Q3 2026 market: private prices gained 1.4% while HDB resale slipped for a third straight quarter
URA’s flash estimate says private home prices rose 1.4% in the third quarter, the biggest quarterly gain I can find since the end of 2024. HDB’s says resale prices slipped 0.2%, a third straight dip. Both headlines are right. They are not the same size of story, and the private number hides a split that matters if you are buying in the city.
The short version. URA’s flash estimate (1 Oct 2026) has the private residential price index up 1.4% in Q3 2026 (219.4 to 222.5), after +0.5% in Q2. Landed rose 2.8% and non-landed 0.9%. Inside non-landed, the Outside Central Region gained 2.2%, the Rest of Central Region 0.2%, and the Core Central Region fell 0.1%. HDB’s flash resale price index is 202.4, down 0.2% from 202.8, after −0.1% in Q1 and −0.3% in Q2. That makes it the third straight quarterly decline, and about −0.6% for the first nine months.
What I would take from it: private prices are rising on thin volume and mostly in the suburbs and on landed homes; HDB resale prices are drifting down while the number of deals goes up. That is a more comfortable market for HDB buyers than the headline sounds, a more selective one for HDB sellers, and a wider gap for upgraders to bridge.
Inside HDB (Q3 against Q2, data.gov.sg): deals rose in every flat type: 3-room 1,581 (+6.1%), 4-room 3,270 (+21.0%), 5-room 1,820 (+28.3%), Executive Apartment (EA) 323 (+41.0%) and Executive Maisonette (EM) 208 (+42.5%). Deals at $1m or more were 4-room 241, 5-room 202, EA 65, EM 81 and 3-room 3. By bedroom: there is no public full-quarter split of private sales into 1 to 5 bedrooms; the verifiable pieces are two July launches, below. Landed: prices +2.8% on the flash, volume roughly steady at 436 to 524 caveats a quarter on third-party counts, and the largest landed deal I can source is $24 million at Oriole Crescent.
When a reader sends me a headline like “private up, HDB down” the next message is usually “so should I buy now or wait?” I cannot answer that from two flash numbers, and neither can anyone else. What I can do is check the claim against the official releases, split it into the parts that moved, and tell you which parts I think matter for which kind of buyer or seller. Facts below come from URA, HDB and data.gov.sg; where I am giving you my own read, I say so.
+1.4%
Private prices, Q3 flash
−0.2%
HDB resale prices, Q3 flash
+2.2%
Outside Central Region (CCR: −0.1%)
3rd
Straight HDB resale dip
Checking the claim
Both halves are true, and I checked the streak
Private prices gained. Confirmed. URA’s release of 1 October 2026 gives +1.4% for the quarter, against +0.5% in Q2. For the first three quarters of 2026 the index has risen by an average of 0.9% a quarter, which URA notes is the same as the 2025 average. Looking back through URA’s own quarterly tables, every one of the eight quarters from Q4 2024 to Q3 2026 was positive, and 1.4% is the largest gain since the 2.3% in Q4 2024.
HDB resale prices dipped for a third straight quarter. Also confirmed, with one wrinkle in the wording. HDB’s resale price index went 203.6 (Q4 2025, unchanged), 203.4 (Q1 2026, −0.1%), 202.8 (Q2, −0.3%) and 202.4 (Q3 flash, −0.2%). So it is three declines in a row, not four, because the quarter before the run started was flat rather than down. Over the year to Q3 the index is down about 0.6%, while the private index is up about 3.4%.
Last four quarters, quarter-on-quarter · Source: URA quarterly releases and HDB Resale Price Index table; Q3 2026 is a flash estimate for both
| Quarter | Private index | Private change | HDB RPI | HDB change |
|---|---|---|---|---|
| 4Q 2025 | 216.4 | +0.6% | 203.6 | 0.0% |
| 1Q 2026 | 218.3 | +0.9% | 203.4 | −0.1% |
| 2Q 2026 | 219.4 | +0.5% | 202.8 | −0.3% |
| 3Q 2026 (flash) | 222.5 | +1.4% | 202.4 | −0.2% |
Swipe to see the full table →
Private homes against HDB resale, both rebased to 4Q 2024 = 100 · Source: URA and HDB index tables; 3Q 2026 flash
Swipe to see the full chart →
Private market
The 1.4% is a suburban and landed story
Start with how URA splits it. Non-landed homes rose 0.9%, after a 0.1% fall in Q2. Landed homes rose 2.8%, after 2.5%. Within non-landed, the three regions did not move together at all.
