Post date 23 Feb 2026 (original LinkedIn date)
Original LinkedIn version· 2,104 words
I love what we do at Rebalance Earth, am super passionate about what we as a nascent industry can actually achieve, and yet I keep coming back to a simple, sometimes uncomfortable question. Are we truly on a path to reverse biodiversity loss, or are we just getting better at describing decline and holding it briefly in check?
The question landed with me after reading two pieces back-to-back accidentally: a Science Advances meta-analysis on global biodiversity scenarios (“Can we bend the curve?”), and an LSE Grantham Institute commentary arguing that nature has to be built into macroeconomic models if we’re serious about shifting outcomes at scale.
Read together, they don’t just critique policy. They quietly critique all of us building nature-based solutions, including those of us inside businesses and platforms like Rebalance Earth.
The core message is pretty blunt and inescapable. Bending the biodiversity curve is possible in theory, but rare in the evidence base and brutally demanding in practice. And if we don’t connect project-level wins to the economic machinery that drives land use, food systems, and capital allocation, we’ll keep winning locally while losing globally.
What ‘bending the curve’ really means
The phrase is fairly commonplace these days. It’s in strategies, investor decks, government announcements, conference panels. It’s become a kind of shorthand for ambition.
The Science Advances paper forces a more specific definition. ‘Bending the curve’ is not “slowing loss”. It’s not “netting off harm in one place with gain in another” (even when that’s legitimate within a regulatory framework). It’s not “improving (ecosystem) condition here and there”.
In their framing, bending the curve means global biodiversity loss is reversed relative to present levels. It means the trajectory turns upward, not merely less downward.
That’s why the results can sting. Most scenario work still projects continued decline under current trends. A minority of scenarios even model interventions in a way that can be compared, and only a small handful qualify as true ‘bending’ scenarios.
The point is not that the modelling and developer community is failing. It’s that the scale of the task is so large that even our scenario tools struggle to represent credible reversal pathways. That alone should be a warning against easy narratives.
Language also matters. Over the past year, I’ve noticed a linguistic pivot that should make all of us a bit wary. Nature restoration has suddenly become fashionable to describe as ‘business resilience’ or ‘green infrastructure’, even when it palpably isn’t (we’re as guilty of its popularisation as anyone, but our use is fully considered). A wetland restoration that isn’t designed against future flood hydrology, doesn’t protect a specific downstream asset, and has no operational plan for extreme events is not infrastructure. A biodiversity uplift that improves habitat condition but doesn’t reduce exposure to heat stress, water scarcity, supply disruption, or regulatory risk for a defined set of actors is not resilience.
These framings aren’t just harmless hype. They can be actively misleading because they smuggle in an assumption of performance and durability that hasn’t been demonstrated, and they shift attention away from the hard work of defining baselines, stress-testing under climate scenarios, and putting governance and maintenance in place.
If we want nature to be taken seriously by CFOs, insurers, and the investment committees of asset owners, we have to be stricter with our own language than we might be tempted to be.
The modelling problem is also a governance problem
A second point in the Science Advances analysis feels technical at first, but it matters for every practitioner: biodiversity modelling is fragmented.
There are many different biodiversity indicators in use and limited standardisation. That makes cross-study comparison hard, and it creates a real-world problem too. If we can’t compare, aggregate, and reconcile outcomes across approaches, we risk building a tower of local metrics that never adds up to a credible global story.
Then there’s what I’ll term the ‘pressure-stack’ issue. Many models focus heavily on land-use change, far fewer incorporate climate change, and almost none include invasive species. Here’s the so what of that: the ‘bending’ scenarios, the ones that look most hopeful, often omit climate impacts.
If you’ve ever tried to build a nature investment case, you’ll recognise the temptation. If you don’t model climate, the future looks tidier. Baselines are cleaner. Recovery looks more plausible. Payback periods look safer. It’s easier to tell a story.
But biodiversity does not live in tidy models. Climate acts on hydrology, temperature stress, disturbance regimes, fire risk, pests and pathogens, species distributions, and the resilience of restored ecosystems. If the most optimistic scenarios underweight that, then practitioners have to lean in the opposite direction and over-correct with realism.
So, yes, scenario modelling needs tightening. But the deeper point is that governance needs tightening. If our best ‘bend the curve’ stories are systematically optimistic because of what they leave out, then decision-makers will underinvest in resilience, and projects will underperform against long-run outcomes.
Which brings me to the LSE point: you can’t fix nature from the margins
The LSE Grantham commentary comes at the problem from a different angle and it’s arguably equally troubling. Even if we perfect our ecological metrics and build excellent restoration projects, it won’t matter at scale if the economic system keeps treating nature as external. GDP-centric baselines can appear healthy while liquidating the underlying natural asset base. That creates a structural bias against investment in resilience and a structural bias towards extraction, conversion, and short-term output. Put simply, if nature is not inside macroeconomic models, then a lot of what we call ‘progress’ is just better reporting of decline.
This is where the two pieces lock together, at least in my head! Science Advances is effectively saying: genuine reversal requires integrated strategies across multiple domains. LSE is saying: the only way those integrated strategies happen at scale is if the macro and finance system changes what it rewards, what it penalises, and what it deigns to call ‘growth’.
That sounds like a systems challenge, not a project challenge alone.
A useful honesty check for nature projects
This is the part I’ve been sitting with in relation to Rebalance Earth. Not ‘are our projects good?’ (they are, I’ve checked, double-checked and turned the world upside down on them), but ‘what kind of good are they delivering?’. This is where our rigorous application of Commonland’s wonderful 4 returns framing is really handy.
