Post date 20 Apr 2026 (original LinkedIn date)
Original LinkedIn version· 1,764 words
“Learning to Fly“ by Tom Petty & the Heartbreakers, released in 1991 on the album Into the Great Wide Open
Risk managers have a useful distinction that many investors never encounter. They differentiate risk, an uncertainty you can quantify, because you can assign a probability to it, from Knightian uncertainty, a risk you can’t assign a probability to, because the situation that underpins it is changing in ways that haven’t been seen before. With a week off work to recharge the batteries, while wandering up Pen y Fan, visiting Lundy Island to see the pufflings, and kayaking the beautiful coasts of Devon and Cornwall, I’ve been thinking about this distinction a lot this week, after reading three very different pieces that, taken together, make a case I think every allocator should sit with.
The first is a Project Syndicate column by Morten Nyboe Tabor, the Co-Founder and Director of the INET Centre on Knightian Uncertainty. His argument sounds straightforward: much of the forecasting profession still defaults to single baselines that treat the future as a probabilistic replica of the past. When the structure of the economy changes in unforeseeable ways, as it is doing now, that approach goes beyond imprecise. It is simply wrong. Tabor points out that more and more central banks now acknowledge this. They’re moving from single-point forecasts to scenarios, recognising that different futures are possible and that the job of the forecaster isn’t to predict which one will happen, but to map out what could happen and assess how exposed you are to each.
I think that shift matters enormously for investors. We’ve been trained on the language of expected returns, central estimates, base cases. But in a world defined by structural breaks, whether geopolitical, technological, ecological, the base case is precisely the thing that’s least likely to materialise. The question to ask therefore isn’t ‘what do we expect?’, but rather ‘what are the scenarios, and how badly do we get hurt in each?’
The second piece comes from the Institute and Faculty of Actuaries. Their Biodiversity Scenarios Working Party published a few months back what I think is one of the most important papers the actuarial profession has produced in recent years: Nature at Risk, Models at Fault: Why Biodiversity Can’t Wait. Their headline conclusion is stark: there is no use-case for climate scenarios that do not explicitly integrate biodiversity. The current NGFS, IEA, and similar frameworks that underpin most regulatory and disclosure efforts have a structural blind spot. They model climate in isolation, and they don’t capture the knock-on effects of ecosystem collapse: food system disruption, land use tensions, pandemic risk, forced migration. And because they don’t, they systematically underestimate both chronic and acute risks.
The IFoA paper identifies three realities that make this integration non-negotiable. Biodiversity risk is equivalent in scale to climate risk. The two are deeply interlinked and mutually reinforcing - biodiversity loss weakens natural climate regulation and can turn carbon sinks into carbon sources. And both have fundamental macroeconomic consequences. Their assessment is that there is a material risk of biodiversity and nature-related macroeconomic impacts equalling the 2008 Global Financial Crisis, or even COVID-19, within the next ten years.
So as an asset owner, if you’re running scenarios that don’t include biodiversity, you’re not running scenarios. You’re running a subset of scenarios that excludes some of the largest and most interconnected risks.
The third piece is from the UK Joint Intelligence Committee’s national security assessment on global biodiversity loss and ecosystem collapse. This is the document that was presented at COBRA (the cabinet committee convened for the most serious national threats) and then quietly buried until a Freedom of Information request forced its partial publication earlier this year. Worth reading that sentence again! Ecosystem collapse was briefed at COBRA … and then suppressed.
The JIC assessment identifies six ecosystems of strategic importance to the UK and they’re not all immediately obvious by any stretch: the Amazon, the Congo Basin, boreal forests, the Himalayas, and Southeast Asia’s coral reefs and mangroves. It concludes that severe degradation or collapse of these systems would be highly likely to result in water insecurity, severely reduced crop yields, fisheries collapse, changes to global weather patterns, novel zoonotic diseases, and the release of trapped carbon that would accelerate climate change further. Without significant increases in food system resilience, the UK would be unlikely to maintain food security if ecosystem collapse drives geopolitical competition for food.
What makes this document worth rereading carefully is less its conclusions than its methodology. It uses the uncertainty frameworks of intelligence assessments, ie reasonable worst-case scenarios, confidence levels, scenario pathways, rather than the single-baseline forecasting that Tabor criticises. Intelligence analysts have long understood what economists are only now beginning to accept: that when you’re dealing with genuine uncertainty, you work with scenarios, not point estimates.
Now add a fourth piece. A couple of weeks’ ago, Absolute Strategy Research published Liquid Assets: Investing in Water, a report that takes the abstract arguments about risk, chain-reaction effects, and Knightian uncertainty and gives them a very concrete, investable form.
The headline numbers are deeply arresting. Global water availability per person has fallen by almost two thirds since 1960. By 2030, water demand is projected to exceed supply by 40%. Even advanced economies like the US, Australia and Canada have seen their water availability cut in half. Within five years, most of Southern Europe (eg Italy, Spain, Greece) is projected to be in extreme water stress, where demand exceeds 80% of available renewable supply. California, Texas, Arizona, New Mexico are all sadly heading the same way.
