How we measure prices and sales

The Realila index compares the price of each home with the price the same home sold for earlier. What is a same home price index? explains how that kind of index works, what it does well, where it falls short, and when to use it.

Which sales the index uses

The index uses sales of private condominiums and apartments across Singapore, as recorded by URA from caveats. A caveat is a legal notice that a buyer's lawyer lodges after a purchase, and it records the price. The sales counts in the Market Snapshot are counts of sales with a lodged caveat.

Each comparison in the index is made of two sales of the same home: one sale, and the very next sale of that home after it. A home that sold three times, for example in 2012, 2018 and 2025, gives the index two comparisons: the 2012 sale with the 2018 sale, and the 2018 sale with the 2025 sale.

The earlier sale in a comparison can be a purchase from the developer, a sub sale or a resale. The later sale is whichever kind of sale came next for that home.

Which sales the index leaves out, and why

• Executive condominiums. Executive condominiums are sold under public housing rules for their first years. Leaving them out keeps the index to the same kinds of homes as URA's index for non-landed private homes.
• Landed homes. The index covers condominiums and apartments only, the same scope as URA's index for non-landed private homes.
• Collective (en bloc) sales. In a collective sale, the price reflects what the land is worth for redevelopment, so it says little about what the home is worth to live in.
• A home's two sales less than 12 months apart. Over such a short time, the price change usually says more about that particular deal than about the market.
• A home whose recorded floor area differs by more than 5% between its two sales. A difference that large usually means the home was altered, or one of the records is wrong.
• A home whose price changed by an extreme amount between its two sales. Very large rises or falls are more likely to come from recording errors or unusual deals than from ordinary sales. The limits are listed under Technical basis below.

The sales that the index leaves out stay in Realila's records. They are only left out of the index calculation.

The two kinds of figure in the Market Snapshot

Figures for the latest 90 days

The headline figures in the Market Snapshot cover the latest 90 days. They compare prices in the latest 90 days with prices in the same 90 days a year earlier, and with prices in the 90 days before the latest 90 days. These figures update as new sales are recorded. The 90-day figures use the same method as the quarterly index, applied to sales in rolling 90-day periods.

Figures for completed quarters

The quarterly figures use calendar quarters. Each quarter is compared with the same quarter a year earlier. A quarter's figure first appears 30 days after the quarter ends. The figure is marked provisional until 180 days after the quarter ends, and is final from then on.

Why the figures can change

The newest figures can change because some sales are recorded weeks after they happen. When those sales are added, the figures for the latest 90 days and the newest quarter are worked out again, and they may move. This is why the newest quarter is marked provisional.

Early readings for all homes have been revised up slightly, by under 0.2 percentage points, as late sales were recorded, in checks of seven readings made between April and August 2026. Figures for the smaller groups have been revised by more.

How the index is broken down

The Market Snapshot also shows the index for smaller groups of homes.

By region: the Core Central Region (CCR), the Rest of Central Region (RCR) and Outside Central Region (OCR).

30 October 2026. A home's region follows URA's rule. A home is in the CCR if it is in postal district 9, 10 or 11, in the Downtown Core planning area, or on Sentosa. It is in the RCR if it is elsewhere in URA's Central Region, and in the OCR otherwise. Before this date we placed homes by postal district alone, but URA splits several districts between regions. The correction moved 105,490 of 718,150 recorded sales, mostly in Bedok, Bukit Timah, Clementi and Bishan. Feed notes that used the earlier regions say so in their Method section, and one is restated as a new note.

By how the owner first bought the home: homes that were first bought new from the developer, and homes that were first bought on the resale market and sold again on it. Homes first bought as a sub sale count only in the figure for all homes.

Figures for a smaller group rest on fewer sales, so they rise and fall more from one period to the next than the figure for all homes.

What the repeat sales count means

The repeat sales count shown beside the 90-day figures is the number of sales in that 90-day period that were matched with an earlier sale of the same home, made before the period began, and that passed every check listed above. The repeat sales count is smaller than the total number of sales in the period. Final quarters show no count.

The URA figures shown alongside

The quarterly figures in the Market Snapshot also show URA's Private Residential Property Price Index for non-landed homes, for the same quarters, as published by URA. What is a same home price index? explains why the two indexes can show different figures.

What this tells you

This note tells you which sales the Realila Same Home Price Index is built from, which sales it leaves out, when its figures appear and become final, and the settings behind each figure, so that every figure in the Market Snapshot can be traced to its method.

What this does not tell you

• How to read or use the index. This note does not explain how to read or use a same home price index; What is a same home price index? covers that.
• Why prices moved. This note does not explain why prices moved in any particular quarter.
• What a home is worth. This note cannot tell you what a particular home is worth.

Technical basis

The Realila Same Home Price Index uses the repeat-sales method published by Bailey, Muth and Nourse in 1963. For each pair of sales of the same home, the natural log of the later price per square foot divided by the earlier price per square foot is regressed on a set of quarterly time indicators, using ordinary least squares. The estimated quarterly coefficients give the index level for each quarter.

Setting	Value
Homes covered	Private condominiums and apartments; executive condominiums excluded
Base	Q1 2009 = 100
Minimum time between the two sales	12 months
Floor area at the two sales	Must agree within 5%
Later price per sq ft divided by earlier price per sq ft	Between 0.5 and 4.0
Yearly rate of change between the two sales	Between −20% and +40%
Market segment (CCR, RCR or OCR) at the two sales	Must be the same
Collective sales	Excluded
Minimum number of pairs for a group	A figure is shown only when at least 75 pairs have one of their two sales in the period the figure covers. Otherwise the page shows that there are too few repeat sales.
Quarter figure first shown	30 days after the quarter ends
Quarter figure final	180 days after the quarter ends
Changes shown	To one decimal place
Source of sales	Private residential sales recorded by URA from lodged caveats (REALIS)

Version history. Version 2, October 2026: executive condominiums were taken out of the index, so that it covers the same kinds of homes as URA's index for non-landed private homes. Version 1 included executive condominiums.