Property Questions for Condo

Ask Gary Anything

Get Instant Answers 

Property FAQs

Ask Your Question

Freehold vs Leasehold Condo: The Endgame Maths

Freehold vs Leasehold Condo in Singapore: What Tenure Is Actually Worth at the Endgame

First, credit where it is due

If you searched this question, you have probably already read Stacked Homes. You should. They have published two genuine data studies on tenure in Singapore, and one of them found something that still annoys people at dinner parties: over a ten-year holding period, leasehold condos can deliver better returns than freehold ones. That is real work, done on real transaction data, and it got there before this article did.

I am not going to pretend otherwise. There is a style of Singapore property content that arrives late to a well-served keyword and writes the same article again with slightly worse numbers. I have no interest in that. If what you want is the transaction-data return comparison between freehold and leasehold condos over a decade, Stacked Homes is the right place to start and I would send a client there without hesitation.

So what is left to say? Quite a lot, as it turns out. Because the ten-year return study answers the question “how have the two tenures performed in the recent past”, and that is not the question that keeps people awake. The question that keeps people awake is the one at the far end of the lease. What happens when the clock gets short. What the tenure difference is genuinely worth when somebody finally has to settle up. And that question has an answer nobody publishes, because the answer is buried in a seventy-five-year-old valuation table and in the one transaction type where a party actually writes a cheque specifically for tenure.

That is the article. 新加坡买房,就找对的团队 — and part of finding the right team is a team that tells you where somebody else did the work better.

What this article adds to the freehold vs leasehold condo question

At a glance — the six numbers that carry this piece.

  • 96.0% — a fresh 99-year lease is worth this fraction of freehold under Bala’s Table. Not 100%, and not far off it either.
  • 0.2811% per year — the compounded cost of tenure in the first nineteen years of a 99-year lease. Our arithmetic from Bala’s Table.
  • 10.6335% per year — the compounded cost of tenure in the last five years before expiry. Same table, same arithmetic.
  • 37.82x — the ratio between those two. The last five years of a lease destroy value nearly thirty-eight times faster per year than the first nineteen. Our figure.
  • $111,813,412 — the lease top-up premium we back out of the Lakeside Towers collective sale. Our inference, reconciled to the reported land rate.
  • 92.7646% — the share of implied freehold value the developer paid for restored 99-year land at that site, against Bala’s 96.0%. Our arithmetic, one site, caveats attached.

Two of those are original contributions and I want to be precise about what they are. The first is the shape of lease decay, stage by stage, expressed as an annual cost. Everyone knows a lease declines in value. Almost nobody knows that the decline is so wildly unevenly distributed across the life of the lease that treating it as a single annual figure is close to useless. The second is a reverse-engineering exercise: taking a real collective sale, backing out the lease top-up premium the developer must have paid, and asking what ratio of freehold to leasehold value that cheque implies. Then comparing it to the table. They do not agree, and the gap is interesting.

Neither contribution is a transaction-price study. We have not measured the freehold price premium in resale data in this piece. I say that again later in its own section because the boundary matters.

Bala’s Table: the numbers the whole market quietly runs on

When a valuer in Singapore needs to know what a leasehold interest is worth as a fraction of the freehold interest in the same land, there is one table that comes out of the drawer. It is called Bala’s Table. It has been in use for roughly three quarters of a century. It was published online by the Singapore Land Authority from 31 July 2000. And nobody knows who wrote it.

I find that genuinely remarkable. Billions of dollars of Singapore land value get apportioned every year using a reference table whose authorship is folklore. The Centre for Liveable Cities commentary by Kwek Sian Choo and Dionne Hoh, published in 2017, puts it about as plainly as an official-adjacent document can: “It is widely believed that the table was prepared by a Land Office employee by the name of Bala.” Believed. Not recorded.

The same commentary is equally candid about the origins. “The exact time and basis for the adoption of Bala’s Table are not known,” it says, with the evidence pointing to somewhere around 1948. And then, in a line I quote to clients regularly because it is more honest than most market commentary manages:

“The Table, and its choice of implicit discount rate, is not an exact science. But it has served… for decades.”

That is the correct posture. The table is a convention. Conventions are enormously valuable precisely because everyone uses the same one, which is why the table survives despite nobody being able to derive it from first principles. But a convention is not a law of physics, and when you see somebody quoting a Bala figure to four decimal places as though it were a measurement, be a little suspicious.

