Haute Lumière
Commerce · I.09 · MMXXVI · daylight
TEN CONCEPT BRIEFS · Chapter I.09 — The Institutional Handshake
One page each. A reader who reads only these ten pages has the chapter.
BRIEF 1 — The Category, Not the Decision
The idea. An institution that cannot change its own rules can still decide, on any ordinary day, which of its existing rules a new thing belongs under. That classification is the transaction. Everything else follows from it.
This is why proposals that are obviously good are refused and proposals that are obviously dull are signed. The refusal is almost never a judgement on the substance. It is a report that the thing did not resemble anything the institution already knows how to do, and a body with no memory of its own reasons protects its consistency by declining what it cannot classify.
Worked example. A practitioner proposes a regenerative grounds-maintenance trial to a university. Framed as "a new sustainability initiative", it requires a new budget line, a new owner and a paper to a committee that meets twice a year. Framed as "a variation to the existing grounds contract under paragraph 7 of the Sustainable Estates Policy adopted in March", it is a decision the Head of Estates already has authority to take. Same intervention, same cost, same people. One took nine months and one took eleven days.
Why it matters. It changes what you write. You are not making a case for a new activity. You are describing an existing authorised activity of which your proposal is an instance — and citing the institution's own policy, by paragraph and date, back to the body that adopted it.
You already know this because you have watched an expense claim sail through when it was filed under the right code and bounce when it was filed under the wrong one, with the same amount and the same receipt attached.
BRIEF 2 — The Scheme of Delegation
The idea. Every institution publishes a document setting out who may commit it to what, alone. It is called a scheme of delegation, a financial regulations schedule, a delegated authorities matrix. It is not confidential. It is the map of every door in the building, and almost no outsider ever asks for it.
How to read it. You are looking for the lowest level of authority that can sign the whole of what you want, with no co-signature and no notification requirement. Note three things: the figure, the role, and the exceptions — most schemes carve out categories (property, employment, anything multi-year) that sit outside the figure regardless of value.
Worked example. A district's scheme gives a headteacher £25,000, a business manager £10,000, and the trust's finance director £100,000 — but it also says any commitment beyond twelve months goes to the audit committee whatever it is worth. A £9,000 three-year arrangement therefore needs a committee; a £24,000 nine-month one does not. The term, not the money, was the constraint.
Why it matters. The figure you find becomes the design constraint on everything else. You do not design the intervention and then look for a signature. You find the signature and design to it.
You already know this because you have asked a manager for something, been told they would need to check, and understood instantly that you had asked the wrong person.
BRIEF 3 — The Delegation Ratio
The idea. One number belongs on the front of your proposal.
delegation ratio = pilot value / the delegated authority of the signatoryBelow 1.00 the thing can be signed. Target 0.60 or below.
Worked example. Against a $10,000 micro-purchase threshold, a $9,400 pilot has a ratio of 0.94. It clears — and then one change order, one currency move or one forgotten delivery charge puts it over, and the whole approval reopens at a level that now needs a committee. The same pilot scoped at $6,000 has a ratio of 0.60 and holds through a 66 percent overrun.
Why it matters. The 40 percent you leave behind is not caution. It is the contingency that protects an approval you have already won, and approvals are the expensive thing. A pilot that has to be re-approved has cost you more than a pilot that was smaller in the first place.
You already know this because you do not fill a suitcase to the airline's exact weight limit if you intend to buy anything at all while you are away.
BRIEF 4 — Cycle Time, and the Papers Deadline
The idea. The cost of an institutional yes is dominated by the calendar, not the price.
A quarterly committee sits every 91.3 days. A proposal that becomes ready at a random moment waits on average 45.7 days for the next sitting, and it must be in the pack before the papers deadline — typically ten working days earlier, so add 14. Each committee layer therefore costs a mean 59.7 days, before anyone has read anything.
The number that matters. The papers deadline, not the meeting date. Miss the deadline by four days and you have not lost four days; you have lost a whole cycle.
