Figures

60

Median hawker age (MSE, Nov 2024)

official2024

Hawker culture's practitioners are among the oldest workers in Singapore, and the Minister restated the median age of 60 in November 2024.

Sections: §1.1 · Hawkers and Social Protection, C.5 · Sources: Response to Motion on Hawker Culture — Dr Koh Poh Koon · Primary source

87.3%

Patrons who would not want to be a hawker (NEA survey, 2019)

official2019

Almost nine in ten hawker-centre patrons surveyed by NEA in 2019 said they would not want to be a hawker, citing no interest and long hours.

Sections: §1.1 · Sources: Workgroup on Sustaining the Hawker Trade Report · Primary source

Lowest 20%

Typical hawker income, the lowest 20% (MSE, Nov 2024)

official2022

A typical hawker's income sits in the bottom fifth of earners; Follow the Plate's model puts a modelled estimate behind this cell.

Sections: §1.1 · Follow the Plate, D.4 · Method, Levers and Data, F.2.3 · Sources: Response to Motion on Hawker Culture — Dr Koh Poh Koon · Primary source

20–30¢

Typical hawker margin per dish (NEA Workgroup, 2020)

official2020

Hawkers typically clear 20 to 30 cents a dish; patrons interviewed in 2018 guessed $1.50.

Sections: §1.1 · Sources: Workgroup on Sustaining the Hawker Trade Report · Primary source

2.8%

Of 566 who finished HDP's classroom training, 16 run a stall in a hawker centre (MSE, Jul 2024)

official2024

Of the 566 who finished the Hawkers' Development Programme's classroom training, 29 started a stall in a hawker centre and 16 still ran one in April 2024: 55% of those who started.

Sections: §1.1 · §1.2 · §3.5 · Sources: Parliamentary reply: HDP outcomes (566 trained, 120 through apprenticeship)

No list

Govt keeps no register of at-risk hawker trades (MCCY, 7 Oct 2026)

official2026-10-07

The Government confirmed in Parliament that it keeps no list of hawker trades or culinary practices at risk of disappearing.

Sections: §1.1 · §3.6 · Hawkers and Social Protection, C.5 · Sources: Written reply to PQ 3030: safeguarding and transmission of traditional hawker trades · Primary source

$19.68M

4-year programme

derivedproposed2026-10

The net funding requirement for Year 0 to Year 3 at three pilot centres, summed from the four ledgers in Part VII.

Sections: §the-plan · §7.1

$40M

endowment target

derivedproposed2026-10

The Hawker Endowment's target by the end of Year 2.

Sections: §the-plan · §7.6

44

Master Fellows

derivedproposed2026-10

Master Fellowships funded across Years 1–3 at the three centres (10, 16 and 18): one for each heritage-worthy stall in each centre's planning count, a ceiling (27 at Chinatown Complex, 8 at Geylang Serai, 9 at Tekka).

Sections: §the-plan · §7.1 · the Budget Book, B.3 · the Budget Book, B.4 · the Budget Book, B.5

88

apprenticeships

derivedproposed2026-10

Apprentices funded across Years 1–3 at the three centres (20, 32 and 36), two for each Master Fellowship.

Sections: §the-plan · §7.1 · the Budget Book, B.3 · the Budget Book, B.4 · the Budget Book, B.5

26

funded landings

derivedproposed2026-10

Graduate Landing Packages funded in Years 2 and 3 (10 and 16), about 30% of each centre's apprentices.

Sections: §the-plan · §7.1 · the Budget Book, B.4 · the Budget Book, B.5

3

pilot centres

derivedproposed2026-10

Two Lighthouses, Chinatown Complex and Geylang Serai, open in Year 1 and Tekka beside them in Year 2, each with the full loop; five at most before the gate criteria.

Sections: §the-plan · §6.3 · §6.5

0

cameras

derivedproposed2026-10

The quality standard uses no cameras, data feeds or league tables.

Sections: §the-plan · §5.1

$906,598

Year 0 net requirement

derivedproposed2026-10

Year Zero at both Lighthouses: the critical-illness pool, listening and co-design, pilot documentation, the stall baseline (a two-week till count at every consenting stall, by dish type), the evaluator's baselines and the Hawker Council's first election.

Sections: §7.1 · the Budget Book, B.2

$4,629,249

Year 1 net requirement

derivedproposed2026-10

Year 1, when the two Lighthouses launch the full loop, Tekka gets its trust deposits and its stall baseline, and the Regulars' List opens at every government hawker centre.

Sections: §7.1 · the Budget Book, B.3

$6,706,257

Year 2 net requirement

derivedproposed2026-10

Year 2, when Tekka opens beside the Lighthouses and their first graduates land.

Sections: §7.1 · the Budget Book, B.4

$7,441,704

Year 3 net requirement

derivedproposed2026-10

Year 3, three centres running the full loop; no part of it waits on the gate.

Sections: §7.1 · the Budget Book, B.5

$19,683,808

The four-year total

derivedproposed2026-10

The programme's total net requirement for Years 0–3 at three centres: $5,430,106 of operating budget, the Seed Bank's small part in the Buying Circles, the national Regulars' List and the stall baseline among it, and $14,253,702 of programme budget that reaches hawkers. The two Lighthouses alone would cost $16,224,542.

Sections: §7.1

$5,586,777

Old Airport Road, the fourth pilot centre, an add-on, over Years 0–3

derivedproposed2026-10

A stated add-on to the three-centre plan, costed at its own size, 168 stalls and 16 heritage-worthy: $274,516 in Year 1 (trust deposits and its stall baseline before it opens), $2,026,842 in Year 2 and $3,285,419 in Year 3.

Sections: §7.1 · §6.3

$4,412,940

Maxwell, the fifth pilot centre, an add-on, over Years 0–3

derivedproposed2026-10

A stated add-on to the three-centre plan, costed at its own size, 103 stalls and 13 heritage-worthy: $213,246 in Year 1, $1,636,325 in Year 2 and $2,563,369 in Year 3.

Sections: §7.1 · §6.3

29,683,525

Four-year programme at five pilot centres, the most before the gate

derivedproposed2026-10

The three-centre plan plus Old Airport Road and Maxwell.

Sections: §7.1 · §6.5

72.4%

Share of spend reaching hawkers directly

derivedproposed2026-10

Each year's ledger splits the operating budget, the institution's own burn, from the programme budget that reaches hawkers, centre by centre; this is the programme budget as a share of the four-year net requirement.

Sections: §7.1

$113,000

Operating burn a month

derivedproposed2026-10

The operating budget over Years 0–3, $5,430,106, spread over 48 months.

Sections: §7.1

about $5.6M

Year-3 spend with the two Lighthouses alone (downside mode)

derivedproposed2026-10

If the endowment campaign stalls at its first close, both Lighthouses keep running and Tekka defers: $5,624,989 in Year 3.

Sections: §7.8

$416,570

The Regulars' List over Years 1–3

derivedproposed2026-10

A coordinator; 800 listing visits over three years, each with a decal, a Registry page and a thank-you note; new decals for stalls that move up a grade; a booklet and map for every government centre; and the listing on the open price board: $132,750, $124,900 and $121,050 before contingency, in the operating budget.

Sections: §3.6 · §7.1

$4.13M

The programme a year from Year 4 at three pilot centres

modelled2026 (model baseline)

From Year 4 each centre funds new Fellowships at a maintenance rate fed by the yearly at-risk count, about 10% of its planning count a year, against about $5.32M from the draw, the operating match and continuing gifts: at the central assumptions no year runs short.

