Condo rental yield in Singapore: all 26 districts ranked, and three things nobody tells you
At a glance — 7 October 2026
- Highest condo rental yield: District 14 at 4.52% gross.
- Lowest: District 11 at 2.82% gross. That is a spread of 1.70 percentage points, or 1.6028 times.
- Median district: 3.71% gross. Mean 3.6954%. If a broker quotes you 5% on a Singapore condo, ask what is in the denominator.
- Correlation between a district’s yield and its rent level: +0.0485. Effectively zero. High rents do not mean high yield.
- Correlation between yield and sale psf: −0.6038 (R-squared 0.3646). Price explains about 36% of the gap, which leaves two thirds unexplained.
- One-bedders vs five-bedders: 3.8759% against 2.5981% on a like-for-like asking basis. A 1-bed yields 1.4918 times a 5-bed.
- Every figure here is gross, not net. Tax, maintenance, agent fees, repairs and voids are not deducted.
The number every landlord quotes predicts almost nothing
I have sat in a lot of living rooms over 17 years. When an owner wants to tell me their condo is a good investment, they almost always reach for the same number. The rent. “This unit rents for $4,800, Gary. Very strong.” And they are right that it is strong. But as a statement about condo rental yield, it is close to meaningless.
Here is the measurement. Across the 26 Singapore postal districts with enough condo rental stock to measure, I pulled the average annual asking rent per square foot and the average transacted sale psf, then ran the correlation between a district’s yield and its rent level. The answer came back +0.0485.
That is not a weak relationship. That is no relationship. It is the statistical equivalent of a shrug. Knowing that a district commands high rents tells you essentially nothing about whether buying there produces a good return.
Look at what that means on the ground, straight from the table below. District 07 — Beach Road, Bugis, Rochor — has the second-highest condo rents in the country at $99.16 per square foot per year. Its yield is 4.16%, fifth best out of 26. Great. Now District 17, out by Changi and Loyang, has the lowest rents in Singapore at $50.12 per square foot per year. Its yield? 3.50%. Dead mid-table. The cheapest rents in Singapore and a perfectly ordinary return.
And then the one that ends the argument. District 11 — Newton, Novena, Thomson, some of the most desirable addresses on the island — has middling rents at $65.43 per square foot per year, and the worst condo rental yield in Singapore at 2.82%. Mid-table rent. Bottom-of-the-table return.
So the practical point is this: “this district rents well” is not an investment case. It is half a sentence. Rent is the numerator. Yield is a fraction. And the number almost every landlord in Singapore quotes me is the one that predicts the least.
That is the lead because it is the mistake that costs the most money. People buy a nice address, get the strong rent they were promised, and then wonder for ten years why the return never showed up. The rent was never the problem. The denominator was.
What follows is the whole table, three original measurements I have not seen published anywhere else, and an honest accounting of what this data cannot tell you. I would rather you leave this page with four reliable numbers than fourteen impressive ones.
How these numbers were built — read this before the table
I am putting this before the data, not after it, because a yield table without its method is just decoration. Five things you need to know, and none of them are buried.
One. These are gross yields. Not net. Annual rent divided by purchase price, full stop. Nothing has been deducted. Property tax has not come off. Neither has maintenance, sinking-fund contributions, agent commission on the lease, repairs, replacement of an aircon system that dies in year four, income tax on the rental, or the void month between tenants that almost everybody forgets to budget for. A 4.52% gross yield is not 4.52% in your pocket. If you want to run that properly, the calculators at listings.sg/tools do the money math with the outgoings included, and I would rather you use them than trust a headline.
Two, and this is the important one. The two legs of the fraction sit on different bases. The rent leg is the average asking rent from live rental listings on Listings.sg as at 7 October 2026. The sale leg is the average transacted psf from URA sale caveats. Asks are not deals. Landlords ask high and settle lower, almost always. So treat the rent leg as a forward-looking ceiling, not as achieved income, and expect the level of every yield in this table to be slightly generous.
Why do I still publish it? Because the bias is broadly in the same direction in every district, so it affects the level of the numbers far more than the ranking. If landlords across Singapore ask somewhere around five to eight per cent above what they eventually sign, then every yield here is inflated by roughly that much, and the order barely moves. For ranking districts against each other, which is what an investor actually needs, the asking basis holds up. It also has one real advantage over transacted-lease data: it tells you what the market is asking today, not what it signed six months ago. That is a forward-looking read, and in a moving market I will take it.
Three. These are averages, not medians. A handful of very large or very expensive units will pull a district’s figure. In a district with a few enormous penthouses and not much else, the average is doing some heavy lifting. This is why I show you the sample sizes in the table. Look at them. D19 rests on 13,023 sale caveats and 2,282 live rental listings — that is a solid number. D26 rests on 369 rental listings. Both are in the table. They are not equally reliable, and I am not going to pretend otherwise.
Four. District averages mix project vintages. A 40-year-old leasehold block with a lift that groans and a 2025 new launch with a sky gym sit in exactly the same cell of this table. Freehold and 99-year sit together. A 400 sq ft shoebox and a 2,800 sq ft family unit sit together. District-level data smooths all of that away, which is the point of it and also its limit. It will tell you where to look. It will never tell you what to buy.
Five. Twenty-six of 28 districts. D06 and D24 are not here. Neither has enough active condo rental stock to produce a figure I would stand behind, so they are excluded rather than estimated. If a yield table shows you all 28 districts, ask what it did about those two.