Private residential price index by segment · Source: URA flash estimate, 1 Oct 2026 (Annex A); 2Q 2025 and 3Q 2025 levels from URA’s 3Q 2025 release; year-on-year is my calculation from the index levels
| Segment | 2Q 26 index | 3Q 26 flash | Q3 change | Q2 change | vs 3Q 2025 |
|---|---|---|---|---|---|
| All private | 219.4 | 222.5 | +1.4% | +0.5% | +3.4% |
| Landed | 258.4 | 265.6 | +2.8% | +2.5% | +8.5% |
| Non-landed | 210.6 | 212.5 | +0.9% | −0.1% | +1.9% |
| Core Central (CCR) | 161.5 | 161.3 | −0.1% | +1.8% | −1.3% |
| Rest of Central (RCR) | 226.2 | 226.7 | +0.2% | −1.2% | +0.5% |
| Outside Central (OCR) | 271.1 | 277.1 | +2.2% | −0.1% | +5.4% |
Swipe to see the full table →
Quarter-on-quarter change by segment, 2Q 2026 final against 3Q 2026 flash · Source: URA, 1 Oct 2026 and 24 Jul 2026
Swipe to see the full chart →
Two things stand out. First, the Core Central Region index at 161.3 is about 1.3% below where it was a year ago (163.4 in Q3 2025). So “private prices are up” is not true of the most expensive districts over twelve months. Second, the Outside Central Region is up about 5.4% on the year, which is where most of the buying by upgraders and young families actually happens.
I would not read 1.4% as a broad-based surge. Landed deals are few and lumpy, so one or two big sales can swing a quarter, and URA’s own flash for landed in Q1 (−1.8%) was later revised to −0.4%. The OCR number is more meaningful to me, because it comes from thousands of apartments. It also coincides with July’s OCR launches, the largest of which I mention below. URA does not attribute the move to any project, so that link is my inference, not a fact.
Flash estimates move
URA builds the flash from stamp-duty transactions and developer sales up to about mid-September, and warns that flash estimates may differ from the final. This year they have: the Q1 flash said +0.3% overall and the final was +0.9%. The Q2 flash (+0.5%) matched its final, though the segments shifted by a few tenths (for example, the CCR from +2.0% to +1.8%). Treat 1.4% as a good first draft. The final Q3 statistics are due on 23 October 2026.
Landed in Q3: volume, type split and the biggest deals
The table above gives the landed index. This is everything else I can verify about landed in the quarter, and where it stops. The index went from 258.4 to 265.6 (+2.8%), about 8.5% above a year ago, after +2.5% in Q2 (flash +2.6%) and −0.4% in Q1 (flash −1.8%). Compounded, that is about +4.9% for the first nine months of 2026.
Volume. URA’s flash gives 4,296 private sales and no landed split, so there is no official Q3 landed count yet; the final tables are due on 23 October. The public counts below differ by definition but all sit between about 420 and 520 a quarter, which is why I describe landed volume as steady while total private volume fell about 30%.
Landed sales volume and type data from public sources · Counts use different definitions (caveats, exclusions), so do not subtract one source from another.
| Source | Period | Landed count | By type and limit |
|---|---|---|---|
| ERA, citing URA caveats | 1Q 2026 and 2Q 2026 | 447 then 502 | No type split in what I read. |
| Huttons, via EdgeProp | 4Q 2025 and 1Q 2026 | 482 then 418 | 1Q 2026: 228 terrace, 133 semi-detached, 57 detached; average psf $2,284, $1,934 and $2,069. Excludes Good Class Bungalows and cluster homes. |
| PropKaki tracker (URA caveats) | 3Q 2026 so far versus 2Q 2026 | 436 versus 524 (−16.8%) | Median land psf in Q3: terrace $2,411, semi-detached $1,989, detached $2,155. Third-party computation I could not reproduce; it does not say how far into the quarter it runs; per-type counts not published. |
| Knight Frank, quoted 1 Oct | Each quarter of 2026 | About 500 to 600 | An analyst’s round range, not a count. |
| Realion (OrangeTee and ETC), via Real Estate Asia | Good Class Bungalows, 1Q and 2Q 2026 | 4 then 7 caveated deals | 2Q average land rate $2,341 psf, largest a $64.9 million Nassim Road sale. No verified Q3 GCB count, and none in the EdgeProp columns I read. |
| URA flash, 1 Oct | 3Q 2026 to mid-September | 4,296 private sales, all types | The release and its Annex give price indices, not landed counts or a type split. |
Swipe to see the full table →
By type. The only verified count by type is Huttons’ first quarter: terrace 228, semi-detached 133, detached 57. For Q3 I have only a third-party tracker’s median land psf by type, shown with that label. A full terrace, semi-detached, detached and Good Class Bungalow split for Q3 needs REALIS or URA’s final tables. For Good Class Bungalows I found Realion’s second-quarter count and nothing for Q3.
