Market Insights
The questions Singapore property buyers actually ask — answered with data from our full transaction database. For the full technical detail behind each answer, see the methodology.
Data-driven answers
Freehold vs leasehold — which performed better?
Freehold / 999yr
not enough data
0 matched pairs
99yr leasehold
not enough data
0 matched pairs
Repeat-sales method · matched buy→sell pairs held >5 years · all districts · Singapore condos + landed
See district-by-district breakdown and hold-period simulator →
Gov-firstURA caveats · matched resale pairs
Measured the honest way — the average annualized return on matched buy→sell pairs (the same physical unit sold twice) — freehold/999yr and 99-year leasehold have performed remarkably close across the whole market. Across all districts, both bucket at roughly 2.5–3% per year; the gap between them is well under a percentage point and flips by district.
Bottom line: tenure is not the lever most buyers think it is for realized return. A young 99-year leasehold bought at the right price has historically kept pace with freehold; the real risk is buying a leasehold late in its life. Use the holding-period and district controls to see your own slice.
Repeat-sales method: each pair’s return is (sell ÷ buy)^(1 ÷ years held) − 1. Freehold and 999-year leasehold are grouped together (economically equivalent). A bucket with too few matched pairs shows “not enough data” rather than a misleading number — the pair count is always shown.
MRT Effect Study — does being near an MRT mean higher returns?
Repeat-sales method · median annualized return on matched buy→sell pairs held >5 years · Singapore condos · distance to nearest MRT
Gov-firstURA caveats × LTA / OneMap MRT map
Measured the same honest way as the tenure study — the median annualized return on matched buy→sell pairs (the same condo sold twice) — bucketed by distance to the nearest MRT. Each bar is the median realized return per year for condos in that distance band, so you can see whether walk-to-MRT stock actually compounds faster.
How to read it: compare the bars, but watch the matched-pair count under each band — a band with a thin sample is shown as “not enough data” rather than a misleading number. Condo-first: landed’s MRT effect is weaker and its tenure data has gaps, so this study covers condos.
Repeat-sales method: each pair’s return is (sell ÷ buy)^(1 ÷ years held) − 1, same methodology as the tenure study above. Distance to the nearest MRT is per project (≈230 geocoded MRT/LRT stations); condos without an MRT distance on record are excluded rather than guessed.
If I go for leasehold, at what remaining lease does it become non-viable?
99-yr leasehold condos · last 10 years · Singapore
Gov-firstURA caveats · 99-yr leasehold condos
The discount accelerates sharply below 60 years remaining. Above 60 years, PSF holds close to freehold parity. Between 40–59 years the discount widens to ~15–25%. Below 40 years, CPF usage is restricted, shrinking the buyer pool and compressing liquidity further.
Practical threshold: 60 years remaining is the floor for a liquid, CPF-eligible asset. Below 40 years, price in the financing constraints on both your offer and your expected exit.
Based on 99-yr leasehold condo transactions, last 10 years. Bands with fewer than 50 transactions excluded.
Which performs better — 1-bedroom, 2-bedroom, 3-bedroom, 4-bedroom, or 5-bedroom?
HDB resale · yearly median PSF · last 10 years · all Singapore towns
Gov-firstHDB resale records (data.gov.sg)
Using HDB resale data (the cleanest apples-to-apples comparison — URA condo unit-type records are freeform): smaller flat types (3-Room and below) deliver stronger PSF growth. 4-Room and 5-Room flats command lower PSF but often deliver better absolute dollar gains due to larger area. Executive flats are compressed by limited liquidity.
For private condos: 1–2 bedroom units command the highest PSF; 3–4 bedroom units offer the best absolute appreciation in prime districts.
HDB resale only. MULTI-GENERATION excluded (very low volume).
Do bedroom-count returns differ by district?
Repeat-sales method · actual buy→sell resales of the same condo unit, held 4–10 years · exact bedroom counts from haio’s unit-level records (unknowns excluded, never guessed) · solid bars need 30+ pairs; 5–29 render muted with the count and a low-sample marker; fewer show the count only
Gov-firstURA caveats · matched resale pairs
Yes — and a district can invert the island-wide story. Island-wide, the larger private-condo buckets (3–4 bedroom) have delivered the stronger annualized returns from actual resales of the same unit. Pick a single district and the ordering can flip — which is why this chart always overlays the island-wide bars when a district is selected. Check the district you are actually buying into, not the national average.
The metric is the average annualized return from actual buy→sell resales of the same condo unit (the repeat-sales method — same maths as the freehold-vs-leasehold chart above), bucketed by the unit’s exact bedroom count from haio’s unit-level records. Units without a known bedroom count are excluded, never guessed.
