§2.2Rent Bands & Ballot

• ASK OF GOVERNMENT · NEA PILOT, ADMINISTRATIVE
• ASK OF GOVERNMENT · A TABLE-CLEANING FEE SUBSIDY AT THE PILOT CENTRES
• NO NEW LEGISLATION
• BUILDS ON · 2012 & NOV 2024 TENDER REFORMS, AND THE FIXED RENTS OF OPERATOR-RUN CENTRES

Government has already acted twice on the same diagnosis. Removing reserve rent in 2012 let bids fall to a dollar: in 2023 about one in five cooked-food stalls was awarded at $500 or less, and 44% of winning bidders paid below the assessed market rent (NEA's valuation of what a stall would fetch). The November 2024 renewal reform brings high tendered rents down in stages towards the assessed market rent precisely to deter overly high bids. Both moves treat the symptom honestly – yet entry price at a contested stall is still set by the single most optimistic bidder in the room, and the five highest bids are still published monthly, so five-figure headlines keep teaching young cooks that the front door is an auction. Granted, those headlines are outliers: only about 4% of cooked-food stalls pay above their own stall's assessed market rent, whose median is about $1,250 (Dr Koh's correction by written statement, Official Report, 13 Nov 2024). The case for bands does not rest on record bids being common. It rests on the door: a qualified cook should know the rent before applying.

A rent known in advance is older than any party's proposal. The Government's own Public Consultation Panel proposed in 2012 that rents be "pre-determined" by food type, at "a certain percentage of the market rent … assessed through professional valuation", and that individuals come before companies (Panel report, ¶17–18; §1.2). The bands are that idea, at three centres, with a test. A ballot has been proposed in Parliament too (the Record, A.7). In the debate of 13 November 2024 Leong Mun Wai and Hazel Poa, then Non-Constituency MPs, both proposed allocating stalls by ballot: Leong within each food category, at a turnover rent with a floor ("$500 or 3% of Gross Turnover, whichever is higher"), and Poa in place of the tender, with a pre-determined rent. A month earlier Melvin Yong, the MP for Radin Mas, asked whether a ballot "at a fixed rental rate targeted at cost recovery" had been considered. I agree that stalls should be allocated by ballot rather than by the highest bid. Price is a poor test of a cook, and the State's own record says so: of HDB's 65 eating-house tenders scored on price and quality since 2018, 40 went to someone other than the highest bidder (MND, 4 Aug 2026). I do not agree with one fixed rent, or with a rent tied to turnover, and the Government's objections, answered one by one below, are fair ones. What has been missing is a design that keeps the tender's virtue. We put one forward here: published fixed bands, allocated by ballot.

• Fixed bands by centre class and stall type, published and reviewed every three years, calibrated to the assessed-market-rent data NEA (National Environment Agency, which runs the public hawker centres) already commissions – for non-subsidised cooked-food stalls, the median of NEA's assessed market rents (AMR) was $1,250 when last given to Parliament (November 2024). Indicative cooked-food bands, a month: Band A (Class A, prime centres) $1,700 · Band B $1,150 · Band C $650. Each pilot centre's class is set from its own tender record, not from its profile: Chinatown Complex's median winning bid from August 2024 to January 2025 was $1,408, the lowest of the centres in DollarsAndSense's analysis of NEA's results. Every applicant knows the rent before applying; no one's tenancy depends on a bidding war.
• Ballot allocation among qualified applicants, as Build-To-Order (BTO) flats are balloted. Qualification is a completed apprenticeship, a Seed Bank programme, or a simple food-business track record, and each applicant enters once a round; the ballot weights first-time operators and Endangered Dish Registry proposals, which gives the Hawker Centre 3.0 Committee's "retention of traditional hawker food" a rule to work by (Recommendation 9, accepted 2017).
• Pioneer-generation rates preserved in full. Nothing changes for any sitting hawker. On succession to a non-family successor, rent steps to the applicable band over three years.
• No rent subsidy from the programme. A stall for an Endangered Dish Registry dish pays its band's rent like any other; its weight in the ballot is its only favour, and a graduate's first years are carried by the landing package, paid to the graduate, never off the rent (§3.4) – visible policy, not private deals.
• Tested rather than argued. The pilot runs at the three pilot centres, five at most, so the tender carries on everywhere else. The evaluator publishes what ballot entrants paid beside what first-time tenderers paid at comparable centres. Let the numbers settle it.