Here are the figures. I have cross-checked them against two sources that agree: the Centre for Liveable Cities 2017 commentary, Appendix 1, and the figures published by Lee Sze Teck, Senior Director of Data Analytics and Research at Huttons Asia, in EdgeProp on 12 June 2025. They agreed at all three marks I tested.

Bala’s Table: leasehold value as a percentage of freehold value, by remaining lease. Sourced figures from the Centre for Liveable Cities commentary (2017, Appendix 1), cross-checked against figures published by Lee Sze Teck of Huttons Asia in EdgeProp on 12 June 2025. Not our arithmetic — these are the published marks.
Remaining lease % of freehold Remaining lease % of freehold
99 years 96.0% 50 years 74.7%
95 years 95.6% 45 years 71.8%
90 years 94.6% 40 years 68.5%
85 years 92.9% 35 years 64.6%
80 years 91.0% 30 years 60.0%
75 years 88.5% 25 years 54.6%
70 years 86.0% 20 years 48.0%
65 years 83.0% 15 years 40.0%
60 years 80.0% 10 years 30.0%
55 years 77.3% 5 years 17.1%

Read the first row again. A brand-new 99-year lease is worth 96.0% of freehold. Ninety-six per cent. Hold that number in your head the next time somebody tells you freehold is a fundamentally different asset class from a new leasehold launch. Under the convention the entire professional valuation industry uses, the difference at the top of the lease is four percentage points.

Now read the last row. Five years left, 17.1%. The same land, the same building, the same address — worth about a sixth of freehold. Somewhere between those two rows is the whole story of freehold vs leasehold condo ownership, and the interesting part is not that the number falls. It is how it falls.

Our ninth self-correction, and why it matters here

Before I build anything on top of that table, a confession that belongs here rather than buried in a footnote.

This site has already published a correction to its own earlier version of these numbers. I wrote a piece some time back that quoted Bala’s Table from memory and from secondary sources, and six figures went into it. Four of the six were wrong. Worse — and this is the part that taught me something — all four were wrong in the same direction. Every single error made leasehold look more pessimistic than the published table actually says.

That is not random noise. That is bias. Somewhere in my own head, and I suspect in the head of a lot of people in this industry, there is a prior that says leasehold decay must be brutal and early. So when recall was fuzzy, recall rounded in the direction of the prior. Four for four in the same direction is not a coincidence; it is a fingerprint.

We published the correction. It was this site’s ninth published self-correction, and I keep the count public on purpose. If you are going to read an article that builds a 37.82x finding and a nine-figure reverse-engineering on top of a table, you are entitled to know whether the author has ever got that table wrong before and whether he told you when he did. I did, and I did.

Which is also why the table above carries two independent sources rather than one. I am not quoting Bala’s Table from memory again.

The cost of a year of tenure, by stage

Here is the first of our two original contributions, and it is the one I would tattoo on the inside of every buyer’s eyelids if the regulations permitted it.

People talk about lease decay as if it were a depreciation rate. A flat percentage a year, like a car. It is nothing like that. Using the industry’s own table, a year of tenure costs roughly a quarter of a per cent a year during the first two decades of a 99-year lease, and more than ten per cent a year during the last five. Same table. Same asset. Thirty-eight-fold difference in the annual cost of time.

I worked it out stage by stage. Every figure in the table below is our arithmetic, derived from the published Bala marks above. The “per year, compounded” column is the geometric annual rate implied by the total change over that span — in other words, the rate at which value is actually compounding downward. The “points per year” column is the simpler arithmetic average of percentage points lost per year, included because some readers find it more intuitive even though it is the less correct way to think about a rate.

The cost of a year of tenure, by stage of lease. All figures our arithmetic, derived from the published Bala’s Table marks in the preceding table. The compounded column is the geometric annual rate of value decline over each span; the points column is the simple arithmetic average of percentage points lost per year.
Lease falls from Over Value goes Total change Per year, compounded Points per year
99 to 80 years 19 yrs 96.0% to 91.0% −5.21% −0.2811%/yr −0.2632
80 to 60 years 20 yrs 91.0% to 80.0% −12.09% −0.6421%/yr −0.5500
60 to 40 years 20 yrs 80.0% to 68.5% −14.38% −0.7730%/yr −0.5750
40 to 20 years 20 yrs 68.5% to 48.0% −29.93% −1.7624%/yr −1.0250
20 to 10 years 10 yrs 48.0% to 30.0% −37.50% −4.5913%/yr −1.8000
10 to 5 years 5 yrs 30.0% to 17.1% −43.00% −10.6335%/yr −2.5800

Look at the compounded column from top to bottom. 0.28, 0.64, 0.77, 1.76, 4.59, 10.63. It is not a line. It is a hockey stick lying on its back, and the blade is the last fifteen years.