Condo rental yield by district: all 26, ranked
Here is the whole thing, sorted highest yield to lowest. Gross yield is average annual asking rent psf divided by average transacted sale psf. Sample columns are there so you can judge how much weight each row carries.
| District | Gross yield | Avg sale psf | Avg annual rent psf | Sale caveats | Live rent listings |
|---|---|---|---|---|---|
| D14 | 4.52% | $1,673.35 | $75.67 | 4,241 | 1,178 |
| D25 | 4.41% | $1,268.25 | $55.97 | 1,889 | 305 |
| D28 | 4.41% | $1,360.25 | $60.02 | 2,071 | 327 |
| D08 | 4.32% | $1,822.55 | $78.80 | 1,442 | 522 |
| D07 | 4.16% | $2,381.82 | $99.16 | 960 | 415 |
| D27 | 4.13% | $1,334.39 | $55.15 | 5,076 | 730 |
| D22 | 4.07% | $1,656.51 | $67.37 | 2,893 | 689 |
| D02 | 4.06% | $2,363.24 | $95.84 | 1,279 | 711 |
| D13 | 3.97% | $1,908.16 | $75.85 | 2,064 | 408 |
| D19 | 3.91% | $1,581.77 | $61.85 | 13,023 | 2,282 |
| D04 | 3.85% | $1,852.27 | $71.29 | 1,606 | 628 |
| D18 | 3.80% | $1,547.43 | $58.81 | 9,208 | 1,356 |
| D23 | 3.74% | $1,512.25 | $56.54 | 7,565 | 1,050 |
| D20 | 3.68% | $1,790.87 | $65.83 | 2,752 | 654 |
| D16 | 3.54% | $1,766.12 | $62.50 | 4,720 | 1,023 |
| D05 | 3.52% | $1,973.68 | $69.43 | 7,604 | 1,614 |
| D17 | 3.50% | $1,433.38 | $50.12 | 2,566 | 435 |
| D12 | 3.49% | $1,953.82 | $68.10 | 3,085 | 654 |
| D03 | 3.46% | $2,442.35 | $84.55 | 5,359 | 1,482 |
| D01 | 3.41% | $2,509.31 | $85.62 | 1,862 | 657 |
| D15 | 3.19% | $2,142.01 | $68.44 | 9,874 | 1,818 |
| D26 | 3.10% | $2,121.45 | $65.82 | 4,655 | 369 |
| D09 | 3.09% | $2,616.66 | $80.95 | 5,688 | 2,451 |
| D10 | 3.00% | $2,569.53 | $77.18 | 7,127 | 3,169 |
| D21 | 2.93% | $2,073.29 | $60.71 | 4,935 | 798 |
| D11 | 2.82% | $2,321.62 | $65.43 | 2,666 | 963 |
| Median district yield 3.71%. Mean 3.6954%. Unweighted mean sale psf $1,922.17; unweighted mean annual rent psf $69.88. D06 and D24 excluded for thin condo rental stock. Gross yields — no outgoings deducted. | |||||
Three things to notice before we get into the analysis.
First, the whole island fits inside a 1.70 percentage point band. From 2.82% to 4.52%. The best district yields 1.6028 times the worst. That is a real difference — on a million-dollar purchase it is roughly $17,000 a year of gross rent — but it is not the chasm people imagine. Nobody in Singapore is quietly collecting 8% on a condo. If someone tells you they are, they have either left something out of the denominator or they are talking about a different asset class entirely.
Second, the top of the table is not where the glossy marketing is. D14 Geylang and Paya Lebar. D25 Woodlands. D28 Seletar and Yio Chu Kang. D08 Little India and Farrer Park. These are working districts. They do not appear in launch brochures with a drone shot and a sunset. They sit at the top of this table anyway.
Third, and I want to be fair about this, the bottom of the table is not a list of bad buys. D11, D10, D09 and D21 are there because their prices are high, and prices are high because people want to own there. A 2.82% yielding condo in Newton may still be the better ten-year hold than a 4.52% yielding one in Geylang. Yield measures income efficiency. It does not measure capital growth, liquidity, or how easy the thing is to sell in a soft market. I will come back to that trade-off properly in the who should chase yield section.
Price predicts condo rental yield — but only about a third of it
Everybody knows the folk rule: cheap districts yield more, prestige districts yield less. Inverse prestige. And the data agrees with it. The correlation between a district’s yield and its average transacted sale psf is −0.6038. That is a genuinely strong negative relationship. The rule is real.
Now square it. R-squared is 0.3646. Price explains about 36% of the variation in yield between districts. Which means roughly two thirds of the yield gap in Singapore is not explained by how expensive the district is at all.
| Relationship | Correlation | R-squared | What it means |
|---|---|---|---|
| Yield vs avg annual rent psf | +0.0485 | 0.0024 | Nothing. Rent level is not an investment case. |
| Yield vs avg sale psf | −0.6038 | 0.3646 | Strong tendency, not a law. Explains about a third. |
| Sale psf vs rent psf | +0.7587 | 0.5756 | They move together, but loosely. The slack is the yield. |
I want to state that carefully, because this is exactly the place where property writing overclaims. The honest version is: a strong tendency, not a law. If I handed you only a district’s price and asked you to guess its yield rank, you would do considerably better than chance and considerably worse than right.
The exceptions are sitting in the table in plain sight, and they are the ones worth memorising.
D07 is the third most expensive district in Singapore by transacted psf at $2,381.82 — and it still yields 4.16%, fifth best on the island. That single row breaks the folk rule on its own. Expensive and high-yielding at the same time. The reason is that D07’s rent leg is enormous: $99.16 per square foot per year, second only to nothing much. Bugis, Beach Road and Rochor sit on top of the CBD fringe, the tenant pool is deep and willing, and the rents have kept up with the prices. That is what a healthy high-price district looks like.
And D17 is the third cheapest district at $1,433.38 psf — and yields only 3.50%, flatly mid-table. Cheap and unremarkable. Here the rent leg is the lowest in Singapore at $50.12. Changi and Loyang are cheap to buy because they are far from everything, and they are cheap to rent for exactly the same reason. The discount shows up on both legs, so it cancels, and the yield lands in the middle.
Put those two rows side by side and the lesson is clean. Cheapness is not yield. Cheapness is a discount that may or may not be bigger than the rent discount that comes with it. You are not hunting for low prices. You are hunting for districts where the price discount is deeper than the rent discount. That is a different search entirely, and it is why a table like this one is worth having.
Why the spread exists: two legs, loosely joined
This is the subtle bit and I think it is the part most worth getting right, so bear with me for a few paragraphs of mechanics.
When you see a 1.70 percentage point spread in yields, the intuitive explanation is that one of the two legs must be swinging much harder than the other. Either prices vary wildly while rents stay flat, or rents vary wildly while prices stay flat. That would produce a spread. It is also wrong.
