Haute Lumière · The reading

Everything the economy runs on is missing from its books.

The pollination, the floodwater, the living soil — and the arithmetic that puts them back into the number.

The cost of the morning was met by something that never sent an invoice.

THE MISSING COLUMN

Somewhere this morning a person paid for a punnet of strawberries, and the transaction was recorded, correctly, as economic activity. The number is not wrong. It covers the picking, the crate, the fuel to the market, the fee for the stall, and a season of the farmer's hours. What it does not cover is the part of the strawberry nobody billed for — the bees that worked the blossom, the soil alive enough to hold water through a dry fortnight, the rain that arrived at roughly the right time. Those were not free. They were unbilled, which is a different thing, and the accounts have no way to tell the difference.

That is the quarrel, and it fits inside one punnet. Gross domestic product is a sum of transactions. A thing nobody transacts over does not appear in it, and in an accounting system invisibility is indistinguishable from worthlessness. So a country can log the timber and not the forest, the catch and not the fishery, the breakfast and not the pollination that made it possible. It is a little like judging a symphony by the cost of the sheet music and the hire of the hall.

The numbers are not small, which is the part people tend to find surprising. Studies of coffee plantations find that healthy pollinator populations raise yields by fifteen to twenty per cent. Take a smallholding producing a thousand kilograms a year at five dollars a kilogram. A twenty per cent lift from pollination is two hundred extra kilograms, or a thousand dollars. A ten per cent shortfall when pollinators decline is a hundred kilograms gone, five hundred dollars off the top. That is a fifteen-hundred-dollar swing on a five-thousand-dollar farm gate, decided entirely by an insect population that appears nowhere in the books.

The usual objection to counting any of this is that nature is priceless, and putting a figure on it cheapens it. The objection has the direction of travel exactly backwards. Nature does not currently lack a price in these models. It has one, and the price is zero, entered by default the moment a column was left out. Zero is the most aggressive valuation available. It says the forest can go and nothing measurable is lost.

A blank in the ledger is not neutrality. It is a price of zero, entered by default and defended by nobody.

What follows from that is practical rather than philosophical. If the number is going to be wrong either way, it is better to be wrong by an argued amount than wrong by an unexamined blank, because an argued amount can be challenged, revised, and improved. A disputed figure is a live thing in a spreadsheet. An omission is not disputed at all; it simply sits there, doing its silent work, and every decision downstream inherits it.

MACHINE OR BODY

The reason the column went missing is older than any spreadsheet. It is a picture we carry without noticing: the economy as a machine. Inputs enter, outputs leave, and the environment is the room the machine happens to sit in — occasionally noisy, occasionally in the way, but structurally elsewhere. The picture is tidy and it is the source of nearly every error in this field. It is not that the machine model is imprecise. It is that it has the containment the wrong way round.

The alternative picture is a forest. Sunlight arrives and is caught by leaves; the sugars made there become wood, and the wood becomes a stock that stands for decades. Leaves fall, fungi take them apart, and what was a leaf becomes soil that feeds the next tree. Nothing in that description is a supply chain with a beginning and an end. It is a loop that has been running long enough to look stable, and its stability is the outcome of the loop, not a property of any tree in it.

Living systems have a small vocabulary worth learning, because each word names a thing the machine model has no slot for. Stocks are the accumulations — the timber, the fish biomass, the topsoil, the water in the aquifer. Flows are the rates at which those stocks fill and drain. Feedback is what happens when the state of a stock changes the rate of a flow: too many deer, then not enough browse, then fewer deer. Coupling is the fact that draining one stock quietly changes a flow somewhere you were not looking.

Two more are worth having. Emergence is the behaviour a system shows that none of its parts contains — a flock turns as one without any bird holding the plan, and a market panics without any trader having decided to. Antifragility is the property of systems that do not merely survive a shock but come out of it better arranged: the fire that clears deadwood, the drought that selects for deeper roots. Fragile things break under stress. Robust things endure it. Antifragile things use it.

Set the two pictures side by side and the consequence for finance is immediate. A machine is optimised by removing slack, so a machine model rewards concentration, specialisation and the elimination of redundancy. A living system is kept alive by slack, so a living-systems model rewards diversity, spare capacity and the ability to lose a part without losing the whole. These are not two moods about the same portfolio. They are two different portfolios, and they behave differently in the year everything goes wrong.

Ecological economics is simply the discipline that takes the second picture seriously enough to do arithmetic with it. It does not begin by asserting that nature is precious — plenty of fields do that and produce nothing usable. It begins by observing that the economy is a subsystem of a larger system it cannot exit, and that a subsystem which ignores the state of its container will eventually be corrected by the container, on the container's schedule rather than its own.