The biggest deals. The ten largest I can source are below. The top is Oriole Crescent at $24 million, from the 2 October column, so a mid-to-late September deal that I cannot date more closely. Four of the ten are over $20 million. Seven are described as detached houses or bungalows, one as semi-detached.
Highest-priced landed deals I can source for Q3 2026 · Partial list assembled from nine EdgeProp “What’s moving the market” columns between 24 Jul and 2 Oct; each names only a few notable sales, so a bigger deal could exist that none of them mentions. Psf is on land area as EdgeProp gives it. Dates are as the column gave them.
| # | Street | Type as stated | Land sq ft | Price | Psf | Source and date |
|---|---|---|---|---|---|---|
| 1 | Oriole Crescent | House (type not stated) | not given | $24,000,000 | $2,155 | EdgeProp 2 Oct (fortnight to 30 Sep) |
| 2 | Trevose Crescent | Freehold detached | 7,782 | $21,500,000 | $2,763 | EdgeProp 24 Jul; dated 8 Jul |
| 3 | Pasir Panjang (street not given) | Freehold bungalow | 15,830 | $21,000,000 | $1,327 | EdgeProp 31 Jul (fortnight to 28 Jul) |
| 4 | Gentle Road | Freehold detached | 8,762 | $20,380,000 | $2,327 | EdgeProp 4 Sep; sold 20 Aug |
| 5 | Gentle Road | Eight-bedroom bungalow | 7,661 | $18,500,000 | $2,415 | EdgeProp 28 Aug; sold 11 Aug |
| 6 | Princess of Wales Road | Detached, 999-year lease | 6,485 | $17,500,000 | $2,698 | EdgeProp 24 Jul; date not given |
| 7 | Cove Drive, Sentosa Cove | Detached, 99-year lease from 2007 | 8,241 | $17,000,000 | $2,063 | EdgeProp 7 Aug; sold 28 Jul |
| 8 | Bournemouth Road | Freehold detached | 6,462 | $16,200,000 | $2,507 | EdgeProp 21 Aug (fortnight to 19 Aug) |
| 9 | Berrima Road | Landed house (type not stated) | 5,304 | $14,880,000 | $2,805 | EdgeProp 21 Aug |
| 10 | Trevose Crescent | Freehold semi-detached | 4,092 | $14,750,000 | $3,604 | EdgeProp 31 Jul (fortnight to 28 Jul) |
Swipe to see the full table →
Highest-priced landed deals sourced for Q3 2026, $ million · Source: EdgeProp columns, my compilation; partial list
Swipe to see the full chart →
Just outside the ten, and worth knowing for the psf point: Clover Avenue $13.9 million, Vanda Drive nearly $13.69 million, Woo Mon Chew Road $13.54 million, Branksome Road $13.168 million, and two semi-detached houses at much higher land prices, Boscombe Road at $9.8 million ($3,610 psf, 10 July) and Fir Avenue at $9.85 million ($3,054 psf).
The index says landed prices keep rising; the deal list says why a headline is slippery. The three highest land prices per square foot among the deals I read were small freehold semi-detached plots at about $3,000 to $3,600 psf, while the $21 million Pasir Panjang bungalow was $1,327 psf. A big cheque usually means a big plot. With roughly 420 to 520 landed deals a quarter, a handful of $20 million-plus sales can lift an average, which is one reason I would wait for the final 23 October numbers before treating +2.8% as settled. I would compare landed homes street by street on land psf.
Volumes and launches
Prices up, deals down: private sales are thin
URA counted 4,296 private sale transactions in Q3 up to mid-September, against 6,148 for the whole of Q2, which it describes as a fall of about 30% quarter on quarter. The comparison is not like for like, because the Q3 count stops two weeks short of quarter-end, so I would read it as “clearly quieter” rather than “exactly 30% lower”. On a year-to-date basis URA has 15,857 sales, against 19,793 in the same period of 2025.
My read is that the quiet is mostly new-launch driven, since the launch calendar was thin after July. Developers sold 731 new private homes (excluding ECs) in July, according to URA data reported by EdgeProp, with two launches taking about two-thirds of that: Lentor Gardens Residences moved 270 units at a median of about $2,357 psf and Dunearn House 212 units at about $3,111 psf. In August the figure was reported at 153 units, with only 116 homes released, all from a later phase of one project and with no brand-new launch in the month. August also sat inside the Hungry Ghost Month, which ran from 13 August to 10 September.