Sections: the Budget Book, B.9.4 · §7.6 · §7.8

$5.7M

Major gifts needed from donors and family foundations to close the $40M endowment on time

modelled2026 (model baseline)

At the central assumptions, this is what major donors must give so that each close (Years 0, 1 and 2) is reached. It is the balancing item.

Sections: the Budget Book, B.9.1 · the Budget Book, B.9.4 · the Budget Book, B.6.1

$1.08

Public money for every private dollar while the endowment is raised (Years 0–2)

modelled2026 (model baseline)

Government carries the campaign. Public money here is the anchor, the match, the operating match, half of the Year 0 NHB grant and the tax forgone.

Sections: the Budget Book, B.9 · the Budget Book, B.9.4

$3.6

Private money for every public dollar in a typical later year (Year 6)

modelled2026 (model baseline)

After the campaign the ratio turns around: in Year 6, $3.43M of private money against $0.94M of public money.

Sections: the Budget Book, B.9 · the Budget Book, B.9.4

121%

The endowment in Year 20, in today's dollars, as a share of its $40M target

modelled2026 (model baseline)

From Year 4 the pilot centres fund new Fellowships at a maintenance rate, so a little under half of continuing gifts is spent and the rest joins the corpus, which ends well above its target in real terms.

Sections: the Budget Book, B.9.2

$0.42M, $1.05M and $1.41M

The endowment's draw in Years 1, 2 and 3

modelled2026 (model baseline)

The model reproduces the draws printed in §7.6: 3.5% of the corpus, then of its three-year average.

Sections: the Budget Book, B.9 · §7.6

59%

Chance the final $40M close is reached by the end of Year 2

modelled2026 (model baseline)

Share of 5,000 modelled futures in which fundraising lands on time, with the adopted investment policy (fixed-income core, 3.5% draw). In the worst tenth the campaign is $3.2M short.

Sections: the Budget Book, B.9.3

71%

Chance the corpus keeps its real value to Year 20: the adopted policy (fixed-income core, draw 3.5%)

modelled2026 (model baseline)

Share of 5,000 modelled futures in which the endowment is still worth at least $40M in today's dollars in Year 20, while diners and partners keep giving.

Sections: the Budget Book, B.9.3 · the Budget Book, B.9.4 · §7.6

34%

Chance the corpus keeps its real value to Year 20 if all giving stops after Year 5, at a 3.5% draw

modelled2026 (model baseline)

The stress test: every diner, partner and company stops giving after Year 5. At 3.5% the corpus keeps its real value in only this share of 5,000 futures.

Sections: the Budget Book, B.9.3 · the Budget Book, B.9.4

45%

Chance the corpus keeps its real value to Year 20 if all giving stops after Year 5, at a 3% draw

modelled2026 (model baseline)

The same stress test with the draw cut to 3%: the reason the draw rule carries a trigger.

Sections: the Budget Book, B.9.3 · the Budget Book, B.9.4

$427 a month

Rent rebate each heritage stall receives from its partner operator (20 hosted stalls)

modelled2026 (model baseline)

Partner operators may meet up to half of their pledge as rent rebates, straight to the heritage stalls they host. Not matched, and no tax deduction.

Sections: the Budget Book, B.9.1 · the Budget Book, B.6.1

$7.56 to $8.86 an hour

What the rent rebate does for an HDB-rented coffeeshop stall's net income per hour

modelled2026 (model baseline)

From $7.97 to $9.27 an hour at about $427 a month; the lift is the rebate spread over the stall's working hours.

Sections: the Budget Book, B.9.1

50, 100 and 250

Endorsements for Silver, Gold and Platinum on the Regulars' List; Bronze, the listing itself, needs 20

assumptionproposed2026-10

Or 10%, 20% and 50% of a stall's weekly orders placed by its regulars, at 1.5 orders a week each, whichever comes first, from Silver up. The unit is the order: a family's takeaway of several plates is one order. Grades only go up and never gate a Fellowship.

Sections: §3.6 · Follow the Plate, D.4.2

800

Stalls on the Regulars' List by the end of Year 3

assumptionproposed2026-10

Ramped from 300 by the end of Year 1 and 600 by Year 2; about 12–15% of the more than 6,000 cooked-food stalls in government centres are assumed to be listed.

Sections: §3.6 · §7.1

120–200

Listed stalls counted at risk by Year 3

assumptionproposed2026-10

Assumed at 15–25% of listed stalls; the count behind the pilot centres flagged 8 of 55 heritage-worthy stalls, a floor. Published only as national and per-centre totals, never naming a stall.

Sections: §3.6

10%

New Master Fellowships a year from Year 4, once a centre's planning count has been offered

assumptionproposed2026-10

Fed by the yearly at-risk count, from hawkers' voluntary signals and public facts: three a year at Chinatown Complex and one each at Geylang Serai and Tekka.

Sections: §7.8 · §7.1

$1,700

Indicative cooked-food band, Class A centre (prime)

derivedproposed2026-10

The proposed fixed rent for a prime-centre cooked-food stall under rent bands and ballot.

Sections: §2.2

$1,150

Indicative cooked-food band, Class B centre

derivedproposed2026-10

The proposed fixed rent for a stall in a Class B centre; also the rent a Heritage Anchor coffeeshop unit would match.

Sections: §2.2 · §2.4

$650

Indicative cooked-food band, Class C centre

derivedproposed2026-10

The proposed fixed rent for a stall in a Class C centre.

Sections: §2.2

$90,750

Total Fellowship value over three years, CPF included

derivedproposed2026-10

Every Master Fellow receives the same package: a teaching year at $2,500 a month ($30,000), a 24-month stewardship at $1,000 a month ($24,000), a $15,000 completion grant, up to $15,000 of health security and employer CPF on the whole stipend, $6,750 at 12.5%, the 2026 rate for a Fellow over 60 (a Fellow under 60 would cost 16–17%).

Sections: §4.2

≈$20,000

Expected cost of the Graduate Landing Package

derivedproposed2026-10

The expected cost of landing each graduate in a stall with rent relief, financed equipment and an income floor.

Sections: §3.4

$2,220 a month

Apprentice pay from July 2026, the Progressive Wage for a food or drink stall assistant ($2,360 from July 2027, $2,500 from July 2028), plus employer CPF

officialproposed2026-10

Seed Bank apprentices do an assistant's work at the master's stall, so they are paid an assistant's wage for all 18 months.

Sections: §3.3 · Sources: Government accepts TCF recommendations on Food Services PWM 2026 to 2028 · Primary source

$80

Hawker's co-pay for a relief cook ($40 for medical leave); the corps member earns the Progressive Wage for a cook for a day of at most 12 hours, with overtime beyond 8, about $212 a day in 2026

derivedproposed2026-10

The programme funds the gap between the hawker's co-pay and the relief cook's day rate.

Sections: §2.7

$15M

Government anchor

derivedproposed2026-10

A one-time Government anchor gift to the Hawker Endowment.

Sections: §7.6 · §7.7

$10M

1:1 matching pool

derivedproposed2026-10

A public pool matching diners' and companies' gifts to the endowment dollar for dollar until the $10M is used.

Sections: §7.6 · §7.7

$15M

corporate & community giving

derivedproposed2026-10

The share of the endowment raised from companies and the public.

Sections: §7.6

$12M

first close at launch

derivedproposed2026-10

The endowment's first close; if the campaign stalls here, both Lighthouses keep running and Tekka defers.

