Why do Singapore property statistics disagree with each other?

Usually because they are medians of two groups that differ in more than the thing being measured. Four examples on this page: freehold resales run 19 per cent dearer pooled across Singapore, and 11 per cent dearer in the core central region, 8 per cent cheaper in the rest of central region, 10 per cent dearer in the outside central region, because freehold stock sits in the dearer districts. The storey premium reads 2.1 times across storey bands and 0.63 per cent a storey inside a single block, because only newer blocks in dearer towns have a high floor to sell. Lease decay has no single national curve at all, because the stock with least lease left is older stock in better locations. And private flat price growth reads 29.9 per cent pooled, 23.7 per cent on resales only and 17.5 per cent on the official mix-adjusted index, because new launches went from a third to over two-fifths of sales.

Is a high floor really worth twice as much in Singapore?

No. The 2.1 times figure comes from comparing $555 psf on storeys 01 to 03, across 1,767 sales, with $1,180 psf on storeys 37 to 39, across 34 sales — and only newer blocks in dearer towns have a 37th storey to sell. Measured inside the same block, across 758 blocks and 4,353 sales, a storey is worth 0.63 per cent of the price per square foot, which compounds to about 1.25 times over 36 storeys.

How fast have Singapore private home prices actually risen?

Between 2022 and 2025, this site’s median price per square foot for private flats rose 23.7 per cent on resales, across 69,678 resales of 118,834 sales. Pooling new launches in takes the same figure to 29.9 per cent, which is not a price rise but a change in what was being sold. URA’s own mix-adjusted non-landed index rose 17.5 per cent over the same two calendar years, and that is the figure to quote for the private market.

Four Singapore property statistics that measure something else

Four of the most-quoted figures in Singapore housing are medians of one group set against medians of another, and in each case the two groups differ in something that sets the price by itself: where the home is, how new it is, which building it is in, or who sold it. Stratify each one and the freehold premium goes from 19 per cent dearer pooled to 11 per cent dearer in the core central region, 8 per cent cheaper in the rest of central region, 10 per cent dearer in the outside central region; the storey premium from 2.1 times to 0.63 per cent a storey; private flat growth from 29.9 per cent to 23.7 per cent on resales and 17.5 per cent on the official mix-adjusted index; and lease decay stops being a single national curve at all.

The test that catches all four is one sentence: before comparing two groups, ask what else differs between them. If the answer is the thing that actually sets the price, the comparison is measuring that instead. This page keeps the score, including the one this site got wrong in public.

The four, side by side

Statistic	Commonly published	Measured properly	What else differed
The freehold premium	+19%Pooled across Singapore, resales	+11% / −8% / +10%core / rest / outside	Freehold stock sits in the dearer districts, so a national ratio is partly measuring location; and almost everything built this century is leasehold, so across all sales the comparison is measuring how new a flat is.
Lease decay	One national curveA single discount per band of lease	−44% at 50 to 59 yearsrest of central region, against 80 to 89 years	The stock with least lease left is disproportionately older blocks in prime locations, so a single national curve measures where a flat is rather than how long it has.
The storey premium	2.1×Storeys 01 to 03 against 37 to 39	0.63% a storey1.25× over 36 storeys, within a block	Only newer blocks in dearer towns have a high floor to sell, so comparing storey bands across the whole island mostly measures which buildings are tall.
Private price growth	+29.9%Median per sq ft, all flat sales, 2022 to 2025	+17.5%URA’s mix-adjusted index. Our resales-only median reads +23.7%	New launches rose from a third to over two-fifths of private sales and are bought at a developer’s list price, so a pooled median rises without any home being worth more.

Three of the four move in the direction that flatters the market. Two were caught here before anything was published, one shipped and was corrected, and the fourth was found while writing up the other three.

1. The freehold premium

Is freehold worth more than leasehold in Singapore? Dearer in two of the three regions, and by about a tenth rather than a fifth. Comparing resales inside the same region, freehold — counting 999-year leases with it — runs 11 per cent dearer in the core central region, 8 per cent cheaper in the rest of central region, 10 per cent dearer in the outside central region. Pool the three and the premium reads 19 per cent dearer, which is higher than any one of them.

That is not a stronger result, it is a different error. 33 per cent of freehold resales are in the core central region against 8 per cent of leasehold resales, so a national ratio sets a group weighted towards the dearest part of Singapore against one weighted towards the cheapest. Most of the difference is a map. The full measurement, by region and district.

