Why Newer Singapore Towns Have Coffee Shops and Not Wet Markets

A visitor who knows Singapore's older estates and then walks around Sengkang or Punggol notices an absence and usually misreads it. The newer towns are not short of cheap food; they are short of one particular building. Understanding why turns on a gap in the government's own construction programme and on a different landlord.

The 26-year gap

The clearest official statement of the gap is retrospective: "In 2011, the Government announced its intention to restart building hawker centres after a 26-year hiatus, focusing on new estates or existing ones that are relatively under-served in terms of a diversity of affordable dining options." The environment agency's own factsheet puts it slightly differently — "20 new centres will be built by 2027, after a hiatus of about three decades" — and the arithmetic of the first phrasing points to about 1985 as the start, though no official document states that year.

What no official source explains is why building stopped. Searches of the agency, the sustainability ministry and the parliamentary record for a stated reason return nothing; the earliest primary statements located are the 2019 factsheet and a 2020 parliamentary reply, both looking back.

The consequence is a generation of towns planned and built with no hawker centre in them. Sengkang's comprehensive planning "begun in the 1990s" and the housing board describes it as "developed in the 1990s". The Punggol blueprint was unveiled in 1996. Both fall squarely inside the gap. The board's own design guide for Sengkang lists its commercial facilities as malls — Compass One, The Seletar Mall, Sengkang Grand Mall — plus a neighbourhood commercial centre, and names no market or hawker centre. For Punggol, the guide names "Precinct Shop Clusters" at light rail stations, designed to give residents "convenience and access to eating houses and super-marts".

That phrase, eating houses, is the answer. It is the board's own term for a coffee shop.

What the older format actually was

It is worth being precise about what the newer towns lack, because the older format was not mainly a food court. Taking the agency's 2016 stall dataset as the baseline: of 107 centres, 77 were combined market and hawker centres, 27 hawker centres only and 3 markets only. Of 14,058 stalls, 7,937 — about 56 per cent — were market produce stalls rather than cooked food.

So the older building was a grocery shop with a food court attached, and it is the grocery half that a coffee shop does not replace. A supermarket does, which is why the board's design guides for the newer towns pair shop clusters with "super-marts" rather than with market stalls.

Sengkang does now have one market-bearing centre. Fernvale Hawker Centre and Market opened with 28 cooked food stalls and appears in the agency's dataset with a completion date of 22 March 2022, co-located with a community club. Its neighbour Buangkok Hawker Centre, 38 stalls, completed 15 February 2023, has no market. Anchorvale Village, 36 stalls, completed 14 December 2023, sits with a supermarket, shops and flats — the newer answer to the same need, assembled from different parts.

The per-dish pricing question has its own guide, the economy rice decoder.

The coffee shop is the replacement, and there are 776

The housing board's figures are specific and have been restated consistently. In 2022 it reported "772 HDB-built eating houses, of which 372 are rented out by HDB and 400 are privately-owned". Joint releases in 2023 and 2024 give a slightly wider count: "a total of 776 coffeeshops in Singapore, of which 374 are rented out by HDB to private operators and 402 are privately-owned", with the note that "since 1998, HDB no longer sells coffeeshops, and only rents them out".

The last line uses a broader category than coffee shops and is not stated to be equivalent, so treat it as the outer bound rather than as an updated coffee shop count. No post-2024 official total for coffee shops specifically could be found.

The stated provision principle is simple: "To ensure that residents have access to affordable cooked food options, HDB provides a good supply of coffee shops in every HDB town and estate." Every town, not every town that lacks a hawker centre.

The supply side has kept moving, and the board publishes its pipeline. In May 2023 it reported having "completed 37 new coffee shops in the last five years" with "another 34 coffee shops slated for completion in the next five years"; by January 2024 the backward-looking figure had been restated as 31 completed over five years, including five in four towns in 2023 — two in Clementi and one each in Punggol, Bukit Batok and Woodlands — with 21 more slated for the next five years, "in tandem with the completion of new BTO developments". New coffee shops arrive with the flats, which is precisely the mechanism a hawker centre lacked during the gap.

The privately managed centres run on different terms, set out in the operator-run model.

The tender method that changed the rents

The mechanism most responsible for coffee shop prices is not a price rule but a tender design, and it changed in 2018.

Before then, board coffee shops were let through an e-bidding system. Since 2018 they have been let by the Price-Quality Method, in which "quality score counts for 50% of the evaluation, and operators with affordable food options will be awarded higher points under this scoring system". The stated purpose is to let the board "better meet the needs of residents while ensuring rents remain sustainable for operators, as PQM tenders are reviewed holistically based on a range of criteria, and not just based on price alone".

The measured effect is the useful part. The board reported in January 2024 that "over the past 5 years, from 2018 to 2023, 97% of HDB coffee shops have not had their rents increase", and that under the quality-weighted method "the rental rates for the winning PQM bids have also generally been lower – by 27% on average – compared to those awarded under the e-bidding system before".

That 27 per cent is the clearest published evidence in Singapore food policy that changing how a tenancy is awarded changes what food costs. The complement of the 50 per cent quality weighting is arithmetically 50 per cent price, but no document located states the price weighting explicitly, and no sub-criteria breakdown is published.