Now consider what that means for an ordinary buyer, because the practical conclusion is not the one people expect.

Take a buyer looking at a new launch on a fresh 99-year lease, who expects to hold fifteen to twenty years and then sell. Under this table, the tenure cost over that entire holding period is in the region of five per cent of value — roughly 0.28% a year compounding. Five per cent. Over twenty years. Now ask what premium that buyer would have to pay today to buy the freehold equivalent in the same location with the same specification. In most of Singapore, that premium is comfortably into double digits, and in some submarkets it is far more than that. The tenure cost over the intended holding period is a rounding error against the price gap. For that buyer, paying up materially for freehold is not risk management. It is prepaying for a risk that will not arrive during their ownership.

Now take a different buyer entirely. Somebody looking at a leasehold condo or flat with around forty years remaining, attracted by the price, planning to hold twenty years. That buyer is sitting in the 40-to-20 row. Their tenure cost is 1.76% a year compounding, nearly thirty per cent cumulative, and it accelerates the whole way. For that buyer, tenure is not a rounding error. It is the dominant term in the return calculation. It will quietly eat a very large share of whatever capital appreciation the market hands them, and if the market hands them nothing, it will eat into their capital directly.

So the question is never “freehold or leasehold”. That framing is a category error and it has cost Singaporean buyers a fortune in both directions — people overpaying wildly for freehold they will own for twelve years, and people buying into the bottom of the curve because the headline price looked like a bargain.

The question is: where on this curve am I buying, and where on it do I expect to sell? Two points on a curve. That is the entire analysis. Everything else is decoration.

The 37.82x finding and how to use it

Divide the last row’s compounded rate by the first row’s. 10.6335 divided by 0.2811. You get 37.82.

The last five years of a 99-year lease destroy value 37.82 times faster per year than the first nineteen. That is our arithmetic, derived entirely from the published table, and as far as I can find, nobody has put that ratio in print before. I would like it to become the number people reach for when this argument starts.

Here is why it is useful rather than merely striking. It reframes tenure from a binary into a position. When a client asks me whether leasehold is risky, the honest answer is: the first two decades of a 99-year lease are about as close to tenure-risk-free as leasehold gets, and the last fifteen are where essentially all of the risk lives. A single asset contains both. Whether you encounter the safe part or the dangerous part depends entirely on which stretch of the lease you happen to own, and that is a choice you make at purchase.

Let me put the practical bands plainly.

How I read each stretch of a lease when advising a buyer. Zones and commentary are my own judgement; the annual tenure cost column is our arithmetic from the preceding table.
Remaining lease Annual tenure cost How I treat it
99 to 80 years about 0.28%/yr Tenure is a near-irrelevance over a normal holding period. Judge the asset on location, layout, rental demand and price. Do not pay a large freehold premium to escape a cost this small.
80 to 60 years about 0.64%/yr Still modest, but now a real line item. Worth a sanity check against the freehold premium in the same micro-market.
60 to 40 years about 0.77%/yr The quiet middle. Tenure cost is manageable, but financing and CPF constraints start shaping your eventual buyer pool. Think about exit liquidity, not just exit price.
40 to 20 years about 1.76%/yr Tenure becomes the dominant term. I want a clear thesis here — redevelopment potential, rental yield that pays for the decay, or a genuinely temporary holding need.
20 to 10 years about 4.59%/yr This is a specialist asset, not a home purchase with upside. Buyers here should be solving a specific problem and know exactly what they are solving.
Under 10 years about 10.63%/yr Effectively a consumption decision with a terminal date. Nothing wrong with that, provided nobody is calling it an investment.

One more observation on that curve, because it explains something people find mysterious. Why does the fall accelerate so violently at the end? It is not because the bricks deteriorate faster. It is because the buyer pool collapses. Financing tenure shortens, CPF usability tightens, banks get conservative, and the set of people who can legally and practically bid for the asset shrinks every year. Value is set at the margin by whoever can bid. Remove the bidders and the price falls faster than any discounting formula would suggest. I come back to that in the CPF section, because it is the most underrated mechanism in this entire subject.

The only real cheque anyone writes for tenure

Here is the problem with every freehold vs leasehold condo study, including the good ones. When you compare transaction prices of freehold and leasehold condos, you are never comparing tenure. You are comparing tenure tangled up with location, age, developer, layout efficiency, facilities, en bloc expectations, and the simple fact that freehold land in Singapore is concentrated in different districts than leasehold land. The statistical work to untangle that is hard, and the honest researchers say so.