Four objections, answered. On 13 November 2024 Dr Koh Poh Koon, the Senior Minister of State, gave Parliament four reasons against a ballot and a fixed rent (the Record, A.7). Each deserves a straight answer.

• "Frivolous applications and excess demand for stalls." The qualification gate answers it. Only a completed apprenticeship, a Seed Bank programme or a food-business record admits an applicant, and each enters once a round. A ballot among qualified cooks is not a lottery open to all comers.
• A fixed rent at the assessed market rent would raise rents for "close to about half of tenderers". The bands are not one rent at the AMR: they follow NEA's own valuations by centre and stall, and Band C sits at about half the median AMR. Granted, Band C's $650 is still above the $500 or less that one winner in five paid in 2023, so a ballot entrant cannot win a stall for a dollar. That is the trade: a known, fair rent for every qualified entrant in place of a bargain for a lucky few. The pilot tests whether it is worth making; the tender carries on everywhere else.
• A rent cap would pull bids up to the ceiling. Inside a band there is no bidding, so there is no ceiling to bid towards.
• Only about 4% of stalls pay above their own AMR. I accept the point, as above: the case for bands rests on the door, not on the outliers.

A fifth objection, older than the others. In 2018 the Minister told Parliament that NEA must award "one vacant stall at a time based on tendered rentals to the highest bidder", and that an allocation that curated stalls "would be complex to execute and be subject to potential audit issues" (Masagos Zulkifli, Official Report, 19 Nov 2018). The answer is in the Government's own centres. In the same speech he praised the operator-run centres for something a price tender cannot do: their operators "conduct food tasting when letting out stalls". The pilot goes further and does it by published rule – a qualification gate anyone can read, bands set in advance, a ballot drawn in public – so the audit has something to check, and the evaluator publishes the result.

Not a subsidy for poor business. The same Minister said it would be "inappropriate for the Government to subsidise a hawker on the basis that business is poor", unfair to the hawker who thrives and to the private food shops nearby. I agree, and nothing here pays a hawker for doing badly. A band is a rent known in advance, not a rebate; the fee subsidy pays for a service at three centres and shrinks if the service costs less; the income top-up runs for two years after a graduate opens his stall, only for a cook who passed the gate (§3.4); and the retirement match pays only on what the hawker saves first (Hawkers and Social Protection, C.8.1).

Not new at home. A fixed rent without bidding already runs at 16 of the 123 centres. At an operator-run centre the operator proposes stall rents in its bid to NEA and may not charge a different rent during the tenancy; the rents are "not subject to bidding" (MSE, 10 Jan 2024 and 5 Feb 2025). Since 2019 NEA has capped them at average independently valued rents, with 80% and then 90% of that in a new centre's first two years (MSE, 24 Sep 2025). So their record is evidence the evaluator can start from now. To be sure, capped is not cheap: the 2025 median rent at operator-run centres was $1,700, against $1,450 at comparable centres (MSE, 8 Sep 2026). The gap is old. In 2018 it was about $2,000 against $1,700 at "comparable existing" centres (Dr Amy Khor, Official Report, 19 Nov 2018): a premium of about 17% both times (our arithmetic), put down both times to bigger stalls and better amenities. Granted, the centres compared are not the same ones each year, seven operator-run centres then and sixteen now. Still, seven years on, the number has not moved. And early terminations there ran at 8% a year against 3% at NEA's own centres (2019–2023), though MSE (Ministry of Sustainability and the Environment, NEA's ministry) notes that the figure includes new centres still settling in (MSE, 13 Nov 2024). A fixed rent alone does not keep a stall open; the package beside the bands is meant to. The ballot itself is older still: registered hawkers drew lots for stalls at Bukit Ho Swee in 1966, and NEA still ballots tied bids today.