2. Lease decay

How much does a shorter lease reduce a Singapore flat’s price? In the rest of central region, private flats with 50 to 59 years of lease left sold for 44 per cent less a square foot than those with 80 to 89 years. In another region the same comparison is worth a fraction of that. There is no single national curve on this site, and that is a finding rather than an omission.

The clearest evidence sits inside one region. In the outside central region the median per square foot turns back up below 70 to 79 years: $1,126 psf on 1,643 sales at 60 to 69 years, against $1,067 psf on 5,903 sales in the band above it. Nobody pays more for less lease. What those sales record is which addresses still change hands when the lease is short — which is why every curve here stops where its evidence does. The curves, and a calculator.

3. The storey premium

How much more does a high floor cost in Singapore? The pooled bands say 2.1 times: $555 psf on storeys 01 to 03 against $1,180 psf on storeys 37 to 39. Measured inside the same block, where the building is held still, a storey is worth 0.63 per cent of the price per square foot — about 1.25 times over 36 storeys.

This one shipped. It was published here as a height premium, in prose that invited the reader to picture two flats in the same block, and the figure beside it compared 1,767 sales drawn from nearly every block in Singapore with 34 sales drawn from the handful tall enough to reach the top band. A competitor’s within-block regression is what prompted the recheck, and it independently landed on 0.6 per cent against the 0.63 computed here. The reference page, and the correction as a story.

4. Private price growth

How fast have Singapore private home prices risen? New launches were a rising share of what sold — they go from about a third to over two-fifths of private flat sales across this window — and they are bought at a developer’s list price. So a pooled median rises without any home anywhere being worth more.

Private price growth, 2022 to 2025, three ways

Our median price per square foot, with and without new launches, against the official mix-adjusted index

Scope	All sales pooled	Resales only	Official index
All private homes81,306 resales of 131,207 sales	+29.8%	+24.2%	+16.9%All residential
Flats and apartments69,678 resales of 118,834 sales	+29.9%	+23.7%	+17.5%Non-landed

Matched on scope: all private homes against URA’s all-residential index, flats and apartments against its non-landed index. The index is compared on calendar-year averages, because our own figure is one whole year’s median against another — a part-year at either end would manufacture a difference out of the calendar. Source: Private Residential Property Price Index (URA), read Oct 2026.

The three columns are not three opinions. The first measures a change in what was sold, the second a change in the price of a resale, and the third a change in the price of a standardised home. Only the third is a price index, and it is the one to quote — which is why the pooled figure is published on this site as a transparency note and never as growth.

What it looks like when a comparison goes wrong

The mechanism is always the same and it is visible in the sale counts. Take the storey bands:

Sales on storeys 01 to 03

1,767

Drawn from almost every HDB block in Singapore. Median $555 psf.

Sales on storeys 37 to 39

Drawn from the few blocks tall enough to have them. Median $1,180 psf.

Blocks where the question can be asked properly

758

4,353 sales, the same address at different heights.

A group of 34 sales and a group of 1,767 sales are not two samples of one population. The second is Singapore; the first is a short list of tall new blocks in expensive towns. Any difference between their medians is mostly the difference between those two kinds of place.

The same shape appears in all four. Freehold against leasehold is a comparison between districts. A long lease against a short one is a comparison between decades of construction. Pooled growth is a comparison between a market where new launches were a third of sales and one where they were over two-fifths. In each case the honest version exists, is computable from the same data, and is smaller.

How these are measured, and how to check them

Every figure on this page is read at render time from the same published files the four detail pages read, so none of it can drift: the private figures from 131,264 Urban Redevelopment Authority caveats covering Oct 2021 to Oct 2026, the HDB figures from 243,611 resale transactions published on data.gov.sg to Oct 2026. Both are refreshed nightly.

The rules this site holds itself to on all four, each of which exists because something went wrong:

Medians, never means, with the sale count beside every one. A median over two sales is reported as what it is.
Never average a set of ratios into a headline. The freehold premium briefly read four per cent here because three regional multiples were averaged. Either compute the pooled figure properly or quote the parts.
A chart must not draw past what its caption claims. A lease-decay chart once showed four bands while its note said two. The note was right.
A figure that appears twice is computed once. The storey premium was asserted in two places with two different values; it is published as one file now and every page reads it.
A median must not round. The helper behind these figures returned a rounded average on an even count, which reported the 0.63 per cent storey gradient as one.
A gap beats a guess. Where a comparison cannot be made honestly — an HDB decay percentage per year, a within-block gradient for condominiums — nothing is published.

