District 9 vs District 10: a data-grounded comparison for CCR buyers

The consensus, and where the data disagrees

Ask any agent in the Orchard belt and you will hear a tidy story. District 9 is the yield play, dense and rental-heavy, where pied-à-terre owners cycle through one and two-bedders. District 10 is the capital-preservation play, leafier, family-sized, school-zone driven. The headline framing has been around for years. The 2023 to 2026 caveat data only partly supports it.

On the part the consensus gets right, the unit mix divergence is real and structural. On the part the consensus gets wrong, the implied yield gap between D9 and D10 has effectively closed at the district-median level. A larger and more decision-relevant divergence sits where most buyers are not looking: incoming supply. District 9 has roughly 1.9x the future-TOP unit count of District 10, almost all of it clustered into the 2029 to 2030 completion arc.

Side-by-side headline numbers

Three years of private non-landed caveats, the same trailing window the district hub pages use:

Implied gross yield is twelve times the latest URA quarterly project-median rental PSF divided by the trailing 3-year median sale PSF. It approximates the district-median investor experience, before vacancy, agent fees, mortgage interest, and the BSD plus ABSD cost stack on the entry ticket.

Yield: the gap has closed

The District 9 yield premium was a real phenomenon for years, mostly because the district leaned on small-format rental stock that commands a higher per-sqft rent. The 2026 Q1 PMI Rental Median release puts D9 and D10 within three cents per sqft of each other (S$5.80 vs S$5.83). At the district-median level, the yield gap has effectively closed, and the headline arithmetic actually tilts marginally to D10 because its median PSF on the sale side runs lower.

Two caveats apply before any reader rushes to rebalance a portfolio. First, the district-median rental PSF is project-weighted, not unit-mix-weighted. A one-bedder in River Valley typically rents at S$7 to S$8 per sqft; a four-bedder at D'Leedon clears closer to S$4.50. If you cherry-pick by unit size, D9 still wins on yield, but only because investor-grade stock is disproportionately small. Second, gross yield ignores vacancy. D9 turns tenants faster than D10, which is good for rental hit-rate and bad for void periods between leases.

The serious takeaway is that the rental thesis for prime D9 stopped being a yield call several quarters ago. It is now a tenant-velocity call. If you cannot reliably re-tenant a one-bedder inside a four-week vacancy budget, the District 9 yield advantage on paper does not survive the year.

Where the divergence is real: unit mix

The consensus framing of D9 as small-format and D10 as family-format holds up cleanly in the caveat data. Across the trailing 3-year window:

District 9 transacts almost twice the absolute volume of one-bedders that District 10 does (523 vs 319), and District 10 transacts almost twice the absolute volume of four-bed-and-larger units (1,053 vs 593). These are not small differences. They drive everything downstream: who lives there, how long they stay, what the resale clearance pool looks like in five years.

The clean mental model is that D9 sells liquidity, D10 sells space. The same money buys you a smaller unit in a faster-moving secondary market in D9, or a larger unit in a thinner secondary market in D10. Both are coherent strategies, but they are different products at the same headline price.

Buyer composition: new sale vs resale

The new-sale share also separates the two districts. Of the 3-year volume, District 9 was 42.7% new sale and 55.6% resale, while District 10 was 31.4% new sale and 66.8% resale, with sub-sale making up the residual roughly 1.7% in each district. District 9 looks like a fresh-launch district because it has absorbed the River Green and River Modern launches in the past year; District 10 looks more like a resale market because its long-running stock (D'Leedon, Leedon Green, Valley Park) carries the bulk of the secondary flow, even as the newer Skye at Holland and Upperhouse cohorts work through their launch arcs.

This matters for entry mechanics. New-sale in D9 means progressive payment schedules and developer-controlled pricing. Resale in D10 means valuation risk on the bank loan and a negotiating dynamic with motivated sellers. Neither is inherently better; they call for different financing strategies.

Top transacted projects: 3-year volume

The five most-transacted private non-landed projects in each district, ranked by caveat count over the past three years:

The shape of each top-five says a lot. District 9's leaderboard is dominated by 2024 to 2025 launches still in absorption mode, with all five entries clearing at or above $2,780 PSF and three of five above $3,100 PSF. District 10's leaderboard mixes two recent launches (Skye, Upperhouse) with three mature projects, and the median entry sits roughly $200 PSF lower. The pricing dispersion within D10 is also wider, with D'Leedon anchoring the value end at $2,000 PSF and 19 Nassim setting the ceiling at $3,396 PSF.

For a buyer looking at like-for-like comparables, the practical implication is that D10 gives you a richer set of resale entry points across vintage and tenure, while D9 forces you into either a fresh launch or one of a handful of older, smaller projects.

