Gary Lim · ERA Senior Division Director · CEA R009877B · 15 November 2026

Property tax Singapore 2026: the one number that matters

If you take one thing from this article, take this. When you stop living in your flat and start renting it out, your property tax does not go up by a bit. It goes up by roughly five to six times. On the midpoint four-room HDB flat in Singapore, property tax goes from $993 a year, after the rebate, to $5,760 a year. That is an extra $4,767 leaving your bank account every single year, and it arrives whether the flat is full or empty, whether the tenant pays on time or not, and whether the market went up or down.

I have been selling and leasing Singapore homes for over seventeen years and have closed more than five hundred deals. In that time I have sat across the table from a lot of owners who had run their rental numbers properly and still missed this one. They had the rent. They had the mortgage instalment. They had the agent’s commission and a rough allowance for repairs. What they did not have was the right property tax line, because they had taken last year’s bill — the one they paid while living in the flat — and assumed it would carry over. It does not carry over. It is a different rate table entirely, and the rebate that softened the owner-occupier bill does not follow you.

So I did the arithmetic for every HDB town that publishes a median rent, for four-room and five-room flats, and I am publishing all of it. The tables below are the part of this article you will come back to. The prose around them explains where each number comes from, what I could verify and what I could not, and what I would actually advise an owner to do with it.

One more framing point before the numbers. A small landlord renting out a flat to cover a mortgage is not a speculator and I am not going to write about them as though they were. Most of the owners I deal with are people who upgraded, kept the old place because the sums roughly worked, and now want an honest view of what the sums are. This article is written for them. There is no fear in it and there is nothing to sell. There is just a bill that is bigger than most people think, and it is better to know the size of it before you sign a tenancy than after.

What annual value actually is, in IRAS’s own words

Everything in Singapore property tax hangs off one figure: the annual value of your property. Not the price you paid. Not what the valuer says it is worth today. Not your outstanding loan. The annual value, and nothing else, feeds the rate table.

IRAS defines it in one sentence on its official answers at ask.gov.sg. I am quoting it rather than linking it, because ask.gov.sg and iras.gov.sg are not on our outbound whitelist and I would rather attribute a source by name than send you to a page I cannot vouch for from here.

“The Annual Value (AV) is the estimated annual rent of the property if it were to be rented out.”

— IRAS, on its official answers at ask.gov.sg

Read that carefully, because the implication is the whole engine of this article. The annual value of your home is an estimate of what it would rent for over a year, whether or not you rent it out. If you live in your four-room flat in Tampines and have never had a tenant, IRAS still assesses what that flat would fetch on the open market across twelve months, and taxes you on that figure. Your flat’s annual value does not change because you chose to live in it. What changes is the rate table applied to it.

That single fact answers a question I get constantly: “Gary, I’m not earning any rent, why am I paying property tax at all?” Because property tax in Singapore is a tax on the property’s letting value, not on your income from it. It is a wealth tax dressed in rental clothing. Whether that is the right design is a question for policy people. For an owner, it is simply the rule.

It also tells you why the annual value figure is harder to argue with than people expect. If IRAS’s own definition is “estimated annual rent”, and the market rents in your block are a matter of public record, then the room for disagreement is narrow by construction. I come back to that later, with a statistic from IRAS that I have not seen published anywhere else.

Why I could not pull the rate tables from IRAS, and what I used instead

I want to be straight with you about something, because this is where most property articles quietly cheat.

I could not retrieve the official rate tables from iras.gov.sg for this piece. The rate pages on that site are built client-side — the numbers are assembled in your browser after the page loads, rather than sitting in the page source — and they did not come back from our environment. I tried. I would rather tell you that than present a reconstructed table as though I had lifted it straight from the authority.

So here is what I did instead. I took two independent, dated secondary sources that publish the 2026 residential rate tables, compared them cell by cell, and only used the bands where they agreed. The two sources are SmartCalculator.sg, whose page is dated 28 April 2026, and PropKaki, whose table is labelled as being current for 2026. They agree on every single band in the owner-occupier table except one, and they agree completely on the non-owner-occupier table, which is the table that matters most for this article.

