---
title: "no single funder, no single supplier"
canonical_url: https://ensurance.app/guide/no-single-funder-no-single-supplier
markdown_url: https://ensurance.app/guide/no-single-funder-no-single-supplier.md
subtitle: "why a book shaped like a watershed can fail more gently than a pipe, and where that claim stops"
category: nature-finance
---

# no single funder, no single supplier

*why a book shaped like a watershed can fail more gently than a pipe, and where that claim stops*

A **resilient financial system** is one that can take a shock (a bank failure, a frozen market) and keep paying, lending, and insuring, without the shock spreading to everything connected to it. That is the usual regulatory framing: not a system where nothing fails, but one where a failure stays local.

After 2008, finance went to ecology for an answer. In 2011 Andrew Haldane, then of the Bank of England, and the ecologist Robert May argued in *Nature* that the banking network before the crisis had become more homogeneous: banks pursued similar strategies and held similar assets, and that sameness made the whole fragile. They also showed that connection cuts both ways, spreading a loss thinner at first and carrying a shock further later. They pointed to modularity and diversity as sources of stability, and to the large, highly connected nodes a failure spreads through.

Take the Crystal River above Marble, Colorado. The headwater forest, the beaver meadow below it, the aquifer, the trout, and the towns downstream at Carbondale and Glenwood Springs already depend on each other, in both directions, whether or not anyone opens an account. Value moves along every one of those relations, and almost none of it is priced. What **[ensurance](https://ensurance.app/?from=guide)** adds is an account for each of them (what a place, a people, or a purpose holds in its own name), so a share of that value can be carried as money. The [Crystal River account](/crystal-river.basin?from=guide) is not the river, and the web of dependence is not a market until people act on it.

## one pipe, one failure

A town can draw water from a pipe off one reservoir, one failure from dry, or from the watershed it sits in: snow melting off different slopes at different times, a meadow holding water into late summer, an aquifer, springs. The whole keeps the river running.

A pipe has one way to fail. A watershed has a thousand small ways not to.

Much money for living places is shaped like the pipe. Here, the book means the accounts in this system and what each holds. Water funds, endowments, and mutual ditch companies already pool many payors, and this builds on them. Where the money is one-ended, it has two single places where it breaks.

## the two single points of failure

**One funder.** A project paid by one grant stops when the grant does, not because the work was bad but because nothing else held it up. [Why conservation projects fail on their own](/guide/why-conservation-projects-fail-on-their-own?from=guide) walks through that cliff.

**One supplier.** A built asset stands on one ecosystem it has never contracted with: an intake below one forest, a ski resort under one snowpack. [Your place depends on a place](/guide/your-place-depends-on-a-place?from=guide) names that exposure.

Read any account two ways and both gaps close: by what it relies on, and by what relies on it. Each reading has three layers, in this order. **First, relation:** the forest is tied to the snowpack, the meadow, and the people who keep the ditch; the trout, the intake, and the ranches below are tied to the forest. These connections exist whether or not anyone pays for them. **Second, value:** what moves along each relation, both ways: water, shade, held sediment, pollination; care, work, knowledge, and meaning, the [relational value](/guide/what-is-relational-value?from=guide) between people and place; and harm too, in smoke and runoff on the same edges. Most of it is never priced, and a price is a bridge to it, never its worth. **Third, money:** the share of that value an account can carry. It buys the certificates and coins of the places and purposes on its first list. A certificate is a share in one named place; a coin is a currency for a purpose wider than one place. The dependents on its second list hold its certificate and its coins.

So "no single funder, no single supplier" is about relations and value first and holdings third. List two names the many funders who could stand where one stood: the dependents who already receive the node's value. List one names the many sources who could stand where one unpriced supplier stood. Whether anyone holds is up to them. [Every account has two lists](/guide/every-account-has-two-lists?from=guide) sets out the idea.

| single point of failure | what it looks like | what the two lists put in its place |
|---|---|---|
| One funder | One grant, one program, one end date | List two: the dependents downstream, downwind, and along the corridor |
| One supplier | One forest, one aquifer, one snowpack, unpriced and unheld | List one: the sources the node depends on, named and held |
| One kind of value | A single output counted for a single payor | Water, shade, soil, pollination, and care, most of it unpriced |

## the ecologist's objection, answered first

A field ecologist will ask: does more connection make a system steadier? Robert May tested it in 1972, and for systems assembled at random the answer was no. In a two-page *Nature* paper he showed that such systems become less stable as they grow in size, in connectance (the share of possible links that exist), and in the strength of each interaction. Complexity does not by itself produce stability. The structure of the links does.

So the claim is not "more links, more safety." A web wired at random is the case May analysed; a network where everyone copies one trade is the case Haldane and May described. The links this book aims for are the landscape's relations, drawn by water, weather, corridor, and soil. That does not prove the structure is stable. The aim is to add no link the landscape did not already make. Nothing in the protocol enforces that today, since an account can hold any certificate or coin, so it is a test to hold the book to, not a property it has. The value on those links is varied long before any of it is money.

David Tilman, Peter Reich, and Johannes Knops ran a decade-long grassland experiment at Cedar Creek in Minnesota. Plots with more plant species had more stable total production over time, while the individual species in them were less stable. The whole steadied while the parts fluctuated. These measure different things. May asks whether a system returns to equilibrium after a small push; Tilman measures how steady the total stays from year to year while the species change.