I think there are two legitimate kinds of effectiveness, and we should be explicit about which one we’re claiming in any given context:
1. Project-level effectiveness (necessary, but not sufficient) means: are we delivering real, additional, durable ecological gains on the ground? Are those gains likely to persist under plausible future climate conditions?
2. System-level effectiveness (the hard bit) means: are we changing the incentive architecture so that business-as-usual becomes irrational? Are we shifting procurement, planning, finance, regulation, or risk pricing in ways that change the baseline?
Both matter. Most organisations do the first, and far fewer do the second. But the second is what ‘bending the curve’ ultimately demands.
Having thought deeply about how to embed this in our work, here’s a practical way I came up with to test where any project or portfolio sits.
Test 1: are we pulling more than one lever?
The bending scenarios in the Science Advances review tend to combine action bundles: expanded conservation and restoration, food system transformation (diet shifts, reduced food waste), sustainable intensification that reduces land pressure, or mitigation choices that don’t simply displace harm into habitat conversion elsewhere. If a project is only doing one lever, it can still be excellent. But it’s more likely to be a ‘reduce-loss’ contributor than a ‘bend-the-curve’ engine.
So ask, plainly: are we only improving habitat supply, or are we also shifting demand and incentives upstream? Are we touching procurement standards, retail supply chains, agricultural inputs, land valuation, or policy instruments that determine conversion pressure?
If not, that’s fine. Just be honest about what we’re achieving. And having both is ok too!
Test 2: is climate reality built into the baseline, not bolted on later?
If you’re restoring a peatland, woodland, river system, saltmarsh, or seagrass meadow, the question isn’t ‘does it work today?’ It’s ‘does it still work in 10, 20, 30 years under likely climate and hydrological change?’
That means forward-looking exposure and adaptation pathways, not as a footnote, but as part of what success means. Species mix, connectivity, drought planning, fire planning, salinity dynamics, sediment regimes, invasive risk management, disease pressure, governance capacity to intervene when conditions change.
If climate realism is missing, your measured gains are at risk of becoming non-durable gains. And if non-durable gains sit inside financial structures, we likely create risk, not resilience.
Test 3: can outcomes be compared and rolled up?
Fragmented indicators don’t just confuse academics; they actually block capital flow. Institutional investors, banks, and regulators need aggregation. They need comparability. They need a small, coherent indicator stack that can be used across sites, asset types, and time.
For a firm like Rebalance Earth, this is a solvable problem and also a strategic advantage. Pick a minimal stack that you will use everywhere, map it to the GBF logic, and resist bespoke reinvention – ‘inspirational return’ in Commonland parlance. A practical pattern is condition, pressure, outcome. Condition tells you what’s there. Pressure tells you what could break it. Outcome tells you what changed in a way that matters. The exact metrics can vary by biome, but the structure ideally should not.
If you can’t roll up outcomes credibly, you’ll struggle to move from project finance to system influence.
Test 4: is there a decision hook that forces uptake?
This is the LSE point in practice. Evidence without a hook becomes a report. A hook changes behaviour. Hooks include planning obligations, procurement contracts, utility price reviews, covenant packages, insurance terms, and public investment appraisals. They can also be softer but still binding in effect, like internal capital allocation rules or risk limits that change how a balance sheet behaves.
If a project produces great evidence but no hook, it can still be worthwhile. But it will remain peripheral to the system that needs to change.
What I think this means for firms like Rebalance Earth, specifically
If Rebalance Earth’s work is primarily about financing the delivery of high-integrity, scalable restoration and building investable natural capital pipelines of both demand (offtakers) and supply (projects), that is valuable. It’s essential groundwork and it’s part of the solution.
But these two articles are a warning against anyone over-claiming. Local projects alone will not bend global biodiversity loss unless they are replicable and connected to the economic drivers and decision loops that determine land use, water quality, input intensity, and climate exposure.
The opportunity, though, is that we can be more than just a project financier. We can be a bridge between ecology and macro constraints, a bridge that has three lanes:
i. Lane one is ecological integrity and durability. If we are more climate-realistic than the modelling literature, we’ll build assets that hold their value and outcomes for longer.
ii. Lane two is indicator discipline. If we standardise an outcome stack that can aggregate across projects, we can speak in the language of capital allocation and genuine business resilience without diluting ecological truth.
iii. Lane three is decision integration. If we design projects around hooks that alter procurement, finance terms, and risk pricing, we are no longer hoping the system changes. We are actually building the mechanisms that make it change.
That’s the difference between ‘good projects’ and ‘system relevance’.
A final thought
‘Bending the curve’ has become a well-used slogan. The Science Advances paper asks whether our modelling efforts meaningfully support the ambition of the Global Biodiversity Framework. The LSE piece asks whether our economic models meaningfully support ecological reality. The combined message for practitioners is sharper: if we want to claim curve-bending, we need to show linguistic honesty, integrated levers, climate realism, indicator discipline, and decision hooks. Otherwise, we should claim what we’re actually doing, which might be just as important, but different.
And maybe that’s the most useful reframing for the next phase of nature finance. Less rhetoric, more honest taxonomy of action and impact. If we do that, we can still be ambitious. We’ll just be ambitious in a way that survives contact with the real world.
References
‘Can we bend the curve: trends in global biodiversity scenarios’, Science Advances (2026) https://www.science.org/doi/epdf/10.1126/sciadv.aeb2277
LSE Grantham Research Institute on Climate Change and the Environment, ‘Time to align economic practice with ecological reality: the critical need to include nature in macroeconomic models’, (2026) https://www.lse.ac.uk/granthaminstitute/news/time-to-align-economic-practice-with-ecological-reality-the-critical-need-to-include-nature-in-macroeconomic-models/
Commonland’s 4 returns framework: https://commonland.com/4-returns-framework/