That is 884 units across July and August, against 2,141 for the whole of Q2 and 3,288 in Q3 2025 (both URA finals). September’s developer-sales figure had not been published when I wrote this, so I cannot give you a Q3 new-sale total or a take-up rate for the quarter. URA publishes the monthly figure in mid-October and the full quarter on 23 October. If you are looking at a specific launch, see the new launch price list and my page on Lentor Gardens Residences.
Quarter-on-quarter change, private homes against HDB resale, 4Q 2024 to 3Q 2026 · Source: URA quarterly releases, HDB RPI table; 3Q 2026 flash
Swipe to see the full chart →
Private by bedroom: what is verifiable for Q3
You may wonder how the 1.4% splits by bedroom count. It does not, at least not in public. URA data has no bedroom field, SRX and 99.co publish resale volume by region without a bedroom cut in what I could read, and a proper split needs a REALIS pull with an area-band rule for turning floor area into bedrooms. I have not made a floor-area guess and labelled it a bedroom figure. What I can verify is how three launches sold, since those developers and their agencies report bedroom detail.
Private new-sale bedroom data I could verify around Q3 2026, by bedroom count · Checked 5 Oct 2026. These are launch-weekend snapshots for three projects, not a quarter total and not a resale split.
| Bedrooms | What I can show | Source and limit |
|---|---|---|
| 1-bedroom | None found | No public one-bedroom sales figure for Q3 from SRX, EdgeProp, 99.co or URA. |
| 2-bedroom | Lentor Gardens Residences: 40% of launch-day sales (18 Jul). Lucerne Grand, early Oct: 145 of 222 units sold (65%). | Lentor: percentage of sales only, from Kingsford Group via EdgeProp. Lucerne Grand is October, just after the quarter. |
| 3-bedroom | Lentor: 42% of launch-day sales. Dunearn House (25-26 Jul): three-bedroom layout fully sold. Lucerne Grand: 157 of 253 (62%). | Dunearn: Frasers Property’s release names the three-bedroom as fully sold; I could not reconcile unit counts by type across reports, so I show none. |
| 4-bedroom | Lentor: 18% of launch-day sales. Lucerne Grand: 48 of 95 (51%). | Same limits as above. |
| 5-bedroom | None found | None of the projects above has five-bedroom units that I could see; no public figure. |
Swipe to see the full table →
The July launches that drove Q3 were weighted to two- and three-bedroom homes: at Lentor Gardens Residences (270 of 499 units sold on 18 July at an average $2,350 psf) three-bedroom homes were 42% of the launch-day sales and two-bedroom homes 40%, and at Dunearn House (212 of 380 sold on the 25 and 26 July weekend at about $3,140 psf) the three-bedroom layout sold out. The Lucerne Grand launch on 3 and 4 October (350 of 570 sold, about $2,480 psf) falls in Q4 but is the only one with full bedroom counts: 145 of 222 two-bedroom, 157 of 253 three-bedroom and 48 of 95 four-bedroom units, as reported by EdgeProp. The fuller Lucerne Grand tables, with units and take-up by layout, are in my September report.
Lucerne Grand is the clearest case: buyers took the standard two-bedroom and the four-bedroom premium first, and the larger four-bedroom layouts are the ones left. The Lentor and Dunearn tallies are too thin for me to say more. This is a pattern in a few projects, not a market-wide finding, and it says nothing about resale, where I could find no bedroom data at all.
HDB resale
Small dips, rising volume: a calmer market, not a falling one
The index is the headline, but it is built to strip out the mix of flats sold, so a 0.2% dip is a tiny move. What the raw transaction file shows is more activity, not less. I re-downloaded the HDB resale dataset from data.gov.sg on 5 October and counted registrations by month:
HDB resale registrations, Jul–Sep 2026 vs earlier · Source: data.gov.sg “Resale flat prices based on registration date from Jan 2017 onwards”, downloaded 5 Oct 2026; counts and medians are my calculation
| Measure | Q3 2025 | Q2 2026 | Q3 2026 |
|---|---|---|---|
| Resale registrations | 6,939 | 6,184 | 7,415 |
| Of which $1m and above | 480 | 490 | 596 |
| $1m and above, share of deals | 6.9% | 7.9% | 8.0% |
| 3-room median price | $450,000 | $440,000 | $435,000 |
| 4-room median price | $630,000 | $630,000 | $628,000 |
| 5-room median price | $740,000 | $738,000 | $740,944 |
| Executive median price | $900,888 | $920,000 | $920,000 |
Swipe to see the full table →
So: the number of resale deals is up about 20% on Q2 and about 7% on a year ago, and the number of million-dollar flats reached 596, the highest quarterly count in the file, which starts in January 2017. Toa Payoh (93 deals), Queenstown (86) and Bukit Merah (56) led. Meanwhile the typical 3-room flat is selling $15,000 lower than a year ago, and larger flats are flat. The pressure on prices is at the lower end of the market, not at the top.
