La Bourse / Volume VI / Nº VI.06 / Ten concept briefs

A man in a pale suit walking through a glass-roofed hall of plants, the low sun behind him. — Plate VI.06 · Ten concept briefsThe 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.

TEN CONCEPT BRIEFS · Chapter VI.06 — Participatory Budgeting, Measured

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

BRIEF 1 — The Participatory Share

The idea. The question is never did people participate. It is how much of the money did they decide.

        s = P / B

  P = the sum actually allocated by the participatory process, per period
  B = total public spending by the same authority, same period

Turnout is a fact about the room. The share is a fact about the money, and the two can move in opposite directions without anybody noticing.

Worked example. New York City's eighth participatory budgeting cycle put $39,000,000 on the ballot and 118,308 people voted — a genuinely impressive room. The city's fiscal 2019 adopted expense budget was $89,160,000,000. So s = 0.0437%, or one part in 2,286. Porto Alegre in 1999 was at 21 per cent. Same words, two instruments 480 times apart in intensity.

Why it matters. It is the one number that separates governance from consultation, and almost nobody prints it. Any authority can compute it in an afternoon, which is why its absence is informative.

You already know this because you have sat on a committee with a real agenda and a committee with a briefing pack, and you could tell the difference within ten minutes without anyone explaining the constitution.

BRIEF 2 — The Execution Rate, and the Executed Share

The idea. A budget line is a promise. Spending is the event. The gap between them is where a participatory process goes to die quietly.

        E = s · x

  s = the nominal participatory share
  x = the execution rate — what fraction of allocated money was spent

Worked example. Porto Alegre, 2005: R$338.1 million budgeted for public investment, R$109.4 million spent. x = 32.4%. In 2006, R$256.6 million against R$121.0 million — 47.2 per cent. Project completion fell from 97 per cent under the earlier administrations to 43 per cent, and to 26 per cent in the top five priority sectors residents had ranked first.

Carry the 1999 nominal share of 21 per cent forward as an upper bound and apply the 2005 execution rate: the executed share was 6.8 per cent — a fall of fourteen percentage points with no assembly cancelled.

Why it matters. This is the failure mode no participation statistic can see. A process can be dismantled without ever being closed. Publish x quarterly, by district, or you will be the last to know.

You already know this because you have watched a budget get approved and then underspent, and you know that the underspend was the decision.

BRIEF 3 — Door and Tiers

The idea. Participatory budgeting has two populations: everyone who votes, and the much smaller group who write what gets voted on. They are demographically different, and almost all reporting describes only the first.

Worked example. In New York's third cycle — 10 districts, 16,642 voters, over $14,000,000 — 39 per cent of participatory voters reported household incomes below $35,000 against 21 per cent of voters in the 2013 local elections. Eleven per cent identified as Asian against four; 24 per cent Hispanic or Latino/a against 14; 66 per cent women against 56. The door was wider than the election.

The tier was not. The same research found the proposal stage dominated by already-organised groups, including parent-teacher associations from higher-income neighbourhoods arriving with an agenda. In Porto Alegre, district assemblies drew disproportionately from the poor while the elected budget councillors — 92 seats, one per 14,130 residents — ranked above the city average on education and income.

Why it matters. A process can report a broad, representative turnout and still route money toward the already-organised. Measure both populations or you are measuring the wrong one.

You already know this because you have been in an organisation where everybody was consulted and four people decided.

BRIEF 4 — Stratified Sortition

The idea. Invite at random, then select so that the group matches the population — age, sex, education, income, geography — against census data. It is the method the OECD documented across 289 deliberative processes, and it is the only known way to make a small decision-making body demographically honest.

Worked example. The Convention Citoyenne pour le Climat spent €282,423 recruiting 150 stratified seats: €1,883 per seat. That figure is published, which means it can be budgeted rather than guessed.

Why it matters. Open participation self-selects for the organised; lottery selection does not. Applied to the proposal tier rather than the ballot, it fixes the gradient exactly where the gradient is, and leaves the wide door alone.

The limit. Acceptance rates for unpaid invitations run 2 to 5 per cent, and the people who decline are not a random sample either. Sortition without pay imports the same bias through a different door — which is Brief 5.

You already know this because you accept that a jury of twelve strangers can decide a serious matter, and you would not accept a jury made of whoever volunteered.

BRIEF 5 — Paid Participation

The idea. Unpaid civic time is a regressive tax. Whoever can afford the evenings decides.

Worked example. The Convention Citoyenne paid the French jury-service rate, €86.04 a day, coming to €1,462.68 per citizen across seventeen sitting days. German-speaking Belgium's permanent citizens' council pays €155 a day and reports an invitation-acceptance rate of 10 per cent, against the 2 to 5 per cent typical of unpaid processes — two to five times the acceptance.

Why it matters. That multiple is the whole argument. It is the difference between a shift worker being able to serve and not, and it costs less than the printing.

Priced for a real process. Sixty stratified seats, eight paid sitting days: 60 × 8 × €86.04 = €41,299. Add recruitment at €1,883 a seat — €112,980 — and a €12,000 care and travel fund, and the whole repair is €166,279, or 0.42 per cent of a €40,000,000 pot.

You already know this because you have watched an unpaid committee slowly fill with the retired and the salaried, and it was nobody's decision.

BRIEF 6 — Cost Per Participant

The idea. A governance process has a unit cost, and stating it converts an argument about democracy into a line a treasurer can approve.