Private condos only. Every bar shows its real pair count: buckets with 5–9 resale pairs render muted with the count on top, and buckets with fewer than 5 show only the count — never a thin, misleading number.
Which district should I buy into?
Select up to 5 districts to compare
Private condos · yearly median PSF · last 10 years
Top 10 districts by latest median PSF
| District | Latest PSF | 3-yr CAGR | Txns (yr) |
|---|---|---|---|
| D6 | $3,167 | +0.1% | 4 |
| D9 | $3,053 | +6.0% | 774 |
| D2 | $3,023 | +9.8% | 264 |
| D1 | $2,798 | +12.2% | 223 |
| D16 | $2,551 | +18.7% | 1,074 |
| D3 | $2,323 | +1.7% | 396 |
| D10 | $2,311 | -3.4% | 482 |
| D7 | $2,305 | -4.1% | 95 |
| D5 | $2,299 | +7.8% | 764 |
| D21 | $2,230 | -1.9% | 363 |
Gov-firstURA caveats · by district & year
By latest PSF: D9, D10, D11 consistently lead. D1–D4 follow with strong recent momentum. By 3-yr CAGR: D19 and D20 offer competitive growth at lower entry prices. By liquidity: D15 and D19 are the most active markets — liquidity matters at exit.
Private condos only (URA caveats). Yearly median PSF across all condo transactions in that district and year. Districts with fewer than 20 transactions in a year excluded.
How do I read a new launch's price against its neighbouring projects?
haio transaction records (URA caveats) — per-project median $ psf across all sale types, 12 months to 12 Jun 2026. Lentor Hills Residences (n=2) and Lentor Central Residences (n=5) are thin counts — read those bars as indicative only.
Gov-firstURA caveats · 12-month medians
Put the actual neighbours on one axis, not the district average. A district aggregate blends old resale stock with new launches and hides the spread that matters. The honest comparison is each comparable project’s own median transacted $ psf over the last 12 months — here, the Lentor enclave’s six launches plus Springleaf Residence in the wider District 26 corridor, with Lentor Gardens Residences’ expected launch band shaded for context.
Sold-out launches keep printing through sub-sales and resales, so the 12-month medians cover all sale types; the transaction count behind every bar is in its tooltip — a thin count means an indicative bar, not a verdict.
haio transaction records (URA caveats), 12 months to 12 Jun 2026. The shaded band is a market-indicative expectation calibrated to unconfirmed agent-flyer anchors — not a released price list.
Does primary school matter?
Data coming
We’re building the school-proximity comparison — realized median-PSF growth for homes within 1 km of a primary school versus those outside, by holding period (all schools). The transaction-to-school join isn’t in the public data layer yet, so we’re not putting a number here we can’t stand behind.
Singapore buyers pay a premium to live within 1 km of a popular primary school. The open question is whether that premium translates into better resale appreciation over a typical holding period, or whether you simply pay more going in.
The data layer currently exposes only a per-project nearest-school distance, not a transaction-level within-1km flag joined to realized returns. This section is a “data coming” shell. See the methodology for what we hold.
How much new residential land has the State released?
2026: 9,320 homes — about 54% above the decade average of 6,060.
| Year | Confirmed-List homes |
|---|---|
| 2016 | 3,730 |
| 2017 | 5,170 |
| 2018 | 5,410 |
| 2019 | 3,740 |
| 2020 | 3,145 |
| 2021 | 3,605 |
| 2022 | 6,290 |
| 2023 | 9,250 |
| 2024 | 10,500 |
| 2025 | 9,755 |
| 2026 | 9,320 |
GLS Confirmed List · private residential (incl. EC) · URA media releases
Gov-firstURA Government Land Sales programme
Each year the Government Land Sales (GLS) programme puts a set number of private homes on its Confirmed List — sites sold on a fixed schedule regardless of demand. That is the upstream tap that feeds new condo and EC launches. The chart shows the Confirmed-List homes released per year across the past decade, straight from URA’s half-yearly releases.
How to read it: the latest year is highlighted, and the dashed line is the average of the years before it. When the highlighted bar towers over that line, the State is releasing unusually heavy supply — a forward-looking read on supply, not on prices.
Source: URA Government Land Sales programme media releases. The figures are the Confirmed-List private-residential supply (including Executive Condominiums) the State published for each year — never an estimate of ours.
How many new private homes get completed each year?
