The Institutional Handshake

Volume I — Transition: From Here to the Living Economy

THE PLATE

Two women on floor cushions with cups of tea, turned toward each other in conversation, soft window light. — Plate I.09Two Hands and a Standing Order.An institution cannot change its own rules. It can, on any Tuesday, decide which of its existing rules a new thing belongs under — and that decision is the whole of the transaction.

THE LETTER

By now you have a result. Chapter I.01 sent you looking for the place where the living-systems approach was already outperforming inside your own numbers, and by Chapter I.05 you had a pilot that paid. If the sequence has worked, you are holding a verified saving, a signed baseline, and a one-page story.

And now you have met the wall that everyone meets. The next move requires a counterparty who cannot simply decide. A procurement department bound by a scheme of delegation it did not write. A regulator whose rules are made elsewhere. An auditor who is not permitted to be persuaded. A pension trustee whose duty is owed to people who are not in the room and mostly not yet born. A university finance office running an indirect cost rate agreed with a federal agency. A school district whose budget is a statute.

Here is the thing worth knowing before you spend another week on this, and it is the finding that this chapter exists to hand you: in almost every case, the institution already has permission to do what you are asking. What it does not have is a category to put you in. The constraint is nearly never risk appetite, and it is nearly never values. It is precedent — the institution's requirement that any new thing be recognisable as an instance of an old thing, because that is how a body with no memory of its own reasons stays consistent.

This is very good news, and it changes what you are writing. You are not writing a case for a new activity. You are writing a description of an existing activity that the institution is already authorised to perform, of which your proposal happens to be an instance.

The rest of this chapter is how to find that category, how to read the law correctly when it turns out to be more permissive than the people inside the institution believe, how to size the thing so that one person can sign it, and what the document says. The document is a memorandum of understanding. It is short. One of its clauses does nearly all of the work, and we will write that clause out in full.

— The Editors

DISCOVERY

What is already working

Start where the record is strongest: with the institutions that have already done this, and with the moments when somebody went and read the rule rather than inheriting a summary of it.

The Law Commission asked what fiduciary duty actually said, and found more room than anyone was using. In 2014 the Law Commission of England and Wales published Fiduciary Duties of Investment Intermediaries (Law Com No 350). The question put to it was whether trustees were legally required to maximise short-term returns and legally barred from considering anything else. The answer was no on both counts, and the Commission set out a structure that has been the working basis for practitioners ever since. Financial factors — including environmental, social and governance factors where they are financially material — must be taken into account. They are not optional; ignoring a material risk is itself a breach. And non-financial factors may be taken into account where two conditions hold: trustees have good reason to think beneficiaries share the concern, and there is no risk of significant financial detriment to the fund.

Read that carefully. It is not a permission slip for anything anyone likes. It is a two-part test, and a two-part test is a door. The practitioner's belief that Cowan v Scargill [1985] Ch 270 forbade all of this was an over-reading of a case about a specific and quite different set of facts — Sir Robert Megarry himself said as much afterwards.

The charity courts said it again, and more warmly. In Butler-Sloss v Charity Commission [2022] EWHC 974 (Ch), two charitable trusts asked the High Court whether they could adopt investment policies excluding holdings that conflicted with the aims of the Paris Agreement. Mr Justice Michael Green held that trustees have a discretion to exclude investments that conflict with the charity's purposes, weighing the risk of financial detriment and of alienating supporters, and that where they properly exercise that discretion the court will not interfere. The line runs back to Harries v Church Commissioners for England [1992] 1 WLR 1241, where Sir Donald Nicholls VC had already said it. The institutions asked, and got a clearer answer than the one they had been frightening themselves with.

In the United States, the statute is more generous than its reputation. The Uniform Prudent Management of Institutional Funds Act, adopted in nearly every state, directs an institution managing an endowment to consider, among other things, "an asset's special relationship or special value, if any, to the charitable purposes of the institution." A university finance office that believes it is forbidden to weigh the relationship between an investment and the university's mission is misreading its own governing statute. And the Department of Labor's 2022 rule under ERISA — Prudence and Loyalty in Selecting Plan Investments and Exercising Shareholder Rights, 87 Fed. Reg. 73822, codified at 29 C.F.R. §2550.404a-1 — states plainly that a fiduciary's risk-and-return analysis may include the economic effects of climate change and other factors where they are relevant. (It has been litigated; read its current status before you rely on it. That habit is the chapter's method in miniature.)