That one disagreement gets its own section further down, with my arithmetic on exactly how much it could move your bill and exactly which readers it could affect. I am not going to bury it in a footnote.

The 2026 rate tables, owner-occupier and non-owner-occupier

Singapore runs two separate progressive rate tables for residential property. One for homes you live in, one for homes you do not. Both are sliced, meaning each slice of your annual value is taxed at its own rate rather than the whole figure being taxed at the top rate you reach. That slicing is the single most misunderstood thing about the system, and it is the reason the arithmetic later in this article produces results that look wrong at first glance.

Notice what the first row does. The first $12,000 of annual value is taxed at nothing. For a flat with an annual value of $36,000, that zero-rated slice is a third of the whole figure. For a bungalow with an annual value of $180,000, it is a rounding error. That is where almost all of the strange behaviour in the later tables comes from.

Now the other table — the one that applies the moment you move out and let the whole place go.

Four bands instead of eight, and no zero-rated slice at all. The very first dollar of annual value is taxed at 12%. Put the two tables side by side and the structural gap jumps out: an owner-occupier pays nothing on the first $12,000 and then 4% up to $40,000, while a landlord pays 12% on the first $30,000 and then 20%. On the same flat, with the same annual value, in the same year.

That is not a loophole or an accident. It is deliberate policy: concessionary rates to support home ownership, full rates on everything else. But it means the step from owner-occupier to landlord is a cliff, not a slope, and the tables I built from the real rent data show just how high the cliff is.

The disputed band: two published tables disagree

Here is the one cell where my two sources part company, and here is how I am going to handle it.

Two published tables disagree, and I could not settle the question at IRAS. So rather than pick one and pretend, let me bound the damage with arithmetic, which is the useful thing to do with a contradiction.

The disputed slice is $5,000 wide — the stretch from $70,001 to $75,000. On one table that slice is taxed at 14%; on the other at 10%. The difference in rate is four percentage points, and four percent of $5,000 is $200. That is the entire exposure. The disagreement changes an annual owner-occupier bill by at most $200, and it only touches homes whose annual value sits above $70,000. At an annual value of $70,000 or below, the two tables produce identical numbers to the dollar.

An annual value above $70,000 means an estimated monthly market rent above roughly $5,833. So if your home would rent for less than about $5,800 a month, this contradiction cannot affect you at all.

Where it does bite is the upper reaches of the private scenarios later on. I flag that in the caption of that table rather than making you remember it. This is, I think, the honest way to publish a source conflict: name it, measure it, and tell the reader precisely which of your own numbers it touches and which it does not. Hiding it would have been easier and would have made the article worse.

The 2026 rebate goes only to owner-occupiers

Now to firmer ground. The rebate figures come from a primary source: the Ministry of Finance newsroom release dated 28 November 2025. mof.gov.sg is also off our whitelist, so again I quote and attribute rather than link. These are the Ministry’s own words.

A “15%” rebate for owner-occupied HDB flats.

A “10%, capped at $500” rebate for owner-occupied private residential properties.

“All one- and two-room HDB flats will continue to pay no Property Tax”.

Three-room and larger HDB owner-occupiers see average monthly increases of “$2 to $3 per month, after the PT rebate”.

“Half of private property owner-occupiers will see a PT increase of less than $6 per month, after the PT rebate”.

— Ministry of Finance newsroom release, 28 November 2025

Those are reassuring numbers if you live in your home. Two or three dollars a month for a typical HDB owner-occupier is not something anyone needs to rearrange their life around, and the Ministry is right to say so. One- and two-room flat owners continue to pay nothing at all, which is the right outcome and worth stating plainly.

But read the qualifier on every one of those lines. Owner-occupied. Owner-occupied HDB flats. Owner-occupied private residential properties. Owner-occupiers seeing increases of two to three dollars. The rebate is a concession attached to living in your own home, and a landlord does not get it. Not a reduced version of it. None of it.

This has a consequence that I have not seen stated anywhere, and it is the structural point of this whole article. Every rebate announcement is reported as relief. For a landlord it is the opposite. The rebate lowers the owner-occupier bill and leaves the non-owner-occupier bill exactly where it was. So the rebate does not narrow the gap between living in your flat and renting it out — it widens it. Every year the government announces a property tax rebate, the financial penalty for letting out your home gets slightly larger, because the comparison point moved down and you did not.