When yearly completions run above ~20,000, new supply tends to weigh on prices; when they fall below ~10,000, scarcity supports them. URA expects 6,955 units in 2026 (forecast).
| Year | Units | Type |
|---|---|---|
| 2014 | 23,298 | Actual |
| 2015 | 22,267 | Actual |
| 2016 | 26,288 | Actual |
| 2017 | 20,568 | Actual |
| 2018 | 13,242 | Actual |
| 2019 | 8,913 | Actual |
| 2020 | 4,061 | Actual |
| 2021 | 6,388 | Actual |
| 2022 | 9,526 | Actual |
| 2023 | 21,284 | Actual |
| 2024 | 10,617 | Actual |
| 2025 | 7,996 | Actual |
| 2026 | 6,955 | Forecast |
| 2027 | 10,021 | Forecast |
| 2028 | 10,701 | Forecast |
| 2029 | 13,028 | Forecast |
Source: URA Real Estate Statistics · private residential (incl. Executive Condominiums) · completions by year. Forecast years are URA’s expected completions — never an estimate of ours.
Gov-firstURA Real Estate Statistics
Completions are the actual handover of finished private and Executive-Condominium homes — the moment new supply hits the market. They move in cycles: a wave of launches today becomes a wave of completions three to four years later.
How to read it: the chart shades three bands. When yearly completions run above ~20,000, the market is absorbing heavy supply and prices tend to soften; around ~10,000 is balanced; below ~10,000 is a supply crunch, where scarcity supports prices. The pale, dash-bordered bars from 2026 on are URA’s expected completions — clearly marked as forecast, not record.
Source: URA Real Estate Statistics — Completion of Private Residential Units (including Executive Condominiums), by year. Actual completions through 2025; 2026–2029 are URA’s expected completions, never an estimate of ours.
What happened to en-bloc (collective) sales?
Peak: 2018 — 45 deals releasing 3,185 units. The pipeline thinned sharply afterwards.
| Year | Deals | Units lost |
|---|---|---|
| 2013 | 10 | 240 |
| 2015 | 1 | 45 |
| 2016 | 7 | 600 |
| 2017 | 33 | 2,764 |
| 2018 | 45 | 3,185 |
| 2019 | 6 | 342 |
| 2020 | 6 | 147 |
| 2021 | 12 | 342 |
| 2022 | 4 | 575 |
| 2023 | 5 | 125 |
| 2025 | 1 | 9 |
Source: collective-sale (en-bloc) transaction history, 2013–2025. Shows deals concluded + existing units released; new replacement-unit counts are omitted (our coverage of those is too thin to report honestly).
Gov-firstURA collective-sale history
En-bloc (collective) sales are the second feeder of new private supply — old estates sold whole to a developer, then redeveloped. The chart shows deals concluded and existing units released each year, 2013–2025.
How to read it: the peak year is highlighted. After 2018 the deal count fell sharply and has stayed low — meaning one of the two pipelines that refill private housing supply has been running near-empty for years. Fewer feeders today implies thinner new supply in the years ahead.
Source: collective-sale transaction history. We chart deals and units lost (which we track in full); we deliberately omit new replacement-unit counts, where our coverage is too thin to report honestly.
Is there a best month of the year to buy or sell?
Quietest month: Dec (14,500 deals) — thinner competition. Busiest: Mar (23,395).
| Month | Transactions | Median PSF |
|---|---|---|
| Jan | 16,501 | $1,564 |
| Feb | 16,077 | $1,524 |
| Mar | 23,395 | $1,546 |
| Apr | 22,648 | $1,586 |
| May | 21,629 | $1,518 |
| Jun | 17,095 | $1,522 |
| Jul | 22,913 | $1,522 |
| Aug | 20,341 | $1,538 |
| Sep | 18,234 | $1,520 |
| Oct | 19,854 | $1,552 |
| Nov | 19,870 | $1,645 |
| Dec | 14,500 | $1,504 |
Source: URA Real Estate Statistics · private residential caveats · trailing 10 years, by calendar month. Median PSF = recorded price ÷ recorded area.
Gov-firstURA caveats · by calendar month
Most buyers ignore the month they act in. The transaction record doesn’t: both volume and price move with the calendar. The chart shows transaction volume (bars) and median PSF (line) by calendar month, averaged over the past ten years.
How to read it: the quietest month is highlighted — thinner competition for buyers. Busier months mean more demand for sellers. Timing within the year is a free optimisation most people leave on the table.
Source: URA Real Estate Statistics, private residential caveats, trailing 10 years, by calendar month. Median PSF = recorded price ÷ recorded area.
Which HDB flat type costs the most per square foot?
HDB resale data · median PSF by flat type · last 12 months · all Singapore towns
Gov-firstHDB resale records (data.gov.sg)
By the dollar, bigger flats cost more — obviously. The sharper question is price per square foot, and there the pattern inverts: smaller flats carry the higher PSF. Over the last 12 months of HDB resale, 1-Room and 2-Room flats sit well above the larger types per square foot, while 5-Room and Executive units are among the lowest.
How to read it: a bigger flat is more money in total but often better value per square foot. The bars show median PSF for each flat type; hover any bar for its median dollar price and transaction count.