Preston found the permission in a statute already on the books. From about 2013 the city council and a group of anchor institutions — the university, the constabulary, the further education college, the housing associations — examined where their combined procurement spend landed and redirected what they lawfully could toward suppliers in the local economy. No rule was changed. The Public Services (Social Value) Act 2012 already required contracting authorities to consider how a procurement might improve economic, social and environmental well-being, and the European procurement directive already permitted award on the most economically advantageous tender, including environmental and social characteristics and life-cycle cost (Directive 2014/24/EU, Articles 67, 68 and 70). Preston's officers used what was there. Independent evaluation records a substantial rise in the proportion of anchor spend retained locally — the same money, routed differently, under powers that already existed.

Cleveland built the handshake into a supply agreement. From 2008 the Evergreen Cooperatives — a laundry, an energy business, a hydroponic greenhouse — were established around procurement commitments from the Cleveland Clinic, University Hospitals and Case Western Reserve University. The institutions were not asked to donate. They were asked to buy laundry, which they already bought, from a supplier constituted differently. That is the shape of every successful institutional handshake: the institution keeps doing what it already does, and one attribute of the counterparty changes.

Five cases, one pattern. In none of them did an institution change a rule. In every one of them, somebody read the rule that already existed and found it wider than the folklore around it.

So: go and read yours. Three documents, all of which are published and none of which anyone will refuse you.

The scheme of delegation. Who may sign what, alone, today.
The standing policy. Social value, responsible investment, community benefit, sustainable procurement, mission-related investment. One of these exists. It is your category.
The committee calendar, with the papers deadlines on it. Not the meeting dates — the deadlines. The deadline is the real date.

THE ARITHMETIC

What the handshake costs, and where it fails

Now the part that decides everything, and it is not the part anyone expects.

The cost of an institutional yes is dominated by cycle time, not by price.

Model an approval as a ladder. Each rung is a body that meets on a calendar. A quarterly committee sits every 91.3 days; a proposal that becomes ready at a random moment waits on average half a cycle, 45.7 days, and it must be in the pack before the papers deadline — ten working days ahead, so add 14. Each rung therefore costs a mean 59.7 days, and each rung has some probability of passing. Assume 0.80 per rung; substitute your own institution's figure, which is in the minute book, because the minute book records every paper submitted and every paper deferred.

  route                         days     p(yes)
  --------------------------------------------
  one delegated signature       18.0      0.900
  one committee layer           59.7      0.800
  two committee layers         119.3      0.640
  three committee layers       179.0      0.512

Now price the two effects together. Weight each route by its probability and discount it for the delay at an 8 percent institutional rate, and ask the question that matters: how much larger must the committee proposal be, just to draw level with the one that fits under a single signature?

  layers of committee    must be worth      i.e. larger by
  -------------------------------------------------------
  one                          1.135x              13.5%
  two                          1.437x              43.7%
  three                        1.818x              81.8%

Read it the other way round, because this is the move. You may pay a 44 percent premium for the version that one person can sign, and still be ahead of the version that needs two committees. Not a 44 percent worse outcome — a 44 percent higher cost for the same outcome, and it still wins.

Which reverses the instinct that every practitioner arrives with. You have been trying to make the proposal better. The arithmetic says make it smaller and sooner, and spend the difference on whatever it takes to fit it under an existing signature. The institution is not deciding whether to say yes. It is deciding which of its existing categories this is — and the category is worth more than the money.

The effect is robust. At a generous 0.95 pass probability per rung it is still there; at a realistic 0.70 a three-layer proposal must be worth 2.71 times as much to draw level.

Worked, on two real-sized proposals. Proposal A saves £48,000 a year and sits under one signature. Proposal B saves £96,000 and needs two committees. Over three years at 8 percent, A's expected present value is £110,909 and B's is £154,406 — B wins, because twice is more than 1.44 times. So B is right. 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. Most of what you will be tempted to propose sits below the indifference point.