Several rents, different things. The NEA figures quoted as "the rent" each measure something different. The median assessed market rent, $1,250 a month when last given to Parliament, is a valuation: what NEA's valuers judge a stall would let for, and the level renewal rents move towards. The median rent that non-subsidised stallholders actually pay has sat near $1,250 across NEA's tendered stalls (won at NEA's open tender, usually by the highest bid) for a decade (NEA, 4 Nov 2024); in 2025 it was $1,450 at centres comparable with the operator-run ones, including tenancies won at tender above the valuation (MSE, 8 Sep 2026). The 2023 median winning bid, about $1,800, is the price of entry at stalls that went to tender. NEA quotes rents and conservancy (the monthly charge for cleaning and upkeep of a centre) before GST, which the stallholder pays on top (NEA tender notice, Aug 2026); MSE's medians do not say whether they include it. Read together, they are a range, not one number. Follow the Plate, D.4.1 ("How much rent can a stall bear?") tests how much rent a stall can carry. Without the package, a typical tendered hawker earns less than his assistant's $11.64 an hour at any rent above about $610 a month. With the package – Part II, including the Buying Circles of §2.12 and the fee subsidy below – he stays above it up to about $2,450 a month, or about $2,150 without the named deputy (§2.7). Across the $1,250–1,800 range, 66–77% of tendered stalls fall below the assistant's pay today and 26–37% with the package; a stall loses money only above about $4,400. Even Band B's $1,150 sits well above that $610 point: without the package, a typical hawker paying Band B rent still earns less than an assistant. That is why the bands come with the package, not instead of it. The tight spot is succession: at an $1,800 winning bid a successor keeps about $1,510 a month, barely above the protected $1,500 (§4.3), and the protection binds from about $1,810.

What a newcomer actually pays. The medians flatter the newcomer's bill. Take a stall won at the 2023 median bid of about $1,800, where NEA's valuation is $1,450. Under the rule NEA adopted in November 2024, he pays his bid for the first three-year tenancy, halfway between bid and valuation for the second, about $1,625, and the valuation only from the seventh year: about $18,900 above the valuation before he reaches it, or $29,700 above the $1,250 system median (our arithmetic; the valuation is illustrative, since NEA publishes none, and the rule applies only to tenancies won after it took effect). On the typical stall, that leaves him about $8.01 an hour in the first tenancy, $8.54 in the second and $9.08 from the seventh year, each below the $11.64 an hour of his own assistant (Follow the Plate, D.4.1). KF Seetoh puts it more sharply, that the benchmark is reached "only after six years of paying whatever rent a stallholder tendered" (The Online Citizen, 15 Sep 2026). He overstates the middle three years, which run at the halfway point, not at the bid. He is right on the direction and on the six years, and a stall that closes before its seventh year never pays the median at all.

