The Developer League Table Is Mostly a Size Ranking

Buyers use the developer's name as a shortcut. It is the one piece of information on a launch brochure that feels like a quality signal rather than a sales claim, and it carries into the resale market years later: it's a City Developments project, that's a UOL block, as though the name were a grade.

It is testable. The Resale Project Scorecard carries a developer field and a single weighted score for 2,357 resale condominiums. Group the scores by developer, and the names do separate — by a lot. Then ask the only question that matters: is the separation about who built it, or about what they chose to build?

It is almost entirely the second one.

The market is more atomised than the brochures suggest

Before the league table, a fact that reframes it. Of the 2,357 scored projects, 2,042 carry a named developer, and those names number 1,090. Nine hundred and twenty-two of them appear exactly once. Only 26 developers have built ten or more of Singapore's scored resale condos, and only twelve have built twenty or more — between them 495 projects, 21% of the market.

So the familiar names are not the market. They are a fifth of it, and the rest is a very long tail of one-off developers whose track record cannot be assessed at all.

The raw table

National mean across all 2,357 projects is 5.946.

A 1.43-point spread, which on this scorecard is roughly the distance between a B and an S. Read straight, it says Sim Lian builds far better condos than Keppel Land.

Read the third column of the next table instead.

What they built, and when

Two things dominate this scorecard that have nothing to do with construction quality. Project size carries 16% of the card directly, and larger projects also tend to sit on suburban sites with better school catchments and stronger measured price growth. Completion decade matters because the historical-performance sub-score, at 25% the heaviest factor, is computed against a project's own vintage cohort, and because newer projects sit on land released more recently — closer to new rail, in newly built-out towns.

Those two variables are not randomly distributed across developers. They are the developers' business models.

The four developers at the top of the raw table are the four with the largest median projects. That is the whole of the first-order story.

Adjusting for it

Every project is assigned to a cell defined by project-size band × completion decade, and each developer's portfolio is compared against the average score of the cells it actually occupies. The difference — call it alpha — is how much better or worse a developer's projects score than other projects of the same scale and vintage.

The 1.43-point raw spread becomes 0.91, and if you set Keppel Land aside it becomes 0.64 — eleven of the twelve developers land within a third of a point either side of zero, which on a ten-point card is nothing.

Three specific reversals are worth naming.

MCL Land goes from +0.950 to −0.034. Its entire apparent advantage is that it builds the largest projects of any developer in the table. Judged against other large projects of the same age, it is exactly average.

Roxy-Pacific goes from 40th-of-a-point above the national mean — dead average, tenth of twelve — to second on alpha. It builds the smallest projects in the table by a distance, median 80 units, mostly in Districts 9, 14 and 15. Small projects score badly on this card almost automatically. Roxy-Pacific's do better than other small projects, and the raw table hides that completely.

CapitaLand and GuocoLand both look mildly above average raw and are mildly below on alpha. GuocoLand's median completion year is 2022 — the newest portfolio in the table — and new suburban-rail projects score well as a class.

The one real outperformer

Sim Lian Group survives the adjustment, and it is the only one of the four raw leaders that does. +0.285 after controls is small in absolute terms but it is the largest positive in the table, and the composition of its portfolio explains why.

Twenty-four of its 44 projects carry an S grade — against 218 S grades in the whole 2,357-project market. Thirty-two of the 44 are 99-year leasehold. Its S grades cluster hard in District 19 (eight of them) and District 18 (four) — Hougang, Sengkang, Punggol, Tampines and Pasir Ris. Its sub-scores read schools 7.95 against a national 6.40, historical performance 7.64 against 5.57, project size 7.93 against 4.64.

This is a developer that has spent two decades building large leasehold projects in family suburbs with deep primary-school catchments, at a time when that was the part of the market that ran hardest. Whether that is skill or a well-held position, the scorecard cannot tell you. What it can say is that the projects score better than other large suburban projects of the same vintage, which is more than the other three raw leaders manage.

Its own floor is instructive too: Floraville at 4.43, a 50-unit 2017 project, is its lowest card. When Sim Lian builds small, it scores like everyone else who builds small.

The one real laggard

Keppel Land is the only genuine outlier in the table, and the gap is wide: −0.622, three-quarters again as large as the next-worst. It is not a small-project penalty — its median project is 327 units, sixth-largest of the twelve. It builds at scale and underperforms other projects built at the same scale.

The sub-scores locate it. Schools 5.09 against a national 6.40, and historical performance 4.76 against 5.57 — the two heaviest factors on the card, 45% of it between them. Its grade distribution is B13 / C9 / S4 / D4 / A3.

The geography explains most of that. Keppel Land's portfolio concentrates in Districts 10, 4, 9, 5 and 11 — prime and prime-adjacent. Its weakest cards are prime-district boutiques: Ardmore 3 at 3.88, 19 Nassim at 4.01, Caribbean @ Keppel Bay at 4.48 despite 969 units. Its best are suburban and large: Waterfront Key 8.48, Carabelle 8.05, The Glades 7.99.

That is the same pattern we have documented district by district — the prime core scores badly on this card, on thin rental yields and weak measured appreciation — showing up in a single developer's book because that is where the developer chose to build. It is a portfolio result, not a construction result, and a buyer should read it as one.

What this is and is not

This measures resale scorecard outcomes, not build quality. Nothing here is about defects, finishes, handover disputes or how a development is managed after handover, and nothing here should be read as a claim about any of them. A developer can build a well-made project on a site that scores poorly, and the card will mark the site.

It is also silent on 79% of the market. The 1,090-name tail cannot be evaluated, and a developer with three projects has no track record in any statistical sense — the twelve here are the only ones with enough projects to say anything about at all.

The takeaway for a buyer

Two things follow.

The brand is not carrying the information you think it is. Once you know a project's size and vintage, knowing who built it moves the expected scorecard result by about a third of a point at most in either direction for eleven of the twelve most prolific developers in Singapore. That is smaller than the gap between two projects on the same street.

The things the brand is standing in for are all directly observable. Project size, completion year, tenure, distance to a station, which primary schools fall inside a kilometre — these are on the listing, and they are what the league table was actually measuring. Read those, and the developer's name adds very little on top.

The exception proves the rule. Where a developer does show a real effect, it is because of a consistent, decades-long choice about where and at what scale to build — and you can see that choice in the project in front of you without knowing the name at all.