I lead a youth organization in Toronto. Nothing in this proposal funds it, or any organization — it names no provider, and I am not seeking to deliver under it. The model is built as one buyer, one published price, one public record, with no preferred providers. My interest is in the policy: how the city buys, and whether the record is published.

Section 1 · The problemMoney without a promise

Ask what the city does for young people and you will get a long list: after-school programs, library hubs, recreation, summer jobs, safety grants, youth employment. The list is real. What Toronto does not have is a single system — one buyer, one count, one promise a family can rely on. The spending that chases one effect moves through 16 named streams, at least 10 funding streams, and 8 authorities. Each runs its own application cycle, its own eligibility rules, its own reporting format, and its own definition of success. No two streams count the same thing. The honest total is about $29–35 million a year, and only $17–22 million of that can be pointed to in a published figure.

Three kinds of numbers — do not mix them. A guaranteed place: one child is promised a spot in a program for the whole program year — a set number of hours, a set number of children per staff member, a program that must exist whether or not any particular child shows up. The place exists because a community organization is under contract to run the program — its staff, its site, its hours — in a school, a community centre, or its own space. A registration: a child signs up for a course; a registration can be a few hours long, and one child can hold many. A drop-in visit: a young person walks into a space that is open and is counted once for that visit. A registration is not a place. A visit is not a place. Only a contracted place — named child, known hours, a staffed ratio, a price — is a promise a family can plan a work schedule around. Most of what Toronto buys today is the second and third kinds. That is not a criticism of the programs. It is a description of the promise.

Section 7 · Three scalesThe same coverage rate, three geographies

About one child in eight — that is the rate this section works from. Toronto has 234,408 children aged 5 to 13 (City of Toronto / Statistics Canada, 2022). New York's number is a rate, not a count: its 102,104 places against 830,714 resident children aged 5 to 13 (US Census Bureau, 2024) give a coverage rate of 12.29% — about one child in eight. That rate is what "New York-equivalent" means for a smaller city.

All three at the modelled blended price of $3,680 per place. Option (a) rests on an assumption, and it is called one every time it appears: that 30% of Toronto's children aged 5 to 13 live in the priority neighbourhoods. No government publishes that share. At 25%, the priority system is 7,203 places and $26.5M a year. At 35%, it is 10,084 places and $37.1M a year. The first job of a design phase is the census count that replaces the assumption with a count. Phase-in: Years 1–3 build the priority system at 8,643 places and $31.8M, with the mid-course evaluation published; Years 4 and 5 move to citywide half-scale at $53.0M only if Council chooses — a decision informed by the Year 3 evaluation, not an escalator built into this plan.

Section 8 · The priceBuilt in the open — a model, called a model

The price in this plan is a model. It is called a model every time it appears, because nobody has tendered it and pretending otherwise would hand critics their easiest win. What a model can do — what this one does — is print every assumption, so anyone can change one and watch the price move. Front-line pay: $22.00/hr elementary; $23.00/hr middle (the City's own ARC (After-School Recreation Care) site-coordinator posting rate; program leaders post at $19.76). Program days: 194.

The build, step by step

Start with what a place is: one elementary-aged child with a guaranteed place in a program a community organization runs under contract, for the whole school year — 596 hours of programming across 194 school days, with one staff member for every 10 children.

People are most of the price. Front-line staffing is $1,840 — each child carries one-tenth of the staff cost for those hours, at wages benchmarked to the City's own ARC grid. The coordinator who runs the site adds $314; the director above them, $132.

Then the day-to-day costs. Food is $340 — a snack at $1.75 per child per day. Supplies are $150. The room is $200, charged at school permit rates — cheap, not free, and not market rent. Together with the staffing lines, the direct cost is $2,975.

Two layers sit on top. The provider's administration, at 12%, adds $357. The City's system layer, at 6%, adds $200 — the buyer unit, the monitors, the data system, the evaluation. The school-year price of one elementary place: $3,532.

The tiers quote $3,680 instead — because that figure is a blend. About two-thirds of places are elementary, and half of those add a summer block of 240 hours: $1,389 more, for a year-round elementary price of $4,921. About one-third are middle-band places at $2,604, at 540 hours and one staff member to 15 children. Every tier is costed on the blended $3,680.

Section 9 · The ledger and the verdictComparable money, categorically more guaranteed output

Before anyone asks whether Toronto can afford a guaranteed system, count what Toronto already spends chasing the same effect. Published — strictly published, like-for-like components only: $17.3–22.3M — printed both ways because no City document states whether the Summer Safety Plan's $5M and SafeTO's $5M youth-violence-prevention component are the same $5M counted once or two counted separately; neither is presented as resolved. Modelled — published plus two labelled modelled components (Youth Hubs at ≈$9.2M; Toronto's share of the Ontario After School Program at $2.2–3.7M): ≈$29–35M. Where the priority system finds its money: Parks, Forestry and Recreation youth envelope conversion ~$12M; Ontario After School Program formalized Toronto allocation ~$3M; SafeTO and CPIP alignment ~$5M; federal streams $0 in the base case; new City money ≈$12M — about 0.06% of the City's $18.9B operating budget. One exclusion is stated flatly: Children's Services' $1,672.5M is not in this stack and cannot be — it is child-care money flowing under federal-provincial rules. Stress the numbers and the verdict bends without breaking: under harsh assumptions the priority system rises toward ~$42M against ~$20M of provable spend — roughly double, for a guaranteed system against an unguaranteed patchwork. In no version does consolidation pay for itself. Anyone who claims it does is overclaiming, and this plan does not.