The fee beside the rent. Rent is not a hawker's only charge for the space he cooks in. In 2025, at centres comparable with the operator-run ones, a non-subsidised cooked-food stallholder paid a median of about $200 a month in conservancy and about $520 for table cleaning (MSE, 8 Sep 2026): together about half as much as the median rent there of $1,450. Rent, conservancy and table cleaning come to about 1.5 times the rent at comparable centres and 1.6 times at operator-run ones ($2,170 on $1,450; $2,750 on $1,700; our arithmetic). MSE did not publish dishwashing, point-of-sale or the other charges it was asked about, and did not say whether it collects them (The Online Citizen, 15 Sep 2026). Hawkers and their advocates say the whole bill can reach double the rent; on the published medians, that would take another $650–730 a month of unpublished charges (our arithmetic). It may be so at some centres. At the median it cannot be tested until the charges are published, which is why the Plate Ledger records each one separately (§2.10). The table-cleaning fee was priced for a service that included clearing every tray, before tray return became the rule in 2021. Granted, tray return did not end the cleaners' work: they still wipe and sanitise the tables and staff the tray-return points (MSE, 14 Sep 2021). The Minister said as much at the time: "We do not anticipate any reduction in cleaning costs to stallholders" (MSE, 6 Jul 2021). So far she has been right. The fee follows the cleaning contractor's price (NEA tender notice, Aug 2026). It is cost recovery, not profit: conservancy and table-cleaning fees fund "a significant part" of a centre's running costs (MSE, 5 Feb 2024). And the Progressive Wage (the pay ladder the Government sets for lower-wage jobs) for a cleaner at a hawker centre has risen from $1,312 a month in July 2022 to $2,080 in July 2026 (MOM), a rise this plan supports. Yet the fee has not kept pace with it. At four NEA-managed centres it was the same before GST (goods and services tax) in December 2022 and August 2024 – $402 a month at Amoy Street, $360 at Chomp Chomp – while the cleaner's Progressive Wage rose by a third, from $1,312 to $1,740 (derived from NEA's tender notices; MOM). A re-tender may find the contract already priced below what the cleaner's rising wage requires, not above it.

So I no longer ask NEA to cut the fee. I ask the Government for a table-cleaning fee subsidy: $200 a stall a month at the pilot centres, three, five at most, where the evaluator can show what it does before anyone prices it nationally. It picks up where NEA's temporary support leaves off: $200 a month for every cooked-food stall from September 2026 to February 2027, meant to offset higher electricity and gas costs (NEA). And it comes with a test. Each pilot centre's cleaning contract is re-tendered for the work left after tray return; if the new price is lower, the subsidy shrinks by the difference, and the evaluator publishes both. The Government pays, as the main landlord. It is a subsidy towards a contractor's price, not rent a landlord gives up. Who re-tenders depends on who holds the contract. At an HDB (Housing & Development Board, Singapore's public-housing landlord) hawker centre, which a Town Council manages, NEA's tender notices say the table-cleaning service "is managed by the respective Hawkers Association" on behalf of the stallholders, and that its charges "depend on the table cleaning contract" (NEA tender notices, Dec 2022 and Aug 2024). Two of the three pilot centres, Chinatown Complex and Tekka, are such centres (NEA list, 17 Aug 2026). There the hawkers' association re-tenders its own contract, NEA convenes the test, the Town Council, which collects the conservancy charge, agrees to it, and the Government's $200 is paid towards the association's contract. At an NEA-managed centre such as Geylang Serai the notices name only "the table-cleaning contractor" (NEA tender notice, Aug 2026), so Year 0 confirms who holds the contract. At an operator-run centre the operator only passes the fee on: NEA's rules say operators may not profit from charges they pass on, such as table cleaning by third-party contractors (NEA, Jul 2026). So any saving there comes from re-pricing the contract, never from an operator's share. The subsidy covers table cleaning only. Dishwashing, where NEA already co-funds up to 70% of the fee for four years at existing centres, stays as it is.

Every cooked-food stall at a pilot centre pays the fee, so the cost follows the centre's size: about $2,400 a stall a year, which is $542,400 a year at Chinatown Complex's 226 stalls, $151,200 at Geylang Serai's 63 and $285,600 at Tekka's 119. That is $979,200 at the three, 408 stalls, and $1,629,600 at five, with Old Airport Road and Maxwell (Follow the Plate, D.7.4). It is an upper bound (the most it could be), because re-tendered contracts may meet part of it. Nationally, across the 6,900 stalls in public centres, it would be at most about $16.6M a year, all of it the Government's. That is a decision for later, on the evaluator's evidence, and not part of this plan. On its own, the subsidy lets a typical tendered hawker pay $200 a month more in rent before his pay falls below his assistant's: dollar for dollar. A fee should buy a service. Where the service costs what it costs, the State can choose to carry part of it, openly, and show what it bought.