On the midpoint four-room flat the owner-occupier bill is $1,168 before the rebate and $993 after it. The landlord bill is $5,760 either way. The rebate pushed the multiple from 4.93 up to 5.80. That is the mechanism, in one line.

How the bands were widened on 1 January 2025

It is worth knowing how we got to the current owner-occupier table, because the change was genuinely significant and it is the reason the owner-occupier side of my tables looks as gentle as it does.

From the Ministry of Finance newsroom release dated 29 November 2024, verbatim: “all AV bands of the OO residential PT rates would be adjusted from 1 January 2025”, including the “revision of the first AV band from $8,000 to $12,000”, “with effect from 1 January 2025”.

That is a fifty percent widening of the zero-rated band, from $8,000 to $12,000. Every owner-occupier in the country benefited, and the owners of cheaper homes benefited most in proportional terms, because for them the zero-rated slice is a bigger share of the whole annual value. It is also, mechanically, one of the reasons the owner-occupier-to-landlord multiple is so extreme at the bottom of the market. The widened band made the owner-occupier bill on a Choa Chu Kang flat very small indeed, and dividing a fixed landlord bill by a very small number gives you a very big ratio.

The same 2024 release also sets out that year’s rebate, and comparing it to this year’s is instructive.

The 2024 release describes the HDB leg as a “one-off Property Tax (PT) rebate of 20% for Owner-Occupied (OO) HDB flats” and the private leg as “15%, capped at $1,000, for OO private residential properties in 2025”.

So the 2026 rebate is smaller than the 2025 rebate on both legs. Twenty percent down to fifteen for HDB. Fifteen percent capped at a thousand dollars down to ten percent capped at five hundred for private. That is a real, under-reported fact, and I say that carefully: both releases were covered when they came out, but I have not seen the year-on-year comparison drawn, and it matters. If you are modelling your own holding costs forward, do not assume the rebate is a fixture at any particular level. It has moved down two years running, and it is announced fresh each year rather than legislated as permanent.

Which brings me to the practical planning note. The 2027 announcement is expected around late November 2026, on the pattern of the last two. If you are deciding today whether to let out a flat from January, the owner-occupier comparison figure you are working with will move within weeks of that decision. The landlord figure will not, because there is nothing there to move.

Our method: turning median rents into an annual value proxy

This is the part of the article that is mine rather than anyone else’s, so let me set out exactly how it was built. You should be able to redo it yourself and get the same answers.

IRAS’s own definition does the heavy lifting. If the annual value is “the estimated annual rent of the property if it were to be rented out”, then the best public proxy for a town’s annual value is the town’s actual median monthly rent multiplied by twelve. That is the whole method:

Annual value proxy = the town’s 2026-Q2 median monthly rent × 12.

Then apply the published owner-occupier and non-owner-occupier bands to that figure.

The rent data is the official “Median Rent by Town and Flat Type” dataset on data.gov.sg, resource d_23000a00c52996c55106084ed0339566. That is the one external link in this entire article, because the whole analysis rests on it and you should be able to check me: Median Rent by Town and Flat Type on data.gov.sg.

Three things you need to know about that data as at today, 6 October 2026.

First, 2026-Q2 is still the latest quarter published. I checked by querying the dataset directly for 2026-Q3 and got zero records back. So if you read a competing article this month quoting third-quarter HDB median rents, be sceptical: as of today there is no third-quarter figure to quote. The next release is one of my update triggers for this piece.

Second, Bukit Timah publishes no four-room or five-room median at all, and Marine Parade publishes no five-room median. The volumes are too thin for a published median. That is why my four-room table covers 25 towns and my five-room table covers 24. I have not estimated the missing rows and I would be wary of anyone who does.

Third, and most importantly: this is a proxy, not an assessment. IRAS assesses every flat individually. Your actual annual value depends on your block, your floor, your flat’s condition, its size within the flat type, and the specific lettings the valuer is looking at. Two four-room flats in the same town can carry different annual values, and both can differ from the town median. So treat the tables below as a very good map of the terrain, not as a quotation of your personal bill. Your IRAS notice is the only document that tells you your actual annual value, and if you want your real number, that is where to look.