STOCKS AND FLOWS

Here is the single most useful distinction in the whole book, and it costs nothing to learn. A stock is an amount. A flow is a rate. Fish in the sea is a stock; fish caught this year is a flow. Trees standing is a stock; timber cut is a flow. Carbon in the atmosphere is a stock; emissions are a flow. Almost every serious argument about sustainability is an argument about a stock being read as though it were a flow.

Picture a bath. The tap is what the economy takes and makes; the drain is what nature regenerates and absorbs. Open the tap wider than the drain and the level rises, which is fine right up until it is not. Nothing about the moment of overflow is announced by the taps, the plumbing, or the sound of the water. The level simply crosses a line it had been approaching all along, and every reading before that moment looked like a bath filling normally.

This is why income and capital get confused so easily. An orchard that yields a crop each year and stands afterwards is producing income. An orchard cut down and sold is producing a very good year followed by nothing, and the accounts record only the very good year. A fishery landing more tonnes each season may be improving or it may be emptying, and the revenue line reads identically in both cases. The difference lives entirely in a stock nobody put on the balance sheet.

Harvesting a stock and earning a return look the same in the revenue line. They stop looking the same later.

The fix is not complicated, which is the encouraging part. It requires only that the stock be written down beside the flow, and that the flow be judged against the rate at which the stock replenishes. Take less than the regeneration rate and the system holds; take more and you are spending principal while reporting interest. Every sustainable-yield rule in forestry, fisheries and groundwater is a version of that one sentence, and all of them predate the modern language by a century or more.

What living systems add to the arithmetic is that regeneration rates are not constants. They depend on the state of the stock itself — a thinned forest regrows differently from a whole one, a depleted fishery spawns differently from a healthy one. The rate you measured last year was a property of last year's stock, not a law of nature. This is precisely where linear extrapolation, the most natural thing in the world to do with a column of numbers, quietly becomes the most dangerous.

Everything in the frame is a stock. What holds it there is a flow, and the flow is not in the picture.

THE ARITHMETIC

All of this becomes tractable with one equation, and it is friendlier than it looks. Write it as dX/dt = rX(1 − X/K). Read from left to right: the rate of change of the stock X over time equals the intrinsic growth rate r, times how much stock there currently is, times a brake that tightens as the stock approaches K, the carrying capacity. That is the whole thing. Three quantities and one brake.

Work it with rabbits, because rabbits are honest about this. Start with a hundred in a meadow that can support five hundred, with an intrinsic growth rate of two-tenths a year. The brake term is one minus a hundred over five hundred, which is eight-tenths. So the change is two-tenths times a hundred times eight-tenths, which is sixteen. A hundred rabbits become a hundred and sixteen, and the meadow has barely noticed.

Now run the same equation on something with money in it. A forest holds a hundred thousand cubic metres of timber, regenerates at five per cent, and could hold two hundred thousand at most. The brake is one minus a half, which is a half. Five per cent of a hundred thousand is five thousand; halved by the brake, that is two thousand five hundred cubic metres of growth this year. Cut two thousand five hundred and the forest is exactly as you found it. Cut four thousand and you have taken fifteen hundred cubic metres of principal while reporting a good year.

Two features of that brake are worth dwelling on. When the stock is small relative to capacity, the brake is close to one and growth is nearly exponential — which is why a recovering fishery can look miraculous and a new market can look unstoppable, for a while. When the stock nears capacity, the brake closes toward zero and growth flattens regardless of how favourable everything else is. The slowdown is not a failure of management. It is the shape of the equation.

The same curve underwrites the finance. Compound growth is the exponential case with no brake at all: a million dollars of revenue at five per cent for five years comes to roughly one million two hundred and eighty-four thousand. Nothing in that expression knows about a meadow, a forest or an aquifer, which is exactly the problem. Exponential growth is not a villain. It is a model with the brake term deleted, applied to a world that has one.

None of these models is the world, and the book is careful never to pretend otherwise. Real ecosystems carry disease, weather, migration, interacting species and shifting capacities, and no three-term equation captures that. What the equation gives you is not prediction but posture: a habit of asking, of any growing quantity, what its K is and how close the thing already is to it. Most projections in most spreadsheets have never been asked that question once.

THE CEILING

K stops being an abstraction the moment someone tries to measure it for the planet. The tool for that is the ecological footprint, which converts a way of living into the area of biologically productive land and water required to supply it and absorb its waste. Set that against biocapacity — the productive area actually available — and you get a ratio rather than an opinion. The Global Footprint Network's estimate is that humanity uses the equivalent of one point seven Earths a year.

