Haute Lumière

Commerce · VI.06 · MMXXVI · daylight

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

A man in a pale suit walking through a glass-roofed hall of plants, the low sun behind him.
Plate VI.06 · Workbook — the executiveThe Hall, Before the Count.A vote is a claim about who decided. A count is the only thing that can test it, and somebody has to stay late to do the counting.

WORKBOOK — THE CORPORATE EXECUTIVE

Chapter VI.06 · Participatory Budgeting, Measured

For the executive. Participatory allocation applied to a P&L, a capital committee and a business unit — where the firm's own numbers already support it, what it costs, and the two figures that turn it from an engagement programme into a governance instrument the board can approve.

THE PREMISE, STATED COMMERCIALLY

Your firm already runs participatory allocation. It calls it something else.

The innovation fund whose shortlist is voted on. The site improvement budget the plant manager lets the shift leads divide. The charitable matching pot, the capex wish list, the internal marketplace for engineering time. Somewhere in your organisation, a sum of money is allocated by a group rather than by a single signature, and nobody has computed what fraction of your spending it represents or how much of it actually got spent.

Those are the two numbers from the chapter, and they work identically inside a firm:

        s = P / B          the participatory share
        x                  the execution rate
        E = s · x          what was actually decided AND spent

The public-sector range is instructive precisely because it is so wide. Porto Alegre ran at 21 per cent of a municipal budget. Scotland's councils reached 1.4 per cent, Paris 1.04 per cent, New York City's eighth cycle 0.0437 per cent — one part in 2,286. Same phrase, four instruments.

Your firm is almost certainly at the New York end and describing itself at the Porto Alegre end. That gap is not a communications problem. It is a governance measurement you can close this quarter.

PART ONE — DISCOVERY

Days 1–30: find the allocation that is already collective

Exercise 1.1 — The inventory (half a day, with your FP&A lead)

List every budget line in the firm that is allocated by a group rather than by one authorised signatory. Five places to look:

  1. Innovation and improvement funds. Almost always group-allocated, almost never measured.
  2. Site and facilities budgets delegated to plant or floor level.
  3. Community and matching funds, including the ones employees nominate.
  4. Capital wish lists where the shortlist is built by a cross-functional group before the committee sees it. The shortlist is the decision. Count it.
  5. Discretionary team budgets — training, tooling, small equipment.

For each, record P for the last full year.

Exercise 1.2 — The share (2 hours)

Sum P. Divide by total operating and capital spend, B. Write s to four decimal places and as one part in n.

Do not react to the number yet. Establish it. Most executives who run this exercise find a figure between 0.1 and 2 per cent, and find that nobody in the firm had ever added the lines together.

Exercise 1.3 — The appreciative case file (half a day)

Before anyone evaluates anything, find five things these funds bought that demonstrably worked — a fixture on a line, a piece of test equipment, a training course, a fix to a process everyone hated. Cost the benefit where you can.

This is the file you will need in Part Three, and it is the file that makes the proposal an extension of something that works rather than an experiment.

PART TWO — THE ARITHMETIC

Days 31–45: the two numbers, and what they cost to produce

Exercise 2.1 — The execution rate (1 day)

For each line, what fraction of what was allocated was actually spent within the period? Then E = s · x.

The public case is the warning. Porto Alegre budgeted R$338.1 million for public investment in 2005 and spent R$109.4 million — an execution rate of 32.4 per cent — while completion fell from 97 per cent under previous administrations to 43 per cent, and to 26 per cent in the priorities residents ranked first. Carry the 21 per cent nominal share forward and the executed share was 6.8 per cent. Nothing was abolished. The meetings met.

Inside a firm the same thing is called an underspend and is often reported as a saving. It is not a saving if it was allocated by people you asked to allocate it. It is a withdrawal of a decision right, and the people who made the allocation will read it exactly that way next cycle.