C. S. Holling gave the word its ecological meaning in 1973. **Resilience** is the amount of disturbance a system can absorb and still persist, as distinct from stability, the return to equilibrium. That is the test for a book of many small links: not that no account has a bad year, but that a bad year at one does not end the rest.

:::johnson
**resilience is a shape, not a count.** A book earns it the way a watershed does: many funders, many sources, links that follow real relations, and the weak nodes kept held through their bad years.

[write your own two lists →](/guide/how-to-write-your-two-lists?from=guide)
:::

## the forest everyone depends on

Large, highly connected nodes concentrate risk as well as service. In a headwater basin the forest is that node: the trout, the intake, the ranches, and the rivers below all sit on its second list. Some acres hold up more than their own ground ([what is a keystone parcel](/guide/what-is-a-keystone-parcel?from=guide)).

The two lists do not remove that concentration; they make it visible. That is an argument for many dependents holding the forest's certificate, rather than leaving it to one funder or to none. And supporting what the keystone depends on is part of supporting the keystone.

## where the claim stops

A basin-wide drought hits every node in the basin; they share one sky and much the same snowpack. A book of accounts all on the Crystal River and all tied by water is a concentrated position carrying many names.

Resilience here comes from structure, not independence: from link types that fail in different ways: water, fire, corridor, and soil. They are not independent: drought feeds fire, and fire sends sediment into water, so a dry year can reach several link types in sequence. Spreading across them adds ways to absorb a shock, not a guarantee that they fail separately. Each has a purpose account addressed to it ([water cycle](/water-cycle.syndicate?from=guide), [wildfire resilience](/wildfire-resilience.syndicate?from=guide)). From regions that need not share the same bad year: a dry year on the [Colorado](/colorado-river.basin?from=guide) does not have to be a dry year on the [Arno](/arno-river.basin?from=guide) in Tuscany, though continental drought years can coincide. And from keeping weak nodes held through bad years rather than dropping them. The same limit for networks of projects is set out in [nature already runs a portfolio](/guide/nature-already-runs-a-portfolio?from=guide). None of that is a forecast or a return claim.

The finance side is in [the fine print of the free lunch](/guide/the-fine-print-of-the-free-lunch?from=guide), [the correlation outside the matrix](/guide/the-correlation-outside-the-matrix?from=guide), and [what systemic risk actually is](/guide/what-systemic-risk-actually-is?from=guide).

## the bet

The series makes one bet: a financial system shaped like nature, diverse and resilient because its structure is nature's structure. [Pay the land, pay the people](/guide/pay-the-land-pay-the-people?from=guide) makes the same bet in the language of credits.

What exists today is the structure. An account can hold another account's certificate and coins, and its page shows what it holds. Proceeds can send a set share onward; the operator sets the split, and anyone can distribute it. A person confirms each action unless a program is switched on.

The trout is not a position. The account is addressed to the living thing; the people who keep the place still keep it, and the account is not title or standing ([who pays for nature](/guide/who-pays-for-nature?from=guide)). Money shaped like a watershed should fail the way one does: in small places, with the rest still standing.

## frequently asked questions

### what is a resilient financial system?

A resilient financial system takes a shock and keeps paying, lending, and insuring without the shock spreading to everything connected to it. Haldane and May (2011) argued from ecology that homogeneity made pre-2008 banking fragile, and pointed to modularity and diversity as sources of stability.

### does diversity alone make a system stable?

No. May (1972) showed that large systems wired at random grow less stable as size, connectance, and interaction strength rise. The claim here is about structure, not count: the aim of links drawn only where a real relation already exists.

### what happens to a book like this in a basin-wide drought?

Every node in the basin takes the hit together; the book is not independent of the weather. Spread across link types and regions, with weak nodes kept held, a book can limit how far the loss travels through it. It cannot limit the drought.

## sources

[Haldane, A. G. & May, R. M., "Systemic risk in banking ecosystems," *Nature* 469, 351–355 (2011)](https://doi.org/10.1038/nature09659) — drawing on ecology, the pre-2008 banking network had become more homogeneous, which reduced systemic resilience; connection cuts both ways, spreading a loss thinner at first and carrying a shock further later; modularity and diversity as sources of stability; the role of large, highly connected nodes.

[May, R. M., "Will a large complex system be stable?" *Nature* 238, 413–414 (1972)](https://doi.org/10.1038/238413a0) — randomly assembled large systems become less stable as size, connectance, and interaction strength grow; complexity does not by itself produce stability.

[Tilman, D., Reich, P. B. & Knops, J. M. H., "Biodiversity and ecosystem stability in a decade-long grassland experiment," *Nature* 441, 629–632 (2006)](https://doi.org/10.1038/nature04742) — at Cedar Creek, plots with more plant species had more stable total production over time, while individual species were less stable.

[Holling, C. S., "Resilience and stability of ecological systems," *Annual Review of Ecology and Systematics* 4, 1–23 (1973)](https://doi.org/10.1146/annurev.es.04.110173.000245) — resilience as the amount of disturbance a system can absorb and still persist, distinct from stability as return to equilibrium.

## the series

1. [every account has two lists](/guide/every-account-has-two-lists?from=guide) (read first)
2. [what you depend on](/guide/what-you-depend-on?from=guide)
3. [who depends on you](/guide/who-depends-on-you?from=guide)
4. [a bid on both ends of every edge](/guide/a-bid-on-both-ends-of-every-edge?from=guide)
5. [no single funder, no single supplier](/guide/no-single-funder-no-single-supplier?from=guide)
6. [how to write your two lists](/guide/how-to-write-your-two-lists?from=guide)