Haute Lumière

Commerce · III.06 · MMXXVI · daylight

La Bourse  /  Volume III  /  Nº III.06  /  Workbook — the executive

WORKBOOK — THE CORPORATE EXECUTIVE

Chapter III.06 · Mycelial Finance

Applied to a P&L, a board paper and a business unit. Ninety days.

---

THE PREMISE, STATED COMMERCIALLY

You almost certainly have a receivables problem, a supplier fragility problem, or both, and you have been treating them as procurement questions. They are financing questions, and the financing is already sitting inside your own company in a form nobody has costed.

Here is the commercial case in four lines.

One. You can see your suppliers' and customers' trading performance every month, without asking, because they invoice you. That is a credit signal a bank would pay for and you obtain it as a by-product.

Two. The reason nobody lends to the small end of your supply chain is arithmetic, not prejudice. A relationship assessment costs roughly $3,000 whether the loan is $20,000 or $2,000,000, and at a 9.0 percent margin on principal that sets a floor of $33,333 below which a person cannot operate. Your best small suppliers sit under that floor.

Three. You already carry their risk. When a sole-source supplier fails, you pay in expedite fees, requalification, line stoppage and lost margin — you simply pay it in operations rather than in treasury, where it would have been visible and priced.

Four. A revenue-indexed facility inside your own supply chain earns a spread, reduces your operational fragility, and requires no new information system. On a $1,200,000 book at a $25,000 average ticket — 48 counterparties — a 6 percent net spread is $72,000 a year.

That last figure is deliberately unimpressive. This is not a profit centre in year one. It is a risk instrument that happens to pay for itself, and it should be presented that way to your board.

---

PART ONE — DISCOVERY

Days 1–30: find the credit you are already extending

1.1 — Inventory the credit you extend and do not call credit. Pull a list of every non-standard payment term granted in the last twenty-four months: early payment, extended terms, advance against order, tooling paid for by you, a deposit waived, a volume commitment. For each, record the counterparty, the amount, the decision-maker and the reason.

You are looking for a number most companies have never produced: the total value of uncosted credit currently extended by your operating units. In most mid-sized firms it is larger than the working capital facility they negotiate about every year.

1.2 — Name the signal you already receive. For each of your top hundred counterparties, answer one question: what do we see about their trading, monthly, without asking them? Invoice volume, order frequency, returns rate, lead-time drift, tonnage, headcount on site. Score each signal: seen automatically, seen on request, or not seen.

Only the first column is usable. A signal seen on request carries an assessment cost, and the assessment cost is what sets the floor.

1.3 — Find the near-misses. List every supplier who has had a liquidity problem in the last three years and survived. Ask what it cost you at the time — expedite, dual-sourcing, the engineer you sent. Then ask what an advance of twenty or thirty thousand dollars at that moment would have cost you instead. This comparison is the whole board paper and you can assemble it in a fortnight.

1.4 — Ask the question out loud. In one operations meeting, ask: which of our suppliers would we protect even if a cheaper quote arrived tomorrow, and why? Write down the answers. People know. That list is your first cohort, and it was produced at an assessment cost of one hour for the whole group.

---

PART TWO — THE ARITHMETIC

Days 31–45: put the numbers on one page

2.1 — Your own cost per credit decision. Time the real process: the credit application, the reference check, the internal approval, the sign-off. Fully loaded, per decision. Then compute:

  margin on principal = net spread x average duration
  breakeven ticket    = cost per decision / margin

  at $3,000 and 9.0 %   ->  $33,333
  at   $500 and 9.0 %   ->   $5,556
  at    $40 and 9.0 %   ->     $444

Then find the median value of the transactions you would like to finance. If it sits below your breakeven, you have a decision-cost problem, not a credit-policy problem, and the three moves are: automate the decision, delegate it to the person who already holds the signal, or stop pretending you will do it.