Exercise 2.2 — Cost to run (half a day)

Cost your allocation process honestly: facilitation time, the platform, the meetings at loaded rates, the finance time to verify. If the firm cannot say, build a model and label every line assumed, as the chapter does — a four-person unit at €62,000 loaded (€248,000), outreach (€130,000), platform and translation (€50,000), venues and counting (€40,000), verification (€32,000): €500,000 a cycle.

Across 118,308 participants that is €4.23 each and 1.25 per cent of a €40,000,000 pot — 80 euros placed under collective decision per euro of process cost. Scale to your headcount and your pot; the ratio is the figure that goes in the paper, not the absolute.

Exercise 2.3 — The comparator, so the board can calibrate (1 hour)

A fully sorted, fully paid deliberative body — the Convention Citoyenne pour le Climat — cost €5,900,000 for 150 seats: €39,333 per participant, 9,307 times the per-head cost of an open process. Recruitment alone ran €1,883 per stratified seat and citizens were paid €86.04 per sitting day.

These are the two ends of a real market. Breadth is cheap; representativeness is expensive; the firm should buy whichever one the decision actually requires, and say which in the paper.

PART THREE — DESIGN

Days 46–60: the instrument and the board paper

Exercise 3.1 — Write the line, not the programme (half a day)

Convert the strongest existing fund into a governed instrument with five terms:

Exercise 3.2 — Price the tier (2 hours)

At the published public-sector rates, a stratified paid proposal tier costs:

  recruitment, 60 seats at EUR 1,883         EUR 112,980
  stipend, 60 x 8 days at EUR 86.04          EUR  41,299
  care and travel fund                       EUR  12,000
  ------------------------------------------------------
  total                                      EUR 166,279
  as a share of a EUR 40,000,000 pot               0.42 %
  increase on the cycle cost                      33.3 %
  cost per participant, after                EUR    5.63

Inside a firm most of this is already paid — your people are on salary — so the real cost is release time and the care fund. Say so in the paper; it makes the number smaller and the argument stronger.

Exercise 3.3 — Where the firm's own numbers already support it (half a day)

Three places the commercial case is usually already made and never written down:

  1. Capital productivity. Compare realised benefit per pound on group-allocated small capital against committee-allocated small capital. Group allocation frequently wins on items under a threshold, because the people who use the equipment know which one is broken.
  2. Retention and voice. Your engagement survey already asks whether people have influence over their work. Cross it against turnover by unit. The correlation is usually there and usually unexploited.
  3. Underspend recovery. The carry-forward converts a chronic year-end underspend into a delivered asset. Quantify last year's lapse; it is often larger than the entire process cost.

PART FOUR — DESTINY AND DELIGHT

Days 61–90: make it survive you

Exercise 4.1 — Put s and x in the pack (2 hours)

Anything reviewed monthly persists; anything reviewed by exception does not. Get both numbers onto the standing management pack this quarter, with the definition printed beside them so a successor cannot redefine the denominator quietly.

Exercise 4.2 — Name the second owner (1 hour)

One executive is a hobby; two is a practice. Recruit a second owner from finance rather than from people or communications, because the instrument is a budgeting instrument and it will be defended in budget language or not at all.

Exercise 4.3 — Go to a handover (2 hours)

Attend the moment a group-allocated purchase arrives on a floor. Watch what happens. The chapter's claim about why these mechanisms survive is that people walk past the results, and you cannot assess that claim from the pack.

THE FAILURE MODES, NAMED

So you can see them coming

The flattering denominator. A share stated without its denominator is not a measurement. New York's process is 0.0437 per cent of total spending and 0.240 per cent of annualised capital spending — five and a half times apart, same money. Fix the denominator in the definition and print it every time.

The hollowing. The meetings continue and the money stops. It presents as an underspend and reports as a saving. The carry-forward and the published execution rate are the only two defences, and they are cheap.

The tier capture. Breadth at the ballot, narrowness at the proposal stage. In New York's third cycle 39 per cent of voters had household incomes below $35,000 against 21 per cent of local-election voters — a wider door than the election — while the proposal stage was dominated by already-organised groups. In a firm this is the same shape: everyone votes, and the same six enthusiasts write the options.