But there is one transaction type in Singapore where somebody writes a cheque specifically, explicitly and solely for tenure. Nothing else. Just years.

When a developer buys a leasehold site through a collective sale, the lease on that site is partly run down. Before the developer can launch a new 99-year project on it, the lease has to be restored to a fresh 99 years. That restoration is not free. The developer pays the State a lease upgrading premium — also called the lease top-up premium — and that payment buys exactly one thing: years of tenure on land the developer already controls. Same location. Same plot. Same plot ratio. Same everything. The only variable being purchased is time.

That makes it the cleanest price signal for tenure that exists in this market. If you want to know what the gap between leasehold and freehold is genuinely worth, you look at what somebody pays to close part of that gap.

And now the inconvenient part, which I want to state as plainly as I can: the lease top-up premium is assessed by the Chief Valuer and the method is not published. There is no public formula. There is no published schedule. Market participants widely assume Bala’s Table sits somewhere in the assessment, and that assumption is reasonable, but it is an assumption. The Chief Valuer does not publish the working.

So we cannot read the price of tenure off a document. We have to infer it. Which is what the next section does, with all the caveats that implies.

Lakeside Towers: reverse-engineering the lease premium

The worked case is Lakeside Towers, in the Jurong West planning area. Here are the site facts.

Lakeside Towers collective sale: reported site and transaction facts as published in market reporting. These are sourced figures, not our arithmetic. The over-built plot ratio pair was published by JLL.
Item Detail
Development Lakeside Towers, Jurong West planning area
Units 144 units across two 20-storey blocks
Tenure Leasehold from 1975, roughly 48 years remaining
Site area 153,237 sq ft
Consent level Above 80%
Tender closed 1 October 2026
Reported sale price $350,000,000
Reported land rate $1,277 per sq ft per plot ratio, inclusive of the lease top-up premium and bonus balcony gross floor area, with no land betterment charge payable
As-built gross plot ratio 2.36, against a base of 2.1 — published by JLL as an explicit number pair

Two things about that table are unusual and both of them are why I chose this site.

First, the reported land rate is stated as inclusive of the lease top-up premium. That is the crucial phrase. The $1,277 psf ppr is not what the owners received divided by the floor area. It is the developer’s all-in land cost, including the cheque to the State for tenure. Which means the premium is hiding inside a published number, and if you know the floor area you can pull it out.

Second, the site is over-built, and JLL published that as an explicit number pair rather than leaving it to inference: as-built gross plot ratio of 2.36 against a base of 2.1. In my experience reading collective sale reporting, that is the only site where the over-built position has been stated as two numbers rather than hinted at. It matters here for two reasons — it tells us why no land betterment charge is payable, and it gives us the baseline gross floor area to divide by.

The baseline GFA follows directly. Site area times the as-built plot ratio: 153,237 × 2.36 = 361,639.3 sq ft. That is our arithmetic, and it is the denominator for everything below.

Now the reverse-engineering. Every step here is our arithmetic and every step is reversible by you with a calculator. I have deliberately kept it to simple operations so you can check me.

Reverse-engineering the Lakeside Towers lease top-up premium. Every figure in this table is our arithmetic, built from the reported sale price, the reported land rate and the JLL-published plot ratio pair. It is an inference, not a published valuation.
Step Figure Working
Baseline GFA 361,639.3 sq ft 153,237 sq ft site area × 2.36 as-built gross plot ratio
Implied lease top-up premium $111,813,412 Backed out so that total land cost divided by baseline GFA equals the reported rate
Reconciliation check $1,277.00 psf ppr ($350,000,000 + $111,813,412) ÷ 361,639.3 sq ft — against the reported $1,277. It reconciles.
Premium as share of owners’ proceeds 31.9467% $111,813,412 ÷ $350,000,000
Bala value at 48 years remaining 73.540% Interpolated between the published 50-year and 45-year marks
Points of freehold value bought 22.460 points 96.0% for a restored 99-year lease, less 73.540%
Implied freehold land value $497,833,535 $111,813,412 ÷ 0.2246
Implied freehold value per sq ft of baseline GFA $1,376.60 psf $497,833,535 ÷ 361,639.3 sq ft
Developer all-in for restored 99-year land $461,813,412 $350,000,000 + $111,813,412
Share of implied freehold value paid 92.7646% $461,813,412 ÷ $497,833,535
Gap against Bala’s 96.0% −3.2354 points 92.7646% less 96.0%

Walk that with me, because the logic is simple even though the numbers are large.