2.2 — Expected loss and the tail. Take your candidate cohort and assume an 8 percent default rate and 65 percent loss given default:

  expected loss                                       5.2 %
  one exposure's loss sd  0.65 x sqrt(0.08 x 0.92) = 17.63 %
  portfolio sd = 17.63 % x sqrt( 1/N + rho(1 - 1/N) )

  rho = 0.00   sd 0.176 %   1-in-100 loss   5.61 %
  rho = 0.02   sd 2.500 %   1-in-100 loss  11.01 %
  rho = 0.05   sd 3.947 %   1-in-100 loss  14.38 %
  rho = 0.15   sd 6.832 %   1-in-100 loss  21.09 %

On a $1,200,000 book that is $62,400 of expected loss a year against $72,000 of spread — and $172,564 in the one-in-a-hundred year at ρ = 0.05. Print those three numbers on one line. They are the paper.

2.3 — Face the correlation honestly. Your suppliers are in the same industry, often the same region, frequently dependent on the same input price and the same end market — as your own company is. That is not ρ = 0. It may be ρ = 0.15. The exercise is not to find a comforting figure; it is to write down a figure, the evidence for it, and what the book does if it is wrong by a factor of two.

2.4 — The instrument, priced. An advance of $250,000 at a 1.5× cap, repaid at 6 percent of monthly revenue from a base of $150,000 a month:

  revenue +2 %/month   31 months   IRR 34.6 %
  revenue flat         42 months   IRR 27.7 %
  revenue -1 %/month   54 months   IRR 23.7 %

A counterparty who shrinks by a quarter over four years repays in full. Nothing about that outcome requires a workout, a lawyer, or a line in your provisions. Model it and put it beside your current bad-debt experience.

---

PART THREE — DESIGN

Days 46–60: the instrument, and the accounting conversation

3.1 — The structure.

• Facilities of $10,000 to $50,000 to counterparties already trading with you for at least six periods.
• Repayment as a fixed percentage of observed trading with you — 4 to 7 percent — until a cap of 1.35× to 1.5× is met. No fixed maturity, no covenant, no personal guarantee.
• Signal read automatically from your own accounts payable or receivable. If it requires a form, the design has already failed.
• First-loss reserve at 6 percent of principal, funded from the origination charge so it accumulates with the book rather than requiring an annual decision.
• Retention. If you syndicate any of this later, retain 10 percent of every facility, pari passu. Write it into the first document; it cannot be retrofitted.

3.2 — The accounting conversation, which you have in week seven or not at all. Under IFRS 9 a financial asset is carried at amortised cost only if its cash flows are solely payments of principal and interest (paragraphs 4.1.1–4.1.4, the SPPI test). A payment indexed to revenue fails that test, so these assets are carried at fair value through profit or loss, and the fair value moves with your counterparties' trading. Under US GAAP the equivalent conversation is ASC 825.

Take this to your audit partner before the first drawdown, with a valuation methodology drafted. Presented in month one it is a paragraph in the accounting policies. Discovered in the year-end audit it is a restatement, and it will kill the programme regardless of how the book performed.

3.3 — Where it sits. Treasury owns the book; procurement owns the relationship; neither owns both. Split it any other way and either the credit standards or the supplier relationship gets sacrificed to the other. The reporting line that matters is that the loss lands in the same P&L as the spread.

3.4 — The capital. Unexpected loss at ρ = 0.05 is 9.18 points over the mean; the fund covers 6.0; the residual 3.18 percent of the book is capital you are holding whether or not you have named it. Name it. A programme with unnamed capital behind it is a programme that will be stopped by the first surprise.

---

PART FOUR — DESTINY AND DELIGHT

Days 61–90: ten facilities, and one page

4.1 — Ten, not one hundred. Ten facilities, one segment, real money. Above noise, below the threshold that requires a second committee, and — the only thing that matters next — it produces a repayment series somebody else can check.

4.2 — The monthly one-pager. Six lines, published internally on the same date every month: book size, facilities outstanding, weighted average repayment share, realised loss to date, fund coverage ratio, and the ρ currently underwritten. That last line is the discipline. A programme that publishes its own decisive assumption cannot drift quietly.

4.3 — What makes it survive you. Three things and only three. It is on the standing reporting pack, so it is reviewed monthly rather than by exception. Somebody's objectives move with the fund coverage ratio — the signal matters more than the size. And it has a second owner in a different function, recruited early and given the credit for the first result.