Check the dispersion of each leg. Sale psf across the 26 districts runs from $1,268.25 to $2,616.66 — a range of 2.0632 times, with a coefficient of variation of 0.2075. Annual rent psf runs from $50.12 to $99.16 — a range of 1.9785 times, coefficient of variation 0.1727. The ratio of those two coefficients of variation is just 1.2018.
In plain English: the two legs are about equally spread out. Prices vary roughly as much as rents vary. Neither leg is the villain.
So the yield spread is not caused by one leg being more variable than the other. It is caused by the two legs being imperfectly correlated with each other. Correlation between a district’s sale psf and its rent psf is +0.7587. Rent and price move together — strongly — but loosely. And the gap between them is the yield.
Think about what a 0.7587 correlation leaves on the table. If sale psf and rent psf were correlated at 1.0, every district in Singapore would have an identical yield and this article would not exist. If they were correlated at zero, yields would be chaos and nobody could plan anything. At 0.7587 you get what we actually observe: a sensible, mostly-ordered table where the cheap districts generally do yield more, punctuated by genuine exceptions like D07 and D17 where the two legs drifted apart.
That slack between price and rent is where the opportunity lives. It is also why I am sceptical of any single-factor rule for picking an investment district. The rule “buy cheap districts” captures about 36% of the story. The remaining 64% is the slack — local tenant demand, project mix, how much new supply landed recently, whether the MRT line opened before or after the launch prices were set, how many shoebox units are in the district’s stock. None of that is visible in a price tag.
Practically, this is why I run the two legs separately for a client rather than looking at a yield number. Yield is a ratio and ratios hide things. If a district’s yield is 4%, I want to know whether that is a cheap price with an ordinary rent, an ordinary price with a strong rent, or both. Those three districts behave completely differently over a ten-year hold, and only one of them is interesting if you also care about resale. For the long-run price side of that question, our analysis of the HDB-to-condo price gap in 2026 is the better companion piece to this one.
The ask-to-transaction gap: a metric nobody publishes
Here is the measurement I am proudest of in this article, because as far as I can tell no one in Singapore publishes it, and it takes about four seconds to use.
The sale leg in the district table above is URA transacted caveats — real deals, real prices, real money that changed hands. But Listings.sg also publishes the average asking psf of live condo sale listings. Same site, same day, same district definitions. So you can put the two side by side and measure, district by district, how far above recent transacted prices sellers are currently asking.
That gap is a negotiation signal. A wide gap means sellers in that district have not yet adjusted to what the market is actually paying. A narrow gap means asks and deals are close, so there is less room and less point in grinding.
| District | Live avg asking psf | Avg transacted psf | Gap |
|---|---|---|---|
| D07 | $2,919.78 | $2,381.82 | +22.59% |
| D08 | $2,220.95 | $1,822.55 | +21.86% |
| D20 | $2,145.08 | $1,790.87 | +19.78% |
| D28 | $1,539.03 | $1,360.25 | +13.14% |
| D15 | $2,411.38 | $2,142.01 | +12.58% |
| D09 | $2,944.15 | $2,616.66 | +12.52% |
| D10 | $2,866.49 | $2,569.53 | +11.56% |
| D14 | $1,859.28 | $1,673.35 | +11.11% |
| D12 | $2,169.94 | $1,953.82 | +11.06% |
| D19 | $1,748.69 | $1,581.77 | +10.55% |
| D02 | $2,602.31 | $2,363.24 | +10.12% |
| D13 | $2,080.45 | $1,908.16 | +9.03% |
| D23 | $1,648.68 | $1,512.25 | +9.02% |
| D22 | $1,799.67 | $1,656.51 | +8.64% |
| D04 | $1,990.62 | $1,852.27 | +7.47% |
| D27 | $1,420.67 | $1,334.39 | +6.47% |
| D17 | $1,517.93 | $1,433.38 | +5.90% |
| D21 | $2,182.41 | $2,073.29 | +5.26% |
| D18 | $1,628.28 | $1,547.43 | +5.22% |
| D25 | $1,333.04 | $1,268.25 | +5.11% |
| D11 | $2,438.88 | $2,321.62 | +5.05% |
| D03 | $2,559.33 | $2,442.35 | +4.79% |
| D05 | $2,062.69 | $1,973.68 | +4.51% |
| D26 | $2,194.32 | $2,121.45 | +3.43% |
| D01 | $2,578.50 | $2,509.31 | +2.76% |
| D16 | $1,744.44 | $1,766.12 | −1.23% |
| Median gap +8.83%. Mean +9.17%. Shaded rows at the top are the only three gaps that survive the noise-floor test below. D16 is the only district where sellers are, on average, asking below recent transacted psf. | |||
The median district has sellers asking +8.83% above recent transacted psf. The mean is +9.17%. That is the normal temperature of the Singapore condo sale market right now: sellers list roughly nine per cent above what is actually closing, and the negotiation closes most of that.
And then there is D16. Bedok, Upper East Coast, Eastwood, Kew. It is the only district in Singapore where condo sellers are, on average, asking below recent transacted psf — minus 1.23%. Every other district on the island is asking up. D16 is asking down.
I will not oversell what that means, because I cannot prove the cause from this data. It could be composition. It could be that the particular projects currently listed in D16 skew older or smaller than the ones that recently transacted. It could be a cluster of owners who want out. But it is the sort of thing worth knowing before you walk into a viewing, and nobody was going to tell you.
The noise floor test: which gaps survive
Now the discipline, because a table of 26 numbers is where honest analysis usually goes to die.
The house rule on my team is that you never interpret a measured effect without first asking how big the ordinary scatter is. We call it the noise floor: how far apart two genuinely comparable units trade for no interesting reason at all. Same project, same stack, same size, different month, different seller mood, different agent. Across eight separate measurements on this project we have recorded noise floors of 2.56%, 3.69%, 4.22%, 4.43%, 4.75%, 5.29%, 5.31% and 5.57%.
For this article I am using the conservative 3.7% figure — the second-lowest of the eight, which is the harshest realistic floor I can justify. Using a low floor is deliberately hard on myself: it lets more of my own findings survive, so if a finding fails even here, it was never real.
