FONAFIFO's published schedule for 2025 pays a Costa Rican landowner 21,000 colones per hectare per year to leave a forest standing. If that same hectare also protects a water supply, the rate rises to 37,000. If it protects a water supply and high biodiversity, it rises to 53,000, or 265,000 colones across the five year contract. The price list is the mechanism. Everything else is the machinery required to make a price list of that kind enforceable.
A payment for ecosystem services is a conditional transaction: somebody who benefits from a functioning ecosystem pays somebody who controls the land, and the payment depends on the land being managed a particular way. The three conditions that make it what it is are that the buyer is voluntary or at least identifiable, that the seller could plausibly have done otherwise, and that the money stops if the condition fails. Remove the third and it is a subsidy. Remove the second and it is a windfall.
Costa Rica's Forest Law 7575, passed in 1996, is unusually explicit about what is being bought. Article 3(k) names four environmental services: mitigation of greenhouse gas emissions, covering fixation, reduction, sequestration, storage and absorption; protection of water for urban, rural or hydroelectric use; protection of biodiversity for conservation and for sustainable, scientific and pharmaceutical use, including research and genetic improvement; and natural scenic beauty for tourism and science. FONAFIFO, the national forest financing fund, repeats the same four on its programme page and describes the shift the law made as replacing the idea of a subsidy or incentive with an economic recognition of services the forest already provides. It is a national scheme that has been running since the late 1990s and publishes its own contracting figures, which is why so much of the argument about whether this mechanism works is in practice an argument about Costa Rica.
Four services, one price
The four services are not sold separately. What the 2025 schedule actually prices is a single bundled hectare of protected forest, with two increments layered on top: one for watershed protection, one for high biodiversity. The base row is labelled for greenhouse gas mitigation and biodiversity together. Scenic beauty, the fourth service the law recognises, is not priced in the schedule at all. This matters when the model is copied. A programme designed around the idea that each service commands its own price will find that the scheme it is copying does not work that way and never has.
The rest of the schedule is a set of different contracts for different land uses rather than different services. Natural regeneration pays 16,513 colones per hectare per year over five years. Forest management pays 28,768 over the same term. Reforestation is paid over a much longer horizon and is heavily front loaded: native species earn 1,325,366 colones per hectare across a sixteen year contract, of which 662,683 arrives in year one and 66,268 in year five. Agroforestry is priced per tree rather than per hectare, at 543 colones in year one and 271 in years three and five. All of these rates are set by ministerial instrument, currently Resolution R-570-2025-MINAE and Executive Decree 44607-MINAE, which means they can be changed without reopening the law. Readers converting these figures should take the rate from the Banco Central de Costa Rica for the date they care about rather than from a secondary source, because the colon has moved substantially over the life of the programme and a single conversion factor applied across three decades produces nonsense.
Comparing two lines of that schedule shows what the programme is actually willing to pay for. Five years of base forest protection is worth 105,000 colones a hectare. Establishing native species forest on a hectare is worth 1,325,366, roughly thirteen times as much, and every colon of it is disbursed in the first five years of a sixteen year commitment. Restoration is expensive, front loaded and slow to verify. Non-destruction is cheap, and the state buys it in five year blocks it can decline to renew.
Where the money comes from
Article 69 of the Forest Law directed one third of the receipts of the selective consumption tax on fuels and other hydrocarbons to compensation programmes for forest and plantation owners. FONAFIFO's own account records what happened next: the Tax Simplification Law replaced that levy with a single fuel tax, of which 3.5 per cent is assigned to the programme. The earmark survived. Its arithmetic did not.
A second stream comes from water users. FONAFIFO tracks contracts financed from the water tariff separately, and between 2015 and 2025 that source paid for 1,263 contracts covering 84,936 hectares, worth roughly 21.8 billion colones. The distribution is lumpy rather than steady. The year 2024 alone accounts for 39,123 of those hectares, close to half the area financed this way across the whole series, on 323 contracts.
Article 47 lists what else may capitalise the fund, and one clause is worth noticing because it connects two mechanisms that are usually described as alternatives. Among the permitted sources are proceeds from the conversion of external debt. Costa Rica wrote the receipts of a sovereign debt conversion into the statute governing its payment programme in 1996, decades before the current wave of such transactions.
Eleven years of contracting
Between 2015 and 2025 the programme signed 7,815 contracts covering 567,740 hectares and 3,488,436 individual trees under agroforestry, on the table that excludes mixed systems. Forest protection dominates: 526,689 of those hectares, against 21,573 for natural regeneration, 15,409 for reforestation and 4,070 for forest management. Whatever else the scheme is, in volume terms it is a payment to keep existing forest intact rather than a payment to create new forest.
Two years break the pattern. In 2021 the programme very nearly stopped. Forest protection contracting fell to 5,870 hectares, against 31,954 the year before and 46,857 the year after. Only 157 contracts were signed on that table, alongside 14 mixed-systems contracts recorded separately. No forest management or natural regeneration contracts were formalised at all, and FONAFIFO's separate table on indigenous territories records that no contracts were signed there either, the only such year in the series. Then in 2024 forest protection contracting reached 116,043 hectares, close to double the next highest year in the window, while agroforestry fell to 112,703 trees, its weakest showing outside the 2021 collapse. FONAFIFO's tables report all of this and explain none of it. Anyone who needs the cause should read the fund's annual reports and financial statements for those years, which are published on the same site, rather than accept an inferred explanation.
The indigenous territory figures are worth separating out. Across the same eleven years the programme signed 153 contracts in indigenous territories worth about 33.4 billion colones. That is one eighth of the contract count of the water tariff stream and roughly one and a half times its value, which means the average contract in an indigenous territory is very much larger than the average contract elsewhere in the programme. That is a consequence of territory size and collective title rather than of a higher rate, since the rates are set nationally.