The developer’s all-in land cost is $461.8 million for land with a fresh 99-year lease on it. Under Bala’s Table, a fresh 99-year lease is worth 96.0% of freehold. The $111.8 million premium bought the move from 73.540% (at 48 years remaining) up to 96.0% — a distance of 22.460 percentage points. If 22.460 points cost $111.8 million, then one point costs about $4.98 million, and a hundred points — the full freehold interest — would be about $497.8 million. That is the implied freehold value of the land.

So the developer paid $461.8 million for something the table says should be worth 96.0% of $497.8 million. But $461.8 million divided by $497.8 million is 92.7646%, not 96.0%. The market paid 3.2354 points less than the table’s number for freshly restored 99-year land.

What might that gap mean? Honestly, several things, and I am not going to pick one and call it proven. It could be a discount for execution risk, holding cost and the time value of a development that will not produce revenue for years. It could be that the Chief Valuer’s actual methodology is not Bala’s Table and my whole chain rests on a wrong premise. It could be site-specific — Jurong West, a particular competitive dynamic, a particular tender. It could be that the “inclusive of bonus balcony gross floor area” phrase in the reported rate means my denominator is wrong, which I address in the caveat box below. Or it could be a real and repeatable finding that developers systematically pay slightly less than the table’s ratio for restored leasehold land because the table does not price the practical frictions of owning a leasehold development pipeline.

I lean towards a combination of the first and the last. But leaning is not knowing.

Required honesty — read this before you quote any number above

This is our inference from one site. It is not a published valuation, it has not been confirmed by any valuer, and it rests on an assumption we cannot verify: that Bala’s Table is the basis the Chief Valuer actually used to assess the Lakeside Towers lease top-up premium. The Chief Valuer does not publish the method. If the method is something else, the chain breaks at the second link and the 92.7646% figure means nothing.

There is also a known ambiguity in the denominator that I want on the record rather than tucked away. The reported land rate is described as inclusive of bonus balcony gross floor area. We have treated the baseline GFA of 361,639.3 sq ft as the divisor. If instead the 7% bonus balcony GFA sits in the denominator, the implied premium is about $144,100,000, the implied freehold value becomes $641,585,040, and the share paid becomes 77.0124%. That is a materially different answer and it would change the interpretation completely.

We prefer the $111.8 million figure because it reconciles with a premium we reverse-engineered independently on an earlier study, which gives us two roughly consistent readings rather than one. But we disclose the alternative and we do not assert either as fact. One site is one site. If you are making a decision with real money attached, engage a licensed valuer and do not rely on a reverse-engineering exercise published on an agent’s website — including mine.

What the cheque implies about freehold vs leasehold condo value

Set the caveats aside for a moment — not forgotten, just parked — and ask what the exercise teaches even if the precise percentage is uncertain.

The first lesson is about magnitude, and it is the one that lands hardest with clients. On a site that sold for $350 million, the cheque for tenure alone was in the region of $111.8 million. That is 31.9467% of what the owners received, paid on top, purely for years. Nothing about the plot changed. No extra floor area, no change of use, no betterment of any kind. Just a lease reset from 48 years to 99.

Think about what that says about the thing you are buying when you buy a leasehold condo. Roughly a third of the owners’ proceeds, again, is what it costs to put the clock back. Tenure is not a footnote on the title deed. It is a substantial fraction of land value, and the market prices it as such when it is forced to transact on it explicitly.

The second lesson is about direction. Whatever the exact figure, the implied ratio came out below the table, not above it. Both of my scenarios — the preferred $111.8 million and the alternative $144.1 million — produce a share paid that is lower than Bala’s 96.0% for a fresh 99-year lease. In one case modestly lower, in the other dramatically lower. I cannot tell you which, but I can tell you that in neither reading did the market pay more than the table says for restored leasehold land. That is a one-directional result across both scenarios, and one-directional results across a plausible range are worth more than a precise number resting on a single assumption.

The third lesson is the one I would actually use in a conversation with a seller sitting on a 48-year-remaining leasehold condo in an en bloc discussion. The premium is the single biggest determinant of whether a collective sale works. Developers underwrite to the all-in land cost, not to your reserve price. If the premium on your site is going to be a third of your proceeds again, your reserve price has to live inside what is left after the developer has paid the State. Owners who do not model the premium end up anchored to a number no developer can justify, and the tender fails. I have watched it happen. For the mechanics of how the collective sale process itself has changed, our write-up of the updated en bloc rules in Singapore covers the consent thresholds and procedural steps in detail.