Here is the consequence, stated plainly because it has to be.
Every gap below about 5% in that table is inside ordinary unit-to-unit scatter and must not be read as a signal. With a 3.7% noise floor, and allowing a modest margin for the composition problem I describe below, a gap has to clear roughly 5% before it means anything at all.
That retires D16, D01, D26, D05, D03, D11, D25, D18 and D21 from interpretation entirely. Nine districts. Including, yes, the D16 negative gap I just spent three paragraphs on — it is interesting, it is worth watching, and at minus 1.23% it is statistically indistinguishable from zero. I am telling you that rather than letting you run with it.
What survives are three genuinely wide gaps: D07 at +22.59%, D08 at +21.86% and D20 at +19.78%. Those three are four to six times the noise floor. They are real.
And now the limitation I am obliged to disclose, because it is the one that could undermine the whole metric. The asking leg and the transacted leg are different mixes of projects. The units currently listed for sale in a district are not a random sample of the units that recently sold there. If the live listings in D07 happen to skew towards newer, bigger or freehold stock, part of that +22.59% is composition, not seller optimism. I cannot separate the two with district-level data, and I am not going to pretend the gap is pure sentiment.
So treat this for exactly what it is: a screening tool for which districts to negotiate hardest in. It is not a valuation. It does not tell you a D07 condo is 22% overpriced. It tells you that if you are viewing in D07, D08 or D20, the asks you are shown are sitting a long way above what has recently closed, you should ask your agent for the transacted comparables before you make an offer, and you should expect that offer to be well below the sticker. In the nine retired districts, do the same homework but do not expect the gap to be your friend.
That is a useful thing to know and a modest claim. I would rather publish the modest version.
One-bedders win, and the reason is not the price
Different measurement, different basis, so I am labelling it clearly. Everything in this section comes from live Listings.sg asking data on both legs — asking rents against asking prices, 7 October 2026. That is internally consistent, and it is not the same basis as the district table above, which pairs asking rents against transacted caveats. Do not mix the two sets of numbers.
| Bedrooms | Rent listings | Avg ask rent/mo | Rent psf/mo | Sale listings | Avg ask price | Ask psf | Gross yield |
|---|---|---|---|---|---|---|---|
| 1-bed | 5,485 | $3,584 | $7.07 | 4,901 | $1,178,591 | $2,188.91 | 3.8759% |
| 2-bed | 9,306 | $4,792 | $5.92 | 8,210 | $1,748,277 | $2,179.96 | 3.2588% |
| 3-bed | 7,703 | $6,908 | $5.06 | 9,103 | $2,537,363 | $1,978.01 | 3.0698% |
| 4-bed | 2,850 | $12,870 | $5.55 | 4,454 | $4,990,745 | $2,280.65 | 2.9202% |
| 5-bed | 387 | $16,422 | $5.36 | 1,331 | $7,250,995 | $2,475.70 | 2.5981% |
| Annual rent psf: 1-bed $84.84, 2-bed $71.04, 3-bed $60.72, 4-bed $66.60, 5-bed $64.32. A 1-bed yields 1.4918 times a 5-bed and beats a 3-bed by 0.8061 percentage points. Gross, before outgoings. | |||||||
The headline is the one you have probably read before: small units yield more. A one-bedder at 3.8759% against a five-bedder at 2.5981%. That is 1.4918 times the yield. Against a three-bedder, the one-bedder wins by 0.8061 percentage points — which, in a market where the entire district spread is 1.70 points, is close to half the available range from a single decision about unit size.
What I have not seen anyone explain is why, and the answer is genuinely surprising. Everyone assumes the one-bedder wins because it is cheap — small ticket, small denominator, big ratio. That explanation is wrong, and the data kills it in one line.
A one-bedder and a two-bedder cost almost exactly the same per square foot to buy — asking psf $2,188.91 against $2,179.96, just 0.411% apart — but they cost 19.426% apart per square foot to rent, $7.07 against $5.92. So the entire yield advantage of a one-bedder over a two-bedder comes from the rent side. The sale market prices those two square feet as identical. The rental market does not.
Sit with that for a second, because it changes how you should think about the whole decision. Four tenths of one per cent apart on purchase psf. Nineteen per cent apart on rental psf. The buying market has decided that a square foot of one-bedroom condo and a square foot of two-bedroom condo are the same thing. The renting market has decided they are emphatically not, and is willing to pay nearly a fifth more per square foot for the smaller one.
The reason is demand per square foot. A tenant renting a one-bedder is usually a single professional or a couple who needs an address, a bedroom and a door that locks, and they are paying for access, not for area. The rent floor for “a liveable private unit in a decent location” does not scale down below a certain point no matter how small you make the box. Squeeze the box and the rent barely falls, so rent psf shoots up. Meanwhile the sale market prices by land and build cost and comparable psf, which does not care how many walls you put inside.
Which means the one-bedroom yield premium is not an affordability artefact. It is a structural mispricing between two markets that value the same square foot differently. That is a much more durable thing to own than a price quirk, and it is the single most useful sentence in this article.
Two caveats so you do not over-rotate. One-bedders have higher tenant turnover, so your void risk and your re-letting agent fees are higher — both of which come off a gross figure. And they are a narrower resale pool: when you sell, you are selling to investors and singles, not to families, and investor demand is more sensitive to interest rates and cooling measures than family demand is. The yield is real. It is not free.
For a related question — whether the unit you want is better bought new or secondhand — we put the numbers side by side in resale condos vs new launch in 2026.
The three-bedroom trough: an honest anomaly
Now the part I could have quietly left out, which is exactly why it is going in.
If small units rent better per square foot, the rent psf ladder should fall smoothly as unit size rises. It does not. Across one to five bedrooms it runs: $7.07, $5.92, $5.06, $5.55, $5.36 per square foot per month.
Read that again. It falls, falls, then rises, then falls. The three-bedroom is the trough — the worst rent psf of any unit size in the Singapore condo market. A four-bedder rents for more per square foot than a three-bedder. So does a five-bedder. That should not happen under a simple small-is-efficient story, and it is the kind of bump that most writers would smooth over or not notice.
