The Bayesian moment for carbon and nature

In Bayesian reasoning, every belief (a “prior”) is a statement about the world before we see new data. Such new data may just have been strengthened by Joseph Romm and Stephen Lezak, whose recent article, “Are carbon offsets fixable?” published in the Annual Review of Environment and Resources (2025)¹, dissects the state of global carbon offsets. Their conclusion is as elegant as it is sobering: most offsets still overstate their impact five- to tenfold. After 25 years, the system that was meant to channel capital from industrial emissions to genuine climate mitigation remains riddled with faulty baselines, temporary storage, and accounting optimism. This should not shake our belief that carbon markets are needed. A uniform, comprehensive Pigouvian price on carbon emissions remains the first-best instrument for decarbonisation. It aligns private cost with public harm, drives capital efficiency, and makes abatement cheaper than inertia. But the Romm–Lezak paper marks a Bayesian update: The evidence now demands that we revise, not abandon, our carbon priors. And rethink our nature priors altogether.

Carbon needs reform

Romm and Lezak’s review is the most systematic to date of the carbon offset industry. They conclude that most projects exaggerate their contribution to emission reductions; additionality is rarely proven; leakage often cancels claimed benefits; and permanence is uncertain. Weak oversight and inconsistent verification mean that a large share of traded credits represent accounting artefacts rather than atmospheric gains.

Their prescription is pragmatic rather than nihilistic. Offsets, they argue, should be limited to rigorously verified carbon removals (CDR) — projects that demonstrably take CO₂ out of the atmosphere and keep it out. The future of carbon markets lies in these removal-based systems, complemented by direct domestic mitigation, not in the trade of hypothetical “avoided emissions.” BCG and others similarly predict that the voluntary and compliance markets will increasingly migrate toward durable removal credits, not avoidance schemes²³. In conclusion: carbon markets still matter — but only when they measure equivalence.

Nature needs reinvention

The Romm–Lezak analysis does not discuss biodiversity or ecosystem finance directly, yet its implications land hardest there: it shows that nature must be more than a “nature-based climate solution” squeezed into the logic of carbon markets. The evidence points to a simple truth: carbon markets will not fund nature – particularly standing nature. Or clearly not at the extent, speed required. Five reasons follow logically from their findings:

1. When success disqualifies itself — protection undermines credibility: Romm and Lezak (2025)¹ demonstrate that carbon offsets rest on unverifiable counterfactuals and inflated baselines — most overstate their impact five- to ten-fold. The more secure a forest becomes, the less “additional” it appears, and the fewer credits it can issue. The Carbon Paradox describes this as the “success penalty” of offset logic: sound stewardship deletes its own revenue stream⁴. The market, therefore, rewards vulnerability, not stewardship, and channels funds toward marginal or threatened land rather than the ecosystems already performing well.
2. Scarcity devalues stewardship — markets punish stability, not neglect: In their review, Romm and Lezak trace how offset systems create value from the risk of loss, not from sustained care. The Carbon Paradox calls this “the economics of endangerment.”⁴ High-risk ecosystems — where deforestation appears imminent — attract credit supply, while well-protected areas yield nothing because destruction is improbable. The paradox is stark: the safer nature becomes, the lower its financial return. Carbon markets thus monetise peril, not preservation, and undervalue the continuous stability on which resilience depends.
3. Capital follows carbon, not risk — finance flows where nature least needs it: Romm and Lezak highlight how market incentives chase low-cost, high-density carbon¹. The Carbon Paradox extends this observation: carbon finance floods humid tropics where tonne yields are cheap and bypasses ecosystems of low carbon content but high ecological importance—grasslands, saltmarshes, coral reefs, alpine meadows⁴. Carbon markets price molecules, not fragility; nature’s real risk map remains unmonetised.
4. One metric, many losses — carbon accounting erases ecological complexity: As Romm and Lezak note, offsets translate climate performance into a single scalar—tonnes of CO₂ reduced—ignoring broader environmental integrity¹. The Carbon Paradox argues that this one-dimensional logic “flattens the ecology of value.”⁴ Biodiversity, soil fertility, hydrology, and cultural services vanish into the tonne, despite attempts to establish co-benefit rewards. The metric that brings clarity to carbon policy, therefore, brings blindness to nature policy. Living systems cannot be reduced to arithmetic without losing what makes them living.
5. Integrity reforms widen the gap — counterfactual finance can’t fund real assets: Romm and Lezak conclude that the integrity crisis of offsets is structural, not incidental: stricter verification cuts eligible supply and “cannot deliver reliable mitigation at scale.”¹ The Carbon Paradox echoes this dynamic — raising standards improves honesty but deepens the financing gap⁴. The cleaner the carbon market becomes, the less it finances nature. The logical remedy is not looser rules but a new asset class: real-asset finance for ecological condition and performance. Counterfactual instruments can’t maintain actual ecosystems.