The thresholds are published, and they are the shape of the door. For institutions spending US federal award money, the Uniform Guidance and the FAR set them out:

  micro-purchase                   $10,000    no competitive quotes required
  micro-purchase, self-certified    $50,000    documented, up to this ceiling
  simplified acquisition           $250,000    simplified procedures only

(2 C.F.R. §§200.1 and 200.320; FAR 2.101. UK thresholds are recalculated every two years under the Procurement Act 2023 — read the current threshold notice, and do not quote a figure from a textbook, including this one.)

From which the single number that should be on the front of your memorandum:

   delegation ratio  =  pilot value / the signatory's own authority

   $9,400 / $10,000  = 0.94   clears, with no room for a change order
   $6,000 / $10,000  = 0.60   clears with room
 $140,000 / $250,000 = 0.56   clears with room

Target 0.60 or below. The 40 percent you leave on the table is the contingency that stops a single change order re-opening an approval you have already won.

What a social-value weighting actually buys. Where you must compete, PPN 06/20 set a floor of 10 percent of the total score for social value in central government contracts. Score a tender 50 quality, 40 price, 10 social value, with price scored relatively. Win the social-value section by six points of ten and you can carry a 17.6 percent price premium. Inverted: a bid 12 percent dearer loses 4.29 price points and needs 43 percent of the entire social-value section just to stand still. The door is real. It is not wide. Plan to be competitive on price and to win on the margin, not to be rescued by the weighting.

Now the three places this fails, honestly.

One: the long-horizon case loses a tiebreaker, and the reason is arithmetic rather than ideology. Where the ERISA rule permits collateral benefits, it permits them only when competing investments equally serve the plan's financial interests (29 C.F.R. §2550.404a-1(c)(2)). Take a benefit landing in year twelve. Discount it at 7 percent and a pound becomes 44.4 pence. Multiply by the probability that the mandate survives twelve years — at an assumed 15 percent annual turnover, 0.1422. Six pence in the pound. A twelve-year regeneration benefit cannot demonstrate equivalence to a three-year benchmark, and no amount of conviction closes that gap. It closes by shortening the horizon of the claim, or it does not close.

Two: the threshold route is bounded by law, and the boundary is criminal in some jurisdictions. A $140,000 requirement cannot become fourteen purchases of $10,000. Dividing a requirement to evade a threshold is expressly prohibited (2 C.F.R. §200.320; FAR 13.003(c)(2)), and the equivalent prohibition exists in every serious procurement regime. Capital-intensive interventions have no small version. A building retrofit, a district heat network, a fleet — for these the committee route is the only route, and the whole of this chapter's speed advantage is unavailable. Say so early and plan for eighteen months.

Three: a rationed door is still a closed door. The Financial Conduct Authority's regulatory sandbox accepted 24 of 69 applicants in its first cohort — just over a third. A regulator can flex a process, a reporting form, a timetable. It cannot flex a capital requirement, and it will not, and an approach that depends on it doing so is not an approach.

DREAM

What becomes ordinary

Describe it in the present tense, because a dream in the future tense is a wish.

In the institution that has learned this, there is a named officer whose job includes classification. When something arrives that does not obviously fit, it goes to them, and their question is not should we do this but what is this an instance of. They know the scheme of delegation the way a librarian knows a catalogue. They find the category in a morning, and the finding is treated as a service rendered rather than a favour granted.

The scheme of delegation itself is a living document. It is reviewed annually, and one of the review questions is: what did we decline in the last year that was smaller than this threshold and would have been signed if it had been classified faster? The answer is a list, and the list moves the thresholds.

The standing policy — social value, responsible investment, community benefit, whatever it is called here — has worked examples attached to it. Not principles: examples, with the memorandum that was signed, the baseline that was agreed and the verified result. A policy with worked examples attached is an instruction; a policy without them is a mood. New officers read the examples and know what good looks like, and a supplier arriving for the first time can be handed the same file.

The committee calendar is published a year ahead with the papers deadlines on it, and people plan backwards from it as a matter of course, the way a farmer plans backwards from a frost date. Nobody misses a cycle by four days any more. The phrase "we'll take it to the next one" has lost its capacity to mean "we have quietly declined."

