A number that stopped a conversation#
In May 1997, a paper appeared in Nature that would go on to become one of the most cited and most argued-about publications in the history of environmental science. Robert Costanza, an ecological economist at the University of Maryland, and twelve co-authors from four countries had spent three years assembling the most comprehensive attempt ever made to answer a deceptively simple question: how much is nature worth?
Their answer was $33 trillion per year. In 1997 dollars. Compared to world GDP that year of approximately $32 trillion.
The paper was not claiming that all ecosystem services could be replaced or bought. It was claiming something more pointed: that the flow of services from natural systems to the human economy — water filtration, pollination, climate regulation, soil formation, nutrient cycling, flood control, coastal protection, and dozens of other processes that underwrite the economic activity we actually measure and count — was worth more, annually, than all the economic activity we measured and counted. And that the price charged for these services by the natural systems that provided them was zero.
The economics profession treated the paper with considerable hostility. Critics argued that the methodological aggregation was invalid, that marginal valuation methods cannot be summed to a total because the marginal value of ecosystem services becomes unbounded as they approach zero. These criticisms had technical merit. The response — that treating the value as zero was somehow more rigorous than treating it as positive and large — did not.
What the economy takes for free#
The central claim of the Costanza paper, and of the literature that followed it, is not actually about the precise number. The precise number is controversial and will remain so as long as there is genuine uncertainty about ecosystem service quantities and unit values. The central claim is structural: the economy receives a continuous flow of services from living systems — services that enable production, protect infrastructure, regulate climate, filter diseases, and maintain the soil chemistry on which agriculture depends — and prices none of them in any market, tax system, national account, or corporate balance sheet.
The Ecosystem Dependency Ratio makes this structural claim operational. EDR = Estimated GDP value dependent on biodiversity-supplied ecosystem services ÷ Annual public + private investment in biodiversity protection. Using updated figures: the World Economic Forum's 2020 Nature Risk Rising report estimated that approximately $44 trillion of global GDP — representing more than half of world GDP — has moderate to high dependence on nature and the services it provides. Annual global biodiversity finance — from public protected area budgets, development aid with biodiversity objectives, and private conservation investment — is estimated at approximately $100–130 billion per year (using the OECD's most comprehensive data as of 2023).
This gives an EDR of approximately 340–440, using the conservative WEF figure for the numerator. Using Costanza's 2014 revised global ecosystem service value of $125–145 trillion, the EDR reaches approximately 1,000–1,300. The specific ratio depends on assumptions about what fraction of GDP is "dependent" on ecosystem services versus produced independently of them — a definitional boundary that reasonable analysts can place differently. The direction and order of magnitude are not substantially in dispute: the economy extracts several hundred to over a thousand times more value from nature than it invests in nature's maintenance.
The taxonomy of services#
The Costanza team categorised ecosystem services into seventeen functional types. The categories remain, with refinements, the standard framework used in subsequent ecosystem service valuations: gas regulation (atmospheric composition maintenance, including CO₂ absorption), climate regulation (temperature and precipitation regulation at local and regional scales), disturbance regulation (storm and flood buffering by coastal wetlands and forests), water regulation (watershed storage, flood prevention), water supply (freshwater provisioning through aquifer recharge and watershed hydrology), erosion control and sediment retention (soil retention against runoff), soil formation (parent material weathering and organic matter accumulation), nutrient cycling (nitrogen, phosphorus, and sulphur cycles), waste treatment (pollution processing by microorganisms and filtering ecosystems), pollination (crop and wild plant reproduction support), biological control (pest population regulation by predators and parasites), habitat provision (refugia for commercial and wild species), food production (fisheries, wild game, wild fruits), raw materials (timber, fibre), genetic resources (pharmaceutical and agricultural breeding stocks), recreation (tourism, aesthetic appreciation), and cultural services (spiritual, intellectual, scientific value).
The size of individual service categories is illuminating. The nutrient cycling category — the set of biological processes that convert atmospheric nitrogen into plant-available forms, cycle phosphorus through soil and water, and decompose organic matter into mineral nutrients — was estimated at approximately $17 trillion per year in Costanza's 1997 figures, larger than any single country's GDP at the time. The global commercial fertiliser industry — the partial industrial substitute for natural nutrient cycling — was worth approximately $200 billion per year in 2022. The ecosystem service that fertiliser partially replaces is worth roughly 85 times the value of its replacement, and the replacement covers only the provisioning function, not the soil aggregate stability, moisture retention, and biological diversity maintenance that the natural nutrient cycle simultaneously provides.
The 2014 revision#
In 2014, Costanza and an overlapping set of co-authors published a methodological update in Global Environmental Change that incorporated two decades of additional ecosystem service studies and adjusted for changes in land cover and ecosystem extent. The revised global annual ecosystem service value was approximately $125–145 trillion per year in 2011 dollars — a figure that had grown from the 1997 estimate not because of inflation adjustment alone but because research had identified additional service categories and improved unit value estimates based on revealed preference, replacement cost, and contingent valuation studies across more ecosystems.
The same 2014 paper estimated that land use change between 1997 and 2011 — primarily deforestation, wetland drainage, and conversion of natural grasslands to agriculture — had reduced global ecosystem service value by approximately $4.3–20.2 trillion per year. This is the rate at which the ecological capital stock that generates the EDR's numerator is being drawn down: approximately $4–20 trillion per year in service value lost to land conversion, permanently, as the ecosystems destroyed cannot be quickly restored.
The $4–20 trillion range brackets the annual global cost of all climate mitigation investment (approximately $750 billion/yr as of 2022, IEA), all development finance spending (approximately $190 billion/yr), and approximately 200 times the annual global investment in biodiversity protection. The loss is happening at this scale quietly, largely unrecorded, distributed across millions of land conversion decisions that each seem individually rational under the economic signals that markets provide.
What the invoice would require#
The practical implication of the Ecosystem Dependency Ratio is not difficult to state. If the economy depends on nature at a leverage ratio of 340–1,300, and if the capital stock generating that dependence is being drawn down at a rate that will reduce future service flows, then the rational financial response is to invest in the maintenance of that capital stock at a level proportional to its contribution to economic output.
No institutional investor would accept a leverage ratio of 1,000:1 on an asset base with a clearly deteriorating maintenance account. No government treasury would allow a public infrastructure system with that leverage ratio to operate on a maintenance budget 1,000 times smaller than the economic output it supported, while documentably degrading. The accounting frameworks that make this situation invisible — the national income accounts that treat ecosystem service consumption as income rather than asset drawdown, the corporate accounting standards that record no liability for biodiversity loss, the investment frameworks that have no required disclosure of ecosystem service dependency — are the proximate mechanisms by which the EDR continues to grow while the denominator falls.
The next post in this series moves from the accounting framework to the biological reality that the framework is failing to register: the rate at which the capital stock is being withdrawn, and what the loss of individual species actually means for the ecosystem service flows on which economic activity depends.