Every tax figure in the tables that follow is our calculation, built on the published bands and that rent-times-twelve proxy. I label it as our calculation each time, because the distinction between what a source said and what we worked out should never be blurred.

The columns run as follows. The town’s median monthly rent for the flat type. The annual value proxy. The owner-occupier tax, computed on the owner-occupier bands. The owner-occupier tax after the 15% HDB rebate. The non-owner-occupier tax, computed on the landlord bands with no rebate because there is none. Then the extra you pay for letting it out, which is simply the landlord figure minus the after-rebate owner-occupier figure. Then the multiple, which is the landlord figure divided by the after-rebate owner-occupier figure. Then the landlord tax as a share of gross annual rent. Then the same thing expressed as months of rent, because that is how every landlord I know actually thinks about costs.

Property tax Singapore by town: the 4-room HDB table

Here is the four-room table, 25 towns, sorted by median rent from highest to lowest. This is the single most useful thing on this page.

Run your eye down the “extra for renting out” column, because that is the real decision number. It never drops below $3,984 and it tops out at $7,552. Even in the cheapest towns on the list — Bukit Panjang and Choa Chu Kang, both at a $3,000 median — letting out your four-room flat adds just under four thousand dollars a year to your property tax. In Central it adds over seven and a half thousand.

Then look at the months-of-rent column, which is the same information translated into landlord language. Across all 25 towns, property tax alone eats between 1.60 and 2.06 months of your annual rent on a four-room flat. Call it one and a half to two months. That is before your agent, before repairs, before income tax on the rental income, and before a single vacant week between tenants.

I want to make the void point properly, because it is where owners get hurt. The property tax bill does not pause when the flat is empty. If you have one month vacant between tenancies — entirely normal, and in a soft leasing market two months is not unusual — then you are paying the full year’s tax out of eleven months of rent, and the share of rent figure in that column goes up accordingly. A Woodlands four-room at a $3,100 median shows 13.55% of gross rent on a full year. Lose a month and it is closer to 14.8%. Lose two and you are at 16.3%.

And notice something about the multiple column which I will unpack properly in a moment: the highest multiples — 5.88x — are not in Central. They are in Bukit Panjang, Choa Chu Kang, Woodlands, Yishun, Sembawang, Hougang, Sengkang, Bukit Batok and Punggol. The lowest multiple on the whole table, 4.97x, belongs to Central, the most expensive town on it. That inversion is real and it has a clean explanation.

Property tax Singapore by town: the 5-room HDB table

Same method, same quarter, 24 towns. Five-room flats carry higher rents and therefore higher annual values, which pushes more of the annual value into the upper landlord bands.

The five-room table pushes the ceiling higher on both ends. Central hits $11,232 a year in landlord property tax against $2,414 if you lived there, an extra $8,818 a year — the largest single figure anywhere in my HDB data. And at 18.35% of gross rent, or 2.20 months, Central five-room is the heaviest tax burden relative to rent in the whole HDB universe I measured.

The reason is the 36% band. Central’s five-room annual value proxy of $61,200 is the only HDB figure in either table that crosses $60,000, which means a slice of it — $1,200 of it — gets taxed at the top landlord rate of 36%. Every other town’s annual value stops short of that threshold. It is a small slice and it adds only a few hundred dollars, but it is the reason Central stands slightly apart from the pattern.

Something else worth noting on the five-room table: the clustering at the bottom. Bukit Panjang, Punggol, Sembawang, Sengkang and Woodlands all show a $3,300 median, which produces an identical $5,520 landlord bill and an identical 5.88x multiple across five different towns. If you are choosing between those towns on tax grounds, there is nothing to choose. The decision has to be made on rent achievability, tenant demand and how long the flat sits empty — none of which the tax table can tell you.

And compare like with like across the two tables. In Choa Chu Kang, a four-room commands a $3,000 median and a five-room $3,200. The four-room landlord pays $4,800 in tax and the five-room $5,280 — $480 more for $2,400 more in annual rent. The five-room keeps more of the extra. At the other end, Central four-room to five-room is $500 more rent a month, $6,000 more a year, and $1,776 more tax. Nearly thirty cents of every extra rental dollar goes to property tax at that level. The bigger flat is still better in absolute terms, but the marginal rate on the step up is steep.

