, and the royalty is the cash flow the facility is advanced against.

The mechanics.

Term. Fifteen years minimum. Below that the establishment period consumes too much of it, and habitat delivers rising service for roughly the first decade.
The measure. Paired blocks inside and outside the service radius, matched on cultivar, age and management, with three seasons of baseline before the first payment. Yield and quality graded separately.
The royalty. Twenty-five per cent of the measured differential, gross of the grower's harvest cost. Twenty-five is a term, not an optimum — it is set where it is because it clears the habitat owner's opportunity cost across the realistic price range without ever exceeding the grower's realised gain.
The floor and the cap. A floor at the land's documented alternative use protects the habitat owner in a trough. A cap at forty per cent of the differential protects the grower in a spike and is what makes the instrument signable.
Attribution. φ stated in the contract, not assumed. Seventy per cent is defensible where managed hives are also present; one hundred per cent is not, and claiming it is how the first dispute is lost.
Verification. The buyer's own intake records are usually sufficient and are already audited. Do not commission ecological monitoring you can avoid; commission the yield measurement you cannot.

The balance-sheet treatment. For the habitat owner, a contracted revenue stream revalues the parcel on an income basis rather than an alternative-use basis — which is frequently the entire argument, because the land was carried at pasture value. For the grower, the royalty is an operating cost matched against a measured yield gain in the same period, which is clean, and it is never a payment for an unmeasured benefit, which is what makes it survivable in an audit. Where establishment is capitalised, depreciate over the service life, not the planting cost's conventional cycle.

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

      royalty per hectare of habitat
   ------------------------------------------   >   1
    opportunity cost  +  monitoring cost

Worked, at L = 4.0, against a pasture opportunity cost of one hundred and fifty dollars and sixty dollars of paired-block monitoring:

  P = $1,320/t    royalty $185 /ha/yr    covers 0.88 ×    does not clear
  P = $3,000/t    royalty $420 /ha/yr    covers 2.00 ×    clears
  P = $6,000/t    royalty $840 /ha/yr    covers 4.00 ×    clears

The royalty clears above one thousand five hundred dollars a tonne — sixty-eight cents a pound. Below that it does not, and the floor is what carries it. State that in the credit paper rather than letting the committee find it: an instrument whose weakness is disclosed and priced is stronger than one whose weakness is discovered.

The first ninety days.

Day	Action	Artifact
1–15	Map habitat, draw the service radius, count crop hectares inside it	The `L` calculation
16–30	Identify paired blocks; pull three seasons of block-level yield and grade	The baseline data pull
31–45	Compute the differential and `φ`; price the royalty at three commodity prices	The sizing memo
46–60	Agree the measure and the attribution with the buyer, in writing	The signed measurement basis
61–75	Draft the covenant, the royalty, the floor and the cap	Draft instrument
76–90	Take it to the lender as a production payment, not as a green facility	The credit paper

The counterparty of last resort. If no buyer will sign a royalty, the compliance market will buy the same hectare as a permit, at a price set by somebody else's planning problem. Take it if you need to — and price it as the permit it is, knowing that the agronomic value is a separate asset you have not sold.

APPRECIATIVE QUESTIONS

Twelve, for a room

Discovery — what is already working

Where on this ground is something flowering that nobody planted, and what has it been doing for us that we have never counted?
Think of a block that has quietly out-yielded its neighbours for years. What is within one kilometre of it that is not within one kilometre of the others?
Who here already knows where the bees are at seven in the morning, and when did anyone last ask them?

Dream — what becomes possible

If our lender underwrote habitat the way it underwrites a supply contract, what would we plant this winter that we are not planting?
Imagine our farm plan carries a leverage ratio on the front sheet, next to soil pH. What decision would it change first?
If the people who maintain the margins were paid a share of the measured uplift, what would they start noticing that nobody notices now?

Design — what we build

What is the smallest strip we could put through the middle of a block this season, and what would we need to measure to know whether it worked?
Which pair of blocks could serve as our control and our treatment, starting this harvest, with no change to anything else?
What would we want written into the contract so that a bad price year never makes the habitat the first thing cut?

Destiny — how it holds

What would have to be true for these strips to still be here in twenty years when everyone in this room has moved on?
Who visits a secured site, how often, and what do they actually check — and if we cannot answer that, what is the smallest version of an answer?
When this works, who else nearby will see it, and what would make it easy for them to copy rather than admire?

WORKS CITED

Bianchi, F. J. J. A., Booij, C. J. H. and Tscharntke, T. (2006). "Sustainable pest regulation in agricultural landscapes: a review on landscape composition, biodiversity and natural pest control." Proceedings of the Royal Society B, 273(1595), 1715–1727.

Bull, J. W., Suttle, K. B., Gordon, A., Singh, N. J. and Milner-Gulland, E. J. (2013). "Biodiversity offsets in theory and practice." Oryx, 47(3), 369–380.

Cardinale, B. J., Matulich, K. L., Hooper, D. U., Byrnes, J. E., Duffy, E., Gamfeldt, L., Balvanera, P., O'Connor, M. I. and Gonzalez, A. (2011). "The functional role of producer diversity in ecosystems." American Journal of Botany, 98(3), 572–592.