Inside the quarter: 3-room, 4-room, 5-room, EA and EM
The all-flat numbers hide a clear pattern, so here is Q3 split five ways. In the data.gov.sg file both kinds of Executive flat share the flat type EXECUTIVE; I separated them with the flat model field, counting any model with Maisonette in its name as an Executive Maisonette (EM) and Apartment, Premium Apartment and Adjoined flat as an Executive Apartment (EA).
HDB resale deals, median price, median psf and $1m+ deals by flat type, Q3 against Q2 2026 · Source: data.gov.sg, downloaded 5 Oct 2026, my calculations from raw rows. EA and EM are my split of flat type EXECUTIVE by flat model. Q2 is April to June, Q3 is July to September. 2-room (208 deals in Q3) and multi-generation (4 deals, all at $1m or more) are not shown.
| Flat type | Deals Q2 | Deals Q3 | Change | Median Q2 | Median Q3 | Psf Q2 | Psf Q3 | $1m+ Q2 | $1m+ Q3 |
|---|---|---|---|---|---|---|---|---|---|
| 3 Room | 1,490 | 1,581 | +6.1% | $440,000 | $435,000 | $594 | $582 | 1 | 3 |
| 4 Room | 2,703 | 3,270 | +21.0% | $630,000 | $628,000 | $619 | $609 | 212 | 241 |
| 5 Room | 1,418 | 1,820 | +28.3% | $738,000 | $740,944 | $580 | $578 | 183 | 202 |
| EA | 229 | 323 | +41.0% | $888,888 | $880,000 | $577 | $574 | 38 | 65 |
| EM | 146 | 208 | +42.5% | $966,500 | $965,000 | $609 | $608 | 55 | 81 |
Swipe to see the full table →
Deals rose in every type, but not evenly. The smaller flats grew least (3-room +6.1%, 4-room +21.0%), and the larger and Executive flats grew most (5-room +28.4%, EA +41.0%, EM +42.5%). Medians barely moved: 3-room and 4-room slipped about 1% and 0.3%, 5-room rose 0.4%, and EA and EM were down 1.0% and 0.2%. The year-on-year picture is similar for the Executive types: against Q3 2025, EA deals are up 22.8% (323 against 263) and EM deals up 20.9% (208 against 172), while $1m+ deals went from 47 to 65 for EA and from 55 to 81 for EM.
Highest-priced HDB resale deal of each flat type in Q3 2026, and the share of each type’s deals at $1m or more · Source: data.gov.sg, my calculations
| Flat type | Block and street | Town | Month | sqm | psf | Price | $1m+ share Q3 | $1m+ share Q2 |
|---|---|---|---|---|---|---|---|---|
| 3 Room | 61 Jln Tenteram | Kallang/Whampoa | Sep | 93.8 | 1,097 | $1,108,000 | 0.2% | 0.1% |
| 4 Room | 1E Cantonment Rd | Central Area | Aug | 93 | 1,473 | $1,475,000 | 7.4% | 7.8% |
| 5 Room | 1C Cantonment Rd | Central Area | Sep | 107 | 1,493 | $1,720,000 | 11.1% | 12.9% |
| EA | 150 Mei Ling St | Queenstown | Sep | 148 | 942 | $1,500,000 | 20.1% | 16.6% |
| EM | 192 Bishan St 13 | Bishan | Aug | 162 | 946 | $1,650,000 | 38.9% | 37.7% |
Swipe to see the full table →
HDB resale deals at $1m or more by flat type, Q2 versus Q3 2026 · Source: data.gov.sg, downloaded 5 Oct 2026, my calculations
Swipe to see the full chart →
The share of a type’s own deals at $1m or more shows where the pressure is. EA went from 16.6% to 20.1%, EM from 37.7% to 38.9%, while 4-room slipped from 7.8% to 7.4% and 5-room from 12.9% to 11.1%. So the extra million-dollar 4-room and 5-room deals in Q3 (241 against 212 and 202 against 183) came mainly with the extra volume, not a bigger share. EA is the type where the $1m line is moving into the mix. The Q3 EA top, $1.5 million at 150 Mei Ling Street, and the EM top, $1.65 million at 192 Bishan Street 13, are each the highest price for that type anywhere in the file since January 2017, which I checked against every row.