Sections: §7.6 · §7.8

$1.4M

a year thereafter, at a 3.5% draw

derivedproposed2026-10

Once fully raised, the endowment yields about this much a year, kept in step with inflation, at a 3.5% draw from a corpus with a fixed-income core.

Sections: §7.6

$66M a year

Makan Vouchers, indicative and tunable

derivedproposed2026-10

$120 a year for about 550,000 households, on the existing CDC voucher system and outside the programme's books.

Sections: §2.5 · §7.7

$120 a year

Makan Voucher per eligible household

derivedproposed2026-10

The indicative yearly voucher for lower-income households, seniors and students.

Sections: §2.5

about 550,000

households receiving Makan Vouchers

derivedproposed2026-10

About 550,000 of Singapore's 1.49 million resident households would receive the voucher.

Sections: §2.5 · Follow the Plate, D.7.3

50 passes

Apprentice Work Pass pilot, pilot centres and their clusters

derivedproposed2026-10

A pilot quota of foreign-apprentice passes tied to training in Singaporean hawker technique.

Sections: §2.8 · §7.7

$63,000

Goodwill at 15× monthly net (worked example: laksa stall, Class B centre)

derivedproposed2026-10

A laksa stall netting $4,200 a month would be valued at $63,000 of goodwill, paid off in about 50 months at the lesser of 8% of gross and 30% of net ($1,260 a month).

Sections: §4.3

≈ $3,200 / mortgage

Expected first-loss cost, simulated (Succession Economics, E.1.4)

modelled2026-10

The expected cost to the programme of one Succession Mortgage for a stall like the paper's worked example. Averaged across every tendered stall with goodwill worth selling, weaker stalls included, it is about $6,500.

Sections: §4.3 · Succession Economics, E.1.4

2.5%

Family Name Royalty, for ten years at most

derivedproposed2026-10

The share of gross revenue paid to a master's family whenever a business trades under the master's name, for ten years from its first month; the business may buy it out at any time at a published multiple.

Sections: §4.4

about $5,500/year

Royalty paid by a heritage stall grossing $220,000/year

derivedproposed2026-10

A busy heritage stall would pay the master's family this much a year for ten years, about $55,000 in all.

Sections: §4.4

6

Completed Hawkers Succession Scheme transfers (4 in progress; Sep 2025)

official2025-09

Nearly four years after launch, few veterans had used the succession scheme.

Sections: §1.2 · §1.3 · §2.3 · Hawkers and Social Protection, C.5 · Sources: Written reply to parliamentary question on hawker stall subletting and Hawker Succession Scheme · Primary source

$951

Succession payment at a typical stall under the adopted cap (§4.3)

modelled2026 (model baseline)

The lesser of 8% of takings and 30% of verified net profit; it retires the goodwill in 4.2 years at 15 months of profit.

Sections: Succession Economics, E.1

4.2

Payoff time at a typical stall under the adopted cap (15 months of profit)

modelled2026 (model baseline)

Gentler payments for an ordinary stall keep the successor near the assistant's benchmark.

Sections: Succession Economics, E.1 · §4.3

about 14%

The 2.5% royalty as a share of a typical stall's profit

modelled2026 (model baseline)

Small beside sales but material to profit, so §4.4 starts it only once the mortgage is retired.

Sections: Succession Economics, E.1 · §4.4

$564

Successor's income if they keep only 90% of the master's customers, under the original terms

modelled2026 (model baseline)

Mortgage and royalty together would leave the successor almost nothing at a typical stall.

Sections: Succession Economics, E.1

$1,250

Median assessed market rent, non-subsidised cooked-food stalls

official2024-11

The median assessed market rent last given to Parliament (13 Nov 2024); the median rent paid across NEA's tendered stalls has also sat near $1,250 for a decade.

Sections: §1.3 · §2.2 · Follow the Plate, D.4.1 · Sources: Response to Motion on Hawker Culture — Dr Koh Poh Koon; New measures to deter high bids for hawker stalls; Parliamentary replies on the tender regime (reserve rent removed 2012; five highest bids published; Nov 2024 staggered renewal adjustment; median AMR ≈$1,200; ~1 in 5 stalls ≤$500) · Primary source

$1,450

2025 median rent paid by non-subsidised cooked-food stallholders

official2025

What sitting tenants actually pay; it includes tenancies won at tender above NEA's assessed valuation. The middle of the three NEA rents the papers read as a range.

Sections: §2.2 · §1.3 · Follow the Plate, D.2.2 · Follow the Plate, D.3 · Follow the Plate, D.4.1 · Sources: Rents and fees, 2025 medians (MSE written reply, 8 Sep 2026) · Primary source

~1 in 5

Cooked-food stalls awarded at $500 or less

official

Since reserve rent was removed in 2012, about one in five stalls goes for $500 a month or less.

Sections: §1.3 · §2.2 · Sources: Parliamentary replies on the tender regime (reserve rent removed 2012; five highest bids published; Nov 2024 staggered renewal adjustment; median AMR ≈$1,200; ~1 in 5 stalls ≤$500)

about $1,800 a month

Median successful cooked-food bid (2023)

official2023

What a typical new tenant bid for a tendered hawker-centre stall in 2023.

Sections: §2.1 · §2.2 · Follow the Plate, D.4.1

$192–$384

Subsidised stall rents

official2023

The pioneer and legacy cohort pays a few hundred dollars a month, held only while they personally operate.

Sections: §2.1 · Follow the Plate, D.2.2 · Sources: Parliamentary reply: subsidised-stall share and rents

about 30%

Subsidised share of stalls in public centres (53% in 2013, about 40% in 2020)

official2023

The subsidised cohort is shrinking as pioneers retire: 53% of cooked-food hawkers paid subsidised rents in March 2013 (Dr Balakrishnan, Committee of Supply), about 40% in 2020 and about 30% in 2023; about 70% of stalls are now tendered. The subsidy passes to immediate family at the same low rent (Dr Koh, 13 Nov 2024).

Sections: §2.1 · Follow the Plate, D.3 · Sources: Parliamentary reply: subsidised-stall share and rents

about 2,000 stalls

Stalls sold in 15 centres under the 1994–97 Stall Ownership Scheme

official1994–2017

The only route to stall ownership; the 20-year leases expired in 2014–17 and no hawker in a public centre owns a stall today. At expiry, those formerly on subsidised rent could return to it, any owner could hand the stall back for an ex-gratia payment, and the rest carried on at market rent (Grace Fu, Committee of Supply, 12 Mar 2013). About 600 of the stalls sold were cooked-food stalls (MTI, 2015).

Sections: §2.1 · Hawkers and Social Protection, C.2 · Sources: Singapore Rental Policy for Stalls in Hawker Centres and Wet Markets (NEA rental policy summary, via WIEGO); Minister's statement in Parliament on the Stall Ownership Scheme outcome at expiry (reported by CNA) · Primary source

near $3,400

Median coffeeshop stall asking rent ($8,000–10,000 at record-priced shops)

derived2025

Coffeeshop stall rents are set by the market and carry both the owner's return and the operator's markup.

Sections: §2.1 · Hawkers and Social Protection, C.7 · Follow the Plate, D.3 · Sources: 2025 hawker stall rentals · Primary source

776

Coffeeshops in HDB estates, 374 of them rented from HDB (July 2024)

official2024-07

Just under half of the coffeeshops in HDB estates have HDB as head landlord; the other 402 were sold by HDB before 1998.

