A high floor is worth a fifth, not a doubling — and this site said otherwise

Pool every 4 room flat sold in Singapore in 2025 and sort the sales by floor, and the pattern looks overwhelming: $555 psf on storeys 01 to 03, $1,180 psf on storeys 37 to 39. 2.1 times.

Every one of those numbers is correct. The ratio between them is not what a higher floor is worth, and this article said it was.

Only certain blocks have a thirty-seventh storey

That is the whole of the problem. The bottom band is drawn from nearly every HDB block in Singapore. The top band is drawn from the handful tall enough to reach it — and those are newer, they have longer leases, and they stand in Queenstown and Bukit Merah and Kallang rather than in Woodlands or Sembawang. Comparing the two is mostly measuring which blocks are tall, not what height is worth. The storey number is standing in for the building.

The way to remove it is to stop comparing buildings. Inside one block the address is its own control: same age, same lease, same town, same model, same walk to the MRT. The only thing that differs between two sales there is the floor and the unit.

Measured inside a block, a storey is worth 0.63 per cent

Across 758 blocks with at least four sales spanning at least six storeys — 4,353 transactions, a median span of 12 storeys — the median gradient is 0.63 per cent of the price per square foot, per storey. Over 36 storeys that compounds to about 1.25 times, not 2.1.

An independent check. TribeSG ran a within-block storey regression on the same public dataset and published 0.6 per cent per storey. Two unrelated routes to one number is the best validation available, and it is the same device that gives the rest of this site’s price work its confidence — our transaction-derived +52 per cent since 2017 against the official index’s +51.5.

What the bands do tell you

They are still worth publishing, as long as nobody subtracts them. “A 4 room flat on the 37 to 39 storeys sold for a median of $1,180 psf” is a true and useful fact about what exists at that height in Singapore. It is an answer to what do high flats cost. It is not an answer to what will a higher floor cost me in this block, which is the question a buyer is actually asking.

Why this matters beyond one statistic

Singapore property figures are quoted constantly and stratified almost never. The same mistake turns up in the freehold premium, which reads nineteen per cent nationally and about ten within a region, because freehold stock sits in dearer districts. It turns up in lease decay, where pooling the whole island makes flats with sixty years left look dearer than flats with seventy, because the short-lease stock is disproportionately 1980s blocks in prime locations. Both of those were caught before publication. This one was not.

The test that catches all three is the same: 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.

What a buyer should take from it

A higher floor is worth paying something for and it is worth less than the market-wide tables suggest — on the order of 6 per cent for ten floors, not a doubling. And when a block median is quoted, here or anywhere else, it is an average over floors: a low-floor unit in an expensive block and a high-floor unit in a cheap one can cost the same, and the median describes neither.

Every block page therefore breaks its own sales down by storey band, and the comparison tool shows the individual transactions, so the floor each sale was on stays visible rather than being averaged away.