The Q3 rise in million-dollar deals is real, but it is not one market. About three-quarters of the 596 are 4-room and 5-room flats (443 of them), in estates where a seven-figure price is becoming normal, and they are rising roughly in line with volume. The change in mix I would watch is EA, where a type that had few seven-figure deals a year ago now has one in five. EA and EM are small samples (323 and 208 deals) so I would not lean on a single quarter’s median for either.
Where medians moved
Among the 22 towns with at least 100 deals in both Q2 and Q3, these had the biggest median swings. Be careful with this table: a town’s median jumps when more newly-MOP flats or larger units happen to sell, so it says “what sold” more than “what prices did”.
Median resale price by town, Q2 vs Q3 2026 · Source: data.gov.sg, my calculation; towns with 100+ deals in both quarters
| Town | Q2 2026 | Q3 2026 | Change | Deals Q3 |
|---|---|---|---|---|
| Geylang | $565,000 | $618,000 | +9.4% | 166 |
| Toa Payoh | $780,000 | $848,000 | +8.7% | 289 |
| Serangoon | $644,000 | $700,000 | +8.7% | 126 |
| Clementi | $646,500 | $682,500 | +5.6% | 132 |
| Woodlands | $600,000 | $588,000 | −2.0% | 519 |
| Bukit Batok | $620,000 | $603,888 | −2.6% | 413 |
| Ang Mo Kio | $515,000 | $500,000 | −2.9% | 226 |
| Jurong East | $540,000 | $519,000 | −3.9% | 144 |
Swipe to see the full table →
Why they diverge
What is pulling the two markets apart
The facts I can source
- HDB supply is rising. HDB’s flash release, as reported, puts the softer prices down to more flats reaching their minimum occupation period, a larger resale inventory, and more flats from Build-To-Order and Sale of Balance Flats exercises. HDB plans about 19,600 BTO flats in 2026, and the November BTO alone offers about 7,970. Reports put the number of flats reaching MOP in 2026 at about 13,480, against 6,973 in 2025.
- Private supply is also high, but arrives differently. URA says 4,745 units will launch under the Confirmed List in the second half of 2026 (9,320 for the full year, over 50% above the ten-year average), and about 60,600 private units are expected to complete over the next few years. That is a pipeline of new homes, not an overhang of resale stock.
- Private volume is thin. When only a few hundred new homes sell in a month, the index is set by a small group of deals, and the mix of projects matters a lot.
HDB buyers have a choice they did not have two years ago. More MOP flats, a long BTO queue and a bigger pool of upgraders listing their flats mean a buyer can say no. That is why I see volumes rising while prices drift. It looks to me like an adjustment after a long climb, not a retreat, and a 0.2% dip does not by itself tell me a trend has turned.
The private market is a market of two speeds. Suburban condos and landed homes are doing the lifting. The central districts, where the CCR index is below last year’s level, are not. If you read “private up 1.4%” as a licence to expect the same from a prime-district unit, I think you will be disappointed.
What it means
For buyers, sellers and upgraders
Everything in this section is my opinion and general commentary, not financial advice. The facts are above.
If you are buying
HDB resale: there is room to negotiate and no need to rush into the first flat you view. Check the last few transactions in the same block before you offer, because town-level numbers are blurry.
Condo: compare like with like. OCR is where the index is rising, so budget for it. In the CCR and RCR, resale and launch pricing has more give than the headline suggests.
If you are selling
HDB: price to the latest comparable sales, not to a neighbour’s peak. Volumes are healthy, so a well-priced flat still sells; an optimistic one sits. Larger flats are holding better than 3-rooms.
Condo or landed: thin volume means each buyer matters. Show a buyer data on your own project, not on the index.
If you are upgrading
The gap is growing. As an illustration only, using the index moves: a $1.8m condo costs about $62,000 more than a year ago at +3.4%, while a $900,000 flat is worth about $5,000 less at −0.6%. Your own unit and project will differ. See my HDB-to-condo upgrade guide and check your MOP date first.
Still open
What happens next
Several of the things I would most like to know have not been published yet. I will update this page as they arrive.
1. The final Q3 numbers, 23 October 2026
URA releases its full real estate statistics for Q3 on this date. That will replace the flash with a final private index and, importantly, give the full-quarter new-sale, resale and sub-sale counts. If the final private number comes in well below 1.4%, as the Q1 one went the other way, the story changes. I will also look for landed counts by type there, which the flash does not give. Out on 23 October: what the final figures added, from vacancy to HDB resale volume, is in Q3 2026 final numbers.