I do not have a proven cause. Let me be explicit: what follows is a hypothesis, not a finding, and I have not tested it.
The plausible explanation is that the three-bedder is the family unit, and family units are not priced by room count. They are priced against the alternatives a family actually considers. A household looking at a three-bedroom condo is often also looking at a four- or five-room HDB flat, an executive condominium, or simply staying put for another two years. That is a wide, price-sensitive, well-informed competitive set, and it caps what a three-bedroom condo can ask. One-bedders have no HDB substitute at all — there is no public-housing product for a single expatriate professional who wants a private lobby and a pool. Four- and five-bedders are up in a thin premium segment where the tenant is often on a corporate housing budget and the comparison set is other large private units, not public housing. So the middle of the ladder is the one squeezed from below.
That story is consistent with the data. It is not proved by it. Testing it properly would mean looking at three-bedroom condo rents against four- and five-room HDB rents in the same towns over several quarters, and that is a separate piece of work. If the hypothesis holds, the practical implication is that the three-bedroom segment is the hardest place in the condo market to push rent, which matters a lot if that is the unit you are about to buy. If it does not hold, something else is going on and I would want to know what.
On the family-alternative point specifically, the cheapest route into a three-bedroom private-style home is often not a condo at all, which is the case we make in why the executive condo is still Singapore’s best value hack in 2026.
What URA’s 3Q2026 flash estimate does to these yields
Yield is a fraction, and in the last quarter the denominator moved. You need to read this table in that light.
On 1 October 2026 the Urban Redevelopment Authority published its flash estimate of the 3Q2026 private residential property price index. The figures, in URA’s own words: overall prices “increased by 1.4% in 3rd Quarter 2026, compared to an increase of 0.5% in 2nd Quarter 2026”. Non-landed prices “increased by 0.9%, compared to the 0.1% decrease in the previous quarter”. Landed “increased by 2.8% in 3rd Quarter 2026, compared to the 2.5% increase in the previous quarter”.
The regional split is where it gets interesting for everything above. Core Central Region “decreased by 0.1%, compared to the 1.8% increase in the previous quarter”. Rest of Central Region “increased by 0.2%, compared to the 1.2% decrease in the previous quarter”. And Outside Central Region “increased by 2.2%, compared to the 0.1% decrease in the previous quarter”.
Follow that through mechanically. OCR rose hardest, by 2.2%, while the prime core actually slipped 0.1%. A rising denominator compresses yield wherever rents did not keep pace. And OCR is precisely where the high-yield districts in my table live — D25 Woodlands, D28 Seletar, D27 Yishun and Sembawang, D22 Jurong, D19 Hougang and Serangoon, D18 Tampines and Pasir Ris, D23 Bukit Batok and Choa Chu Kang.
So the honest reading is that the top of my table is under mild pressure. If OCR prices rose 2.2% in a quarter and OCR asking rents did not rise by a similar amount, the yields shown for those districts are the high-water mark rather than a trend. A 4.41% in Woodlands measured today could be a 4.3% measured in January without a single thing going wrong — just prices doing what URA says they did.
The flip side is equally mechanical and slightly cheering for the bottom of the table. CCR prices fell 0.1% and RCR barely moved at +0.2%. If prime rents hold steady while prime prices stay flat or dip, the 2.82% in D11 and the 3.00% in D10 have a little room to improve from here, purely arithmetically. Not much. But the direction is the opposite of the OCR direction, and that is worth noticing if you have been assuming the gap between prime and suburban yields only ever widens.
One date to put in your calendar. These are flash estimates, computed from transaction prices in the first ten weeks of the quarter. URA states that its “statistics will be updated on 23 October 2026 when URA releases its full set of real estate statistics”. 23 October 2026 is when these numbers get revised, and it is also the first date on which a genuine 3Q2026 OCR median exists rather than an estimate of one. If you are making a decision that turns on whether OCR really rose 2.2%, wait two weeks. I will be refreshing this page shortly after that release.
How I read this table when a client calls
Let me show you the actual workflow, because a table is only as good as the question you bring to it.
When somebody calls me and says “Gary, where should I buy for rental yield”, I do not start at the top of the table. I start by asking what they are optimising for, because about half the time the answer is not yield at all. It is a parent who wants a unit their child can live in later. It is a business owner who wants an asset that is easy to liquidate in six weeks. It is somebody who has already decided on a district and wants me to validate it. Yield is the right lens for maybe one caller in three.
For the ones where it is the right lens, here is the sequence.
First, I decompose the yield. As I said earlier, I never look at the ratio alone. For every shortlisted district I look at the price leg and the rent leg separately, because a 4% built from a cheap price and a weak rent is a completely different asset from a 4% built from a fair price and a strong rent. D07 at 4.16% with a $99.16 rent leg is the second kind. D17 at 3.50% with a $50.12 rent leg is a thin market on both sides. One of those has pricing power. The other has distance.
Second, I check the sample size. The table shows it for a reason. If a district rests on a few hundred live rental listings, the average is fragile and a single large project’s asking prices can move it. I weight D19’s 2,282 rental listings and 13,023 caveats far more heavily than D26’s 369 rental listings, and I say so out loud to the client rather than presenting all 26 rows as equally solid.
Third, I cross-check the ask gap. If a district made the shortlist on yield and also shows a wide gap — D07 and D08 are both in this situation, high yield and a gap above 20% — then the yield in the table was computed on transacted prices, but the price you will actually be shown is the asking price. Which means the yield you personally achieve depends entirely on how well you negotiate. In a district with a 22% gap, buying at the ask instead of near the transacted comparable can wipe out the entire yield advantage that got the district onto your shortlist in the first place. That is not a small point. It may be the most actionable sentence in this whole article.
Fourth, I drop to project level, and I stop using this table. District averages mix a 1985 leasehold walk-up-adjacent block with a 2024 launch. The moment a client is choosing between two actual projects, district data is the wrong instrument and I switch to project comparables, stack, floor, facing, lease remaining and the specific tenant pool within walking distance. That is where the real money is made or lost, and no district table will get you there.
