The freehold premium looks like it collapsed. Hold building age constant and it is bigger than ever.

A freehold two-bedroom cost 34.4% more per square foot than a leasehold one in 2017. In 2025 it costs 7.9% more. The premium looks like it has almost gone.

It has not. Hold building age constant and freehold is worth more than it was, and the older the building, the more it is worth. What changed is not the premium. It is which freehold homes reach the resale market.

Key takeaways

• Pooled, the freehold premium fell from 34.4% to 7.9%. The number that looks like a collapse
• Within a single region it was never above 7.8%, so most of that 34.4% was never about tenure. Was it geography
• Hold building age constant and freehold is worth 8.6% to 34.3% more, and the premium grows with age. Was it age
• Freehold resales have aged from a median of 8 years old to 14. Leasehold resales have stayed young. What actually changed

The number that looks like a collapse

Median resale price per square foot for a two-bedroom:

• Freehold
• Leasehold

Over the nine years a leasehold square foot rose 65.4% and a freehold one 32.8%. The gap between them falls from 34.4% to 7.9%.

Read on its own, that is a straightforward story: buyers have stopped paying for freehold. It is also wrong, and the rest of this note is the two tests that show why.

Was it geography

Freehold does not sit evenly across the island. It is concentrated in prime and on the city fringe, where everything costs more. So a national comparison of freehold against leasehold is partly a comparison of expensive places against cheaper ones.

Comparing within each region:

Region	2017 freehold	2017 leasehold	2017 premium	2025 freehold	2025 leasehold	2025 premium
Core Central	$1,848	$1,818	+1.7%	$2,272	$2,156	+5.4%
Rest of Central	$1,289	$1,224	+5.3%	$1,739	$1,981	-12.2%
Outside Central	$1,034	$959	+7.8%	$1,594	$1,532	+4.0%

Six premium figures, and the largest is 7.8%. Most of the national 34.4% was never about tenure at all. It was about where freehold happens to be.

And on the city fringe the sign has flipped. A leasehold two-bedroom there now costs 12.2% more per square foot than a freehold one.

That is a real finding and it is not the explanation. Region and building age are tangled together: city fringe leasehold is five years newer than city fringe freehold, and suburban leasehold is nine years newer. So this test rules out geography as the whole story without telling us what is left.

Was it age

Comparing homes of the same age, across the whole market in 2025:

Building age	Freehold	Leasehold	Freehold premium
0 to 5 years	$2,386	$2,039	+17.0%
6 to 10	$1,868	$1,720	+8.6%
11 to 15	$1,881	$1,565	+20.2%
16 to 25	$1,910	$1,555	+22.8%
26 years and over	$1,726	$1,285	+34.3%

The premium is not gone. In every band freehold costs more, and in the oldest band it costs a third more.

It also does what a lease should make it do. A 99-year lease with 90 years left is very nearly as good as owning outright, and priced accordingly. With 70 years left it is not, and the gap widens. The one break in the pattern is the newest band, where freehold carries 17.0%: new freehold two-bedrooms are few, 97 sales in 2025 against 805 leasehold, and they are a different kind of project.

What actually changed

If the premium is intact at every age, and the pooled number fell anyway, then what moved is the mix of ages being compared.

It is. Median age of the homes sold:

	2017	2021	2025
Freehold resales	8 years	11 years	14 years
Leasehold resales	7 years	7 years	9 years

Freehold resales have aged six years in eight. The share of them under ten years old fell from 65.4% to 28.0%.

Leasehold resales have barely aged at all. The share under ten years old rose, from 57.1% to 61.8%, because 99-year projects keep completing and reaching the resale market a few years later.

So the pooled comparison has been quietly drifting. Each year it sets older freehold against leasehold that stays new. Older homes cost less per square foot, newer ones more, and the two effects push the measured premium toward zero without anything happening to what freehold is worth.

The premium did not collapse. The comparison did. This is the same mechanism we found in the regional cut, where a suburban two-bedroom lost a sixth of its floor area while prime kept nearly all of its: in both cases the homes reaching resale changed, and the change showed up as if it were a change in price.

What this means

Tenure is standing in for remaining lease. That is why the premium grows with building age: what a buyer is paying for is not the word on the title but how many years are left, and a new 99-year lease has almost all of them. Treating freehold and leasehold as two fixed categories with a fixed gap between them will misprice both.

A national tenure comparison is close to meaningless. Ours moved 26 percentage points in eight years without the underlying premium moving at all. Any figure quoted for the freehold premium needs to say at what age and in what region, or it is measuring the mix.

And the city fringe inversion is worth watching rather than explaining. Leasehold there costs 12.2% more per square foot than freehold. That is a real number about a real market, and we cannot separate how much of it is age, how much is which projects happen to be selling, and how much is something else.

What this does not tell you

• Region and age together. We hold each constant separately. Holding both at once splits 2025 into thirty groups, several of which are too small to read, so we have not done it. The two tests point the same way and neither is a substitute for the other.
• Why new freehold breaks the pattern. The 0 to 5 band carries 17.0% where the band above it carries 8.6%. With 97 freehold sales in that band, that may be the projects rather than the tenure.
• What any particular lease is worth. These are age bands, not remaining-lease years. Two 99-year homes completed the same year can have different lease commencement dates and different amounts left.
• New launches. This is the resale market throughout.

Method

Figures are transacted prices from URA caveat data for private condominiums and apartments, resale only, excluding collective sales and sub-sales. Bedroom count is worked out from the unit record rather than stated in the caveat, and only transactions where it resolves are counted.

Tenure is the project's tenure type, freehold or leasehold. Every two-bedroom resale in the period carries one or the other; none is unclassified.

Building age is the transaction year minus the project's completion year, so a home sold in 2025 in a project completed in 2015 is ten years old. Transactions in projects with no recorded completion year are excluded from the age tables and included everywhere else.

Records with a floor area below 200 sqft or above 6,000 sqft are treated as corrupt and excluded.

Medians are used rather than averages, so a small number of unusual transactions does not move the figures.

Sale data to 4 August 2026.