But I think the more interesting question for investors isn’t the headline scarcity data. It’s the transmission channels, the ways water stress propagates through systems in exactly the chain-reaction fashion that Tabor, the IFoA, and the JIC are all warning about. Water is the connective tissue of the Water-Food-Energy Nexus. Over 70% of global freshwater goes to agricultural irrigation. Shocks to water supply spill immediately into food production, then into food prices, then into social stability. The JIC report’s central concern, that the UK can’t maintain food security if ecosystem collapse drives geopolitical competition for food, runs directly through water.
But it doesn’t stop at food. Water-related conflicts have surged to ten times their historical average. The US has already threatened Mexico with additional tariffs over unpaid ‘water debts’ from the Rio Grande basin. Water scarcity is becoming a geopolitical weapon. And there’s a new dimension that few investors have yet internalised. Datacentre cooling and power generation can be enormously water-intensive. An average 100MW datacentre uses two million litres of water per day, the equivalent of 6,500 households. Yet 58% of US datacentres sit in areas of high or extremely high water stress. In China it’s 82%, in Australia, 88% and in Spain, 87%. The AI investment thesis and the water crisis are on a collision course, and almost nobody is modelling it.
ASR identifies a $7.5 trillion global water infrastructure financing gap between now and 2040, or roughly $500 billion per year. That’s the investment required to build resilience against the scenarios the other three pieces describe.
What ties all of this together, and the reason I think it matters for anyone making long-term capital allocation decisions, is this: when you’re dealing with risks that are coupled and self-reinforcing, where ecological tipping points trigger food insecurity, which triggers geopolitical competition, which triggers supply chain disruption, which feeds back into economic instability, a single-baseline forecast doesn’t just lose precision. It assumes a stability that doesn’t exist. In essence it gives you false comfort.
Water is, in many ways, the canary in the mine for the kind of systemic, interconnected, non-linear risks that all four of these pieces are warning about. It is a commodity with a fixed stock that cannot be substituted or transported, where higher prices do not generate new supply, and where the normal equilibrating mechanisms of market economics simply do not apply. If ever there was a domain where Knightian uncertainty, rather than quantifiable risk, is the right frame, this is surely it?
Let me draw this to a close with what I think are the three practical implications for investors.
First, if your climate scenario analysis doesn’t integrate biodiversity and water, it is structurally incomplete. The IFoA has said this plainly. The tools exist, and yes the data is imperfect but it is also improving. There is no longer a credible excuse for climate-only modelling.
Second, water infrastructure and nature-based solutions are not ‘ESG themes’ sitting in a silo. Rather they are systemic resilience investments. They truncate the left tail of the cascading risks that every portfolio is exposed to (read last week’s thought piece for more on that), through food prices, energy costs, supply chains, and through the companies you hold. This is the argument I’ve been making about nature as portfolio insurance: it doesn’t just offer returns, it protects compounding by reducing the probability and severity of the worst outcomes.
Third, the investment opportunity is real and growing. Water-thematic indices have delivered strong, risk-adjusted performance over the past decade, comfortably outperforming many other climate strategies. As policymakers begin to treat water security as a matter of national security, which is exactly what the JIC report signals, capital will flow towards solutions: efficiency, recycling, desalination, monitoring, sensors, data analytics. The $7.5 trillion gap doesn’t need to be a problem - it is a market.
Frank Knight published Risk, Uncertainty and Profit in 1921. A century later, his distinction between risk and uncertainty has never been more relevant. The world is undergoing seismic structural change. The past is not a reliable guide to the future – everyone writes that on their investment material but it’s never been more true. And the investors who understand this, who think in scenarios rather than forecasts, who model cascading systemic risks rather than isolated variables, who invest in the resilience of the natural systems that underpin everything else, they will be the ones whose compounding survives what comes next.
References
Tabor, M.N. (2026) ‘Is Economic Forecasting Still Possible?’, Tabor, M.N. (2026) ‘Is Economic Forecasting Still Possible?’, Project Syndicate, 6 April. Available at: https://www.project-syndicate.org/commentary/economic-forecasting-under-radical-uncertainty-by-morten-nyboe-tabor-2026-04
IFoA Biodiversity Scenarios Working Party (2026) ‘Nature at Risk, Models at Fault: Why Biodiversity Can’t Wait’, Institute and Faculty of Actuaries. Available at: https://actuaries.org.uk/events/sessional-meeting-climate-models-have-a-fatal-flaw-actuarial-risk-management-does-not-have-to/
Joint Intelligence Committee/DEFRA (2026) ‘National Security Assessment: Global Biodiversity Loss, Ecosystem Collapse and National Security’, HM Government. Available at: https://www.gov.uk/government/publications/nature-security-assessment-on-global-biodiversity-loss-ecosystem-collapse-and-national-security
Penn, M. (2026) ‘Liquid Assets: Investing in Water’, Absolute Strategy Research, 10 April. Available at:
https://www.absolute-strategy.com