The midpoint flat: what a typical HDB landlord pays

When people ask me for one number, this is the one I give. I want to be precise about what it is, though, because the label matters. It is the median of the town medians — the middle town in the list above — not a national median rent. A true national median would weight each town by how many flats are actually let there, and that weighting is not in the published data. So read this as “a typical town”, not “the average Singapore flat”.

So: a typical four-room HDB flat costs you $993 a year in property tax while you live in it, and $5,760 a year the moment you let the whole thing out. A typical five-room goes from $1,161 to $6,420.

Put that against the rent. On the four-room, $5,760 of tax against $40,800 of gross annual rent is a touch over fourteen percent of the money the tenant hands you, gone before you have paid anyone to find them, fixed anything they broke, or declared a cent of it as income. On the five-room it is $6,420 against $44,100, which works out about the same.

Here is how I would frame that conversation with an owner who is weighing up whether to let out the old flat after upgrading. The rent is $3,400 a month, so $40,800 a year. Take off $5,760 in property tax. Take off roughly half a month to a month of rent for the leasing commission, depending on the tenancy length. Allow something honest for repairs, aircon servicing and the replacement of whatever gives up during the term — most owners I work with land somewhere between one and two thousand a year on a flat in reasonable condition. Allow for a vacant month between tenants, because you will have one sooner or later. And then remember that the net figure after all of that is rental income, and rental income is taxable at your personal rate on top.

None of this means letting out your flat is a bad idea. For a lot of owners it is clearly the right move, and I say which ones further down. It means the gap between the gross rent and what reaches your pocket is wider than the headline figure suggests, and property tax is the single largest line in that gap. If you are going to be surprised by one number, be surprised by it now rather than when the January notice arrives.

The counter-intuitive finding: the multiple is biggest in the cheapest towns

This is my favourite finding in the whole exercise, because it runs directly against intuition and because once you see the mechanism it is obvious.

You would expect the tax penalty for letting out your home to be harshest at the top of the market. It is not — not as a multiple. The multiple is biggest in the cheapest towns. Choa Chu Kang, Bukit Panjang, Woodlands and Yishun all sit at 5.88x on a four-room flat. Central, the most expensive town in the dataset, sits at 4.97x, the lowest multiple on the table.

The reason is the shape of the owner-occupier table. The first $12,000 of annual value is taxed at zero and the next $28,000 at only 4%. A Choa Chu Kang four-room with an annual value proxy of $36,000 sits almost entirely inside those two gentle slices — its owner-occupier bill is $960, and $816 after the rebate. That is sixty-eight dollars a month to own and live in your own home. It is a genuinely light tax.

Now divide a landlord bill of $4,800 by that very small $816. You get 5.88. The multiple explodes not because the landlord bill is large but because the owner-occupier bill it is being compared against is tiny. The denominator, not the numerator, is doing the work.

Central’s annual value proxy of $55,200 reaches up into the 6% and 10% owner-occupier slices, so its owner-occupier bill is a much more substantial $1,904 after the rebate. Dividing $9,456 by that gives you a smaller ratio, even though Central’s landlord in absolute terms is paying nearly twice as much tax as Choa Chu Kang’s.

Which brings the important correction: in actual dollars the order reverses completely. Central pays the most property tax of any town on the four-room table, $9,456 a year. Choa Chu Kang pays the least, $4,800. If you are writing a cheque, Central is nearly twice as expensive. The multiple is a statement about the shape of the two rate tables; the dollar figure is a statement about your bank balance. Do not confuse them.

I labour this because I have watched the multiple get misused. An owner in a mature estate sees “5.88x” and concludes that letting out a cheap flat is punitive, when in fact that owner has the lowest absolute tax bill in the country on a let flat and the lowest share of gross rent, 13.33%. Meanwhile the Central owner sees “4.97x”, feels comparatively fine about it, and is handing over 17.13% of the rent — the heaviest burden on the four-room table. The ratio and the burden point in opposite directions, and the burden is the one that pays for things.

