The public registers behind every HDB resale figure here

This site publishes four separate registers, not one dataset, and the difference matters more than it sounds: they record different events, at different moments, under different licences. The one most people arrive for is the Housing & Development Board’s Resale Flat Prices datasets, published on Data.gov.sg under the Singapore Open Data Licence. As of August 2026 that is 985,666 HDB resales going back to January 1990.

What does each row of the register carry?

Month of registration, district, flat type, block, street name, storey range, floor area in square metres, flat model, lease commencement year and resale price. That is unusually complete. Most countries publish a price and an address and leave you to find the floor area somewhere else; here the two arrive together, which is why this site can compute a price per square foot directly rather than matching two registers and losing the resales that fail to match. The flat type field is also the axis of the flat-type pages here: the same register, cut by product instead of by place.

What does it not carry?

No unit number, which is what keeps a resale from being tied to a household. No condition, no renovation, no orientation, no view. No private property: condominiums, executive condominiums and landed housing are recorded separately by URA and are not in this register at all. No BTO prices, because a first-hand sale from HDB is not a resale.

What are the other three registers?

HDB also publishes its renting-out approvals on Data.gov.sg: the flats approved for subletting whole, which is where the rent figures and the gross yields on this site come from. The private market arrives separately, from the URA Data Service, in two registers of its own: sale caveats lodged when an option is exercised, and private lease contracts. Both URA registers are rolling five-year windows rather than a full history.

Keeping them apart is not pedantry. They record different events at different moments, so each one has its own anchor month, and a figure from one may not be averaged with a figure from another: there is no single “Singapore price per square foot” spanning HDB and private, and this site never prints one. The licences differ too. The HDB registers may be reused commercially with attribution; URA’s terms permit derived statistics only, which is why you will find aggregates here and never a caveat-level record.

How do the four registers compare?

Unevenly, and the unevenness is information. The resale register is the deep one: a full history back to 1990, which is what makes the long-run charts on this site possible. The other three are shallower: the two URA registers arrive on a rolling five-year window, so their earliest months fall away as new ones are lodged, and the renting-out approvals series reaches back only as far as its own publication does. The table reads each span live from the file.

The four counts cannot be summed or averaged, because each one counts a different event: a caveat is a step in a purchase rather than a completion, and an approval to rent out a flat is a licence rather than a tenancy. Recorded under different rules over different windows, the counts are statements of what this build holds, not a league table of the four markets’ sizes, and reading them as market shares is the same category error as averaging their prices.

Which month should you ignore, and why?

Resales appear as they are registered, not as they are agreed, so the most recent month in the file is always partial. This site excludes it from every figure and anchors every window on the last complete month. A site that includes the partial month will show a jumpy final data point and a median that moves for no market reason. The URA registers have the same problem from the other direction: lodging a caveat is not compulsory, so recent months fill in afterwards.

How do you use it yourself?

Every area page here links to its own JSON, and the dataset page sets out which licence covers which register. If you want the raw HDB records rather than these aggregates, go to the source: the Resale Flat Prices collection on Data.gov.sg has a bulk CSV download and a REST API, no key required. The URA records are not ours to pass on. Every count and span above is read live from the build’s own index and is current to August 2026.