And the memoranda accumulate. There is a register of them, numbered, searchable, with the verified result recorded against each. It is the institution's own evidence base, and it is consulted, and a new proposal that resembles memorandum number thirty-one is approved in an afternoon because the institution can see what happened to memorandum number thirty-one. The precedent that was the obstacle has become the asset. Nothing about this required anyone's heart to change. It required somebody to keep a list.

DESIGN

The handshake, built

Five moves, in order. The order is not negotiable, because each one determines what the next one can say.

1 — Read the scheme of delegation before you write a word. It is a published document; ask for it by name. You are looking for the lowest level of authority that can sign the whole of what you want, alone, with no co-signature and no notification requirement. Write down the name, the role, and the figure. That figure is now the constraint on your design, and everything else bends to it.

2 — Find the category. The institution has a standing policy that already authorises the substance of what you propose. It will be called something unpromising. Read it, find the sentence, and quote that sentence in your recitals, by paragraph number, with its date of adoption. You are not asking for an exception. You are citing a policy back to the body that adopted it. An officer can approve the application of an existing policy on their own authority. Approving an exception to it is a different act with a different signature.

3 — Size to the ratio. Design the pilot so that its total cost, including contingency, sits at or below 0.60 of the signatory's authority. If it does not fit, do not split it — that is unlawful and it will end the relationship. Reduce the scope, shorten the term, or accept that you are on the committee route and plan accordingly.

4 — Write the memorandum, and make exactly three clauses binding. The baseline, the data and intellectual property, and the exit. Everything else is expressly non-binding, and saying so in clause 1 is what allows a cautious signatory to sign at all. This is treated in full in the next movement.

5 — Work backwards from the calendar. Find the review gate — the committee at which the verified result will be heard — and count backwards: papers deadline, verification period, pilot period. That arithmetic gives you a start date, and the start date is not negotiable in the way that the amount is.

Here is why. A ninety-day pilot needing six weeks to verify produces a result on day 132, in the pack by day 146, heard at the sitting on day 182. Start forty days later and the result lands on day 172, misses the pack deadline, and is heard on day 274. Forty days of delay at the front cost ninety-two days at the back — 2.3 days lost for every day deferred. That ratio is the reason a practitioner who understands institutions seems to move faster than one who does not. They are not moving faster. They are landing on the deadline.

Governance, in one paragraph. The signatory owns the decision. A named finance officer owns the baseline and signs it separately. Internal audit, or its equivalent, owns verification — it is sufficient, it is independent enough for this purpose, and it is usually free. The review gate is a named committee on a named date. Nobody else is required, and every additional name you add to the memorandum costs you the 13.5 percent from the table above.

DESTINY

How it holds when you stop pushing

The memorandum is signed and you leave the room. Three things decide whether it still exists in two years.

It is in the register. Lodged in the institution's contracts register with a reference number, so that a successor officer searching for precedent finds it. An agreement that lives in two inboxes dies with the first reorganisation. This costs one email and it is the highest-return action in the whole chapter.

The verification method is written into the institution's own manual, not yours. A method the institution owns survives your departure. A method that requires your spreadsheet does not.

The role, not the person, is named. The Director of Estates rather than a name. People move; posts persist. Where the memorandum must name a person, add the post in brackets after it.

And the honest part. Here is where this fails. It fails when the category you cited is refreshed in the annual policy review and the sentence you quoted is gone — so diary the review date and be in the consultation. It fails when the signatory is promoted in month four and the successor inherits a document without the conviction, which is why the successor clause names a committee and not a person. It fails when the verified result is genuinely ambiguous and no one agreed in advance what would count as success — which is the baseline clause doing its job, or not doing it. And it fails, most often and most quietly, when the result was good and nobody put it on an agenda, so the memorandum simply ran its term and lapsed. Institutions rarely refuse. They far more often just never reach the item. Everything in the next movement is built against that one failure mode.

DELIGHT

What it feels like

There is a specific pleasure in this work and it is not the signature. It is the moment, usually about twenty minutes into a first meeting, when the procurement officer stops being an obstacle and becomes a collaborator — because you asked the right question, which was what is this an instance of, and they have been waiting years for somebody to ask them something they are actually expert in.

