Published 1 August 2026
Do smaller condo units still sell at higher psf? We checked.
By Realila
For decades the answer was yes: smaller units sold at higher per-square-foot prices than larger units in the same building. Between 2021 and 2025, that premium fell to zero. Here is the evidence.
Analysis date: 1 August 2026. Data through 19 July 2026 (private caveats) and 1 July 2026 (HDB resale).
Ask anyone in Singapore property and you will hear the same rule: smaller units carry higher psf, larger units carry lower psf. Agents price with it. Buyers use it to judge whether a listing is expensive or cheap. Developers built their price lists around it. It is probably the most repeated rule of thumb in this market.
It was true for a long time. It is not true anymore.
Inside the very same condo developments, the price-per-square-foot premium that compact units commanded over large units has gone from roughly 16% to zero in four years. In the most recent readings it has tipped negative: large units in the same building now transact at equal or higher psf than small ones.
Nobody announced this. The market's most familiar pricing rule quietly died, and pricing habits have not caught up.
The rule, and why it existed
Three standard explanations have always been offered for the size-psf slope. First, affordability: far more buyers can pay S$800,000 than S$2.4 million, so the cheapest ticket into a location gets bid up per square foot. Second, construction: kitchens, bathrooms and building services cost roughly the same whether the unit is 450 sqft or 1,400 sqft, and spreading those fixed costs over fewer square feet raises the psf. Third, investors: compact units rent out at better per-square-foot yields and are easier to exit, so investors concentrated in them and paid up for the privilege.
Keep the third one in mind. It matters later.
What the data shows, inside the same buildings
Market-wide averages are a poor test of this rule, because they compare different buildings in different locations of different ages. The rule as people actually use it is a within-building rule: in this development, the 2-bedder trades at a higher psf than the 4-bedder. So that is where we measured it.
For every condo development, in every year, we compared the median resale psf of compact units (400–600 sqft) against large units (1,200–1,600 sqft) in that same development, wherever both had at least three transactions. The table below shows the median premium across all qualifying developments.
| Year | Small-unit psf premium (same development, resale) | Developments measured |
|---|---|---|
| 2017 | +15.5% | 24 |
| 2018 | +21.0% | 27 |
| 2019 | +12.6% | 8 |
| 2020 | +15.7% | 24 |
| 2021 | +9.8% | 62 |
| 2022 | +7.0% | 51 |
| 2023 | +6.4% | 39 |
| 2024 | +3.5% | 43 |
| 2025 | −4.1% | 52 |
| 2026 (to 19 Jul) | −11.1% | 15 |
Through 2020 the premium sat where folklore says it should: mid-teens. From 2021 it compressed every single year. By 2025 it crossed zero.
Maybe different buildings are just selling now?
A fair objection: the set of developments with enough transactions changes over time. So we ran a stricter test. We took only the developments that had both small and large units transacting in both periods, 2017–2019 and 2023–2025, and compared each development with itself.
Sixty developments qualified. Their median small-unit premium fell from +15.9% to +1.3%. The premium compressed in 56 of the 60.
Same building. Same location. Same tenure. Aged equally. The premium still collapsed.
Maybe it's a floor effect?
Larger units sometimes sit on higher floors, and higher floors carry higher psf. Could a shift in which floors were transacting explain it?
No, and we checked three ways. Floor data covers 99.2% of the relevant transactions, so nothing is hiding in gaps. The floor mix barely moved: small and large units transacted at nearly identical average floors in both periods, and what little drift existed had small units moving slightly higher, which should have propped their psf up, not down. And in the strictest version of the test, comparing small against large units within the same development and the same five-storey floor band in both periods, the premium fell from +15.7% to +0.2%, compressing in 60 of 65 measurable cases.
Floors explain nothing.
Maybe Singaporeans just want bigger homes now?
The popular story since the pandemic: work-from-home made everyone value space, so bigger units re-rated. If that were the cause, it should show up wherever Singaporeans buy homes to live in.
Singapore happens to run a near-perfect test for this. The HDB resale market is the one housing market with essentially no investors: you must live in the flat, a five-year minimum occupation period applies, and a family can own only one. Same population, same years, same pandemic, but investor demand switched off.
If space preference re-rated big homes, HDB 5-room flats should have gained on 3-room flats after 2020. They did the opposite:
| Year | HDB 5-room psf vs 3-room psf (resale, median) |
|---|---|
| 2014 | −12.0% |
| 2017 | −8.4% |
| 2020 | −0.3% |
| 2022 | −2.9% |
| 2024 | −3.8% |
| 2026 (to 1 Jul) | −2.6% |
HDB's size gap converged for its own reasons up to 2020, then drifted slightly apart again. The private market's within-building premium moved 15 to 20 points in the opposite direction over the same years. A population-wide taste for space cannot produce opposite movements in the two markets. That explanation fails. (For the broader HDB picture this quarter, see our HDB resale brief for Q2 2026.)
Maybe developers stopped building small units?
