Haute Lumière

Commerce · III.03 · MMXXVI · daylight

La Bourse  /  Volume III  /  Nº III.03  /  Ten concept briefs

TEN CONCEPT BRIEFS · Chapter III.03 — Complementary and Local Currencies

One page each. A reader who reads only these ten pages has the chapter.

---

BRIEF 1 — The Cost Identity

The idea. A currency has a running cost, and the only figure that decides whether it survives is that cost divided by the money it moves.

        c  =  C / T  =  C / (M · V)

  C = annual operating cost      M = stock outstanding
  V = velocity (turnover per unit of stock, per year)
  T = annual turnover = M · V    c = cost per unit of turnover

c is the transaction fee that would exactly break even. Compare it to the fee a member already pays for card acquiring, and you have your answer in one line.

Worked example. The Bristol Pound cost £150,000 a year to run — £1.2 million of grant funding over eight years — and moved about £952,381 a year. c = 15.75 percent. No member would pay fifteen percent to settle a transaction, and no grant funder funds anything forever. The scheme closed after nine years.

Why it matters. It is computable from figures schemes already publish plus one they usually do not, and it turns "is this viable?" from a matter of belief into a division you can do on a train.

You already know this because you have never once wondered whether Visa would still exist next year, and the reason is that its cost per unit of turnover is smaller than what it charges — which is the same test, applied to a bigger circuit.

---

BRIEF 2 — Velocity, Measured

The idea. Velocity is turnover divided by outstanding stock, per year. It is not a metaphor and it is not a mood. It is a division.

        V  =  T / M

The numbers, from the record.

The reading. Sardex's unit works 9.4 times harder than Bristol's. Since c = C/(M·V), every unit of velocity you fail to get multiplies your cost ratio directly. A currency at V = 1 is a deposit with a picture on it.

What drives it. Whether the holders are businesses (who buy from many and sell to many) or consumers (who buy from a short list and sell to nobody). Consumer schemes accumulate credits at the retail end with nowhere to go. That is Bristol's 1.36 in one sentence.

You already know this because you have watched a gift card sit in a drawer for two years, and you know without being told that the shop's money was doing something more useful in the meantime.

---

BRIEF 3 — Mutual Credit

The idea. Issue the unit at the moment of a trade and destroy it at the moment of the matching trade. Stock becomes endogenous; nobody has to buy it in.

In a mutual credit circuit every member starts at zero. A sale moves the seller positive and the buyer negative by the same amount. The sum of all balances is always exactly zero. There is no issuer, no reserve, no float and nothing to speculate on.

Worked example. Sardex, 2015: €4 million of credits in circulation against €80 million of goods and services members had undertaken to supply — 20× backing. Each member's credit line was about one percent of their annual turnover; each undertook to accept up to a tenth of it in credits.

Why it beats a voucher. A voucher currency's stock is whatever people happened to buy and its velocity is whatever they happened to spend. A mutual credit circuit's stock appears exactly when trade wants it. That structural difference is worth a factor of two to nine in velocity, and velocity is the whole cost equation.

You already know this because you have offset an invoice against a bill from the same counterparty rather than sending two payments. That is mutual credit with two members.

---

BRIEF 4 — Demurrage, and Why It Is a Fee

The idea. A carrying charge on holding money rather than on spending it. Silvio Gesell proposed it; Keynes devoted part of a chapter to it; Wörgl and the Chiemgauer are the two best-run field tests.

The identity that makes it legible.

        d*  =  c · V  =  C / M

A demurrage and a transaction fee are the same instrument priced off different bases, and velocity is the exchange rate between them. A scheme at V = 7 needs seven times the carrying charge to raise what a one-percent transaction fee would raise.

The settings, from the record. Wörgl charged 1 percent a month — 12.68 percent a year compounded. The Chiemgauer charges 3 percent every six months on paper (6.09 percent a year) and 0.02 percent a day on the electronic form after thirty days, which is 6.48 percent a year.

A caution, caught on re-reading. Where the demurrage is the financing, as at Wörgl, d* = C/M returns the demurrage rate by construction, not by discovery. The informative figure there is c, which was 4.32 percent.

You already know this because you understand why a supermarket discounts bread at six in the evening. Perishability is demurrage in a different aisle.

---

BRIEF 5 — Countercyclicality

The idea. A parallel circuit is used more when the main one contracts — which is precisely when it is worth most.

The evidence. James Stodder, using Swiss data back to 1948, found WIR turnover to be countercyclical, and more reliably so than the official money supply. Stodder and Lietaer then separated the mechanism: large firms supply the swing through their balances, small firms through their velocity.

Why it happens. SMEs lose bank credit first in a downturn and rely most on supplier credit. A closed circuit is supplier credit that settles finally and nets multilaterally, so it fills in behind the retreating banks for exactly the firms the banks retreat from.

The scale, honestly. WIR balances in 2007 were CHF 612 million — a quarter of one percent of Swiss M1. Stodder and Lietaer argue the leverage on that is high, but the direct share is small, and a chapter that claimed WIR steadies the Swiss economy on its own would be overstating a good result.

You already know this because when cash is tight you pay the supplier who will still deliver next month before the one who will not, and that reordering is a credit system operating.

---

BRIEF 6 — Breakage, Float and Where Schemes Actually Earn

The idea. A large part of a voucher currency's income comes from money that is never spent.