Cardinale, B. J., Duffy, J. E., Gonzalez, A., Hooper, D. U., Perrings, C., Venail, P., Narwani, A., Mace, G. M., Tilman, D., Wardle, D. A., Kinzig, A. P., Daily, G. C., Loreau, M., Grace, J. B., Larigauderie, A., Srivastava, D. S. and Naeem, S. (2012). "Biodiversity loss and its impact on humanity." Nature, 486, 59–67.

Curran, M., Hellweg, S. and Beck, J. (2014). "Is there any empirical support for biodiversity offset policy?" Ecological Applications, 24(4), 617–632.

Dainese, M. et al. (2019). "A global synthesis reveals biodiversity-mediated benefits for crop production." Science Advances, 5(10), eaax0121.

Dasgupta, P. (2021). The Economics of Biodiversity: The Dasgupta Review. HM Treasury, London.

Gallai, N., Salles, J.-M., Settele, J. and Vaissière, B. E. (2009). "Economic valuation of the vulnerability of world agriculture confronted with pollinator decline." Ecological Economics, 68(3), 810–821.

Garibaldi, L. A. et al. (2013). "Wild pollinators enhance fruit set of crops regardless of honey bee abundance." Science, 339(6127), 1608–1611.

Garibaldi, L. A. et al. (2016). "Mutually beneficial pollinator diversity and crop yield outcomes in small and large farms." Science, 351(6271), 388–391.

Hooper, D. U. et al. (2012). "A global synthesis reveals biodiversity loss as a major driver of ecosystem change." Nature, 486, 105–108.

Hough, P. and Robertson, M. (2009). "Mitigation under Section 404 of the Clean Water Act: where it comes from, what it means." Wetlands Ecology and Management, 17(1), 15–33.

IPBES (2016). The Assessment Report of the Intergovernmental Science-Policy Platform on Biodiversity and Ecosystem Services on Pollinators, Pollination and Food Production. Potts, S. G., Imperatriz-Fonseca, V. L. and Ngo, H. T. (eds). IPBES Secretariat, Bonn.

Isbell, F. et al. (2011). "High plant diversity is needed to maintain ecosystem services." Nature, 477, 199–202.

Karp, D. S. et al. (2018). "Crop pests and predators exhibit inconsistent responses to surrounding landscape composition." Proceedings of the National Academy of Sciences, 115(33), E7863–E7870.

Klein, A.-M., Vaissière, B. E., Cane, J. H., Steffan-Dewenter, I., Cunningham, S. A., Kremen, C. and Tscharntke, T. (2007). "Importance of pollinators in changing landscapes for world crops." Proceedings of the Royal Society B, 274(1608), 303–313.

Losey, J. E. and Vaughan, M. (2006). "The economic value of ecological services provided by insects." BioScience, 56(4), 311–323.

Maron, M., Hobbs, R. J., Moilanen, A., Matthews, J. W., Christie, K., Gardner, T. A., Keith, D. A., Lindenmayer, D. B. and McAlpine, C. A. (2012). "Faustian bargains? Restoration realities in the context of biodiversity offset policies." Biological Conservation, 155, 141–148.

Moreno-Mateos, D., Power, M. E., Comín, F. A. and Yockteng, R. (2012). "Structural and functional loss in restored wetland ecosystems." PLoS Biology, 10(1), e1001247.

National Research Council (2001). Compensating for Wetland Losses Under the Clean Water Act. National Academy Press, Washington DC.

Ricketts, T. H., Daily, G. C., Ehrlich, P. R. and Michener, C. D. (2004). "Economic value of tropical forest to coffee production." Proceedings of the National Academy of Sciences, 101(34), 12579–12582.

Ricketts, T. H. et al. (2008). "Landscape effects on crop pollination services: are there general patterns?" Ecology Letters, 11(5), 499–515.

Spash, C. L. and Hache, F. (2021). "The Dasgupta Review deconstructed: an exposé of biodiversity economics." Environmental Values, 30.

United States Army Corps of Engineers and Environmental Protection Agency (2008). Compensatory Mitigation for Losses of Aquatic Resources. Final Rule, 33 CFR Parts 325 and 332; 40 CFR Part 230.

United States Government Accountability Office (2005). Wetlands Protection: Corps of Engineers Does Not Have an Effective Oversight Approach to Ensure That Compensatory Mitigation Is Occurring. GAO-05-898.

zu Ermgassen, S. O. S. E., Baker, J., Griffiths, R. A., Strange, N., Struebig, M. J. and Bull, J. W. (2019). "The ecological outcomes of biodiversity offsets under 'no net loss' policies: a global review." Conservation Letters, 12(6), e12664.

zu Ermgassen, S. O. S. E., Marsh, S., Ryland, K., Church, E., Marsh, R. and Bull, J. W. (2021). "Exploring the ecological outcomes of mandatory biodiversity net gain using evidence from early-adopter jurisdictions in England." Conservation Letters, 14(6), e12820.

Note on figures. Every figure in this chapter is computed in lib/verify/VII_03.py and printed there with its inputs, its intermediate terms and its source. Published effect sizes are marked INPUT; figures with no computation behind them are marked ILLUSTRATIVE or ASSUMED and are not counted as computed. The English biodiversity metric multipliers are those of the statutory metric; the temporal multiplier used here is stated as an assumption.