The same method now operates on the hawker side too. The sustainability ministry noted in November 2024 that "we also started the price-quality method at our SEHCs where rent is pre-determined and stalls are awarded by the operators".

There is one more reason the tender design carries so much weight here. Because the board no longer sells coffee shops — it has only rented them since 1998 — every new unit that comes onto the market arrives through a tender the board controls, and every existing rental unit passes through one again at renewal. Over a three-year cycle that gives the landlord a recurring lever on 374 premises without setting a single price, which is a different instrument from the hawker centre side, where the environment agency states flatly that it does not regulate food prices.

How the market half of the same building is regulated is covered in the wet market guide.

Budget meals, and what changed in January 2026

Attached to the coffee shop tenancy is a price commitment, and it was restructured with effect from 10 January 2026. The previous and current requirements differ in both directions — fewer items, more money.

For a new rental coffee shop, the requirement was "at least six budget meals and two budget drinks" with no funding support. It is now "one economy rice (with two vegetables dishes and one meat), one halal meal and one breakfast item, as well as two budget drinks", and the operator "will now receive 5% rental discount for the entire new tenancy term". Tenderers are also no longer required to propose budget meals as part of the tender bid.

For an existing rental coffee shop, the discount previously ran for one year from renewal against a requirement of four budget meals and two drinks. It now runs for the whole three-year term, and an operator not currently in the scheme takes the new three-item requirement, while an existing participant may keep four meals and two drinks or switch.

For a privately owned coffee shop, which previously got nothing, the new support is "a discount on the Temporary Occupation Licence (TOL) for Outdoor Refreshment Area (ORA) that is equivalent to 5% of the assessed market rent of the coffee shop over a 3-year period, capped at 100% of TOL fee", against the same three-item requirement.

The definition of a budget meal is published in two parts. It must be a full meal: "these budget meals must be full meals, and not side dishes, snacks, kids meal or half-portion meals", priced affordably against nearby establishments. And a 2024 release describes them as "typically priced at $3.50 and below" with drinks at "$1.20 and below" — wording that is descriptive rather than a stated contractual cap.

The scheme's history shows it scaling by crowdsourcing rather than by inspection. It launched with about 40 coffee shops. Between late October 2023 and January 2024 the public "submitted over 1,600 recommendations of budget meals and verified over 2,500 budget meals across some 310 coffee shops islandwide", and by mid-January 2024 over 250 coffee shops were listed. In June 2024 a further 180 coffee shops under nine private chain operators joined, bringing participation to roughly 330. The stated target was budget meals at all 374 board rental coffee shops by 2026. No participation figure as at the January 2026 revision could be found.

Sengkang and Punggol, sized

The reason all this matters is scale: these are not small towns waiting for amenities. They are among the largest in the country.

Sengkang houses more people in board flats than Bedok does, on more flats, and had no purpose-built hawker centre until 2022. The board's classification framework, in place since 1992, puts Tampines and Bedok in the mature category — "estates where land for public housing development was limited yet demand was high" — and Sengkang, Punggol, Jurong West and Bukit Batok in the non-mature category, "usually less central, with fewer amenities and are therefore more affordable priced". Fewer amenities is the official description, not an editorial one.

The building programme is still running. The agency's dataset lists six centres under construction with target years: Yew Tee, Telok Blangah and Chong Pang in 2027, Bukit Timah Market and Tanglin Halt in 2029, and Tampines North in 2030 — the last co-located with a community club, a bus interchange and private residential and commercial units. Which is the newer pattern in miniature: the hawker centre returns, but as one floor of something larger.

One further asymmetry is worth naming. A coffee shop operator is a business letting stalls to sub-tenants, while a hawker stallholder rents directly from the state and must operate the stall personally. That difference in structure, rather than any difference in the food, is what makes a coffee shop plate measurably dearer than a hawker centre one, and the sustainability ministry has put the comparison on the record: "generally, stall rentals at all hawker centres are lower than nearby eating establishments, such as coffee shops, food courts and small eateries."

What this guide does not cover

No stall, coffee shop or budget meal is named or recommended here. The portal is the live list and this guide explains the system behind it.

Several figures were sought and not found. No official reason for the roughly 1985 stoppage in hawker centre building exists in the published record. No post-2024 total for coffee shops could be confirmed, so the 776 figure carries a 2023 and 2024 date. No participation count for the budget meal scheme as at its January 2026 revision is published, and the nine private chain operators that joined in 2024 are not named in the material found.

Seat counts are given only where the agency published them, which is for a minority of centres. Exact years in which the first flats were completed in Sengkang and Punggol could not be sourced; the board's own guides say "the 1990s" and give a 1996 blueprint date.

Two numbers in circulation should be treated carefully. The count of operator-run centres is 14, 16 or 18 depending on the definition and the date, for reasons set out in the agency's own footnotes. And the agency's long-run stall series jumps from about 3,011 stalls in 2003 to 14,708 in 2004, which is a coverage change rather than growth and is not annotated on the dataset.