It is worth being exact about what that sentence means, because it is easy to hear it as apocalyptic when it is something stranger and more ordinary. There is only one Earth, so the ratio does not describe a second planet being consumed. It describes drawdown: stocks falling while flows continue as though they were not. Aquifers lower, soils thinner, forests smaller, carbon accumulating where it is not absorbed. The overshoot is not a prophecy about the future. It is a description of a balance sheet that is already running down, quietly, on schedule.

Overshoot is not a forecast. It is a stock being spent at a rate the income statement has no field for.

Carrying capacity, properly understood, is not a fixed number waiting to be discovered. It moves with technology, diet, settlement pattern, energy source and institutional arrangement — a hectare of land supports a different number of people depending on what is grown, how it is grown and what is wasted. This makes the figure hard to pin down and easy to dismiss. It also means the ceiling is partly a design variable, which is a far more interesting fact than the dismissal allows.

The practical use of a footprint is not planetary at all; it is local and it fits inside an investment decision. Suppose a farm produces a kilogram of salmon at an ecological footprint of ten square metres of productive land for a year, and it sits in a coastal region with fifty square metres of biocapacity per person and a population of a thousand. The region's total is fifty thousand square metres. A farm producing ten thousand kilograms a year needs a hundred thousand. It is asking for twice what the place has.

That calculation does not end an investment conversation. It reframes it, and the reframing is where the value sits. Three responses become visible that were not visible before: decline the position and keep the capital dry; stay in and negotiate the footprint down through feed efficiency, energy source and waste handling; or spread the same capital across operations in regions with headroom. All three are ordinary portfolio decisions. What is new is that a real constraint is now a term in them, rather than a risk that arrives later without a name.

PUTTING A NUMBER

So a column is missing and it should be filled. With what? This is the question that stalls most conversations about natural capital, usually on the grounds that the thing cannot be measured. It can. There are several established methods, none perfect, each with a known bias, and a practitioner's real skill is choosing which bias to accept for the question in hand.

Cost avoidance is the bluntest and often the most persuasive, because it converts an ecosystem into an invoice that did not arrive. Wetlands that filter water save the price of a filtration plant. Mangroves that absorb a storm surge save the rebuild bill. A forest above a coffee plantation that holds a slope together prevents ten thousand dollars of landslide damage every five years, which is two thousand dollars a year of avoided cost — unglamorous, defensible, and available to anyone who can find out what the repair actually costs.

Hedonic pricing reads value out of what people already pay. Compare houses near green space with otherwise similar houses further from it, and the premium that persists after the other variables are controlled is a revealed measurement of what access to that green space is worth to buyers. Nobody was asked their opinion; they voted with a mortgage. The weakness is equally clear: the method measures what the market will pay, which tracks who has money as faithfully as it tracks what is valuable.

Contingent valuation asks people directly what they would pay to protect something, which is the only method that can reach a value nobody currently trades — a species, a view, a river that no one lives beside. Its exposure is well documented. Hypothetical money is looser than real money, and answers drift with how the question is framed. Production function analysis avoids that by working through yields instead of opinions: if a ten per cent rise in bee density lifts apple yield by five per cent, and apples sell at two dollars a pound, pollination is contributing ten cents a pound, which scales to a region as soon as someone counts the acres.

Benefit transfer is the practical shortcut and deserves its warning label. Value a wetland carefully in one place, then apply that value per hectare to a similar wetland elsewhere. It is often the only affordable option, and it is only as good as the word similar. A wetland protecting a city protects more than an identical wetland protecting a field, and a transfer that ignores this will produce a confident number that is wrong in a way nothing downstream can detect.

The honest summary is that every one of these produces a range rather than a figure, and that an ecosystem's worth is not exhausted by any of them. The point was never to establish what a forest is truly worth. It was to stop entering zero. A number with a stated method and a stated uncertainty can be argued with, improved, and defended in a room where decisions are made. That is the entire ambition, and it is enough.

She is not deciding whether to care about this. She is deciding what to count.

THE DISCOUNT RATE

Having a number is only half of it. Ecosystem benefits arrive over decades, and finance has a standard machine for handling that: net present value. Sum the future cash flows, divide each by one plus the discount rate raised to the power of how many years away it is, and add up what remains. A dollar in thirty years is worth less than a dollar today, and the discount rate says how much less. It is the most quietly consequential number in any model that touches the long term.