Sections: Hawkers and Social Protection, C.7 · §2.1 · Sources: More coffee shops to offer budget meals to boost affordable food options; Data on HDB coffee shops tenanted to F&B chains and to independent operators (written answer, via the Office of the Leader of the Opposition) · Primary source

about 5,800

Cooked-food stalls in coffeeshops (776 coffeeshops at about 7.5 stalls each; our arithmetic)

derived2026-10

No official count exists; leaving coffeeshops out would leave out nearly half of Singapore's hawkers.

Sections: Hawkers and Social Protection, C.7 · §2.11.1 · Follow the Plate, D.3

5,600–7,200

Range for the number of cooked-food stalls in coffeeshops

derived2026-10

The range around the central estimate of about 6,000 coffeeshop stalls.

Sections: Hawkers and Social Protection, C.7 · Follow the Plate, D.3

about 6,900

Cooked-food hawker stalls in public centres, incl. operator-run centres (2026; our estimate)

derived2026

Our estimate of stalls in NEA and operator-run centres, not an official count; NEA gives about 15,000 stalls, cooked food and market together. One hawker may hold two.

Sections: Hawkers and Social Protection, C.3 · §2.11.1 · Hawkers and Social Protection, C.7 · Follow the Plate, D.3

2–4%

Stall vacancy rates

official2022-08

Well over a hundred stalls fall vacant each year, so succession demand is not the bottleneck.

Sections: §2.3 · Follow the Plate, D.7.4 · Sources: Parliamentary reply: stall vacancy rates of 2–4% a year

566

Trained under the Hawkers' Development Programme (120 completed apprenticeship)

official2024-07

The programme proved people will train to be hawkers; few stayed in the trade.

Sections: §1.2 · §3.3 · §3.4 · Sources: Parliamentary reply: HDP outcomes (566 trained, 120 through apprenticeship)

~1,150

Stallholders helped by the Hawkers' Productivity Grant (80% co-funding, capped $7,000)

official2026-03-31

The grant funded new automation, never repairs, and closed to applications on 31 March 2026.

Sections: §1.3 · §2.9 · Sources: Hawkers' Productivity Grant (80% co-funding capped $7,000; ~1,150 beneficiaries; applications closed 31 Mar 2026)

~230

Termination notices for stalls not personally operated, 2012–2024

official2024

Where relief cover is unavailable, illness ends up looking like subletting.

Sections: §2.7 · Hawkers and Social Protection, C.2 · Sources: Written reply to parliamentary question on hawker stall subletting and Hawker Succession Scheme · Primary source

24,000+

Hawkers and heartland merchants accepting CDC vouchers

official2026

The voucher system the proposed Makan Vouchers would use already exists.

Sections: §1.3 · §2.5 · §7.7 · Sources: CDC Vouchers Scheme (Jan 2026 $300 and Jun 2026 $500 household tranches; 24,000+ participating hawkers and heartland merchants)

26

Hawker centres with centralised dishwashing (Productive Hawker Centres)

official

The shared service that exists in hawker centres today; the plan builds no shared prep kitchen of its own (§2.6).

Sections: §1.3 · §2.6 · Sources: Productive Hawker Centres: centralised dishwashing

7% → 17%

Platform operators' CPF share, 2026 → 2029

official2026

Under the Platform Workers Act, platforms pay a rising share of CPF for riders born in 1995 or later and older riders who opt in, reaching employee-level rates by 2029; every rider has work-injury cover.

Sections: Hawkers and Social Protection, C.1 · Hawkers and Social Protection, C.4 · §2.11.1 · Sources: CPF contribution rates for platform workers; Platform Workers Act 2024; Written answer to PQ on platform workers' CPF opt-in rate · Primary source

0%

Anyone's CPF share for a hawker

official2026

No employer, platform or landlord contributes to a hawker's CPF.

Sections: Hawkers and Social Protection, C.1 · Hawkers and Social Protection, C.4

~25%

Self-employed in MediSave arrears (end-2024)

official2024-12

About a quarter of all self-employed people were in MediSave arrears at end-2024. No hawker figure is published, and the stall licence cannot be renewed with arrears unless the hawker is on a GIRO plan.

Sections: Hawkers and Social Protection, C.1 · Hawkers and Social Protection, C.4 · Hawkers and Social Protection, C.6 · Sources: Written answer to PQ on self-employed persons with MediSave contributions in arrears; Update, renew or cancel licence or permit for food retail: conditions for renewal; Can I apply for or renew my licence if I have outstanding MediSave payable? · Primary source

≈ $0

CPF LIFE for most MediSave-only hawkers

official2026

Only MediSave above the Basic Healthcare Sum ($79,000 in 2026) flows into retirement savings, so most hawkers who have only paid MediSave reach 55 with little or nothing to annuitise.

Sections: Hawkers and Social Protection, C.1 · Hawkers and Social Protection, C.6 · Sources: How much CPF payouts can I get every month?; MediSave obligation for self-employed (reply to ST letter) · Primary source

about $1,470

Median gross monthly income of full-time own-account hawkers and stall holders, 2021 (MOM)

derived2021

MOM's labour force survey, in the Advisory Committee on Platform Workers' report (2022), Fig. 1.1: below the poorest fifth of employees ($2,200) and below the delivery riders and private-hire drivers the Platform Workers Act now protects. Hours, savings and injuries are still not measured for hawkers.

Sections: Hawkers and Social Protection, C.1 · Hawkers and Social Protection, C.5

≈ $8.3M

Cost of July 2026's rent support for cooked-food stalls: $200 a month for six months across about 6,900 stalls (our arithmetic)

derived2026-10

The same sum each year would match every stallholder, at any age, up to $1,200 a stall a year. The match as proposed, for stallholders under 55 only, costs at most $4.1M a year.

Sections: Hawkers and Social Protection, C.1 · §2.11.1 · §2.11.2 · Sources: Support for stallholders (rent support) · Primary source

up to $1,200 a year

Retirement match per stallholder under 55 a year, on the stallholder's own CPF top-ups

derivedproposed2026-10

From 55 the existing Matched Retirement Savings Scheme already matches top-ups for those within its means tests.

Sections: §2.11.1

$4.1M / year

The match as proposed, for public-centre stallholders under 55 only: the most it costs a year (base case)

derivedproposed2026-10

Base scope of the Hawker Protection Floor: $1,200 a stall for stallholders under 55 in NEA and operator-run centres, all taking it up. From 55 the existing Matched Retirement Savings Scheme already matches top-ups.

Sections: §2.11.1 · §2.11.2 · Follow the Plate, D.7.5 · Hawkers and Social Protection, C.8.1

≈ $7.9M

Retirement match and hospital-cash cover, stretch case: every stallholder under 55, coffeeshops and food courts included, at 70% take-up

derivedproposed2026-10

The stretch scope of the Hawker Protection Floor, against at most about $4.1M a year for the base case (public centres only). Both are ceilings: fewer than half of stallholders are under 55.

Sections: §2.11.2 · §2.11.1 · Follow the Plate, D.7.5 · Hawkers and Social Protection, C.8.1

$200 / stall / month

Rent support announced July 2026, for 6 months (Sep 2026–Feb 2027)

official2026-07

One-off rent support for stallholders in markets and hawker centres.

Sections: Hawkers and Social Protection, C.4 · §2.11.1 · Hawkers and Social Protection, C.7 · Sources: Support for stallholders (rent support) · Primary source

2,376,400

Employed residents (June 2025)

official2025-06

The labour force against which the hawker bloc is measured.