Up the market, the multiple falls and the burden rises

I ran the same arithmetic on private property, as a set of clean annual value scenarios rather than real projects, because private rents vary far too much within a single development to make a median useful. Each row is a round monthly rent turned into an annual value the same way: rent times twelve. Then the owner-occupier bands, then the 10%-capped-at-$500 private rebate, then the landlord bands with no rebate.

Look at the two right-hand halves of that table pulling in opposite directions. The multiple falls from 5.56x to 1.71x as you climb. The share of gross rent rises from 13.33% to 30.00%. Same nine rows, same arithmetic, two conclusions that feel like they belong to different articles.

At the bottom of the table, a $3,000-a-month private unit behaves almost exactly like a four-room HDB flat at the same rent — $4,800 in landlord tax, 1.60 months of rent, a 5.56x multiple. The flat type is irrelevant; only the annual value matters. At the top, a $15,000-a-month house hands over $54,000 a year in property tax, which is 3.6 months of rent, every year, and nearly a third of every dollar the tenant pays.

The honest summary of that table is this: the ratio comforts the expensive landlord and the burden crushes him. An owner letting out a large detached home can look at 1.71x and conclude the tax system treats them mildly compared to the HDB owner at 5.88x. In reality they are handing over thirty percent of their gross rent while the HDB owner hands over thirteen. Whichever lens you pick tells you something true, and picking only one is how people end up with the wrong answer.

For anyone choosing between holding a private unit to let and other uses of the capital, the rows above $60,000 of annual value are the ones to read twice. Once more than a fifth of gross rent is going to property tax before any other cost, the yield maths on a rental hold changes character. That is a different calculation from the one most buyers run at purchase, and it is worth running alongside the usual comparisons — I have written separately about how resale condos compare with new launches in 2026, and holding cost belongs in that comparison just as much as price per square foot does.

Where the $500 private rebate cap starts to bite

The private rebate has two parts and only one of them is generous. It is 10% of your owner-occupier bill, capped at $500. So there is a crossover point: below it you get the full ten percent, above it you get five hundred dollars and nothing more however large your bill.

Our arithmetic puts that crossover at an annual value of about $79,200, which corresponds to a market rent of about $6,600 a month. That is where 10% of the owner-occupier bill first passes $500. Above that figure the cap, not the percentage, is what you are actually receiving.

You can see it operating in the table above. At an annual value of $72,000 the owner-occupier bill is $4,000 and the rebate is $400 — a true ten percent. At $84,000 the bill is $5,680, ten percent of which would be $568, so the cap trims it to $500. By the time you reach $180,000 of annual value the bill is $32,020 and the rebate is still $500, which is one and a half percent. The rebate has effectively stopped existing as a percentage and become a token.

Two practical consequences. First, if you own an owner-occupied private home with an annual value comfortably under $79,200, the rebate is a real if modest saving and it scales with your bill. Second, if you own one above that, treat the rebate as a flat five hundred dollars in your planning and stop thinking of it as a percentage, because it is not behaving like one.

And note once more what the cap does to the landlord comparison. The cap limits a concession that landlords do not receive at all. So it narrows the owner-occupier advantage at the top of the market — which is part of why the multiple falls as you climb the private table — but it does nothing whatsoever for the person letting the property out. For a landlord the entire rebate discussion, cap and all, is a discussion about somebody else’s bill.

If you want to sanity-check your own position against live market numbers rather than medians, the tools on Listings.sg will give you a current read on what comparable units are actually asking in your project or your town, which is the input this whole calculation turns on.

Renting out a room is not the same as renting out the flat

This is the question I get asked most often and the one where the stakes are highest relative to how simple the answer is. It matters to a lot of readers, so let me state it as cleanly as I can.

If you still live in the flat and rent out a room, you remain an owner-occupier. The concessionary owner-occupier rates continue to apply and the rebate continues to apply. Your property tax does not jump.

If you rent out the whole flat and move out, you become a non-owner-occupier. The landlord rate table applies and the rebate is gone. Your property tax multiplies by roughly five to six times on the figures in this article.