A powerful alternative — towards first-best regimes for nature funding

The conversation cannot end with what doesn’t work. And no, there is no textbook solution yet for how to govern a functioning nature market. But after decades of experimentation, we know enough to outline its first-best design principles — a framework that treats nature as a real, measurable, and investable asset rather than a by-product of carbon arithmetic.

1. Clear political guidelines: Nature finance begins with law, not ledger. The most effective regimes start with hard, measurable restoration and condition targets, enforced through “no-deterioration” backstops such as the EU Nature Restoration Law⁵. They must be paired with Ecological Fiscal Transfers (EFTs) that reward sub-national governments for verified ecosystem health⁷ — paying for outcomes rather than promises. This creates a fiscal architecture that values maintenance and resilience.
2. Environmental markets: Permitting systems can embed mandatory “no-net-loss” or “net-gain” rules, as in England’s Biodiversity Net Gain requirement⁶. These instruments translate ecological condition into regulatory currency: a right to develop is conditional on improving or restoring nature elsewhere. Whilst also an instrument of compensation, they move beyond some limitations of carbon markets – and accrue about 10 times more funds than VCMs provide to nature, even today. However, done well, such markets set a floor of obligation that private capital can build on as it searches for new assets.
3. Financial markets: The next frontier is to treat nature outcomes as a trusted asset class. Healthy ecosystems mitigate operational and credit risk, reduce liabilities, secure ecosystem services, and store long-term value. When those attributes are recognised on company, bank, and central-bank balance sheets, nature gains access to trillions, not millions, in capital — a shift from marginal philanthropy to mainstream finance⁸. This is the powerful alternative: a rules-based, multi-market system that prices what nature provides and protects what it sustains.

Romm and Lezak, and many others before them, have given us new data. Bayes’ great insight was that belief is not a fortress but a function — it must change with the evidence. The man who proved that died before anyone believed him. Three hundred years later, his equation still asks the same question we face in climate and nature finance: They are one connected problem. But it may require more than one solution.

This article is also published on LinkedIn. illuminem Voices is a democratic space presenting the thoughts and opinions of leading Sustainability & Energy writers, their opinions do not necessarily represent those of illuminem.

Track the real‑world impact behind the sustainability headlines. illuminem’s Data Hub™ offers transparent performance data and climate targets of companies driving the transition.

Endnotes

1. Joseph Romm and Stephen Lezak, “Are Carbon Offsets Fixable?” Annual Review of Environment and Resources 50 (2025): forthcoming.
2. Boston Consulting Group, “Boosting Demand for Carbon Dioxide Removal,” BCG Publications, June 2024, https://www.bcg.com/publications/2024/boosting-demand-for-carbon-dioxide-removal.
3. Boston Consulting Group, “The Voluntary Carbon Market Is Thriving,” BCG Publications, January 2023, https://www.bcg.com/publications/2023/why-the-voluntary-carbon-market-is-thriving.
4. The Carbon Paradox (2024), Illuminem Voices Series, “Nature Paradox,” https://illuminem.com/illuminemvoices/20-nature-paradox.
5. European Commission, Nature Restoration Law, Regulation (EU) 2024/1323 of the European Parliament and of the Council, June 2024.
6. DEFRA (UK Department for Environment, Food and Rural Affairs), Biodiversity Net Gain Regulations 2023 (England), effective February 2024.
7. Sven Wunder et al., “Ecological Fiscal Transfers: Concept, Practice, and Lessons for Environmental Finance,” Ecological Economics 204 (2023): 107689.
8. Taskforce on Nature-related Financial Disclosures (TNFD), Final Recommendations, September 2023; European Central Bank, “Climate-related and Environmental Risks: Guide for Banks,” 2020.