Fifth, I run the net. Always. Gross is for ranking. Net is for deciding. More on that below.
Want this run on your actual numbers?
If you have a shortlist, send it over and I will decompose both legs, pull the project-level comparables and show you the ask gap on the specific units you are looking at. Free valuation and a straight answer at buycondo.sg, or message me directly on WhatsApp at +65 8986 1688. If you would rather watch than read, I walk through this kind of analysis on video at buyers.sg, and the calculators that handle the net-of-everything math live at listings.sg/tools.
Who should chase condo rental yield, and who should not
This is the section I wish more yield articles included, because the number is not equally relevant to everybody who reads it. Yield is one objective among several, and for some buyers it is actively the wrong one to optimise.
| Your situation | Chase yield? | Why | Where to look first |
|---|---|---|---|
| Buying purely for income; will not live in it | Yes, primary lens | Income efficiency is the whole point. Every basis point compounds over the hold. | Top of the table plus small units: D14, D25, D28, D08; 1-bed and 2-bed stock |
| Highly geared; servicing the loan from rent | Yes, but net not gross | Cash flow decides whether you sleep. Gross yield can look fine while net goes negative. | High-yield districts, then stress-test voids and rate moves at listings.sg/tools |
| Buying a home you will live in | No | You are not collecting rent. Yield is irrelevant to your actual return. | Location, commute, schools, layout, lease remaining |
| Capital growth is the goal | Secondary at most | Yield and growth often trade off. The lowest-yield districts here are low-yield because demand to own is high. | Supply pipeline, infrastructure timelines, CCR and RCR price cycles |
| Need to be able to exit fast | No, liquidity first | A high yield in a thin market is a trap on the way out. Watch the caveat counts. | Deep-volume districts: D19, D15, D10, D05, D23 |
| Buying for a child to live in later | No | Rent is a bridge, not the return. Suitability in 2036 is what matters. | Lease remaining, district permanence, unit size for a future household |
| First investment property, modest budget | Yes, with care | Yield plus liquidity plus a manageable ticket. Do not buy the thinnest market for 20 extra basis points. | D14, D19, D22, D18 — decent yield with real transaction volume |
The row I argue about most is the capital growth one. A client will show me the table, point at D11 at 2.82% and ask why anyone buys there. The answer is that yield measures how efficiently an asset converts price into income, and that is not the only thing an asset does. D11, D10 and D09 are expensive because demand to own there is persistent, deep and global. That demand is what produces price appreciation, and it is also precisely what depresses the yield. The low yield is not a flaw in those districts. It is the receipt for the thing people are actually buying.
The reverse holds at the top of the table. D25 at 4.41% is yielding well partly because demand to own in Woodlands is thinner than demand to own in Newton. That is a real income advantage and it is also a real liquidity and growth question. You are being paid 1.59 extra percentage points a year to hold an asset with a shallower buyer pool. Whether that trade is good depends entirely on your holding period and how certain you are about not needing the money.
And the liquidity row deserves more respect than it usually gets. Look at the caveat counts in the main table. D19 has 13,023 sale caveats behind its figure. D07 has 960. Both are legitimate districts. But if you need to sell inside eight weeks in a soft market, the district with thirteen thousand recent transactions is going to behave very differently from the one with under a thousand. High yield in a thin market is a position that is pleasant to hold and unpleasant to exit, and I have watched clients discover that at the worst possible moment.
Gross is not net: what comes off the top
Third time I am saying it, and that is deliberate. Every yield figure in this article is gross. Annual rent over purchase price, with nothing deducted. The number you actually keep is meaningfully lower, and the gap is not small.
Here is what comes off, in the order it usually bites.
Maintenance and sinking fund. This is the big one and it does not scale with your rent. A condo with a 50-metre pool, a gym, two tennis courts, a function room and 24-hour concierge costs real money every month to run, and your share of it is fixed whether your unit is rented or empty. On a smaller unit this can be a surprisingly large fraction of the gross rent.
Property tax. Non-owner-occupied residential property is taxed at materially higher progressive rates than owner-occupied, assessed on annual value. If you have only ever paid owner-occupier rates, the first investment-property tax bill is a shock. Budget for it before you buy, not after.
Agent commission on the lease. Paid at the start of each tenancy, so the more often your tenant turns over, the more often you pay. This is the hidden cost of the one-bedroom yield premium: smaller units churn faster, so the leasing cost recurs more frequently.
Voids. The month or two between tenants where the rent is zero and the maintenance is not. One void month a year is roughly an eight per cent haircut on gross rent on its own. Most landlords I meet have simply not modelled this.
Repairs and replacement. Aircon servicing on a schedule, then the compressor that dies in year five, the water heater, the oven, the fridge, repainting between tenants. On a furnished unit, furniture has a life of a few tenancies and then needs replacing. Treat this as an annual provision, not a surprise.
Income tax on the rental. Net rental income is taxable at your marginal rate, with deductible expenses. Depending on your bracket this can be the second-largest deduction on the list.
Add them up and the relationship between a gross figure and your actual cash return is a substantial discount, varying enormously by unit, by project, by furnishing level and by your own tax position. I deliberately will not quote you a single conversion rule, because every landlord who has ever repeated “just take 25% off” to me has been wrong about their own unit in one direction or the other. Run your own. The calculators at listings.sg/tools take the outgoings as inputs and give you the number that matters.
What gross yield is genuinely good for is ranking. Because most of those deductions apply broadly similarly across districts, a district that ranks well on gross will usually rank well on net. That is why a gross table like this one is a legitimate screening instrument. It is simply not a decision instrument.
Five mistakes I see landlords make
Seventeen years, more than 500 transactions, and the same five errors keep walking through the door.
One: quoting the rent as the investment case. I opened with this and I will close the loop on it. The correlation between a district’s rent level and its yield is +0.0485. If the first number out of your mouth about your investment property is the monthly rent, you are describing the numerator and calling it the fraction. Always ask what was paid for it.