Worked example. A ninety-day pilot needing six weeks to verify produces a result on day 132, is in the pack by day 146, and is heard at the sitting on day
- Start the same pilot forty days later and the result lands on day 172,
misses the pack, and is heard on day 274. Forty days of delay at the front cost ninety-two at the back — 2.3 days lost for every day deferred.
Why it matters. It means the start date is decided by the committee calendar, working backwards, and not by when you happen to be ready.
You already know this because you have stood at a station and watched a train leave, and understood that arriving two minutes late cost you thirty.
BRIEF 5 — The Break-Even Markup
The idea. Weight each approval route by its probability of passing and discount it for the delay, and you can ask precisely: how much larger must the committee version be, merely to draw level with the version one person can sign?
layers of committee must be worth larger by
----------------------------------------------------
one 1.135x 13.5%
two 1.437x 43.7%
three 1.818x 81.8%(Assumes 0.80 pass probability per layer, 8 percent discount rate, three-year horizon. Substitute your own; the minute book has the pass rate.)
Read it the other way, which is the point. You may pay a 44 percent premium for the version that fits under one signature and still be ahead of the version that needs two committees.
Worked example. Proposal A saves £48,000 a year under one signature; B saves £96,000 through two committees. B wins — twice beats 1.44 times. But the indifference point is £68,956 a year, and below that the smaller, faster, already-authorised thing is worth more than the better one.
Why it matters. It reverses the practitioner's instinct. You have been trying to make the proposal better. Make it smaller and sooner instead, and spend the difference on fitting it under an existing signature.
You already know this because you have taken the flight that leaves today over the cheaper one that leaves on Thursday.
BRIEF 6 — Fiduciary Duty, Read Correctly
The idea. Fiduciary duty is more permissive than its folklore. The Law Commission set the structure out in 2014 (Law Com No 350):
- Financial factors must be taken into account — including environmental, social and governance factors where they are financially material. Ignoring a material risk is itself a breach.
- Non-financial factors may be taken into account where two tests hold: trustees have good reason to think beneficiaries share the concern, and there is no risk of significant financial detriment to the fund.
The parallel authorities. For charities, Harries v Church Commissioners [1992] and Butler-Sloss v Charity Commission [2022] give trustees discretion to exclude investments conflicting with the charity's purposes. In the United States, UPMIFA §3 directs institutions to consider "an asset's special relationship or special value, if any, to the charitable purposes of the institution", and the Department of Labor's 2022 rule (29 C.F.R. §2550.404a-1) confirms that a risk-and-return analysis may include climate and similar effects where relevant.
Worked example. A trustee board believes it is barred from considering supplier labour practices. It is not: where those practices are financially material to the holdings, the board is required to consider them, and the minute recording that it did is the defence.
Why it matters. Most of what practitioners experience as a legal prohibition is a cautious summary of a case somebody read a précis of. Go to the primary text.
You already know this because you have been told "we're not allowed to" by someone who had never seen the rule they were quoting.
BRIEF 7 — Prudence Is a Process
The idea. ERISA requires a fiduciary to act "with the care, skill, prudence, and diligence under the circumstances then prevailing" that a prudent person would use (29 U.S.C. §1104(a)(1)(B)). Read the qualifier. Prudence is judged on the process followed at the time, not on the outcome that arrived later.
What follows. The defensible act is the documented deliberation: the options considered, the evidence weighed, the advice taken, the reasoning minuted. A decision that turned out badly after a good process is defensible. A decision that turned out well after no process is not.
Worked example. Two boards make the same allocation. One minutes the analysis it commissioned, the alternatives it compared and why it chose. The other approves it in four words. Three years on, the allocation underperforms. The first board has a defence and the second has an exposure — from the same decision.
Why it matters. It tells you what to put in front of a fiduciary: not enthusiasm, but the materials that let them run a good process — a comparison, a source, a risk, a named method. You are supplying a defence, and that is a gift.
You already know this because you keep receipts, and you keep them whether or not you expect to be audited.