2. September developer sales, mid-October
This is what will tell us whether Q3 new-launch take-up was only the July launches, and how quiet September was. Out on 15 October: my read is in September 2026 new home sales.
3. The November BTO and the H2 launch pipeline
About 7,970 BTO flats in November, and 4,745 private units on the H2 Confirmed List. A big take-up on the BTO would reinforce the HDB-resale softness I describe above; strong sales for the next private launches would test whether the suburban strength has legs.
4. The Q4 flash estimates, early January 2027
If HDB’s index is flat or up in Q4, the three-quarter run ends. If it falls again, it becomes four. I would not call a bottom on one quarter either way. Mine is a read, not a forecast.
For the month-level HDB detail behind this, including the ten highest-priced flats, see my September 2026 HDB and private market report.
Questions people are asking me about the Q3 numbers
Yes, on the official flash estimates. URA's flash estimate of 1 October 2026 puts the private residential price index up 1.4% quarter on quarter, from 219.4 to 222.5. HDB's flash resale price index is 202.4, down 0.2% from 202.8. Both are flash estimates built from data up to about mid-September, and URA publishes its full Q3 statistics on 23 October 2026.
Yes. The HDB resale price index fell 0.1% in 1Q 2026 (to 203.4), 0.3% in 2Q (to 202.8) and 0.2% in 3Q (to 202.4, flash). The quarter before that, 4Q 2025, was flat at 203.6, so this is three declines in a row rather than four. Over the first nine months of 2026 the index is down about 0.6%.
Landed homes (+2.8%) and non-landed homes in the Outside Central Region (+2.2%). The Rest of Central Region edged up 0.2%, and the Core Central Region slipped 0.1%. The CCR index is still about 1.3% below where it stood in 3Q 2025.
No. Both are flash estimates. URA's release says flash estimates may differ from actual changes, and it has happened this year: URA's 1Q 2026 flash said +0.3%, and the final number was +0.9%. URA publishes the full 3Q 2026 statistics on 23 October 2026, and I will update this page when they land.
Prices are moving in opposite directions, so the gap you have to bridge is widening: over the year to 3Q 2026 the private index is up about 3.4% while the HDB index is down about 0.6%. Whether that matters depends on your equity, your loan limit, your timeline and which segment you are targeting. The Core Central Region is lower than a year ago and the Outside Central Region is about 5% higher. This is general commentary, not financial advice.
Nobody can promise either way. What I can say is that three small dips in a row sit alongside rising resale volumes, a large Build-To-Order pipeline and a bigger group of flats reaching their minimum occupation period, all of which add choice for buyers. I would watch the 23 October final numbers, the fourth-quarter flash estimate in early January and the November BTO take-up.
In the data.gov.sg file, deals rose in every type: 3-room 1,581 (from 1,490), 4-room 3,270 (from 2,703), 5-room 1,820 (from 1,418), EA (Executive Apartment) 323 (from 229) and EM (Executive Maisonette) 208 (from 146). Median prices were little changed: $435,000, $628,000, $740,944, $880,000 and $965,000. Deals at $1m or more were 3 for 3-room, 241 for 4-room, 202 for 5-room, 65 for EA and 81 for EM, against 1, 212, 183, 38 and 55 in Q2. EA and EM are my split of the EXECUTIVE flat type by flat model. Source: data.gov.sg, downloaded 5 October 2026, my calculations.
Not a full one. URA data has no bedroom field and a split needs the paid REALIS database. For Q3 the verifiable pieces are two July launches: Lentor Gardens Residences sold 270 of 499 units on 18 July, with three-bedroom homes 42% of the sales, two-bedroom 40% and four-bedroom 18% (Kingsford Group, via EdgeProp), and Dunearn House sold 212 of 380 units with its three-bedroom layout fully sold. Lucerne Grand's October launch gave the only full counts: 145 of 222 two-bedroom, 157 of 253 three-bedroom and 48 of 95 four-bedroom units. I found no one-bedroom or five-bedroom figure.
URA's flash estimate does not split landed volume, so I cannot give a verified count. A third-party tracker of URA caveats shows 436 landed deals so far in Q3 against 524 in Q2, and Knight Frank describes volume as steady at 500 to 600 a quarter. The largest landed deals I can source from EdgeProp's columns are $24 million at Oriole Crescent, $21.5 million at Trevose Crescent, $21 million for a Pasir Panjang bungalow and $20.38 million at Gentle Road. That is a partial list, not a ranking. Landed prices rose 2.8% on URA's flash, after 2.5% in Q2.
