The 19% problem

”I am a shadow, I am cold And now I seek for warmth Stitch your skin on to my skin And we won't be alone

Don't forget who you are even though you're in need Like a bird in the night, your emotions deserve to be freed

You can cry Drinking your eyes Do you miss the sadness when it's gone? (Gone) And you let the river run wild (gone) And you let the river run wild

You can cry (you can cry, you can cry) Drinking your eyes I don't miss the sadness when it's gone (gone) And the feeling of it makes me smile (gone) As I let the river run wild”

"The River" by Aurora, released in 2019 on the album A different kind of human.

I put Goldman Sachs Asset Management's April note on water to one side as something that I should go back and read at some point - but as ever, stuff got in the way. But here we are, and I think that maybe it is one of the cleanest physical-risk analyses a major bank has put out ever (and I slightly regret not diving into it properly 2 months or so ago!). It is also, on a closer read, a fairly clean illustration of why the paradigm that sustainable finance is currently operating in remains a constraint, not the solution.

The headline number is 19% of revenue across the MSCI ACWI IMI sitting on companies with high or very high water dependency. The $13.2 trillion infrastructure gap to 2040 is in there too, as is the 30% network-leakage figure. Agriculture as 70% of freshwater withdrawals, three quarters of the world living in water-insecure countries, the AI-driven water demand story - in fact, it's all there! The diagnosis is deliciously calibrated, the case studies are all too real, and their framing of 'no longer whether but how' seems to me to capture something true about where institutional capital has actually moved on water in the last two years.

I think the report should be read as a milestone by everyone in our nascent natural capital industry. It shows that water has crossed into mainstream thematic allocation territory just the way climate did between 2016 and 2018, following the Paris Agreement. To some that will represent genuine progress, but to others it is also a near-perfect statement of the operating limits of the lens we are currently using. It is this inconsistency that struck me as I read it.

Goldman defines water dependency as reliance on four ecosystem services. Flow regulation, supply, purification, and flood control are the named services. The 19% is the share of revenue that is dependent on them. And then every one of those services, in the analysis the bank runs and on the balance sheets the bank cites, is sadly valued at zero. A failure to recognise the glaring truth, that wetlands regulate flow, that watersheds and aquifers supply, that soils and forests purify, and that coastal wetlands and floodplains control floods. None of them appear in the $13.2 trillion that needs to be spent, and none of them appear as infrastructure on any balance sheet referenced. The largest, oldest, most widely distributed utility on the planet sits at zero in every accounting frame the report cites. The framework being used to flag the risk treats the assets producing the service as a costless externality. I thought we'd moved beyond that.

That inconsistency is not just a quirk of one report tho, rather it seems to be the operating condition of how the financial system currently sees nature. A useful piece of work has just been published in Cell Reports Sustainability by Paul Dingkuhn, Jeanne Nel, Dirk Schoenmaker and Francisco Alpizar that puts a name to this and embeds it in a wider argument. They call it the "external" view of finance, and they identify it as one of three paradigms blocking what they describe as a nature-positive financial system. The other two are the dominance of "market-fixing" over "market-shaping" approaches to financial governance, and the persistence of shareholder-primacy logics in corporate decision-making. Each paradigm comes with suggested concrete interventions across macro, meso and micro scales. The paper is open access and well worth the read in full (see below).

Let's apply that same lens to the Goldman piece. The whole report sits in the market-fixing register ie identify risks, disclose them, surface investment opportunities in pick-and-shovel companies, and then trust capital to allocate appropriately. Sadly there is no taxonomy of harmful water exposures, no discussion of constraint, only of opportunity. Agricultural lending into aquifers in terminal decline, data-centre buildouts in regions the report itself flags as water-stressed, growth in water-intensive sectors of emerging markets - all of these clear challenges stay inside the 'opportunity set' rather than crossing into the 'constraint set'. Peter Newell's recent paper on transformative climate finance makes the broader version of this argument (and is also well worth a 10-min read): sustainable finance is overwhelmingly about incentivising the green and rarely about restricting the brown. Until both halves are in the room, the financial system absorbs each new theme without changing its underlying logic. We all know the ratio of nature-harmful subsidies to nature-positive efforts: a shocking 30:1.

I think that there is an even deeper second-order problem, too, which a recent Indoeden Substack piece raised in response to the Goldman report. If you take the water-stress map seriously, then a meaningful portion of the biological assets currently on balance sheets (orchards, vineyards, palm plantations, managed forests, agricultural land mortgaged against declining aquifers) is being valued on yield curves calibrated to hydrological conditions that no longer hold. That is not a future risk-pricing problem, it is instead a present mispricing problem. IAS 41 assumptions across multiple crop systems would need real revision before quarterly earnings start reflecting that change appropriately. Goldman's framing makes this invisible because the lens points at the solution providers, not at the incumbent biological asset values. The 19% is what is visible; the real mispriced biological asset stack underneath it is what is not.