BRIEF 8 — Social Value as an Existing Door
The idea. Where you must compete rather than be commissioned, most public procurement regimes already allow award on more than price. The Public Services (Social Value) Act 2012 requires contracting authorities to consider economic, social and environmental well-being; Directive 2014/24/EU permits award on the most economically advantageous tender including social and environmental characteristics (Art. 67) and life-cycle cost (Art. 68); PPN 06/20 set a floor of 10 percent of the score for social value in central government contracts.
The arithmetic, so you do not overestimate it. Score a tender 50 quality / 40 price / 10 social value, with price scored relatively.
win the social-value section by 4 of 10 -> carries an 11.1% price premium
win it by 6 of 10 -> carries a 17.6% premium
win it by 8 of 10 -> carries a 25.0% premiumInverted: a bid 12 percent dearer loses 4.29 price points and needs 43 percent of the entire social-value section just to stand still.
Why it matters. The door is real and it is not wide. Plan to be competitive on price and to win on the margin. A weighting is a tiebreaker, not a subsidy.
You already know this because you have won something on the strength of a better answer to one question, and only after being close enough on the rest.
BRIEF 9 — The Agreed Baseline
The idea. The baseline is the figure that says what was true before you changed anything, agreed and signed by the institution's finance function before deployment. In the memorandum it is one of only three binding clauses, and it is the one that determines whether the result can be argued away.
What it contains. The figure. The method by which it was derived. The period it covers. The date. Two signatures — yours and the finance officer's. And the threshold at which the result is treated as met, agreed in the same breath, so that success is defined by both parties while neither knows the answer.
Worked example. A waste-diversion pilot reduces tonnage by 19 percent. With a signed baseline covering the previous four quarters and a threshold of 10 percent agreed in advance, that is a verified result and it goes to the review gate. With no baseline, it is a conversation about whether last year was unusual — and that conversation is unwinnable, because last year is always unusual.
Why it matters. An unagreed baseline is not a baseline. It is a future dispute with a delay attached, and the delay is a whole committee cycle.
You already know this because you photograph a rental car before you drive it away.
BRIEF 10 — The Successor Clause
The idea. The single clause that carries the weight of the memorandum. It binds nobody to adopt anything. It binds an agenda and a date.
If the verification report records a result at or above the threshold, the parties shall procure that the report, together with this memorandum, is placed on the agenda of the [named committee] at its first ordinary meeting falling not less than fourteen days after the date of the report, for consideration of a successor instrument. Nothing in this clause obliges either party to enter into a successor instrument, or fetters the discretion of the [named committee] in any respect.
Why that shape. A public body cannot lawfully fetter its own future discretion, and a clause that tried to would be struck out and would make the rest of the document suspect. The last sentence is what makes the first one signable.
The failure mode it is built against. Institutions rarely refuse. They far more often just never reach the item — the pilot works, the term runs out, nobody puts it on an agenda, and the thing lapses without anyone deciding anything. This clause makes reaching the item the default and requires an affirmative act to prevent it.
Why it matters. It is the only promise an institution that cannot change its own rules is genuinely able to make. It is also enough.
You already know this because you have had a good idea agreed in a corridor and then never heard of again, and you know the problem was not that anyone disliked it.
THE HONEST BOUNDARY — one page that is not a concept
Three places where everything above stops working, kept together so they cannot be skipped.
You may not split a requirement to get under a threshold. Dividing a requirement to evade a threshold is expressly prohibited (2 C.F.R. §200.320; FAR 13.003(c)(2)), and equivalently in every serious regime. Capital-intensive interventions — a retrofit, a heat network, a fleet — have no small version, and for them the committee route is the only route. Plan eighteen months and say so early.
A long-horizon benefit cannot win a tiebreaker. Where the ERISA rule permits collateral benefits, it permits them only when competing investments equally serve the plan's financial interests. A benefit landing in year twelve, discounted at 7 percent, is worth 44.4 pence in the pound; multiplied by the probability that the mandate survives twelve years at 15 percent annual turnover, 6.3 pence. The gap closes by shortening the horizon of the claim, or it does not close.
A regulator can flex a process, not a capital rule. The FCA's regulatory sandbox accepted 24 of 69 first-cohort applicants. A rationed door is still a closed door for two applicants in three, and an approach that depends on entering it is not an approach.