They did build fewer, and by regulation. URA first capped unit counts in 2012 precisely to moderate what it described as excessive development of shoebox units. In October 2018 it tightened the cap: outside the Central Area, the maximum number of units became the building's gross floor area divided by 85 sqm, up from 70 sqm, with a stricter 100 sqm divisor in nine areas under infrastructure strain, effective for development applications from 17 January 2019 (URA circular URA/PB/2018/06-DCG). The October 2022 revision went further: under the guidelines applying to applications from 18 January 2023, developments outside the Central Area must keep units of 50 sqm or less to at most 20% of the project, and provide at least 20% of units at 100 sqm or larger (URA circular URA/PB/2022/10-CUDG).
But notice what restricting the new supply of compact units should do to existing compact units: make them scarcer, and therefore relatively more expensive. The premium should have risen.
It fell. Wrong direction. That explanation fails too.
The explanation left standing
Strip away composition, floors, space preference and supply, and one candidate remains: the buyers who paid the premium stopped showing up.
Compact units were the investor product. Lower entry price, better rental psf, easier resale. Then the maths broke, in two blows. From 16 December 2021, ABSD for a Singapore Citizen's second residential property rose from 12% to 17%, and for foreign buyers from 20% to 30%, while the Total Debt Servicing Ratio threshold tightened from 60% to 55% (MOF/MND/MAS joint announcement, 15 December 2021). From 27 April 2023, the second-property rate for Citizens rose again to 20%, and the foreign-buyer rate doubled to 60%. On top of that, interest rates climbed steeply through 2022 and 2023, crushing the yield spread that made small units work as an income play.
The shape of the collapse fits this mechanism precisely. It is not a one-year shock. It is a steady four-year bleed, exactly what cumulative policy pressure produces as holding periods roll over and marginal buyers exit one by one.
One honesty note: caveat records do not identify who the buyer is, so this is an inference by elimination and timing, not a headcount of investors. But every rival explanation has now been tested against the data and failed, and this one fits both the timing and the shape.
Even the showflat has stopped believing it
The final tell. Developers' own launch pricing has historically built the size-psf slope in. Within the same new launch, the small-unit premium ran +12% to +15% in 2016–2017. By 2023–2024 it was low single digits. In 2025 it was −0.3%, and in the first half of 2026, across the seven launches measurable, −9.6%.
When the people who set prices for a living stop pricing the rule, the rule is gone.
What this means for you
If you are buying a compact unit: "the psf is high because it's small" is no longer, by itself, a reason a price is fair. Compare against what large units in the same building actually transact at. The old discount structure has inverted.
If you are selling a compact unit: pricing off your building's psf history from three or four years ago will leave you above the market, and sitting. The premium your unit type used to command has gone.
If you own a large unit: relative to your own building, you have quietly re-rated upward. Your comparable set has shifted in your favour.
If you are an agent: any comparative market analysis that still assumes the size-psf slope is now wrong inside the very buildings it compares. The slope has to be measured per development, per year, not assumed.
Methodology and caveats
Population: private non-landed resale transactions (condominiums and apartments; Executive Condominiums excluded; collective sales excluded) from URA caveat data, and HDB resale transactions, as held in Realila's database on 1 August 2026. Latest private caveat in sample: 19 July 2026. Latest HDB resale: 1 July 2026. 2026 figures are part-year.
Definitions: compact = 400–600 sqft; large = 1,200–1,600 sqft; premium = median psf of compact units divided by median psf of large units, minus one, computed within each development (and, where stated, within the same five-storey floor band, matching the granularity at which floors are reported), then taken as the median across developments. A development-year qualifies when both size bands have at least three transactions (two, in the floor-matched cells).
Known limits: some annual readings rest on small development counts (2019: 8; 2026: 15) and should be read as part of the trend, not alone. The 2026 new-launch reading covers seven projects. Market-wide psf averages were deliberately not used as evidence, because they mix buildings, locations and vintages; the within-development design exists to remove exactly that. The related folklore claim that "2-bedders carry the highest psf of all" is a separate hypothesis and has not yet been tested; it is next on our list. Buyer identity is not observable in caveat data; the investor-withdrawal mechanism is supported by elimination and timing, as stated above.
Every figure in this article traces to a query against Realila's transaction database and carries the dates above. On why we work in medians rather than averages, and how medians differ from price indices, see Property Price Indices, Explained.
Policy sources: URA circulars URA/PB/2018/06-DCG (17 October 2018) and URA/PB/2022/10-CUDG (18 October 2022); MOF/MND/MAS joint announcement of 15 December 2021; ABSD rates effective 27 April 2023 as held in Realila's policy reference table.
Related reading on Realila:
Every number in this post is drawn from Realila Research and dated to when it was true. In Research you can change the period, compare segments, and drill into projects on the same data.
A matching brief on the other main segment is on the Feed.
Also read: HDB Resale Prices Q2 2026: The Median Held at S$604 psf. Here's What Moved Underneath.Create a free account to use Research and Lila. We will use your account as the place for product updates as more of the platform opens. There is no separate email newsletter for the Feed yet.