Not sure what the numbers mean for your plan? Ask me.
Tell me whether you are buying, selling or upgrading and roughly when. I will tell you what I would do in your position, using real recent transactions rather than index averages.
Sources and dates
The numbers below are the agencies’, restated in my own words and combined with my own calculations. Opinion is labelled “my read”. Figures are as at 5 October 2026 and the Q3 numbers are flash estimates.
- URA, Release of flash estimate for 3rd Quarter 2026 private residential property price index, 1 October 2026, with Annex A (index levels and changes by segment) — the +1.4% overall, landed +2.8%, non-landed +0.9%, CCR −0.1%, RCR +0.2%, OCR +2.2%, volume of 4,296 against 6,148, the year-to-date 15,857 against 19,793, the Confirmed List and the 23 October release date.
- URA, Release of 2nd Quarter 2026 real estate statistics, 24 July 2026 — final Q2 segment changes, Q2 developer sales of 2,141 units.
- URA, 1st Quarter 2026 real estate statistics (24 April 2026) and its flash estimate (1 April 2026), and the Q2 flash estimate (1 July 2026) — the flash-to-final revisions quoted above. Earlier index levels from URA’s 1Q 2025 and 3Q 2025 releases.
- HDB, Resale Price Index table, 1Q 1990 to 3Q 2026 — the index levels and changes through 3Q 2026 (flash). HDB’s flash estimate page, Flash Estimate of 3rd Quarter 2026 Resale Price Index and Upcoming Flat Supply, confirms the 202.4 figure and the decline; I could not load its full text, so the explanation of the dip, the 7,528 transactions as of 29 September and the 5.2% year-on-year comparison come from press reports of it (for example Business Today, 1 October 2026, and Little Big Red Dot).
- data.gov.sg, Resale flat prices based on registration date from Jan 2017 onwards (HDB), downloaded 5 October 2026 — all registration counts, million-dollar counts, flat-type and town medians. My calculations. For EA and EM I split flat type EXECUTIVE by the flat model field: models containing Maisonette are EM, and Apartment, Premium Apartment and Adjoined flat are EA.
- EdgeProp Singapore, developers’ July 2026 sales report (URA data), and press reports of the August 2026 figure (EdgeProp and Stacked Homes) — 731 units sold in July, project splits and median psf; 153 units sold in August and 116 released. I read the July report in full and the August figure from a search-result excerpt only. September data is not yet published.
- EdgeProp Singapore, “HDB resale market: Entering calmer waters?” — 6,973 flats reaching MOP in 2025. The 2026 MOP figure of about 13,480 and the 19,600 BTO flats in 2026 are from press reports (PropertyGuru, EdgeProp) and are not checked against HDB.
- Bedroom and landed sources: EdgeProp on Lentor Gardens Residences (270 of 499, $2,350 psf, bedroom shares from Kingsford Group) and Lucerne Grand (bedroom counts); Frasers Property, Dunearn House sales release; CDL’s Lucerne Grand launch release. Limits: launch snapshots, shares not counts for Lentor, and I found no one-bedroom, five-bedroom or resale bedroom data.
- EdgeProp “What’s moving the market” columns read for landed deals: 24 Jul, 31 Jul, 7 Aug, 21 Aug, 28 Aug, 4 Sep and 2 Oct, plus the 11 and 25 September columns through Yahoo Singapore reposts. I read these through an automated page reader. Limits: a few notable deals per column, dates mostly not given, columns I did not read (for example the early-July ones) may hold larger deals, and no Good Class Bungalow deal appeared in the ones I read.
- Landed volume sources: ERA Singapore, 2Q 2026 URA quarterly report (447 and 502 caveats); EdgeProp, Huttons’ 1Q2026 landed report; PropKaki landed tracker (436 versus 524 and median land psf by type, a third-party computation I could not reproduce); Yahoo Singapore, landed property as the main driver of the 3Q2026 flash, 1 October 2026 (Knight Frank’s 500 to 600 a quarter); Real Estate Asia, Realion’s Q2 Good Class Bungalow count. URA’s own release (item 1) has no landed volume or type split.
- My own analysis, not reported fact — the EA and EM split and the flat-type readings, the landed deal compilation, the rebased index chart, year-on-year changes and gap, the upgrader illustration, every “my read” box and all of the buyer, seller and upgrader suggestions.
This article is general commentary on public data. It is not financial advice and not a forecast of any property’s price. Flash estimates may differ from the final figures. VionaPropnex is not affiliated with, endorsed by, or speaking for URA or HDB; for the authoritative figures always use their own releases.