Sections: Hawkers and Social Protection, C.3 · Sources: Labour Force in Singapore 2025 · Primary source

About 280,000

Self-employed · 11.8% of employed residents (2025; our arithmetic)

derived2025

The wider self-employed bloc that hawkers belong to.

Sections: Hawkers and Social Protection, C.3 · Sources: Labour Force in Singapore 2025 · Primary source

About 90,000 a month

Platform workers receiving CPF (2025)

official2025

About 90,000 platform workers a month had CPF contributions in 2025, most of them MediSave-only; about 26,500 older workers had opted in to the operator's share (September 2025).

Sections: Hawkers and Social Protection, C.3 · Sources: Written answer to PQ on whether MOM assessed impact of the Platform Workers Act; Written answer to PQ on platform workers' CPF opt-in rate · Primary source

About 26,500

Older platform workers who had opted in to the operator's CPF share (27%, Sept 2025)

official2025-09

Only riders born in 1995 or later, and older ones who opt in, have a platform paying into their CPF.

Sections: Hawkers and Social Protection, C.3 · Sources: Written answer to PQ on platform workers' CPF opt-in rate · Primary source

$950

CPF LIFE standard payout from 65 on the Basic Retirement Sum ($110,200), turning 55 in 2026

official2026

What an employee with the basic retirement sum receives for life; a MediSave-only hawker receives $0.

Sections: Hawkers and Social Protection, C.6 · Sources: How much CPF payouts can I get every month?; What is the CPF retirement sum? · Primary source

76%

Active CPF members reaching the Basic Retirement Sum at 55 (2025), against 65% of all members

official2025

The headline statistic excludes self-employed members who are not also employees, so it does not see hawkers.

Sections: Hawkers and Social Protection, C.6 · Sources: Retirement sum attainment rates · Primary source

$79,000

Basic Healthcare Sum, 2026

official2026

Once a self-employed member's MediSave reaches it, further MediSave contributions flow into retirement savings; below it, they do not.

Sections: Hawkers and Social Protection, C.1 · Hawkers and Social Protection, C.6 · Sources: MediSave obligation for self-employed (reply to ST letter); More Singaporeans eligible for MMSS and MRSS · Primary source

$456M

Matched Retirement Savings credited to more than 250,000 members for 2025

official2025

The Government already matches the voluntary top-ups of members from 55, at scale: the precedent for the plan's match under 55.

Sections: Hawkers and Social Protection, C.6 · §2.11.1 · Hawkers and Social Protection, C.8.1 · Sources: More Singaporeans eligible for MMSS and MRSS; Matched Retirement Savings Scheme: what you need to know · Primary source

Up to $1,080 a quarter

Silver Support

official2026

A long-serving hawker may retire on Silver Support alone, up to $360 a month, if means tests allow.

Sections: Hawkers and Social Protection, C.6 · Sources: Silver Support Scheme · Primary source

≈ $0.30

Profit on a $3.50 plate of chicken rice

derived2025

At this margin a hawker would need to sell about 642 plates a day to match the 2025 median monthly income of $5,775 for full-time employed residents, including employer CPF.

Sections: Hawkers and Social Protection, C.5

About 13 a day

Hours worked, across 21 interviewed hawkers

official2025

Hawkers work the longest hours of any bloc; 70-hour weeks recur in reporting.

Sections: Hawkers and Social Protection, C.5 · Follow the Plate, D.3 · Follow the Plate, D.4 · Sources: Hawker interviews (Hack 2025: hawker-thing) · Primary source

56%

Raw materials' share of operating costs (manpower 20%; rent under 10%; NEA, 2022)

official2022

Ingredients, not rent, are the biggest cost at a hawker-centre stall.

Sections: Hawkers and Social Protection, C.2 · Hawkers and Social Protection, C.5 · Follow the Plate, D.3 · Sources: Written reply to PQ on hawker food prices (hawker cost structure) · Primary source

8% a year

Cooked-food tenancies ended early in operator-run centres, against 3% in NEA-run centres (2019–2023)

official2019–2023

Operator-run centres look more like platforms, and their stallholders leave early more often.

Sections: Hawkers and Social Protection, C.2 · Sources: Written reply to PQ on rental contract termination in SEHCs · Primary source

$3.5bn

Cooked-food sales a year, 14,400 stalls (model)

modelled2026 (model baseline)

What diners spend a year at the cooked-food stalls the model covers.

Sections: Follow the Plate, D.1 · Follow the Plate, D.2

45¢

Of each dollar to ingredient suppliers

modelled2026 (model baseline)

Ingredients take the largest share of every plate in every kind of stall.

Sections: Follow the Plate, D.1 · Follow the Plate, D.5

3¢–26¢

To the landlord, by type of stall

modelled2026 (model baseline)

The landlord's cut ranges from a subsidised NEA stall to a mall food court.

Sections: Follow the Plate, D.1 · Follow the Plate, D.5

7¢ to 21¢

reaches the hawker's family

modelled2026 (model baseline)

The hawker household keeps the least where the rent is set by the market.

Sections: Follow the Plate, D.1 · Follow the Plate, D.3

66%

Tendered NEA hawkers earning less per hour than their assistant

modelled2026 (model baseline)

In 5,000 simulated tendered NEA stalls, two-thirds of owners earn less per hour than the Progressive Wage benchmark for their assistant.

Sections: Follow the Plate, D.1 · Follow the Plate, D.4

$9.69

Typical tendered NEA hawker's earnings per hour, at a 13-hour day

modelled2026 (model baseline)

Less than the hourly Progressive Wage benchmark for the assistant beside them.

Sections: Follow the Plate, D.1 · Follow the Plate, D.7.1

$11.64

Progressive Wage benchmark for a food-stall assistant, per hour

derived2026-07

$2,220 a month over a 44-hour week, the July 2026 benchmark.

Sections: Follow the Plate, D.1 · Follow the Plate, D.4 · Sources: Government accepts TCF recommendations on Food Services PWM 2026 to 2028 · Primary source

25%

Income a private-coffeeshop hawker loses to a 10% rise in rent and fees

modelled2026 (model baseline)

Against 6% for a 10% rise in ingredient prices: the landlord's squeeze is the larger in five of six kinds of stall.

Sections: Follow the Plate, D.1 · Follow the Plate, D.5

6%

Income a private-coffeeshop hawker loses to a 10% rise in ingredient prices

modelled2026 (model baseline)

Ingredient rises hit every stall at once, so more of them reach the diner.

Sections: Follow the Plate, D.1 · Follow the Plate, D.5

8% to 21%

Income lost to a 3% fall in custom, by type of stall

modelled2026 (model baseline)

The hawker keeps only what is left, so a small fall in custom cuts income sharply.

Sections: Follow the Plate, D.1 · Follow the Plate, D.6

up to 76%

Income lost to a 30% rent rise at renewal

modelled2026 (model baseline)

The worst case is a private coffeeshop, where a rent rise cannot be passed on to diners.

Sections: Follow the Plate, D.1 · Follow the Plate, D.6

$34.1M

Cost of $200 a month per stall, stall-attached and made permanent, across 14,400 stalls

modelled2026 (model baseline)

Landlords would recapture $12.8M of it at renewal, $10.4M of that by operators and private owners.

Sections: Follow the Plate, D.1 · Follow the Plate, D.7.2

$12.6M

Recaptured by landlords at renewal

modelled2026 (model baseline)

About 37% of stall-attached support returns to landlords nationally (capture rates are assumptions).