What the evaluations found
The early Costa Rican record has been assessed directly, and the finding is not flattering. Robalino and Pfaff, publishing a nationwide analysis in Land Economics in 2013, estimated the impact of the programme's initial years on deforestation between 1997 and 2000 at between zero and one fifth of one per cent per year. In their own phrasing, deforestation was avoided on at most two out of every thousand enrolled hectares. Their explanation was not that the payments failed to change behaviour but that there was very little behaviour left to change: the national deforestation rate was already low. They also found an enrolment bias toward land facing a lower clearing threat, and concluded that enrolling higher-threat land would raise the impact of the same payments.
Set against that, the strongest positive evidence for the mechanism comes from a randomised trial rather than an observational study. Jayachandran and colleagues, publishing in Science in 2017, ran a payment scheme across 121 villages in western Uganda, of which 60 received the programme for two years. Forest-owning households were offered 70,000 Ugandan shillings per hectare per year, about twenty eight dollars at 2012 rates, to conserve their forest. Measured from high-resolution satellite imagery, tree cover fell by 4.2 per cent in treatment villages against 9.1 per cent in control villages. The authors found no evidence that enrollees moved their tree-cutting to nearby land. Valuing the delayed carbon dioxide emissions using the social cost of carbon, they put the benefit at 2.4 times the programme's cost.
These two results are usually presented as a disagreement about whether payments work. They are better read as the same finding stated twice. In Uganda the counterfactual was rapid clearing, so a small payment bought a large change. In Costa Rica in the late 1990s the counterfactual was mostly standing forest, so the same instrument bought a small one. The variable that moves the answer is not the price. It is the threat facing the enrolled hectare.
The selection problem is the design problem
This is why additionality is not a technical footnote in these schemes but the whole of the design. A programme that pays whoever applies will be applied to by the landowners for whom the payment is easiest to earn, which is to say those who were going to keep their trees anyway. Costa Rica's response has been to weight applications by conservation priority, and FONAFIFO publishes a separate statistical table on contracts ranked by priority criteria. The Uganda trial took the opposite route and simply randomised, which is why its estimate is credible and also why it is not a policy that any ministry could run.
Thirty years in, the useful question about the Costa Rican programme is no longer whether payments changed land use in the 1990s. It is what a national price list, hard-wired to a fuel tax and a water tariff and adjustable by ministerial resolution, does to a country's forest politics over decades. The 2021 collapse and the 2024 surge in the same series suggest the answer is that it makes conservation funding as steady as the budget process behind it, and no steadier.
Sources
Every figure, date and deal name above is drawn from one of the documents below. Each was fetched and cached on the retrieval date shown.
- Costa Rica, Ley Forestal No. 7575 (1996). Full text, including article 3(k) on recognised environmental services, article 47 on the sources of FONAFIFO funds, and article 69 on the fuel tax allocation. Hosted by CONAGEBIO https://www.conagebio.go.cr/sites/default/files/2023-08/leyforestal-7575.pdf Retrieved 2026-08-30
- FONAFIFO, Pago de Servicios Ambientales. The programme page, including the four services recognised under Law 7575 and the financing section describing the move from one third of the fuel consumption tax to 3.5% of the single fuel tax https://www.fonafifo.go.cr/es/servicios/pago-de-servicios-ambientales/ Retrieved 2026-08-30
- FONAFIFO, Estadisticas de PSA. Index of the programme statistics tables, each carrying its own cut-off date https://www.fonafifo.go.cr/es/servicios/estadisticas-de-psa/ Retrieved 2026-08-30
- FONAFIFO, Cuadro 1: hectareas y arboles contratados por actividad de PSA, 2015 to 2025. Cut-off 3 August 2026. Source given as Departamento de Gestion de Servicios Ambientales, SiPSA https://www.fonafifo.go.cr/media/4735/1-distribuci%C3%B3n-de-las-hect%C3%A1reas-y-%C3%A1rboles-por-actividad-psa.docx Retrieved 2026-08-30
- FONAFIFO, Montos a reconocer: the 2025 payment schedule per hectare and per tree, set by Resolucion R-570-2025-MINAE and Decreto Ejecutivo 44607-MINAE https://www.fonafifo.go.cr/media/4736/2-montos-a-reconocer.docx Retrieved 2026-08-30
- FONAFIFO, Contratos PSA financiados con recursos del Canon de Agua, 2015 to 2025. Contracts, hectares and colones financed from the water tariff https://www.fonafifo.go.cr/media/4744/10-contratos-psa-financiados-con-recursos-del-canon-de-agua.docx Retrieved 2026-08-30
- FONAFIFO, Monto total contratado en PSA en los Territorios Indigenas, 2015 to 2025, including the note that no contracts were formalised in indigenous territories in 2021 https://www.fonafifo.go.cr/media/4742/8-contratos-psa-formalizados-en-territorios-ind%C3%ADgenas.docx Retrieved 2026-08-30
- Robalino, J. and Pfaff, A. (2013), Ecopayments and Deforestation in Costa Rica: A Nationwide Analysis of PSA’s Initial Years. Land Economics 89(3), 432 to 448. Abstract https://le.uwpress.org/content/89/3/432 Retrieved 2026-08-30
- Jayachandran, S. et al. (2017), Cash for carbon: a randomized trial of payments for ecosystem services to reduce deforestation. Science 357(6348), 267 to 273. Published article, author-hosted copy, because science.org refuses automated fetch https://seemajayachandran.com/cashforcarbon.pdf Retrieved 2026-08-30
- Jayachandran, S. et al., Cash for Carbon. NBER Working Paper 22378, whose abstract reports the earlier and wider effect ranges https://www.nber.org/papers/w22378 Retrieved 2026-08-30