It is worth meeting the obvious objection head-on, because it is already circulating. A recent piece by Joseph Fournier, built on six decades of data, argues that the El Niño drying narrative is overblown, that global tropospheric humidity at 700 mb rises, not falls, during El Niño, and that fears of a drier planet are unsupported by the aggregate moisture budget. On the narrow physics, that may be true at some aggregate theoretical level: a warmer global atmosphere holds more water vapour, and El Niño years run warm. But I don't think that it changes anything here, and seeing why is critcal. ENSO was never a story about how much moisture sits in the global column. Rather it is a story about where that moisture falls as rain. The warm pool migrates east, convection follows, and the Maritime Continent, eastern Australia, the Amazon, southern Africa and the monsoon belt dry hard while the planetary average ticks up. However a balance sheet is not exposed to the planetary average. Instead it is exposed to the rainfall over one orchard, the recharge of one aquifer, and/or the flow in one river basin. The global mean is the exact resolution at which water stress vanishes from view and the exact resolution at which a financial system trained to take comfort in aggregates keeps mispricing the assets beneath it. That 'the planet is getting wetter' and 'this aquifer is in terminal decline' can both be true at once is not a paradox. It is the exact statement of our mispricing engine, stated in a single line.

Where things get more interesting, in my view, is at the edges of the discourse where people are actually building institutional alternatives. Take the Sustainable Investor podcast with my boss, Robert Gardner of Rebalance Earth, or the Global Infrastructure Hub work on private-sector roles in nature-related infrastructure, or the growing body of evidence that wetland and forest restoration deliver benefit-cost ratios in the range of six to fifty times across most settings. All of it is in service of a different operating reality: treat nature as performing infrastructure with cash flows attached to the upstream ecosystem service, not only the downstream engineered solution. That then flows thru to the idea that we should have corporate off-takers pay for outcomes (cleaner water, restored reefs, sponge landscapes upstream of railway lines) rather than for just sustainable risk reports. So why not allocate two to five per cent of institutional portfolios to natural capital as a hedge against the physical climate exposure on the rest of the book? I've shown in previous articles what that does to the long-term geometric mean that governs the growth in wealth.

That direction is closer to what Dingkuhn and colleagues would call market-shaping. It is also the only place I see the four ecosystem services Goldman identifies actually being repriced from zero to something positive. Whether the institutional plumbing can scale fast enough remains an open question, and the governance issues around who ends up owning the repriced assets are real and as yet unresolved, but the direction of travel has to be right.

My honest read of the stack of pieces I have been working through this month is this: Goldmans is genuine progress and a clean statement of the paradigm's limits at the same time, and the commentary that pointed at the unpriced ecosystem services is exactly the right critique. Equally the Substack piece that pointed at biological-asset mispricing is the right critique that we must take heed of, and the contrarian read that the planet is getting wetter is, at the global mean, probably right but beside the point, because nothing on a balance sheet is priced at the global mean. The restoration-as-infrastructure work is where the alternative is actually being built and I absolutely love being part of it.

Two questions I would put to anyone involved with an investment committee at an asset owner reflecting on this: One: where is the brown side in your water exposure mapping? If your framework can flag opportunity in pick-and-shovel water tech but not exposure in lending to permanent water-insecurity, you are only playing one half of the game. Two: where on your balance sheet does nature appear as infrastructure rather than as scenery? If the answer is nowhere, the $13.2 trillion is not a gap, instead it is the bill for a service the planet has been providing free of charge for the entire history of commerce, finally landing on the desk.

References

Dingkuhn, Nel, Schoenmaker and Alpizar, "Pathways toward a nature-positive financial system," Cell Reports Sustainability (March 2026)

Goldman Sachs Asset Management, "Finding Investment Opportunities in the Global Response to Water Stress" (April 2026)

Indoeden, "The Unpriced Biological Collapse Beneath the Goldman Sachs Water Report"

Fournier, "El Niño Does Not Dry the Planet: Global Humidity Rises with Warming Tropics," Substack (June 2026)

Sustainable Investor, "Restoration: An Investment in Resilience" (November 2025)

Global Infrastructure Hub, "Extreme Weather and Nature Loss: Can the Private Sector Be Part of the Infrastructure Solution" (2026)

Newell, "Towards a more transformative approach to climate finance," Climate Policy (2024)