Two: assuming cheap means high-yielding. The inverse-price rule explains about 36% of the variation, and D17 is the standing refutation — third cheapest district on the island, perfectly ordinary 3.50% yield, because its rents are discounted just as hard as its prices. You are not looking for low prices. You are looking for a price discount that is deeper than the matching rent discount.
Three: paying the ask in a wide-gap district. Median seller in Singapore is asking +8.83% above recent transacted psf. In D07, D08 and D20 it is close to 20% or above. If you buy a high-yield district at the sticker price, you have personally converted it into a mid-yield district on your own balance sheet, because your denominator is not the market’s denominator — it is whatever you agreed to pay. Yield is set on the day you buy, and almost nothing afterwards changes it.
Four: treating gross as spendable. Covered above at length. The landlords who get into trouble are almost always the geared ones who budgeted the loan against gross rent and then met maintenance, property tax at non-owner-occupied rates, a leasing commission and a void month in the same twelve months.
Five: buying the thinnest market for a few extra basis points. This is the subtlest one. Somebody shows me two districts, 4.4% against 4.1%, and wants the 4.4%. Then I look at the transaction depth and the 4.4% district has a fraction of the volume. You are taking on a real exit risk for 30 basis points of gross income — which, after the outgoings, is a rounding error. Liquidity is worth more than it looks, and it is worth exactly the most at the moment you discover you need it.
If you only take one thing from this article: your yield is decided on the day you buy, by the price you agree. The rent is the part you can barely influence. The price is the part you can. Negotiate like it matters, because arithmetically it is the only lever you hold.
Credit where it is due: the other sources worth reading
House rule on my team: name the people doing good work, especially the ones competing with us. Anybody can look authoritative by pretending no one else has covered a topic. Here is who has.
DollarsAndSense published “Guide To Condo Rental Yields In Singapore (2026 Edition)” on 22 August 2026. It runs roughly 4,500 words, covers all 27 postal districts with median rents, median prices and gross yields, and goes further than I have here in several directions: bedroom-count breakdowns, a top-15 ranking at project level, and a comparison against executive condominiums. It is good work and on project-level granularity it is straightforwardly more detailed than this article. I am not going to pretend otherwise.
The difference between their piece and mine is the basis, and it is a real difference rather than a marketing one. They use transacted leases on the rent leg. I use live asking rents. Transacted leases tell you what was signed, typically over a trailing window, which makes them accurate and backward-looking. Asking rents tell you what landlords are demanding today, which makes them forward-looking and slightly generous. In a flat market the two converge. In a moving market they diverge, and which one you want depends on whether you are measuring history or forming a view. Plus the three metrics above — the yield-to-rent correlation, the ask-to-transaction gap, and the identical-sale-psf mechanism behind the one-bedroom premium — which as far as I can find, they do not publish. An investor should read both. If you want project-level medians on a transacted basis, go and read theirs. Then come back here for the basis and the three metrics. That is not false modesty, it is just how research works.
PropertyNet has published that one-bedders win on yield. My data corroborates it independently, on a completely separate basis: 3.8759% against 2.5981% across the bedroom ladder. I want to be precise about what that is worth. I did not discover that one-bedders yield more. PropertyNet got there, and probably others did too. What I can add is the mechanism — the 0.411% sale-psf gap against the 19.426% rent-psf gap between one- and two-bedders — and independent confirmation of their number from different data. Agreeing with a rival’s finding from your own data is a stronger statement than claiming the finding. Two independent measurements landing in the same place is how a result becomes reliable. One person asserting it is just an assertion.
Stacked Homes has done project-level rental yield ranking for East Coast condos, and for that kind of work — this project versus that project, stack by stack — they are the better source. District data cannot answer a project-level question, and I would rather point you somewhere useful than stretch my own dataset past what it can carry.
What I will claim for this article is the basis, the three original metrics, the sample sizes shown openly, and the noise-floor discipline applied to my own findings. That is the moat. It is not that nobody else writes about condo rental yield.
What I would actually do next
Concrete steps, in order, if you are serious about buying for income in the next six months.
Decide whether yield is really your objective. Go back to the buyer profile table and find your row honestly. If you are buying a home, or buying for a child, or buying for growth, yield is a secondary consideration and optimising for it will cost you on the thing you actually care about.
Shortlist on both legs, not the ratio. Pick three or four districts and write down the price leg and the rent leg separately for each. You are looking for a district where the price is discounted more than the rent is. If both legs are discounted equally, you have found a far-flung district, not an opportunity.
Weight by transaction depth. Cross out any district whose figure rests on a thin sample unless you have a specific reason to accept the exit risk. The caveat and listing counts are in the main table for exactly this.
Decide your unit size deliberately. This is worth up to 0.8061 percentage points — one-bedroom against three-bedroom — which is nearly half the entire district spread, from a single choice. And remember the mechanism: the premium is in the rent leg, not the price leg, so it is durable but it comes with faster turnover and a narrower resale pool.
Check the ask gap before you make an offer. Then get the transacted comparables for the specific project and make your offer against those, not against the listing price. In D07, D08 or D20, assume the ask is a long way above the last close and behave accordingly.
Run the net. Maintenance, property tax at non-owner-occupied rates, leasing commission, a void provision, a repairs provision, income tax. Use listings.sg/tools. If the net does not work, the gross was never relevant.
Wait for 23 October 2026 if your decision hinges on OCR prices. That is when URA replaces the 3Q flash estimates with the full statistics and the first real 3Q OCR median appears. Two weeks is a cheap price for a solid number.
Then go look at the actual units. Floor, facing, stack, lease remaining, what is being built next door, how far the walk to the MRT really is in August at two in the afternoon. No table replaces that, including mine.
Gary’s bottom line
Singapore condo rental yield runs from 4.52% in D14 down to 2.82% in D11. The whole island fits inside 1.70 percentage points. There is no secret 7% district, and anybody who shows you one has left something out of the denominator.
The rent level tells you nothing — correlation of +0.0485, which is statistical silence. Price tells you something, but only about a third of the story. The other two thirds is the slack between two markets that price the same square foot differently, and that slack is where the work is.