HDB resale data only. Median PSF of all resale transactions in each flat type over the last 12 months, all Singapore towns.
Which HDB towns are the most and least expensive to buy into?
HDB resale data · median price by town · last 12 months · towns with 30+ resale transactions
Gov-firstHDB resale records (data.gov.sg)
Ranked by median HDB resale price over the last 12 months, the central and mature estates sit at the top while the newer outlying towns anchor the affordable end. The chart ranks every town most-to-least expensive; hover a bar for its median PSF and transaction count.
How to read it: headline price reflects flat-size mix as much as location — a town heavy on larger flats reads pricier even at a similar PSF. Use the PSF in the tooltip to compare like-for-like.
HDB resale data only. Towns with fewer than 30 resale transactions in the last 12 months are excluded.
What ABSD rate will I pay?
Source: IRAS · Additional Buyer’s Stamp Duty rates in force
Gov-firstIRAS · stamp-duty schedule
Additional Buyer’s Stamp Duty (ABSD) is the headline cooling measure on Singapore residential purchases. The rate depends on who you are and how many residential properties you already own. Singapore Citizens pay nothing on a first home and step up sharply from the second; Permanent Residents pay from the first; foreigners and entities sit at a flat top rate regardless of count.
How to read it: the chart shows every buyer profile side by side, grouped by property count — so the full ABSD ladder is visible at once. ABSD is charged on top of the standard Buyer’s Stamp Duty and is payable within 14 days of purchase.
Source: IRAS. Rates in force at time of publication; the Government revises them periodically as a cooling-measure lever.
About haio
What is haio?
A free, public database of every recorded Singapore residential property transaction since 1995 — HDB, condo, and landed. No signup, no paywall, no ads. The home page search and district pages are the two main ways to explore it.
How accurate is the price estimate?
The estimate on each property page uses ValueAI’s AVM, trained on 1,658,309 Singapore residential transactions. Every estimate is shown as a 25th–75th percentile range (not a single number) with the count of comparable sales it’s based on. We never hide a thin sample. It is not an appraisal — banks and licensed valuers run their own underwriting. See methodology §3 for the gating rules.
Why is the chart empty for some properties?
The PSF chart is drawn from the last 3 years of comparable transactions in the same district + tenure cohort, within ±15% of the subject’s area. When fewer than 5 such rows exist, or when every transaction in scope lacks a recorded area_sqft (common on older HDB rows), the chart falls back to an empty state. The underlying transactions still appear in the table below — it’s the chart that can’t draw them.
Does freehold vs leasehold actually matter at resale?
In Singapore, tenure affects resale pricing — but less than most buyers expect, and only predictably once a leasehold property falls below ~60 years remaining. Young leaseholds (99-year with 80+ years left) transact at near-parity with freehold in the same district and building class. The gap widens sharply as lease shortens. See the Freehold vs Leasehold deep-dive for the district-by-district data.
What does "cadastral-only" mean?
Some landed properties are catalogued from SLA OneMap polygons but have no recorded transaction in URA’s database. They are real properties; their addresses are accurate; but we have no price history. These pages carry a caveat banner labelled “cadastral-only”. Treat them as “this address exists”, not “this address has been on the market”. See methodology §7.1 for the exact count.
Can I download the data?
Not yet through haio. The methodology page lists every public source we pull from — URA caveats, HDB resale from data.gov.sg, SLA OneMap, MAS SORA — so you can replicate the same dataset yourself. A bulk export endpoint is on the roadmap.
Is this site sponsored or advertising?
No. haio shows no listings, takes no agent fees, runs no ads. The estimate uses ValueAI’s AVM purely for the price model — no commercial relationship affects what you see. It is built and maintained by one person in Singapore; contact details are in §9 below.
How often is the data updated?
URA private caveats: pulled daily, though URA itself lands them roughly weekly behind the actual transaction. HDB resale: daily delta from data.gov.sg. SLA landed registry: weekly. SORA rates: daily. Every source surfaces its last refresh time on the methodology page.
Can I see homes that are currently for sale?
Yes. Alongside the transaction database, haio shows new-launch condos and landed homes that are currently for sale — natively, right here in haio. Browse them on the Listings page, filter by condo or landed, and open any listing for its full details, unit mix, and price history. No need to leave haio.
I think a number on this page is wrong. What do I do?
Email wttg72@gmail.com. Cite the URL and the field. We’ll trace it back to the source dataset (URA, HDB, SLA, or MAS) and either fix it or document why we can’t.
Is my data being collected?
haio uses no third-party analytics and no cookies beyond what your browser stores locally. Starred properties live in your browser’s localStorage — they are never sent to a haio server. The site is hosted on Vercel.
Still curious? The methodology document is the canonical reference for how every figure on this site is computed.