In 1997 a team led by the ecologist Robert Costanza published a paper in Nature estimating the monetary value of the goods and services that the world's ecosystems provide to humans — water purification, pollination, soil formation, climate regulation, fisheries, recreation, the cleaning of polluted air. Their figure was $33 trillion per year, larger than global GDP at the time. The methodology was widely criticized; the order-of-magnitude conclusion was difficult to dismiss. An updated 2014 valuation reached ~$125 trillion per year. The paper opened a debate that has run through environmental economics, conservation policy, and public-good valuation ever since: what are ecosystems worth, and how should that worth enter human decisions?
The Millennium Ecosystem Assessment (2005), a five-year UN-led synthesis drawing on the work of more than 1,300 researchers, organized ecosystem services into four categories. Provisioning services are the marketable outputs: food, water, fiber, fuel, medicinal compounds. Regulating services are the unmarketed processes ecosystems perform: water purification by wetlands, pollination by insects, climate regulation by forests and oceans, flood and storm-surge moderation by mangroves and reefs. Cultural services are the meanings humans attach to landscapes — recreation, spiritual significance, aesthetic experience. Supporting services underpin the others: soil formation, nutrient cycling, primary production. The categorization was as much political as scientific: making the regulating services visible as services rather than as background conditions was the intellectual move that allowed economists and ecologists to begin pricing them.
Pricing the priceless is the methodological problem. Provisioning services trade in markets and have prices; regulating services must be priced indirectly — by the cost of replacing them (a water-treatment plant equivalent to the wetland that purified the water for free), by avoided damages they prevent (storm surge buffered by mangroves), or by contingent valuation surveys asking what people would pay to preserve them. Contingent valuation is the most contested method: respondents often answer the question they wish they were being asked, and the values produced are sensitive to question wording. Costanza's number was assembled by aggregating heterogeneous estimates from many studies, and critics argue that ecosystems are not substitutable in the way the aggregation assumes — an Amazon-sized hole in the global ecosystem cannot be replaced by spending Amazon-sized dollars on something else. What survives the critique is the directional claim: ecosystems perform work that, if it had to be replaced by industrial substitutes, would carry costs at least comparable to global economic output, and the failure to price that work into normal economic decisions is a systemic externality of the kind environmental economics exists to correct.
Payment for Ecosystem Services (PES) schemes have implemented the valuation logic. Costa Rica's national PES program, established in 1997, pays landowners directly to maintain forest cover; deforestation rates fell sharply in the years following implementation. The REDD+ framework under UN climate negotiations applies the same logic at international scale, paying tropical countries to keep forests standing as carbon sinks. The EU's Corporate Sustainability Reporting Directive (effective 2024) now requires companies to report their biodiversity exposure. The valuation move that began as a thought experiment in Nature in 1997 has, three decades on, found its way into the cost stack of every actor that depends on functioning ecosystems — which, when the audit is finally done, turns out to be all of them.