If you want the single highest-leverage decision: choose your unit size on purpose. One-bedroom against three-bedroom is worth 0.8061 percentage points, and it comes almost entirely from the rent side, not the price side. The buying market thinks a one-bedroom square foot and a two-bedroom square foot are the same thing. The renting market pays 19.426% more for the smaller one. That mispricing is yours to own if you want it.
Then negotiate. The median Singapore seller is asking +8.83% above what is actually transacting, and in D07, D08 and D20 it is near 20%. Your yield is fixed on the day you sign, by the price you agreed. Everything after that is admin.
And please — run the net before you commit, not after. Gross yield ranks districts. Net yield pays your loan.
新加坡买房,就找对的团队 · We Serve with Heart. If you want a straight answer on your own shortlist, message me on WhatsApp.
Frequently asked questions
The seven questions I get asked most about this data.
What is a good condo rental yield in Singapore in 2026?
On this data, the median district sits at 3.71% gross and the mean at 3.6954%, so anything from roughly 3.7% upwards is above average and anything above 4.2% is genuinely strong for a condo. But that is gross, before maintenance, property tax, leasing commission, repairs, voids and income tax. Judge a specific unit against the median of its own district rather than against an island-wide benchmark, and judge your decision on net, not gross. If someone quotes you 6% or 7% gross on a Singapore condo, something is missing from the denominator — most often that the price being used is not the price you would actually pay.
Which Singapore district has the highest condo rental yield?
District 14 — Geylang, Eunos, Paya Lebar — at 4.52% gross as at 7 October 2026, on an average sale psf of $1,673.35 against an average annual asking rent of $75.67 psf. D25 Woodlands and D28 Seletar follow at 4.41% each, then D08 at 4.32%. The bottom is District 11 at 2.82%. Worth noting that D14 is a deep market, with 4,241 sale caveats and 1,178 live rental listings behind that figure, so it is not a thin-sample artefact. Also worth noting that OCR prices rose 2.2% in 3Q2026 per URA, which puts mild downward pressure on the top of this table going forward.
Is gross rental yield the same as the money I actually keep?
No, and the gap is substantial. Every figure in this article is annual rent divided by purchase price with nothing deducted. Off the top come maintenance and sinking fund contributions, property tax at non-owner-occupied rates, agent commission on each new lease, repairs and furniture replacement, void periods between tenants, and income tax on the net rental at your marginal rate. One void month a year alone is roughly an eight per cent haircut on gross rent. I deliberately will not give you a single conversion rule, because it varies far too much by project, unit size, furnishing and tax position. Run your own numbers at listings.sg/tools.
Do one-bedroom condos really yield more than larger units?
Yes, and by a wide margin. On a pure asking basis, one-bedders yield 3.8759% against 2.5981% for five-bedders — 1.4918 times — and they beat three-bedders by 0.8061 percentage points. PropertyNet has published the same conclusion and my data corroborates it independently. The surprising part is the reason. It is not that one-bedders are cheap: a one-bedder and a two-bedder have almost identical asking psf, $2,188.91 against $2,179.96, just 0.411% apart. But they differ by 19.426% on rental psf, $7.07 against $5.92 per square foot per month. The entire advantage comes from the rent side. Offset against that: faster tenant turnover, so more frequent voids and leasing fees, and a narrower resale pool.
Why do you use asking rents instead of transacted leases?
Because asking rents tell you what the market wants today rather than what it signed months ago, which is a forward-looking read and more useful in a moving market. The trade-off is honest and I state it plainly: asks are not deals, landlords settle below their asking price, so every yield level in this article is slightly generous. It works for ranking districts because the bias runs in broadly the same direction everywhere, so the order barely shifts even though the level is inflated. DollarsAndSense uses transacted leases in their 2026 guide, which is the accurate backward-looking basis, and an investor should genuinely read both pieces rather than picking one.
Does a wide ask-to-transaction gap mean a district is overpriced?
No, and I want to be careful here. The gap measures how far above recent transacted psf sellers are currently asking. The median district is at +8.83%, and D07 at +22.59%, D08 at +21.86% and D20 at +19.78% are the only three gaps wide enough to clear our 3.7% noise floor and count as real signal. Nine districts, including the D16 negative reading, sit inside ordinary unit-to-unit scatter and should not be interpreted at all. There is also a genuine limitation: the listings currently for sale are a different mix of projects from those that recently transacted, so part of any gap is composition rather than seller optimism. Treat it as a screening tool for where to negotiate hardest. It is not a valuation.
Should I buy in a high-yield district or a high-growth district?
It depends on your holding period and whether you need the rent to service the loan. The low-yield districts in this table — D11 at 2.82%, D21 at 2.93%, D10 at 3.00% — are low-yielding precisely because demand to own there is deep and persistent, and that same demand is what drives appreciation. The high-yield districts pay you more income for accepting a shallower buyer pool. If you are geared and the rent pays the mortgage, yield wins. If you can hold comfortably for ten years and want capital growth, yield is secondary. If you might need to exit inside two months, look at transaction depth before either — D19 has 13,023 recent sale caveats behind its figure and D07 has 960, and that difference matters enormously on the way out.
Disclaimer. All figures in this article are current as at 7 October 2026 and will change. Rental yield figures are gross, not net — they exclude property tax, maintenance and sinking fund contributions, agent commissions, repairs, furniture replacement, void periods, financing costs and income tax. The rent leg is drawn from live asking rents on Listings.sg and represents advertised asking prices, not achieved rents; the sale leg is drawn from URA transacted caveats. The two legs sit on different bases, as explained in the methodology section. District averages mix project vintages, tenures and unit sizes. D06 and D24 are excluded for insufficient condo rental stock. URA 3Q2026 figures quoted are flash estimates and will be revised when URA releases its full set of real estate statistics on 23 October 2026.
This article is general information and market commentary only. It is not financial advice, not investment advice, and not a recommendation to buy or sell any particular property. It does not take account of your personal financial situation, objectives or needs. Property values and rents can fall as well as rise. Please obtain independent financial, tax and legal advice before making any property decision.
Gary Lim · CEA Salesperson Registration No. R009877B · ERA Realty Network Pte Ltd · CEA Estate Agent Licence No. L3002382K. Regulated by the Council for Estate Agencies.