They will get a pen. They will turn the scheme of delegation around so you can see it. They will say something like "if you took the training out of it and called it a variation, I could do this myself" — and in that sentence they have just saved you a hundred and twenty days and told you the shape of the whole deal. People inside institutions are not guarding the gate. Most of them have never been asked to open it in a way that would let them.

And then the second pleasure, some months later: a document you wrote is quoted back to you by someone who does not know you wrote it, as the way this is done here. That is the moment your work stops being a project and becomes infrastructure, and it is quieter and better than winning an argument.

OPERATIONALIZE THIS

At the level of finance — the memorandum of understanding

The instrument is a memorandum of understanding, and its whole design problem is that it must be light enough for one person to sign and heavy enough to survive that person leaving. It resolves that by being non-binding in general and binding in three specific places.

The structure.

Form. A memorandum of understanding, not a contract and not a grant agreement. Two to four pages. No annexes except the baseline.
Recitals. Name the institution's own authority: "pursuant to the [Policy], adopted [date], paragraph [n]." This is the category, and it is the reason the signature is within delegation.
Clause 1 — Nature. Non-binding and creating no legal relations, save for clauses 6, 9 and 11, which are expressly binding. State it in the first clause. A cautious signatory reads clause 1 and relaxes, and a relaxed signatory signs.
Clause 2 — Classification. One sentence naming what this is an instance of: a variation to an existing arrangement, a trial under the standing policy, a research collaboration under an existing framework.
Clauses 3–5 — Scope, term, value; and what each party will do. State explicitly what the institution is not undertaking: no new budget line, no headcount, no exclusivity, no endorsement, no obligation to proceed.
Clause 6 — The baseline (binding). The agreed figure, the method, the period, the date, signed separately by the finance officer. An unagreed baseline is not a baseline; it is a future dispute with a delay attached.
Clause 7 — Measurement and verification. Named method (IPMVP option and number where it applies), named period, named verifier — internal audit is sufficient.
Clause 8 — The review gate. The named committee, the date of its sitting, and at 8.3 the threshold: the verified figure at or above which the result is treated as met.
Clause 9 — Data, intellectual property, confidentiality, publication (binding).
Clause 10 — Exit. Termination by either party on thirty days' notice without cause. Nothing buys a signature like a cheap exit; a commitment an institution can leave is one it can enter.
Clause 11 — The successor clause (binding). Below, in full.

The clause that carries the weight.

11. Consideration of a successor instrument. If the verification report prepared under clause 7 records a result at or above the threshold at clause 8.3, 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.

Read what it does and does not do. It binds no one to adopt anything. It fetters no discretion, which is what makes it signable — a public body cannot lawfully bind its own future discretion, and a clause that tried to would be struck out and would poison the rest. What it binds is an agenda and a date. That is the only promise an institution that cannot change its own rules is actually able to make, and it is sufficient, because the failure mode is not refusal. It is never reaching the item. This clause makes reaching the item the default and requires an affirmative act to prevent it.

Balance-sheet treatment. Below threshold and within the period, operating expenditure — which is precisely why it is signable. Where a successor instrument creates or improves a long-lived asset, capitalise it then and depreciate over its regenerated life. Do not attempt that conversation at memorandum stage; it belongs to the successor, and raising it early converts a signature into a committee.

The counterparty. The lowest office with sufficient delegated authority, never the highest office that would be flattered to receive it. Seniority above the threshold adds risk aversion without adding authority.

The first ninety days.

Day	Action	Artifact
1–10	Obtain the scheme of delegation and the committee calendar	The threshold figure and the papers deadlines
11–20	Find the category; quote the paragraph	The recitals
21–30	Size to a delegation ratio of 0.60 or below	The costed scope
31–45	Agree and sign the baseline with finance	The signed baseline
46–55	Draft the memorandum; secure the single signature	The signed memorandum
56–90	Deploy; measure; lodge the memorandum in the register	Register reference number

The number that decides it. One figure, on the front page:

                     pilot value
   ---------------------------------------------   <=  0.60
    the delegated authority of the one signatory

Below 0.60, the handshake is a signature. Above 1.00 it is a committee, and everything a committee costs — 13.5 percent for one layer, 43.7 for two, 81.8 for three — should be priced into the proposal before it is written, not discovered afterwards.