That is the entire distinction, and it is binary. There is no partial rate for an owner who lets out two of three bedrooms. The test is whether the property is your owner-occupied home, not what fraction of the floor area a tenant uses.

Which means room rental is, on the property tax side alone, dramatically more efficient than whole-flat rental for an owner who can live with it. Take the midpoint four-room flat. Let out the whole thing and you collect $3,400 a month and pay $5,760 in tax. Stay put, let out a room, collect rather less — and pay $993. The income is smaller; the tax is the same as if you had no tenant at all. For an owner whose main objective is covering a mortgage instalment rather than maximising gross income, that trade is worth modelling properly before defaulting to a whole-flat tenancy.

I will add the obvious caution, and I will add it without theatrics. HDB has its own rules on subletting rooms and on subletting whole flats — eligibility, minimum occupation requirements, tenant quotas, registration, maximum occupancy — and those rules are entirely separate from the tax question I am answering here. Property tax treatment is not permission. Get the HDB side right first, because that is where the serious consequences live, and the tax saving on an unapproved arrangement is not a saving at all. We have a companion piece on HDB illegal subletting coming on buycondo.sg — it is not live yet, so I am naming it in plain text rather than linking you to a page that does not exist.

Objecting to your annual value is a 1-in-50 shot

Every year, some owners look at their annual value, decide it is too high, and file an objection. I understand the instinct completely. I also think almost nobody who files one knows the odds, because the odds are published and nobody repeats them.

IRAS does publish them, in its official answers at ask.gov.sg, in response to a question about objecting to your annual value. These are its words.

“98% of residential AV objections are unsuccessful with no amendments made to the original assessment.”

“The AV for the vast majority of residential objections filed are found to be in line with market rents.”

“IRAS adhere to robust valuation practices to ensure a fair and accurate assessment.”

— IRAS, on its official answers at ask.gov.sg

Ninety-eight percent unsuccessful, with no amendment at all to the original assessment. That is a one-in-fifty shot. I have not seen that number in a single Singapore property article, and it is the most actionable fact in this entire piece.

It gets sharper. IRAS also lists the grounds it does not accept for an objection. Here they are, as rejected grounds:

• “High tax rates”
• “Financial difficulties”
• “You are not receiving any rental income as the property is either vacant or owner-occupied”
• “Issues relating to the property tax apportionment with developers for newly completed properties”

Read that list against the reasons owners actually want to object. In seventeen years I have heard three complaints about property tax, over and over. The rates are too high. I cannot afford this right now. I am not even earning any rent from the place. Those are the first three items on the list of grounds IRAS explicitly will not hear.

That is not a trap and it is not unfair. It follows directly from what annual value is. If the annual value is the estimated annual rent of the property, then the tax rate applied to it is a separate question from whether the estimate is right, your personal cash position has nothing to do with the property’s letting value, and whether you have chosen to collect rent is irrelevant to what the property would fetch. The definition does the work. Every one of those three objections is an argument about the tax, not about the valuation — and an objection is a valuation process.

So here is the genuinely useful advice, which almost nobody gives. An objection only has a chance if you can show that the annual value is out of line with actual market rents for comparable properties. That is the one thing the process is designed to consider. If you have evidence — real lettings in your block or a closely comparable one, at rents meaningfully below the annual value IRAS has assigned you, for units of similar size, floor and condition — you have something to say and you are in the two percent’s territory. If your case is “this feels too expensive”, you are in the ninety-eight.

Before you file anything, do the comparable-rent work. Pull the actual lettings, not the asking prices. Match the flat type and the size properly. If the evidence does not clearly support you, save yourself the effort. And if you are building a valuation case, the same discipline applies as when you are establishing a sale value — I have written about that process in detail in my guide to HDB valuation in 2026, and the evidential standard is the same: real transactions, properly matched, nothing else.

Who should rent out, and who should not

None of the arithmetic above tells you what to do. It tells you the size of one cost. So here is my actual read, after seventeen years of having this conversation across a lot of kitchen tables. I have tried to be useful rather than diplomatic.