Run it on carbon. A reforestation project sequesters ten tonnes a hectare a year; carbon credits trade at twenty dollars a tonne; the discount rate is five per cent. The first year's two hundred dollars is worth a hundred and ninety dollars and forty-eight cents today. The second year's identical two hundred is worth a hundred and eighty-one dollars and thirty-six cents. Keep going for thirty years and sum, and you have the present value of a benefit that arrives slowly, expressed in a currency an investment committee already reads.

Notice what the rate is doing across that arithmetic. It is not describing the forest. It is describing impatience — the opportunity cost of capital, plus a judgement about how much risk sits between now and the payment. At five per cent, a benefit thirty years out retains roughly a quarter of its face value. At ten per cent it retains under a sixteenth. Move one parameter and a mature forest becomes rounding error, without a single fact about the forest having changed.

The discount rate is where the ethics sit, wearing the clothes of arithmetic and rarely asked to identify itself.

This is why the choice of rate is the most contested question in the field, and why the argument is so often mistaken for a technical one. A high rate is an assertion that the future should count for very little, made in a form that looks like a market convention rather than a claim. A low rate asserts the opposite. Neither is neutral, neither is derivable from the data, and pretending otherwise is how an ethical position gets smuggled into a model as a default.

The workable response is not to abolish discounting, which would make every model useless, but to declare the rate, defend it, and test the conclusion against alternatives. Run the project at three per cent, at five, at eight, and see which conclusions survive all three. A recommendation that holds across the range is robust. A recommendation that depends on the rate is not a finding about the world; it is a finding about the assumption, and it should be reported as one.

WHEN MODELS CHANGE

Put the pieces together and see what actually moves in a real comparison. A fund weighs two seafood operations. The first catches ten thousand tonnes of tuna a year by conventional methods, selling at five thousand dollars a tonne for fifty million in revenue, emitting five thousand tonnes of carbon dioxide with significant bycatch. The second is an artisanal shellfish fishery: two thousand tonnes a year, ten million in revenue, five hundred tonnes of carbon, negligible bycatch. On revenue alone the choice is not close.

Apply a shadow price — a figure the fund sets for damage the market does not charge for — of fifty dollars a tonne of carbon dioxide. The tuna operation carries two hundred and fifty thousand dollars of environmental cost, leaving forty-nine million seven hundred and fifty thousand. The shellfish fishery carries twenty-five thousand, leaving nine million nine hundred and seventy-five thousand. The larger operation is still, by this measure, far more profitable, and the book says so plainly rather than manufacturing a reversal that the arithmetic does not support.

That refusal to overclaim is what makes the rest of the analysis worth trusting. The gap narrows; it does not close. What tips the decision is not the carbon line but the things the carbon line points at — regulatory exposure on bycatch, the stock status of the tuna fishery itself, the difference between an operation whose costs are about to be internalised by law and one already operating inside the limits that law will impose. The shadow price did not decide anything. It made visible a risk that had been sitting in the position unpriced.

The same arithmetic runs in the other direction when the ecosystem is an asset rather than a liability. A five-hundred-acre organic quinoa farm plants hedgerows for pollinators and estimates fifty dollars an acre of additional yield, or twenty-five thousand dollars. No-till practice and cover crops sequester about a tonne of carbon an acre annually, worth ten thousand dollars at twenty dollars a tonne. Efficient irrigation cuts water use by thirty per cent, around fifteen million gallons, which at half a cent a gallon is seventy-five thousand dollars. That is a hundred and ten thousand dollars a year that a conventional model would never have gone looking for.

And there is a risk story underneath the revenue story that matters more over a holding period. Diverse soil biology reduces exposure to a single pest. Water efficiency reduces exposure to a drought year and to the price of water rising. Practices already inside future regulation reduce exposure to that regulation arriving. None of these show up as a line of income; all of them show up as a lower probability of the year that ruins the position.

Which returns the argument to where it started, with the container. A portfolio built on the machine model is optimised against a world assumed to be stable and separate. A portfolio built on the living-systems model is optimised against a world that is neither. The second is not the sentimental choice. It is the choice that has priced in the thing the first one left out, and pricing in what you left out is the ordinary definition of good analysis.

THE FIRST MOVE

None of this is reserved for people who manage other people's money. The methods scale down without losing their shape, and the first move is the same at every size: find out what you are actually holding. Where does the income come from, what is it doing, and what does it depend on that nobody is charging for? Where does the spending go, and which stocks are being drawn down at the far end of it? The audit is unglamorous and it is the part that changes the most.