4.4 — Delight, and it is a real commercial asset. The first time a counterparty has a thin quarter and the repayment simply falls, with no letter, no breach and no phone call, your buyer will hear about it. What you have bought is not goodwill in the sentimental sense: it is first refusal on capacity in a shortage, and first call when they are about to have a problem. Both of those have a price and neither is on your invoice.

---

THE FAILURE MODES, NAMED

So you can see them coming

The fee drift. If any part of this is compensated on volume rather than on outcome, volume is what you will get. The correction is retention: a slice that loses first.

The form. Six months in, somebody adds a two-page application "for consistency". Assessment cost moves from $40 toward $3,000, the minimum viable ticket climbs from a few hundred toward tens of thousands, and the programme quietly stops serving the counterparties it was built for. Nothing on the P&L shows this. The symptom is a rising average facility size with a falling facility count.

The mean-sized reserve. A fund built against 5.2 percent expected loss meets a 14.38 percent correlated year and is recapitalised out of somebody's return. RateSetter's — a decade-old, well-run mutual reserve — met exactly this in May 2020 and the answer was to halve lender rates for the rest of the year and divert half of all returns into the fund. Size against the tail or say plainly who pays when it arrives.

The unstated ρ. Every figure in this workbook moves with the correlation assumption and nothing else moves as much. An unstated assumption is not a conservative one.

---

THE NINETY DAYS ON ONE PAGE

---

BOARD PAPER TEMPLATE

One page. In this order.

1. What we already do. The value of uncosted credit currently extended by operating units, and by whom.
2. What it costs us today. The near-miss comparison: operational cost of supplier liquidity events over three years, against the advance that would have prevented them.
3. The instrument. Revenue-indexed facility, cap 1.35×–1.5×, first-loss reserve at 6 percent of principal, 10 percent retained if syndicated.
4. The arithmetic. Book $1,200,000 · 48 counterparties · spread $72,000 · expected loss $62,400 · one-in-a-hundred loss $172,564 at ρ = 0.05.
5. The assumption that decides it. ρ, the evidence for the figure chosen, and the outcome if it is wrong by a factor of two.
6. Accounting treatment. FVTPL under IFRS 9; methodology agreed with the auditor on [date].
7. The ask. Ten facilities, one segment, capped at [amount], reviewed in ninety days against a published six-line pack.

---

---

THREE OBJECTIONS YOU WILL MEET, AND THE ANSWER TO EACH

"We are not a bank." Correct, and this is not banking. You are already extending credit — the inventory in Exercise 1.1 proves it — and the proposal is to price and record what you already do, in an instrument that self-liquidates out of trading you already see. The alternative is not "no credit exposure". It is the same exposure, unpriced, sitting in operations.

"Procurement will use it to avoid renegotiating prices." A real risk and worth a hard control: no facility to a counterparty whose commercial terms were agreed in the same quarter, and the facility is never a substitute for a price conversation. Put that line in the policy on day one, because it is unarguable on day one and contentious in month nine.

"What if a supplier we have financed fails anyway?" Then you lose the outstanding principal and keep the operational information you bought — which is that you had a fragile sole source and now know it. Size the first cohort so that the total possible loss is smaller than one line-stoppage event you have already survived. If that number is not available to you, the first ninety days should produce it, and producing it is valuable whether or not you ever write a facility.

APPRECIATIVE QUESTIONS FOR YOUR LEADERSHIP TEAM

Six, for a room, in the 4D order.

1. Discovery. Where are we already extending credit that we do not call credit, and who decided it?
2. Discovery. Which counterparty do we see so clearly that we could predict their next quarter without asking? What exactly are we seeing?
3. Dream. If our minimum viable financing ticket fell from thirty thousand to five hundred, who would we be doing business with next year that we are not doing business with now?
4. Design. What signal do we already receive, automatically, that a fair lender would pay for — and what are we currently doing with it?
5. Design. What is the smallest slice we could retain in something we originate that would change how carefully we originated it?
6. Destiny. If our correlation assumption turned out to be wrong by a factor of two, who finds out first, and by what route?