The three terms, distinguished. Seigniorage is the gap between a note's face value and its production cost. Breakage is vouchers sold and never redeemed, where the seller keeps the proceeds. Escheatment is abandoned property passing to the state. Bristol Pound staff commonly called their income seigniorage; Petz and Finch record that the majority of sterling income into the scheme was in fact breakage.

The extreme case. In Japan's fureai kippu, NALC's 2010 returns show 198,091 time credits earned and 10,548 redeemed — a redemption rate of 5.32 percent. 94.68 percent were never spent. The scheme is cheap to run and thirty-one years old.

The uncomfortable reading. Income from unspent money is income from the currency failing to circulate. It is the one revenue line that grows as V falls, which means it pays best exactly when the scheme is working least. Know which part of your income this is before you celebrate it.

You already know this because you have a gift voucher in a drawer that expired, and you know who kept that money.

---

BRIEF 7 — The Half-Life, Computed

The idea. You cannot measure how long schemes last by averaging the ones that ended, because that throws away every scheme that contradicts you.

The method. Kaplan and Meier's 1958 estimator treats a scheme still running as censored — known to have lasted at least this long — rather than as missing data.

The cohort. All eight sterling local currencies of the transition wave, 2007–2018: Totnes, Lewes, Brixton, Stroud, Bristol, Kingston, Exeter, Lake District. Six closed, two running.

   t     at risk   closures    S(t)
   2         8          1     0.8750
   3         7          1     0.7500
   7         6          1     0.6250
   9         5          1     0.5000   <- median

The answer: a half-life of nine years, an implied annual hazard of ln2/9 = 7.70 percent, about one closure in thirteen scheme-years. The naive median of the six closures alone is 8.0 years and is biased downward.

The finding that changes the conclusion. Run the same estimator over sixteen schemes including WIR, Sardex, Chiemgauer, BerkShares, Banco Palmas and fureai kippu: the survivor function bottoms at 0.5469 and never reaches one half. The median is not reached. Nine years is the half-life of the retail town pound, not of complementary currency.

You already know this because you would not estimate how long marriages last by surveying only divorce courts.

---

BRIEF 8 — Scope Is the Product

The idea. The boundary of who may join is not a limitation on the currency. It is the thing that makes acceptance predictable, which is the only property a currency has.

How the durable ones drew it. WIR: Swiss SMEs only, by by-law since 1972 — 60,703 client enterprises in 2005, 16.3 percent of Swiss firms in its sectors and 37.0 percent of construction firms. Sardex: vetted Sardinian businesses, contract-based, no independent individual members. Banco Palmas: one neighbourhood. Chiemgauer: consumers who nominate a local association on joining.

Why it works. Inside a tight boundary a member can predict who will accept the unit, which means they will take it in payment, which means the next member can predict the same. Acceptance is self-reinforcing only inside a boundary. Outside one it has to be bought with marketing, and marketing is C.

The corollary most founders get backwards. Widening the scheme to be more inclusive usually lowers velocity, because it adds holders who buy from a short list and sell to nobody.

You already know this because you trust a members' club's IOU more than a stranger's, and the reason is not the paper.

---

BRIEF 9 — The Balance-Sheet Attachment

The idea. The question is never what should the currency be? It is what balance sheet is it bolted to, and does that balance sheet earn somewhere else?

The four proven attachments.

The uncomfortable cut. Bank WIR's commission and service income in 2023 was CHF 19.202 million against net interest income of CHF 67.8 million — the interest business is 3.53× the fee business. The world's most successful complementary currency is a mortgage bank that runs a currency on the side.

What is not on the list. Grant funding. Bristol raised £1.2 million over eight years, 58 percent from trusts and foundations, and generated no surplus in its life. Instituto Palmas, by contrast, is 72 percent earned income and 28 percent grant.

You already know this because every free service you rely on is paid for by something else, and you can usually name what.

---

BRIEF 10 — The Viability Line

The idea. Write the four numbers that decide the scheme before you launch it, and write down the number at which you would close it.

  target velocity        V
  annual operating cost  C
  stock at full draw     M
  implied fee            c = C / (M · V)

Worked example, sized for a 400-member trade association. Members with a median turnover of 900,000 take credit lines of one percent (9,000 each), giving a credit mass of 3,600,000 against acceptance undertakings of 36,000,000. A fee of 900 a member raises 360,000, which funds six operators.

  V =  2.0   turnover   7,200,000   c = 5.00 %   fails a 3 % fee test
  V =  4.0   turnover  14,400,000   c = 2.50 %   clears
  V =  8.0   turnover  28,800,000   c = 1.25 %   clears

  break-even velocity at a 3 % fee test:  V* = 3.33 per year

The sentence that belongs in the board paper. Below 3.33 turns a year, we close it. Written in advance, closing becomes a decision rather than a defeat, and the people who ran it keep their reputations for having said so first.

Why it matters. The field's largest information failure is that almost nobody publishes c, so every founder re-learns the same lesson from inside at a cost of roughly eight years. Publishing three numbers — cost, turnover, and their quotient — would end that.

You already know this because you have never opened a shop without knowing the rent and the margin, and this is the same two figures wearing different names.

---

All figures in these briefs are computed in lib/verify/III_03.py, printed with their inputs and their sources, and cited in the chapter's Works Cited. Estimates are labelled as estimates at the point of use.