APPRECIATIVE QUESTIONS

Twelve, for a room

Discovery — what is already working

Think of a time this institution said yes to something unusual and it went well. What category did it end up being filed under, and who found that category?
Which of our standing policies is more permissive than the way we currently use it — and who here has read it all the way through this year?
What have we approved under delegated authority in the last twelve months that a neighbouring institution would have taken to a committee?

Dream — what becomes possible

If a supplier arrived tomorrow with something genuinely new, what would have to be true for us to be able to say yes within thirty days and be proud of how we did it?
Imagine our register of agreements three years from now, with verified results recorded against each. What would we use it for that we cannot do today?
If our scheme of delegation were rewritten by the people who use it most, which single threshold would move, and what would that unlock?

Design — what we build

What is the smallest thing we could sign this quarter under an authority we already hold, without asking anyone for an exception?
Which of our policies deserves a worked example attached to it, and what would the first example be?
Who here is expert in what we are allowed to do, and how could we make it easier for people to ask them early rather than late?

Destiny — how it holds

What would have to be true for an agreement signed this year to be found, read and used by someone who has not yet joined us?
When we review this policy next year, how will we know which sentences people have actually been relying on?
What is the first sign we would see that a good result had quietly failed to reach an agenda — and who would notice it first?

WORKS CITED

Alperovitz, G., Howard, T. and Williamson, T. (2010). "The Cleveland Model." The Nation, 1 March 2010.

Butler-Sloss v Charity Commission for England and Wales [2022] EWHC 974 (Ch).

Cabinet Office (2020). Procurement Policy Note 06/20: Taking Account of Social Value in the Award of Central Government Contracts. September 2020.

Cooperrider, D. L. and Whitney, D. (2005). Appreciative Inquiry: A Positive Revolution in Change. Berrett-Koehler.

Cowan v Scargill [1985] Ch 270.

Directive 2014/24/EU of the European Parliament and of the Council of 26 February 2014 on public procurement, Articles 67, 68 and 70.

Efficiency Valuation Organization. International Performance Measurement and Verification Protocol (IPMVP), Core Concepts. Successive editions.

Employee Retirement Income Security Act of 1974, 29 U.S.C. §1104(a)(1).

Federal Acquisition Regulation, 48 C.F.R. §2.101 (thresholds) and §13.003(c)(2) (prohibition on dividing requirements).

Financial Conduct Authority (2017). Regulatory Sandbox Lessons Learned Report. October 2017.

Harries v Church Commissioners for England [1992] 1 WLR 1241.

Kay, J. (2012). The Kay Review of UK Equity Markets and Long-Term Decision Making: Final Report. Department for Business, Innovation and Skills.

Law Commission (2014). Fiduciary Duties of Investment Intermediaries. Law Com No 350. HMSO.

Law Commission (2017). Pension Funds and Social Investment. Law Com No 374. HMSO.

Manley, J. and Whyman, P. (eds) (2021). The Preston Model and Community Wealth Building: Creating a Socio-Economic Democracy for the Future. Routledge.

Occupational Pension Schemes (Investment) Regulations 2005, SI 2005/3378, regulation 2(3).

Ostrom, E. (1990). Governing the Commons: The Evolution of Institutions for Collective Action. Cambridge University Press.

Procurement Act 2023 (c. 54).

Public Services (Social Value) Act 2012 (c. 3).

Uniform Law Commission (2006). Uniform Prudent Management of Institutional Funds Act (UPMIFA), §3.

Uniform Guidance, 2 C.F.R. Part 200, §§200.1 and 200.320.

U.S. Department of Labor (2022). Prudence and Loyalty in Selecting Plan Investments and Exercising Shareholder Rights, final rule, 87 Fed. Reg. 73822 (1 December 2022), codified at 29 C.F.R. §2550.404a-1.

Note on figures. The delegation ladder, the break-even markup, the delegation ratio, the social-value weighting, the tiebreaker discount and the review-gate calendar are all computed in lib/verify/I_09.py and are reproducible there. Parameters that are assumptions rather than published facts — per-layer pass probability, papers deadline, discount rate, manager turnover — are printed as ASSUMED by that module, with the institutional record from which the reader should substitute their own. Thresholds are quoted from the cited regulations and should be re-checked against the current text, which is revised on a schedule.