Sections: Follow the Plate, D.1 · Follow the Plate, D.7.2 · Method, Levers and Data, F.2.3

$10.3M

Of which to operators and private owners

modelled2026 (model baseline)

The part of stall-attached support that leaks to private landlords, about 30% of the total.

Sections: Follow the Plate, D.1 · Follow the Plate, D.7.2 · Method, Levers and Data, F.2.3

$15.60

Tendered NEA sitting tenant's earnings per hour with the plan's package

modelled2026 (model baseline)

Up from $9.42 an hour, lifting most such hawkers past the assistant's benchmark.

Sections: Follow the Plate, D.1 · Follow the Plate, D.7.1

26%

Tendered NEA stalls below the assistant's wage with the package (from 67%)

modelled2026 (model baseline)

The package, with the Buying Circles, the fee reset and the named deputy, cuts the share of tendered NEA hawkers earning less per hour than their assistant by more than half.

Sections: Follow the Plate, D.1 · Follow the Plate, D.4

3.5 to 5.3 years

Time for the 8% rule to retire a busy stall's goodwill (12–18 months of profit)

modelled2026 (model baseline)

The Succession Mortgage's 8% of takings fits a stall with a queue.

Sections: Follow the Plate, D.1 · Succession Economics, E.1 · §4.3

$1,756

Successor's monthly income at typical takings under the original 8% rule, before the family royalty

modelled2026 (model baseline)

Below the $2,220 assistant benchmark, which is why Succession Economics, E.1's cap was adopted in §4.3.

Sections: Follow the Plate, D.1 · Succession Economics, E.1

$3,542M

Diners pay

modelled2026 (model baseline)

Annual spending by diners at the 14,400 cooked-food stalls in the model.

Sections: Follow the Plate, D.2

$1,577M

Ingredient suppliers

modelled2026 (model baseline)

The largest single flow: ingredients take about 45% of every dollar.

Sections: Follow the Plate, D.2

$193M

Landlords: Government (NEA, HDB)

modelled2026 (model baseline)

Rent and fees the state collects as landlord, in the model.

Sections: Follow the Plate, D.2 · Follow the Plate, D.2.2

$148M

Landlords: centre and coffeeshop operators

modelled2026 (model baseline)

Rent and fees kept by operators who sublet stalls.

Sections: Follow the Plate, D.2

$251M

Landlords: private property owners

modelled2026 (model baseline)

Rent paid to private owners of coffeeshops and malls.

Sections: Follow the Plate, D.2

17%

Landlords together take

modelled2026 (model baseline)

Government, operators and private owners combined.

Sections: Follow the Plate, D.2

$522M

Assistants' wages

modelled2026 (model baseline)

Wages for stall assistants, modelled at one assistant on the PWM benchmark plus part-time help per stall.

Sections: Follow the Plate, D.2

$170M

Utilities

modelled2026 (model baseline)

Electricity, gas and water across the modelled stalls.

Sections: Follow the Plate, D.2

$92M

Delivery platforms and payment providers

modelled2026 (model baseline)

Delivery commissions, payment fees and app discounts.

Sections: Follow the Plate, D.2

$127M

Centre services and other running costs

modelled2026 (model baseline)

Cleaning and conservancy services, licences, insurance, packaging and equipment upkeep.

Sections: Follow the Plate, D.2

$462M

Hawker households

modelled2026 (model baseline)

What the families at the wok keep, before MediSave and tax.

Sections: Follow the Plate, D.2

about $89M

Rent and conservancy fees from cooked-food stalls attributed to NEA

modelled2026 (model baseline)

Consistent with published medians of $1,450 non-subsidised and $192–$384 subsidised rent, with 2025 S&CC of about $200.

Sections: Follow the Plate, D.2.2 · Sources: Rents and fees, 2025 medians (MSE written reply, 8 Sep 2026) · Primary source

about $94M

HDB-rented coffeeshop stall rents attributed to HDB

modelled2026 (model baseline)

Less than HDB's tender receipts, which imply roughly $230M a year if typical of all 400 HDB coffeeshops.

Sections: Follow the Plate, D.2.2

roughly $215 million a year

HDB coffeeshop tender receipts, if the 22 recent tenders (about $48,000 a month) are typical of all 374 HDB-rented coffeeshops

derived2026-10

Operators fund the gap between tender rents and stall rents from their own drinks and food counters.

Sections: Follow the Plate, D.2.2 · Sources: Eating-house PQM tender results · Primary source

between $193M and about $314M a year

What Government collects from the trade as landlord

derived2026-10

Against about $13 million in direct rent support in July 2026.

Sections: Follow the Plate, D.2.2

about $13 million

Direct rent support in July 2026

official2026-07

What the state paid back to hawkers as rent support, set against the rent it collects.

Sections: Follow the Plate, D.2.2 · Sources: Support for stallholders (rent support); July 2026 rental support for hawkers · Primary source

roughly $3–4 billion

Operating receipts of cafés, food courts and other eating places (2021)

derived2021

One of two official totals the model's $3.5bn is checked against; neither isolates hawker stalls.

Sections: Follow the Plate, D.2.1 · Sources: Food & beverage services infographic · Primary source

$491 a month

Resident households' spending at hawker centres, food courts and coffee shops (HES 2023)

official2023

About $8.8 billion a year across 1.49 million households, drinks included.

Sections: Follow the Plate, D.2.1 · Follow the Plate, D.7.3 · Sources: Household Expenditure Survey 2023 · Primary source

$11.45

Tendered NEA hawker's hourly earnings at an 11-hour day

modelled2026 (model baseline)

The per-hour finding depends on hours, the least measured number in the trade; at 10 hours it is $12.25, and the break-even is about 10 hours.

Sections: Follow the Plate, D.4

$610

Rent above which a typical tendered hawker earns less than the assistant's benchmark, without the package

modelled2026 (model baseline)

Below the $11.64-an-hour benchmark at any rent above this, so the finding holds at all three quoted NEA rents; the rent only decides by how much.

Sections: Follow the Plate, D.4.1 · §2.2

$2,450

Rent up to which a typical tendered hawker with the package still earns more than the assistant's benchmark

modelled2026 (model baseline)

Well above the highest quoted NEA rent, so with the package the typical stall stays sound across the range; about $2,300 without the named deputy.

Sections: Follow the Plate, D.4.1 · §2.2

$2,150

Rent up to which a typical tendered hawker with the package, but without the named deputy, still earns more than the assistant's benchmark

modelled2026 (model baseline)

If NEA does not allow the named deputy (§2.7), the package still carries the typical stall past the benchmark across the quoted range.

Sections: Follow the Plate, D.4.1 · §2.2 · §2.7

$1,810

Rent above which the successor's $1,500 protected income starts to bind

modelled2026 (model baseline)

Below this, a typical successor keeps more than $1,500 a month after the Succession Mortgage payment; above it the payment shrinks and the term lengthens.

Sections: Follow the Plate, D.4.1 · §2.2 · §4.3

$2,450

Rent at which the successor's payment stops altogether

modelled2026 (model baseline)

At this rent the successor's own profit is down to the protected $1,500, so no payment is due.

Sections: Follow the Plate, D.4.1

$4,420

Rent above which a typical tendered stall makes a loss

modelled2026 (model baseline)

The stall itself is robust: it loses money only far above any quoted NEA rent.

Sections: Follow the Plate, D.4.1 · §2.2

66–77%

Tendered stalls whose hawker earns less than the assistant's benchmark today, across the $1,250–1,800 rent range

modelled2026 (model baseline)

66% at the $1,250 median rent, 70% at the $1,450 paid in 2025 at centres comparable with the operator-run ones, and 77% at the $1,800 winning bid.