The underlying transactions are all public. The method page names every source, every derivation and what each one cannot tell you.

Questions people ask

Why do Singapore property statistics disagree with each other?: Usually because they are medians of two groups that differ in more than the thing being measured. Four examples on this page: freehold resales run 19 per cent dearer pooled across Singapore, and 11 per cent dearer in the core central region, 8 per cent cheaper in the rest of central region, 10 per cent dearer in the outside central region, because freehold stock sits in the dearer districts. The storey premium reads 2.1 times across storey bands and 0.63 per cent a storey inside a single block, because only newer blocks in dearer towns have a high floor to sell. Lease decay has no single national curve at all, because the stock with least lease left is older stock in better locations. And private flat price growth reads 29.9 per cent pooled, 23.7 per cent on resales only and 17.5 per cent on the official mix-adjusted index, because new launches went from a third to over two-fifths of sales.
What is Simpson’s paradox in property prices?: It is when a comparison holds inside every group and reverses or inflates when the groups are pooled. The freehold premium is the clean example: freehold resales run 11 per cent dearer in the core central region, 8 per cent cheaper in the rest of central region, 10 per cent dearer in the outside central region, yet 19 per cent dearer across all three together — higher than any single region. The reason is the mix: 33 per cent of freehold resales are in the core central region against 8 per cent of leasehold resales, so a national ratio is partly measuring a map.
Should the Singapore freehold premium be quoted as a single national figure?: No. Pooled across the island it reads 19 per cent dearer on this site’s data, which is higher than any one region and should not be quoted as the premium. Measured inside each region on resales only, freehold is 11 per cent dearer in the core central region, 8 per cent cheaper in the rest of central region, 10 per cent dearer in the outside central region — it changes sign between regions. Freehold here includes 999-year leases, and both sides of the comparison exclude new launches.
Is a high floor really worth twice as much in Singapore?: No. The 2.1 times figure comes from comparing $555 psf on storeys 01 to 03, across 1,767 sales, with $1,180 psf on storeys 37 to 39, across 34 sales — and only newer blocks in dearer towns have a 37th storey to sell. Measured inside the same block, across 758 blocks and 4,353 sales, a storey is worth 0.63 per cent of the price per square foot, which compounds to about 1.25 times over 36 storeys.
How fast have Singapore private home prices actually risen?: Between 2022 and 2025, this site’s median price per square foot for private flats rose 23.7 per cent on resales, across 69,678 resales of 118,834 sales. Pooling new launches in takes the same figure to 29.9 per cent, which is not a price rise but a change in what was being sold. URA’s own mix-adjusted non-landed index rose 17.5 per cent over the same two calendar years, and that is the figure to quote for the private market.

The four corrections, and where each is published

The confound in one sentence apiece

Show the figures

Statistic	What else differs between the groups
The freehold premium	Freehold stock sits in the dearer districts, so a national ratio is partly measuring location; and almost everything built this century is leasehold, so across all sales the comparison is measuring how new a flat is.
Lease decay	The stock with least lease left is disproportionately older blocks in prime locations, so a single national curve measures where a flat is rather than how long it has.
The storey premium	Only newer blocks in dearer towns have a high floor to sell, so comparing storey bands across the whole island mostly measures which buildings are tall.
Private price growth	New launches rose from a third to over two-fifths of private sales and are bought at a developer’s list price, so a pooled median rises without any home being worth more.

Is freehold worth more than leasehold in Singapore?Freehold stock sits in the dearer districts, so a national ratio is partly measuring location; and almost everything built this century is leasehold, so across all sales the comparison is measuring how new a flat is.
How much does a shorter lease reduce a Singapore flat’s price?The stock with least lease left is disproportionately older blocks in prime locations, so a single national curve measures where a flat is rather than how long it has.
How much more does a high floor cost in Singapore?Only newer blocks in dearer towns have a high floor to sell, so comparing storey bands across the whole island mostly measures which buildings are tall.
How fast have Singapore private home prices risen?New launches rose from a third to over two-fifths of private sales and are bought at a developer’s list price, so a pooled median rises without any home being worth more.

Citing this. Figures computed on Oct 2026 from 131,264 URA caveats and 243,611 HDB resale transactions, and recomputed nightly as the archive grows, so quote the date you read them. Permanent link: https://www.dailystudysg.com/corrections.