The supply pressure question

This is where the headline-number reading and the data part company. The URA pipeline (PMI Resi Pipeline, latest release) lists future-TOP private non-landed projects in each district. We covered the broader pipeline methodology in how to read URA's pipeline data, including why already-TOPped projects must be filtered out. Applying the same future-TOP filter and excluding the duplicate Robertson Opus row that URA carries twice:

Supply pressure ratio = total incoming units / 3-year annualised caveat volume. A ratio above 1.0 means incoming TOP supply exceeds one year of current absorption. D10 Jervois Mansion unit count is null in the source, so the D10 unit total excludes it. Including any plausible Jervois contribution would not change the cross-district verdict.

The District 9 future-TOP slate is loaded into 2029 and 2030, when River Green (524 units) and River Modern (455 units) complete on Kim Seng Road. Combined with The Collective at One Sophia (367 units, 2029) and The Robertson Opus (348 units, 2027), the district will absorb close to 1,700 new private units in a three-year arc, into a market that historically clears just over a thousand units per year. District 10's arc is meaningfully lighter, dominated by Skye at Holland (666 units, 2029) and Upperhouse at Orchard Boulevard (301 units, 2028).

For an investor buying today with a 5-year hold, this is the single most decision-relevant difference in the data. The TOP wave will pressure both rental and resale markets when it lands. District 9 buyers should expect the 2029 to 2030 cohort to compete hard for the same tenant pool. District 10 buyers face roughly half the supply shock.

Capital trajectory: who's been pulling the median

Quarterly median PSF over the past three years shows two different stories. District 9 traded in a $2,250 to $2,650 PSF band for most of 2024, then broke up sharply through 2025 Q3 ($3,081) and 2026 Q1 ($3,140) as the River Green and River Modern launches dominated the caveat flow. Strip those two launches from the 2026 Q1 cohort and the district median falls back into the high $2,300s. The headline appreciation is real but heavily concentrated in two projects.

District 10 ran flatter through 2024 (mid-$2,200s), broke higher in 2025 Q3 to Q4 around $2,900 PSF as Skye at Holland and Upperhouse cleared, then retreated to $2,326 in 2026 Q1 as the new-sale balance shifted back toward resale stock. Early Q2 2026 caveats (partial quarter to mid-May) sit near $2,260, which is consistent with the post-launch normalisation rather than a fresh price signal. Both districts show the same launch-driven median-PSF lift; District 10 has already given some of it back, District 9 has not.

For a buyer trying to read forward direction, the honest answer is that district-median PSF is mostly a mix-of-vintage indicator at the CCR scale, not a true price-action signal. Project-level CAGR is the cleaner measure on a per-product basis.

The contrarian read

If a property agent leads with "D9 yields better, D10 holds better" in 2026, ask which dataset they are reading. At the district-median level, D10 gross yield (2.77%) edges out D9 (2.55%). The yield premium narrative is a holdover from a unit-mix-weighted reading of the market that has not been updated as the launch pipeline shifted.

The capital-preservation half of the consensus stands, but for a different reason than usually argued. D10 is not safer because it is leafier or more family-friendly. It is structurally less exposed to the 2029 to 2030 TOP wave, which means rental and resale clearance through the back half of any 5-year hold should be smoother than in D9. Lower supply pressure, not higher prestige, is the underwriting reason.

The two genuine reasons to pick District 9 over District 10 today are entry-ticket flexibility (more one and two-bedder inventory, lower absolute quantum at the small-format end) and exposure to the Singapore River and Robertson Quay regeneration arc as CapitaLand and Wing Tai complete their concentrated bets. Neither of those is a yield call.

Who fits where

District 9 fits the buyer who wants liquidity above all else (one or two-bed turnover stock that sells in any market), who is comfortable underwriting the 2029 to 2030 supply wave, and who values the River Valley and Robertson regeneration specifically. Owner-occupier singles and couples without kids in primary school are the natural fit.

District 10 fits the buyer who wants larger floor area in the CCR for the same money, who is anchoring around the tier-1 school catchments along Bukit Timah and Holland (ACS Primary, Nanyang Primary; Singapore Chinese Girls' School sits in D10 at the secondary level), and who prefers a deeper resale market over a heavier new-sale pipeline. Families with school-going children are the natural fit.

Neither district fits the yield-maximising investor in 2026. Both clear under 3% gross at the district median. The CCR thesis is capital preservation and cycle resilience, not income; if the rental return is the deciding number, run the same data slice for OCR districts before committing capital.

Related

• District 9 hub with the live project list, transaction feed, and per-project medians.
• District 10 hub with the live project list, transaction feed, and per-project medians.
• How to read URA's pipeline data for the supply-pressure methodology used in the table above.
• Mortgage and TDSR calculator with BSD, ABSD, joint-buyer IWAA, and full ROI for the typical D9 or D10 ticket.
• Save projects to your watchlist to track new transactions, rental data, and pipeline changes weekly.