Sections: §2.2 · Follow the Plate, D.4.1

26–37%

The same share with the package, across the $1,250–1,800 rent range

modelled2026 (model baseline)

26% at the $1,250 median rent, 30% at $1,450 and 37% at the $1,800 winning bid.

Sections: §2.2 · Follow the Plate, D.4.1

0.7% to 2.3%

Tendered stalls that lose money at the quoted NEA rents ($1,250 to $1,800)

modelled2026 (model baseline)

Few stalls lose money at any quoted rent; the risk is below the benchmark, not below zero.

Sections: Follow the Plate, D.4.1

$18,900

What a newcomer pays above NEA's valuation before he reaches it, under the November 2024 renewal rule

derived2026 (model baseline)

A stall won at the 2023 median bid of $1,800 against a valuation of $1,450: $1,800 a month for three years, $1,625 for the next three and $1,450 from the seventh. Against the $1,250 system median the sum is $29,700.

Sections: Follow the Plate, D.4.1 · Method, Levers and Data, F.2.3 · §2.2

$372

What each day shut for illness costs the hawker at the typical tendered stall

modelled2026 (model baseline)

Rent, centre fees, the assistant, utilities and other running costs continue while the stall is shut. Six weeks closed cost about $13,400, 35% of the year's net; a rent waiver saves about $48 a day of it, and a Relief Corps cook on medical leave turns it into the $40 co-pay.

Sections: Method, Levers and Data, F.2.3 · §2.9 · §2.7

2.17

Leverage of the Relief-Hawker Corps, tendered NEA stall

modelled2026 (model baseline)

Relief cooks return the most per programme dollar, because a covered day keeps takings a closed day loses.

Sections: Follow the Plate, D.7 · Follow the Plate, D.7.1

$78.7M

The plan's programmes at national scale, a year

modelled2026 (model baseline)

Most of it is the Makan Voucher line, and the next largest the Protection Floor; the table-cleaning fee subsidy is counted at the three pilot centres only, and its national figure is shown apart as a later decision. The four-year $19.68M budget funds the pilots, not this roll-out. The HDB tender lever pays for the Protection Floor's coffeeshop match and is not added on top.

Sections: Follow the Plate, D.7.4

$0.98M

Table-cleaning fee income the State would give up a year at the three pilot centres, the plan's ask ($200 a stall a month)

modelled2026 (model baseline)

Every cooked-food stall at the three pilot centres pays the fee: 408 stalls (Chinatown Complex 226, Geylang Serai 63, Tekka 119), each centre managed by NEA or a Town Council; $1.63M at five, with Old Airport Road and Maxwell. An upper bound: re-priced cleaning contracts would meet part of it.

Sections: Follow the Plate, D.2.2 · Follow the Plate, D.7.4 · §2.2 · §7.7

$16.6M

Table-cleaning fee income the State would give up a year if the reset were later made national ($200 a stall a month): a later decision, not part of the plan

modelled2026 (model baseline)

An upper bound for a decision after the pilot centres: the model books the whole cut as income the fee collector forgoes; operators at operator-run centres would bear about $0.5M more, and re-priced cleaning contracts would meet part of it.

Sections: Follow the Plate, D.2.2 · Follow the Plate, D.7.4 · §2.2 · §7.7

1.06

Return to hawkers and diners per programme dollar, income and savings (1.00 in cash)

modelled2026 (model baseline)

Leaving out price freedom, which diners rather than taxpayers fund, the public programmes return 1.24.

Sections: Follow the Plate, D.7.4

1.24

Return of the public programmes, leaving out price freedom

modelled2026 (model baseline)

Public programmes return about 30 cents more than each programme dollar to hawkers and diners.

Sections: Follow the Plate, D.7.4

2 to 5 times

How far a $120 voucher covers an eligible household's extra cost if one stall in five raises prices 5%

modelled2026 (model baseline)

If every stall raised prices 5%, the voucher would cover 41% to 99% of the extra cost.

Sections: Follow the Plate, D.7.3 · §2.5 · Method, Levers and Data, F.2.3

41% to 100%

Voucher cover if every stall raised prices 5%

modelled2026 (model baseline)

A heavy user spending $900 a month would be covered 22%.

Sections: Follow the Plate, D.7.3 · §2.5

$4.14M

Protection Floor, base: a $1,200 retirement match for every public-centre stallholder under 55 (at most 3,450 of 6,900 stalls)

modelled2026 (model baseline)

The base scope of the one list, the match alone. Hawkers and Social Protection rounds it to about $4.1M. A ceiling: the model counts half of stallholders as under 55. At every age it would be $8.28M.

Sections: Follow the Plate, D.7.5 · §2.11.2 · Hawkers and Social Protection, C.8.1

$7.92M

Protection Floor, stretch: every stallholder under 55 at 70% enrolment, with pooled hospital-cash cover ($6.05M match, $1.98M cover)

modelled2026 (model baseline)

The stretch scope; Hawkers and Social Protection rounds it to about $8.0M, and Follow the Plate, D.7.4's total uses it. A ceiling: the model counts half of stallholders as under 55. At every age it would be $14.08M.

Sections: Follow the Plate, D.7.5 · Follow the Plate, D.7.4 · §2.11.2 · Hawkers and Social Protection, C.8.1

$1.7M

HDB's possible contribution: tender rent forgone under the tender lever, at half take-up of 3,000 HDB-rented stalls

modelled2026 (model baseline)

The upper bound, if operators price in the whole commitment; less if HDB's quality scoring lets them absorb part. It pays for coffeeshop stallholders, so it offsets the stretch only.

Sections: Follow the Plate, D.7.5 · Follow the Plate, D.7.4 · §7.7 · §2.4

$14.7M

Retirement match, base: shortfall over Years 4–10 after gifts set aside and Heritage Plate licences

modelled2026 (model baseline)

Giving and licensing cover about 40% of the base; $3.3M a year is still short in Year 10. The remainder is the optional ask of Government.

Sections: the Budget Book, B.9.5 · Hawkers and Social Protection, C.8.1

$37.0M

Retirement match, stretch: shortfall over Years 4–10 without HDB's contribution

modelled2026 (model baseline)

Giving and licensing cover about 21% of the stretch.

Sections: the Budget Book, B.9.5 · Hawkers and Social Protection, C.8.1

$32.4M

Retirement match, stretch: shortfall over Years 4–10 if HDB agrees to count its forgone tender rent

modelled2026 (model baseline)

HDB forgoes up to about $1.7M a year of tender rent once all coffeeshops are re-tendered; only its under-55 part, about $0.84M a year, counts towards the match. That raises the stretch's cover to about 31%; the stretch is still about $6.6M a year short in Year 10.

Sections: the Budget Book, B.9.5 · Hawkers and Social Protection, C.8.1

40%

Share of the base retirement match that giving and Heritage Plate licences could pay, Years 4–10

modelled2026 (model baseline)

At the central assumptions, with a third of Heritage Plate licensees failing in any year; a Heritage Plate twice the central size would raise it to about 68%, and with no licensee failing it would be 54%.

Sections: the Budget Book, B.9.5

21%

Share of the stretch retirement match that giving and Heritage Plate licences could pay, Years 4–10

modelled2026 (model baseline)

Without HDB's contribution; 31% with it.