For an institution the protocol has four steps and they run in order. Identify which ecosystem services are material to the holding — carbon and watershed for forestry, pollination and soil for agriculture, land use and hydrology for renewables. Quantify them using the methods above, with the bias of each method stated rather than hidden. Integrate them: as revenue where they are earned, as cost where they are consumed, as a discount-rate adjustment where the benefit is genuinely long-lived. Then publish the workings, because a valuation nobody can check is a marketing claim wearing a model's clothing.

For a company the first step is measurement rather than intention. Audit the actual footprint — water, land, materials, waste, energy — instead of estimating it, because estimates of one's own impact have a reliable direction of error. Set targets against what the surrounding systems can carry, not against last year's figure, since improving on last year is compatible with overshooting by a wide margin. Then move the operations, which is the slow part and the only part that counts.

For an individual the same logic fits inside ordinary decisions. Calculate the footprint once, honestly, and notice which two or three categories carry most of it — for most people the list is shorter and less surprising than expected. Choose funds whose holdings you have actually looked at rather than whose names sound correct. Support businesses whose supply chains they are willing to describe. Buy fewer things and keep them longer, which is the oldest ecological economics there is.

One caution deserves stating plainly, since it is the most common way this whole approach fails. A sustainability figure that nobody can audit is worse than no figure, because it occupies the space where a real one would go and makes the absence look filled. Ask what method produced a number, what it excluded, and who checked it. A firm that can answer those three questions is doing the work. A firm that cannot is selling the appearance of it, and the appearance is cheaper to produce.

The other caution is about pace, and it runs the opposite way from the guilt that usually attends this subject. Nothing here requires anyone to get the whole picture right before making the first move. Ecosystems do not reward sudden total conversions; they reward small persistent changes to the flows, sustained long enough to change a stock. The same is true of a portfolio, a household and a firm. Begin with the one number you can actually measure, and measure it again in a year.

PROSPERITY RECOUNTED

Step back from the arithmetic and the argument turns out to be about a single word. Prosperity, in the accounts we currently keep, is a flow: how much moved through, how fast, this quarter. But a flow can be high precisely because a stock is being emptied, and the reading gives no sign of which of the two is happening. A household selling its furniture has an excellent month by that measure. So does a country selling its forests.

Read prosperity as a stock instead and the picture reorganises without any new mysticism being required. Soil that can still grow food. Water that can still be drunk without treatment. A climate inside the range the cities were built for. Species enough that no single failure cascades. Skills, institutions, trust, health. These are accumulations, they took time to build, they can be spent, and unlike a quarter's revenue they cannot be recovered by trying harder next quarter.

A good year and a liquidation look identical in the revenue line. The difference is in a stock nobody was asked to report.

What ecological economics adds to that intuition is not the intuition — most people already have it — but the arithmetic to act on it. The logistic curve gives a ceiling and a brake. The footprint gives a ratio of use to capacity. The valuation methods give a defensible number in place of a blank. Net present value, with its rate declared, carries a benefit across decades into a form a committee can read. None of these is exotic. All of them are the ordinary tools of finance, pointed at the container rather than only at the contents.

This is why the field is more hopeful than its subject matter suggests. If the problem were that nature is unmeasurable, there would be nothing to do but argue. But the problem is narrower and far more tractable: the measurements exist, they are improving, and they are not yet in the models. Getting them in is a long unglamorous task of the kind that institutions are genuinely good at, once the case is made in the language they already use.

The strawberry is still the whole thing in miniature. Its price is honest about the hours and the fuel and the crate, and silent about the bees, the soil and the rain. Nothing is wrong with the price. Something is missing from it. Put the missing part in, with a method anyone can check and an uncertainty anyone can argue with, and the number begins to describe what actually happened — which is, in the end, all any accounting system was ever asked to do.

Free to read

Free to read, and free to hear. Every chapter of every book in this house, and every narration of it, is open to anybody. No account, no card, nothing to cancel.

What is in it

• Introducing Living Systems: Ecology Meets Economicsopen this in the house search
• Valuing Nature's Capital: Beyond GDPopen this in the house search
• The Limits to Growth: Ecological Footprints and Carrying Capacityopen this in the house search
• Sustainable Finance: Principles and Practicesopen this in the house search
• Integrating Ecosystem Services into Financial Modelsopen this in the house search

Nature is not unpriced in these models. It is priced at zero, by default, and nobody had to sign for it.

Every projection assumes a ceiling. Most have never been asked to name it.

Spending principal while reporting interest is the oldest accounting error, and the largest one now running.

A measurement with a stated bias can be improved. A blank cannot be argued with at all.

Keep looking

Every phrase on this page opens into the house search. The shelf holds Living Systems Economics and six other shelves, and the reading is free.