For roughly what Toronto already spends — money that today guarantees a place to 1,800 children — the City can guarantee 8,643 places, publish the attendance, pay the providers on time, and be audited on all of it; the evidence supports exactly that promise, and this proposal makes no promise the evidence doesn't support.

The benefitsWho the plan serves

Section 3 · The evidenceNew York buys after-school by the place

New York City does not run after-school programs. It buys them — the way it buys road repair: a written specification, a published price, a contract, an inspection. COMPASS, the Comprehensive After School System of New York City, is New York's after-school system for children from kindergarten to Grade 8 — free to the family, every school day. The buyer is DYCD, the city department for youth and community programs. DYCD runs no programs itself: it writes the specification, runs a competition, and signs contracts with community-based organizations to deliver. The current contracts run six years (2026–2032). The system began in 2005 as Out-of-School Time, at US$46 million, and was renamed COMPASS in 2014. It now comprises more than 820 programs, at about US$447.6 million a year for 102,104 guaranteed places. Elementary contracts require 1,020 hours of programming a year, and staff-to-child ratios are set in the contract. Four mayors carried this system — Bloomberg created it, de Blasio scaled it, Adams funded the expansion, and Mamdani is executing it. A baselined contract system survives elections, because each successive budget re-enacts it.

SONYC, the middle-school model inside COMPASS (Grades 6 to 8), launched in 2014 as the centrepiece of a US$145 million expansion. Its contract requires 540 hours of programming a year at a published price of US$3,900 per child. AIR (the American Institutes for Research) evaluated the expansion year independently: registration went from 18,702 young people to 58,745 in one year, participants attended 13 million hours in total — an average of 236 hours per participant — and program quality was rated high across 2,300 hours of direct observation. In the same evaluation, 98% of surveyed parents said their child liked the program. That last figure needs its label: it is perception-grade evidence. It proves families valued the service; it proves nothing about outcomes, and satisfaction must never be allowed to stand in for impact. New York also got documented things wrong: it pays its providers late, badly; children attend about half the hours their places offer; and its re-bidding rounds have cut loose the neighbourhood organizations the system depends on. This plan designs those flaws out, using fixes Toronto has already proven somewhere in its own system: payment in advance, the way the City already pays its child-care agencies; contract sizes small organizations can hold; a set price with competition on quality; published results as a contract term.

This plan does not promise test scores. No system-scale evaluation anywhere has shown that after-school provision raises them. New York's own evaluator found participants' test-score gains no different from matched non-participants. What structured programs with trained, stable staff do improve, on the evidence, is engagement, school-day attendance and course progression. Those are the outcomes the funded evaluation will measure and publish.

This plan does not promise a percentage reduction in youth violence from after-school places, because no such estimate exists. What the record does establish is timing: juvenile violent crime and victimization peak in the risky hours — 3 to 4 p.m. on school days. Supervision in those hours is the intervention aimed at them. Aim is not an effect size, and this plan does not dress it up as one.

The summer-jobs evidence is stronger, and this plan claims it at exactly its strength, and only for older youth. Assignment to a supported summer job cut violent-crime arrests by roughly two-fifths over 16 months in the original Chicago trial (Heller, 2014). Later and larger trials found smaller, less certain reductions, and across all the studies together, the effect after the program ends could be zero — so what holds up is the direction, not the size. New York's lottery evidence found incarceration about 10% lower across all applicants — about 54% only among 19-to-21-year-olds, a qualification that travels with the figure every time — and deaths from external causes about one-fifth lower (Gelber, Isen and Kessler, 2014). None of it is evidence about ten-year-olds, and none of it is presented here as if it were.

Attendance is promised at levels that can be proved. Children attend roughly half the hours a place offers — in New York's evaluated middle-school expansion, an average of 236 of 540 hours. So the contract sets an 80% participation benchmark, publishes attended hours, and is judged on them. The arithmetic is honest about the gap between an offer and the hours a kid actually showed up for: $5.60 per contracted hour, and about $7 per attended hour at the benchmark the contract enforces. This is not a savings claim: comparable money, categorically more guaranteed output, and under harsh assumptions roughly double the strictly published spend. The case rests on what the money buys, not on a promise that it costs less.

Two labels stay on, in plain sight. The price is a model — $3,680 per place, built line by line from published wage postings in 2026–27 dollars, not a tendered or observed price. And the 30% priority share is an assumption — no government publishes it, the sensitivities at 25% and 35% are printed in Section 7, and replacing it with a census count is the design phase's first job. Toronto has adopted this idea before, in words: the 2012 Middle Childhood Strategy named the same patchwork, in the same language, and stopped at eighteen sites. What this design adds is not another diagnosis. It is a named buyer, a costed unit, committed money, and an owner whose name is on the results.

The companionWhat would let Toronto answer

Four published measures would let Toronto answer, each year, whether its young people are better off. None of the four requires a new program. Each is a field, a series, or a total that an existing publisher could add to what it already releases.