Post date 24 Aug 2026 (original LinkedIn date)
Original LinkedIn version· 1,178 words
[This is one you REALLY need to listen to!:
]
‘Troubled so hard (Natural blues)’ performed by Paolo Nutini, (a massively underrated artist in my less than humble opinion!), at the Isle of Wight music festival in 2007, and later released on an eponymous EP that year
For today’s musings, I am deeply grateful to Leela Raina over at the World Bank, who challenged me to take a look at what Outcome-based finance (OBF) and parametric insurance might have to inform our work at Rebalance Earth looking at nature outcomes wrapped up in Resilience Purchase Agreements (RPAs). Outcome-based finance has clearly come a long way and for me there’s a real opportunity for Nature finance to copy its homework ... and most importantly to learn from its mistakes.
The World Bank Group (WBG)’s proposed Indonesia Coral Bond would pay investors only if coral-reef fish biomass improves in three marine protected areas - not if activities happen, not if money is spent, but only if the fish come back. It’s that simple! The Bank is admirably candid about the hard part: outcome-based models have “never been tested at scale for marine ecosystems, due to challenges in defining and measuring robust outcomes.” I think there’s a big lesson in there for nature finance - the financial instrument is the easy bit. The outcome definition is where these structures live or die.
At Rebalance Earth we are building Resilience Purchase Agreements, long-term contracts under which a company pays for verified flood-risk reduction delivered by upstream natural flood management. As we’ve drafted the legal architecture, I had negelected to go back through fifteen years of outcome-based finance and thirty years of parametric insurance, asking two questions, but I’m delighted that I’ve fixed that now! So what did they get right that we should unashamedly use? And where are the scars we should learn from rather than (re-)earn ourselves?
What outcome-based finance got right
Three things transfer directly. First, the independent verification agent: every credible impact bond names a third party, separately contracted and separately paid, whose only job is to certify whether the outcome was achieved. Appointment, replacement, disputed certification, cost allocation: these are all clearly specified before signing. This machinery is proven and reusable, and we ought to lift it wholesale. Second, the rate card: mature outcomes funds stopped renegotiating a price for every outcome and started publishing standardised metrics with agreed prices attached, so many transactions reference one tested definition. That is exactly what a standardised resilience contract becomes over time: a priced, recognised outcome definition per intervention type. The trajectory from bespoke to rate card is proven, we just have to design for it from the start. Third, the pre-agreed counterfactual: the impact-bond field learned painfully that naive before/after measurement gets gamed, and it moved to validated baselines and attribution rules fixed in the contract, before signing, not bolted on afterwards. This makes additionality is a drafting question, not a measurement afterthought.
Where outcome-based finance scarred itself
Two challenges seem to recur, however, across the academic evaluations, and both appear avoidable. The transaction-cost trap: the most consistent criticism of impact bonds is that metric design, verification and multi-party coordination cost so much that most stayed small - and critically the cost never fell with repetition, because every bond was substantially bespoke. The field of OBF proved the outcome-payment idea and then failed to industrialise it throught scaling an replicability. I think that one fix is documentary architecture: a standard master agreement written once, a small library of outcome-type supplements, and a short deal-specific schedule. The derivatives market solved precisely this problem in the 1980s with the ISDA documentation stack; nature finance has simply never applied it.
Metric fixation: pay narrowly for a measured outcome and providers may default to gaming the metric, even sometimes against the bigger goal it means to proxy. Employment programmes learned this as ‘creaming and parking’. The nature-finance equivalent is optimising the river gauge while the wider catchment degrades. And one answer is a headline metric plus monitored guardrails: leakage checks across a defined catchment boundary, so the gauge cannot be gamed in isolation.
What parametric insurance teaches about measurement
Parametric insurance pays when an objective, pre-agreed index crosses a trigger (a river level, a rainfall total, a wind speed), regardless of loss assessed after the event. It exists precisely because measuring true loss is slow, expensive and disputable. That is the same measurement problem every outcome contract faces, and parametric practice has the discipline nature finance needs:
Measure an index, not a loss: an index both parties can read off an instrument beats a modelled counterfactual for speed, cost and disputability. For flood resilience: river height at a named gauge under defined rainfall events.
Own the basis risk explicitly: basis risk, ie the gap between what the index says and what actually happened, is the central discipline of parametric design. Every outcome contract should have it, but few name it. The question “does the chosen gauge actually track the resilience delivered at the buyer’s site?” is a basis-risk question, and the contract should always say who bears the residual.
Demand objective, independent, timely data: the index should come from a source neither party controls, a public gauge or a met-office dataset, and hence disputes can resolve quickly. This is an even stronger standard than ‘an independent verifier certifies’: it removes discretion from the measurement itself.
Graduate the trigger: modern parametric products use tiered or linear payout curves rather than binary cliffs, reducing basis risk on both sides. Might we build an outcome contract to scale payment with achievement in much the same way?
The synthesis
Putting the two fields together, I think a design brief emerges for any serious resilience contract. Borrow outcome-based finance’s verification machinery, rate-card trajectory and counterfactual discipline; plus borrow parametric insurance’s index-against-trigger measurement, basis-risk honesty, independent data and graduated payouts. Then add the one thing neither field has yet built: a layered, reusable documentary architecture so the second deal is cheaper than the first and the hundredth is routine.
That last element matters more than it sounds. Outcome-based finance did not fail on ideas; it stalled on unit economics. The instruments worked and the market stayed small. If nature markets are to move from credits (abstracted, decontextualised, fungible) to contracts that match the geometry of the underlying ecological dependency, the contracts have to be cheap to repeat. Standardisation is not the enemy of ecological integrity; it is the only way integrity scales.
One caution with which I will bring this to a close. It is tempting to read the Coral Bond and conclude the destination is a tradable security for everything. But sequencing matters massively - a bond wraps cashflows; the cashflows come first, from contracts that survive contact with an auditor, a treasurer and a decade of weather. Build the contract properly, measured on parametric principles, verified on outcome-finance principles, documented on derivatives-market principles, and the capital-markets wrapper follows naturally. If you prefer, build the wrapper first and we have a label in search of a product.