How much does the private market price lease, MRT and schools?

The main drivers of private non-landed property prices in Singapore include tenure (freehold versus leasehold), distance to MRT, remaining lease, building age, primary school catchment, and access to bus stops, hawker centres and shopping malls — but the value of each varies by location and project type. PropKaki's Private Property Pricing Model is a free tool, with no account needed, that quantifies each of these per-feature premiums from URA-registered transaction data covering 1995 to today: pick a slice (planning area, property type and time window) and it computes live how much the market is paying for each feature.

PropKaki's Private Property Pricing Model covers non-landed private residential property in Singapore: Condominium, Apartment and Executive Condominium (EC) transactions. It excludes landed property — Terrace, Semi-Detached and Detached houses — because the landed market has a different pricing structure (overwhelmingly freehold, no walk-to-MRT premium pattern, much smaller volume per slice). It also excludes new-launch primary sales from developers, focusing on subsale and resale transactions where market pricing is most observable. A separate landed tool may be added in future.

For the slice of the Singapore market you define, PropKaki's Pricing Model runs a linear regression of price per square metre against each feature individually. The slope of that regression is the per-unit premium — for example, the dollar change in price per 100m closer to MRT, or per extra floor higher. An r² value is reported alongside each premium as a signal-strength indicator, so you can tell whether the relationship is robust in that slice or just noise.

PropKaki's Pricing Model uses price per square metre (PSM) because total price varies enormously with floor area, which would dominate every regression and produce misleading slopes, including wrong-sign results in some slices. Working in PSM controls for unit size automatically and gives a clean read of how each feature is priced. Headline numbers are then scaled back to a dollar figure by multiplying by the typical floor area for the property type you have chosen.

On PropKaki's Pricing Model, r² is the share of price-per-sqm variance that a single feature explains within the slice you have selected; a higher r² means the feature has a clearer, more consistent relationship with price. The tool bands r² into four signal strengths: strong (0.25 or higher), moderate (0.08 to 0.25), weak (0.03 to 0.08) and no clear pattern below 0.03 or with fewer than 30 transactions. The signal strength appears as a coloured chip on every card.

A feature in PropKaki's Pricing Model shows "no clear pattern" for two main reasons. First, the slice may be too uniform for the feature to matter — for example, if every building in a Singapore planning area is within walking distance of MRT, then MRT distance does not vary enough to predict price. Second, the slice may be too small (fewer than 30 transactions) for any reliable slope. Widening the time window or removing the sub-type filter usually restores a signal.

All three are non-landed private homes in Singapore but URA classifies them differently. A Condominium is a development of at least 4,000 sqm with full strata-titled common facilities. An Apartment is a non-landed private development that does not meet the Condominium definition, typically a smaller project without the full facility set. An Executive Condominium (EC) is a public-private hybrid built and sold by private developers under HDB eligibility rules for its first 10 years, then fully privatised at year 10, and it usually prices between HDB flats and pure private condos. PropKaki's Private Pricing Model lets you analyse each segment separately or pool them as 'all non-landed'.

PropKaki's Private Pricing Model measures the freehold premium directly for the slice you select rather than quoting a fixed figure, since it varies by area and project. It splits transactions into Freehold (including leases of 500 years or more, which covers the 999-year and 9,999-year tenures that price like freehold) and Leasehold (typically 99-year), computes the mean price per square metre for each, and reports the difference scaled to a typical unit size. A two-sample t-statistic shows significance: above 4 means the two groups are clearly priced differently in that slice, below 1.5 means the distributions overlap too much to call.

Singapore land records distinguish strict-freehold titles from very long leaseholds such as 999 and 9,999 years, but in practical market pricing they are indistinguishable, so PropKaki's Private Pricing Model groups them together. A 999-year lease that began in the 1870s still has over 850 years remaining, far beyond any owner's horizon, and the market does not discount it for tenure expiry. Grouping them with strict-freehold gives a cleaner read of the genuine freehold-versus-99-year premium without splitting an already-small group.

For leasehold private property in Singapore (typically 99-year tenure), each additional year of remaining lease tends to add a per-year premium that varies with the slice. PropKaki's Private Pricing Model computes its remaining-lease card only on the leasehold subset of the slice — freehold and effective-freehold transactions are excluded since they have no meaningful 'remaining lease' to compare. If the leasehold subset is too small (fewer than 30 transactions), the card shows 'not enough data'.

PropKaki's Private Property Pricing Model uses URA's caveat database, the official record of every private property transaction in Singapore, where each caveat includes price, area, property type, tenure, completion year and full address. Amenity locations come from official open-data sources including the Land Transport Authority (LTA, for MRT and bus stops) and data.gov.sg (schools, hawker centres, malls), with property-to-amenity distances pre-computed using each building's postal code as the join key. Because the amenity table currently covers a subset of private projects, amenity-based cards (MRT, schools, bus stops, hawkers, malls) may run on fewer transactions than the tenure, building-age and remaining-lease cards; the filter bar shows both the total in the slice and the amenity-covered subset.