The dose error in reverse. Do not claim the public-sector outcome evidence. Brazil's health results were measured at 17 to 21 per cent intensity, about 480 times New York's. Claim what your own arithmetic supports: better small capital decisions, recovered underspend, and measured voice.

THE EVIDENCE BASE, FOR THE SCEPTIC ON YOUR BOARD

You will be asked whether any of this is real outside a Brazilian city council, and the honest answer is more useful than the enthusiastic one.

What the strongest study actually did. A panel of 3,651 comparable areas and 47,707 area-year observations across Brazil, 1990 to 2004, of which 228 adopted participatory budgeting at least once. Adoption was a mayoral choice, reversible every four years, which produced variation in whether and in when — and that variation is what made estimation possible. Adopting municipalities moved 2 to 3 percentage points of budget into health and sanitation and registered 1 to 2 fewer infant deaths per 1,000, against a Brazilian national rate of 48 per 1,000 in 1990. The effect held with controls for the governing party, and an independent team reached the same direction on the same country by a different method.

What its author says it cannot carry. Adoption was not randomly assigned. Matching assumes selection on observables only. Unobserved differences between adopting and non-adopting places remain possible.

Why that matters commercially. It is exactly the evidentiary position you occupy with most internal initiatives: a large, well-instrumented, non-randomised comparison, with a known and stated limit. A board that accepts that standard for a pricing change should accept it here, and a board that does not should be told plainly that the randomised version has never been run by anybody.

And the scale caveat, offered before anyone else raises it. Between 11,690 and 11,825 processes are documented worldwide across 71 countries — and the great majority allocate well under one per cent of a budget. Portugal's national programme, the first in the world, allocated €5,000,000 on 119,703 votes: €41.77 per vote, 49 cents per resident. This is a widely adopted instrument operating almost everywhere at low intensity. That is an argument for being precise about what you are buying, not an argument against buying it.

THE BOARD PAPER, IN ONE PAGE

  1. What we already do. P, the lines, the five things they bought.
  2. What it is, measured. s, x, E, with the denominator defined.
  3. What it costs. Cycle cost, cost per participant, and the ratio of money governed to money spent governing.
  4. What we propose. A percentage line, an execution covenant, a carry-forward, a stratified tier, audited annually.
  5. What it costs to fix the fairness problem. 0.42 per cent of the pot.
  6. The number that decides it. E, printed beside the variance, annually.

THE INSTRUMENT, IN THE FORM A TREASURER RECOGNISES

Name. A ring-fenced collective allocation line with a published execution covenant.

Structure. A standing percentage of the unit's controllable spend, appropriated annually, allocated by an open vote from a shortlist written by a stratified proposal tier, with unspent balances rolling inside the line.

Balance-sheet treatment. Items are capitalised or expensed exactly as they would be under any other allocation route. Ask for no special treatment; asking for one invites the finance function to treat the whole instrument as special pleading. The single accounting decision that matters is the carry-forward, which must be a rolling reserve rather than a lapsing appropriation.

Counterparty. Internal audit, certifying two arithmetic quantities annually. Not a consultancy, and not the engagement function.

First ninety days. Days 1–30, the inventory and s. Days 31–45, x, E and the cost per participant. Days 46–60, the line, the covenant and the priced tier. Days 61–75, the audit sign-off on both definitions. Days 76–90, s and x in the standing pack for the first time.

The number that decides it. E = s · x, published beside the variance. If E is not printed, the firm is running consultation. If it is printed and rising, the firm is running distributed governance, and the difference is worth precisely what the two numbers say it is worth.

APPRECIATIVE QUESTIONS FOR YOUR LEADERSHIP TEAM

  1. Where in this firm has a group already made a better allocation decision than a committee would have, and what did they know that the committee did not?
  2. If s were printed in the pack every month, what number would we want it to be in three years, and what would we have moved to get there?
  3. Who in this business has never been asked to allocate anything, and what would we learn in the first hour if they were?