Sections: the Budget Book, B.9.5

$2.0M

Heritage Plate licence income at the central case, from Year 7

modelled2026 (model baseline)

About $60K in Year 3 (15 licensed outlets); 40% goes to the retirement match, 40% to masters' families and 20% to the Seed Bank. A third of licensed outlets and packaged lines are assumed to fail and pay nothing in any year (most licensed hawker expansions shrank or closed: Chua, IPS, 2024). The outlet numbers, plates an outlet, packaged-goods sales and the failure rate are assumptions.

Sections: the Budget Book, B.9.5 · the Budget Book, B.9.2 · the Budget Book, B.6.1 · §5.5

101%

Endowment in Year 20, as a share of its target, if half of the surplus gifts are earmarked for the match

modelled2026 (model baseline)

Setting half of the surplus gifts aside for the match leaves the corpus just above its target in Year 20, against 121% with none set aside and 86% with all of it.

Sections: the Budget Book, B.9.5

34%

Successors who quit within five years when there is no payment and no royalty (the control)

modelled2026-10

Every simulated hand-over is also run with no payment and no royalty. Some successors quit anyway, for money or personal reasons.

Sections: Succession Economics, E.1.4 · Succession Economics, E.1

49%

Successors who quit within five years, 26 October terms

modelled2026-10

The lesser of 8% of gross and 30% of net, a $600 floor, the royalty after payoff, and no protected income. The simulation tracks each successor for five years.

Sections: Succession Economics, E.1.4 · Succession Economics, E.1

16%

Extra successors who quit within five years because of the payment, 26 October terms

modelled2026-10

The share who quit beyond the control, where no payment is charged. It measures the burden the payment itself adds.

Sections: Succession Economics, E.1.4 · Succession Economics, E.1

4.7

Median years to retire the goodwill, 26 October terms

modelled2026-10

Half of simulated hand-overs pay off sooner than this, and half later.

Sections: Succession Economics, E.1.4 · Succession Economics, E.1

84%

Share of the goodwill sum the retiree recovers on average, 26 October terms

modelled2026-10

Counts the successor's payments and the first-loss guarantee.

Sections: Succession Economics, E.1.4 · Succession Economics, E.1

$7,500

Expected first-loss cost per mortgage, averaged across all simulated stalls, 26 October terms

modelled2026-10

What the first-loss guarantee costs the programme on average, weaker stalls included.

Sections: Succession Economics, E.1.4 · Succession Economics, E.1 · §4.3

34%

Successors who quit within five years, 26 October terms with the $1,500 protected income

modelled2026-10

The same terms, plus a payment and royalty that never take the successor's monthly profit below $1,500. The simulation tracks each successor for five years.

Sections: Succession Economics, E.1.4 · Succession Economics, E.1

0.4%

Extra successors who quit within five years because of the payment, 26 October terms with the $1,500 protected income

modelled2026-10

The share who quit beyond the control, where no payment is charged. It measures the burden the payment itself adds.

Sections: Succession Economics, E.1.4 · Succession Economics, E.1

4.9

Median years to retire the goodwill, 26 October terms with the $1,500 protected income

modelled2026-10

Half of simulated hand-overs pay off sooner than this, and half later.

Sections: Succession Economics, E.1.4 · Succession Economics, E.1

85%

Share of the goodwill sum the retiree recovers on average, 26 October terms with the $1,500 protected income

modelled2026-10

Counts the successor's payments and the first-loss guarantee.

Sections: Succession Economics, E.1.4 · Succession Economics, E.1

$5,900

Expected first-loss cost per mortgage, averaged across all simulated stalls, 26 October terms with the $1,500 protected income

modelled2026-10

What the first-loss guarantee costs the programme on average, weaker stalls included.

Sections: Succession Economics, E.1.4 · Succession Economics, E.1 · §4.3

$104,000

Reserve needed to be 95% sure of covering a book of 10 Succession Mortgages (P95 payout)

modelled2026 (model baseline)

Expected payout for 10 mortgages is $65,000; the reserve must cover the bad tail, not the average.

Sections: Succession Economics, E.1.5 · Succession Economics, E.1

$180,000

Reserve needed to be 95% sure of covering a book of 20 Succession Mortgages (P95 payout)

modelled2026 (model baseline)

A realistic pilot book is 5 to 20 mortgages; 20 sits well inside the $310,000 left for losses.

Sections: Succession Economics, E.1.5 · Succession Economics, E.1

$22,000

First-loss payout per mortgage if a 2020-style shock makes every mortgage in its first two years fail at once

modelled2026 (model baseline)

Against about $6,500 expected; the guarantee would pay up to 24 of the mortgage's scheduled payments.

Sections: Succession Economics, E.1.5 · Succession Economics, E.1.3

$440,000

Crisis payout for a book of 20 mortgages, all in their first two years

modelled2026 (model baseline)

More than double the book's P95 payout, which is why the reserve should also be sized to the crisis payout where affordable.

Sections: Succession Economics, E.1.5

$560,000

Part of the $660,000 Years 1–3 budget left for losses after escrow set-up and administration

modelled2026 (model baseline)

The set-up and administration split ($100,000) is an assumption; the plan does not divide the $660,000.

Sections: Succession Economics, E.1.5

about 74

Mortgages the Years 1–3 reserve can safely carry at P95

modelled2026 (model baseline)

82 if all $660,000 were reserve; growing past about 68 mortgages needs more capital.

Sections: Succession Economics, E.1.5 · Method, Levers and Data, F.2.3

about 25

Mortgages the Years 1–3 reserve can carry through a 2020-style crisis

modelled2026 (model baseline)

30 if all $660,000 were reserve; enough for a pilot, where the gate needs six hand-overs.

Sections: Succession Economics, E.1.5 · Method, Levers and Data, F.2.3

3

of the 21 Plate Tracker indicators that Singapore publishes regularly (11 in part, 7 not at all)

derived2026-10

Most of the missing indicators are already held by an agency.

Sections: Follow the Plate, D.8 · §2.10

0.56%

Hawker price rise with a 1% rise in raw-material costs (against 0.03% for a 1% rise in rent)

official2012–2014

Ingredient costs reach diners; rent costs mostly stay with the hawker.

Sections: Follow the Plate, D.5.1 · Method, Levers and Data, F.2 · Sources: Examining the cost drivers of hawker food prices (Economic Survey feature article, 1Q 2015) · Primary source

102.4

Import price index for food, August 2026 (down 2.6% on a year earlier; meat +5.4%)

official2026-08

More than 90% of Singapore's food is imported.

Sections: Follow the Plate, D.5.2 · Sources: Import and export price indices, August 2026 · Primary source

−0.5

Assumed demand elasticity for hawker meals (varied −0.3 to −1.0)

assumption2026 (model baseline)

Singapore has only one early, indicative estimate (Poh, 2026: 90 diners, not a random sample, more sensitive than assumed); restaurant studies abroad find −1.2 and below.

Sections: Method, Levers and Data, F.2.1 · Method, Levers and Data, F.2 · Sources: Demand for food away from home; Decoding the Consumers of Hawker Food · Primary source

25%

Food-court turnover rent, the weakest parameter in the model

assumption2026 (model baseline)

The only data point is about 20% (2014).

Sections: Follow the Plate, D.3 · Method, Levers and Data, F.2

10 days a year

Unwanted closures to illness or exhaustion (varied 4–16)

assumption2026 (model baseline)

Days a stall loses when the hawker cannot work, the gap the Relief-Hawker Corps fills.

Sections: Method, Levers and Data, F.2