That last row deserves a sentence more, because I see it often. A flat held out of reluctance rather than strategy is usually the worst-performing asset in a household, and the annual property tax bill is the part of it that quietly compounds the problem. If you are at or past your minimum occupation period and genuinely unsure whether to hold or sell, work through the sale side properly rather than drifting into a tenancy by default — my guide to selling an HDB flat after MOP walks through that decision with the numbers laid out.

And one point of fairness, since this article is about a tax on landlords. Letting out a flat to cover a mortgage is a perfectly reasonable thing for a family to do, and the people doing it are not the cause of anything. Equally, the tenant in that flat is paying rent that has to cover this bill, and both sides deserve to be dealt with straight. The clearer the landlord is about their own costs at the start, the fewer mid-tenancy rent arguments there are, which serves the tenant too.

My verdict

Frequently asked questions

How much is property tax in Singapore if I rent out my HDB flat?

On our calculation, between $4,800 and $11,232 a year depending on your town and flat type. A four-room flat in Bukit Panjang or Choa Chu Kang comes out at $4,800; a four-room in Central at $9,456; a five-room in Central at $11,232. The midpoint four-room is $5,760 and the midpoint five-room is $6,420. Those figures use each town’s 2026-Q2 median rent multiplied by twelve as an annual value proxy and apply the published non-owner-occupier bands. Your actual bill depends on the annual value IRAS has assessed for your specific flat, which is on your notice.

Does renting out a single room increase my property tax?

No. If you are still living in the flat, you remain an owner-occupier, and the concessionary owner-occupier rates and the 15% rebate both continue to apply. Only letting out the whole flat and moving out shifts you to the non-owner-occupier rates and loses you the rebate. On the midpoint four-room that is the difference between $993 and $5,760 a year. Do check HDB’s own subletting rules separately, though — tax treatment is not the same thing as permission.

Do landlords get the 2026 property tax rebate?

No. The Ministry of Finance release of 28 November 2025 attaches the rebate to owner-occupied properties only — 15% for owner-occupied HDB flats and 10% capped at $500 for owner-occupied private residential. A landlord gets nothing, which means each rebate widens rather than narrows the gap between living in your home and letting it out. Worth adding that the 2026 rebate is smaller than 2025’s on both legs: 20% down to 15% for HDB, and 15% capped at $1,000 down to 10% capped at $500 for private.

Can I object to my annual value, and will it work?

You can, but the odds are poor. IRAS states on its official answers at ask.gov.sg that “98% of residential AV objections are unsuccessful with no amendments made to the original assessment” and that annual values for the vast majority of objections filed are found to be in line with market rents. It also lists grounds it will not accept, including high tax rates, financial difficulties and the property being vacant or owner-occupied. An objection has a real chance only if you can show the annual value is out of line with actual market rents for genuinely comparable properties. Bring lettings evidence or do not bother.

Why am I paying property tax when my flat is sitting empty?

Because property tax is charged on annual value, which IRAS defines as “the estimated annual rent of the property if it were to be rented out”. The word is if. Your flat has a letting value whether or not you choose to collect it, and that letting value is what is taxed. IRAS specifically lists not receiving rental income because the property is vacant or owner-occupied as a ground it does not accept for an objection. The upside, if there is one: while the flat is genuinely your owner-occupied home you are on the much gentler owner-occupier rates and you get the rebate.

Which HDB town costs a landlord the most in property tax?

Central, on both flat types, and by a clear margin. A let four-room in Central comes out at $9,456 a year on our calculation and a let five-room at $11,232, which is the largest figure anywhere in our HDB data. Central five-room is also the heaviest relative burden at 18.35% of gross rent, or 2.20 months of rent a year, partly because its $61,200 annual value proxy is the only HDB figure that reaches the top 36% landlord band. The cheapest are Bukit Panjang and Choa Chu Kang four-room flats at $4,800.

Is there 2026-Q3 HDB rent data I should be using instead?

Not as at 6 October 2026. We queried the official “Median Rent by Town and Flat Type” dataset on data.gov.sg directly for 2026-Q3 and got zero records back, so 2026-Q2 remains the latest published quarter. If you see a third-quarter median quoted this month, treat it with suspicion. Note also that Bukit Timah publishes no four-room or five-room median and Marine Parade publishes no five-room median, which is why our tables cover 25 and 24 towns rather than the full set.