Worked example — a modelled city cycle. A four-person unit at €62,000 fully loaded (€248,000), outreach partners (€130,000), platform and translation (€50,000), venues and counting (€40,000), verification and evaluation (€32,000): €500,000 a cycle. Across 118,308 participants that is €4.23 per participant, and 1.25 per cent of a €40,000,000 pot — 80 euros of public money placed under public decision for every euro of process cost. Every input in that model is an assumption and is labelled as one in the chapter's figures module.

Worked example — a published comparator. The Convention Citoyenne cost €5,900,000 for 150 seats: €39,333 per participant, or 9,307 times the cost per head of the open process.

Why it matters. Both are defensible. They answer different questions — one buys breadth cheaply, the other buys representativeness expensively — and a board paper that does not distinguish them will fund the wrong one.

You already know this because you already price a customer acquisition and a board seat differently, and you would not be surprised that one costs four figures more than the other.

BRIEF 7 — Treatment Intensity, and Why Evidence Does Not Travel

The idea. An effect measured at one dose is not evidence at another dose.

Worked example. The Brazilian health findings come from processes carrying 17 to 21 per cent of a municipal budget. New York's cycle 8 carried 0.0437 per cent — 1/480th of the intensity. Quoting the infant-mortality result in support of a 0.04 per cent process is an argument about a different treatment.

The underlying result, stated properly. Gonçalves finds adopting municipalities moved 2 to 3 percentage points of budget into health and sanitation — reported as 20 to 30 per cent of the 1990 category mean, which recovers a mean of 10 percentage points, consistently from both ratios — and registered 1 to 2 fewer infant deaths per 1,000, or 5 to 10 per cent of a base rate of 20 per 1,000, against a Brazilian national rate of 48 per 1,000 in 1990.

Why it matters. Citing a strong result at the wrong dose is the fastest way to lose a sceptical finance director permanently.

You already know this because you would not quote a clinical trial's dosage results while prescribing a four-hundredth of the dose.

BRIEF 8 — What the Identification Can and Cannot Carry

The idea. Know what kind of evidence you are holding before you spend it.

The design. Gonçalves uses a panel of 3,651 comparable areas and 47,707 area-year observations, 1990 to 2004, of which 228 adopted participatory budgeting at least once. Adoption varies in whether and in when, which is what makes estimation possible. Fixed effects for area and year; matching on observables; restriction to adopters only; party controls.

The limit, in the author's own words. Adoption was a mayoral decision, not a random assignment. Matching assumes selection on treatment is based on observables alone. Unobserved differences between adopting and non-adopting places remain possible, and the paper says so.

What it can carry. That participatory budgeting is associated with a real, robust, party-independent shift of budget toward sanitation and health, and with a real fall in infant mortality, at Brazilian intensities, across a very large panel — corroborated independently by Touchton and Wampler on municipalities using the process eight years or more.

What it cannot carry. A causal claim of the strength a randomised trial would give, and any claim at all about a process governing a fraction of a per cent of a budget.

You already know this because you already distinguish between a controlled pilot and a well-instrumented rollout, and you price them differently.

BRIEF 9 — The Diffusion, and Its Denominators

The idea. Participatory budgeting is, on any honest count, the most widely adopted democratic innovation of the last forty years — and the count itself needs reading carefully.

The figures. The 2019 Participatory Budgeting World Atlas reports between 11,690 and 11,825 processes across 71 countries: Europe 4,577–4,676, South America 3,061–3,081, Asia 2,773–2,775, Africa 955–958, North America 178, Central America and the Caribbean 134–142. Peru alone runs 1,869 by statute — 61 per cent of South America's total, which is what the arithmetic gives where the summary says "two-thirds". Japan accounts for about 15 per cent of the global total, roughly 1,754 processes. The World Bank put over $280,000,000 behind related programmes between 2002 and 2016.

The denominator note. The continental columns sum to 11,678–11,810 against the atlas's own stated totals — 12 and 15 cases unaccounted for. This is a small discrepancy in a heroic piece of counting, and it is worth printing because it is the same discipline the chapter asks of everyone else.

And the uncomfortable finding. About 85 per cent of cases occur in countries with immature democracies, leaving roughly 1,754 in full ones. A process can legitimate as easily as it can empower.

You already know this because you have seen a consultation used as evidence of consent, and you knew at the time which it was.

BRIEF 10 — The Statutory Line

The idea. A participatory budget that depends on an administration's enthusiasm has the half-life of that administration. A statutory one does not.

Worked examples. Sicily has required its municipalities to put at least 2 per cent of their budgets through participatory allocation since 2014. Peru runs 1,869 processes under national regulation. Poland's 2018 statute made it mandatory for 66 city governments. Portugal became the first country in the world to run a national participatory budget — €3,000,000 in 2017 and €5,000,000 in 2018, on 119,703 votes, which is €41.77 per vote and 49 cents per resident. Scotland's councils agreed a 1 per cent target with COSLA and reached 1.4 per cent — £154,000,000 decided by more than 110,000 people, £1,400 per participant.

Why it matters. A percentage written in statute survives inflation, survives an election, and is much harder to quietly stop executing than a discretionary fund. Note that the national programmes are the smallest per resident — Portugal's 49 cents against Paris's €45.66 — because national scale buys legitimacy, not intensity.

You already know this because you know the difference between a policy and a covenant, and you know which one your lenders read.