The fourth lesson is a warning about freehold. If leasehold’s weakness is a clock, freehold’s weakness is the absence of one. A leasehold site has a built-in forcing mechanism: as the lease runs down, the economics of redevelopment eventually become overwhelming, and that is why ageing leasehold estates get collective-sale interest. A freehold development has no such clock. It can sit, ageing, with owners who will never agree, for decades longer than a leasehold block would. Freehold protects your tenure beautifully and does nothing whatsoever to protect your building. I have seen more freehold owners trapped in unsellable 1970s walk-ups than leasehold owners caught at the bottom of a lease.

The land betterment charge is a different instrument

Now let me clear up the most common mix-up in this entire subject, because I hear it monthly from people who should know better, including some agents.

The land betterment charge and the lease top-up premium are two different instruments doing two different jobs.

The land betterment charge prices an uplift in intensity or use. If a site is permitted to be built more densely than it currently is, or put to a more valuable use, the State captures part of that uplift. The charge is about what you may do with the land.

The lease premium prices an uplift in tenure. If a site’s lease is restored to a fresh 99 years, the State is selling you years. The premium is about how long you may do it for.

These are orthogonal. A site can owe a large lease premium and zero land betterment charge at the same time, and that is exactly Lakeside Towers’ position — roughly $111.8 million of lease premium on our inference, and $0 of land betterment charge because the site is already built beyond its baseline plot ratio. There is no intensity uplift to capture, so there is no charge. There is a very large tenure uplift, so there is a very large premium.

On the charge itself, it is worth knowing that it is not confiscatory. The Singapore Land Authority states that the charge is “not a 100% charge on the value increase” — it runs at roughly 70% of the value increase, with the landowner retaining about 30%. So an intensity uplift is still worth having; you simply do not keep all of it. Our dedicated page on the land betterment charge in Singapore goes through the rate tables and how the assessment works.

The teaching point: over-built saved them the charge, leasehold cost them more than the charge would have

This is the sentence I would put on a slide. At Lakeside Towers, being over-built saved the site the land betterment charge entirely — $0 payable. And being leasehold cost it roughly $111.8 million in lease premium. Same site, same transaction, same day. The thing everyone worries about cost nothing. The thing nobody mentions in the marketing cost nine figures.

Owners in collective sale discussions obsess over the land betterment charge because it has a scary name and a published rate table. Meanwhile the lease premium, which is frequently the much larger number and whose method is not published at all, gets a shrug. Get your priorities the right way round.

The owner’s thirty-second over-built test

If you own a unit in an older development and you want to know, in about thirty seconds, whether your site is protected from the land betterment charge, here is the test. I have not seen a competitor publish it, and it is embarrassingly simple.

Take your development’s as-built gross floor area and divide it by the site area. That gives you the as-built gross plot ratio. Compare that against the gross plot ratio allowed for your site under Master Plan 2025.

The owner’s thirty-second over-built test. Interpretation is my own judgement, applied to the publicly available comparison between as-built gross plot ratio and the Master Plan 2025 gross plot ratio.
If as-built GPR is Your position What it means in a collective sale
Above the Master Plan GPR Over-built, protected No intensity uplift to capture, so no land betterment charge. A developer keeps the existing bulk. This is a genuine advantage and it should be in your marketing brief.
Equal to the Master Plan GPR Protected on intensity only No charge on intensity, but no bonus bulk either. A developer rebuilds to roughly the same size. Neutral.
Below the Master Plan GPR Exposed There is an intensity uplift available, and the State will capture roughly 70% of its value through the charge. The uplift is still worth having, but model the charge before you set a reserve price.

And now the perverse incentive, which is my favourite fact in this whole area and worth repeating until it sticks.

The cramped 1970s and early-1980s blocks with failing facades, dark corridors, surface car parks and lifts that stop at alternate floors are frequently the over-built ones. Why? Because they predate tightened plot ratios. They were built under rules that let them pack more bulk onto the site than today’s Master Plan would permit, and that historical over-build is now a baseline the developer inherits.

So: the uglier your block, the better your baseline may be. I have stood in lobbies that would embarrass a car park and quietly worked out that the site had a materially better collective sale case than the handsome 1990s development across the road, purely because of when it was built. If you own in an unloved old block, do not assume you own a problem. Run the test.

The honest counterweight: over-built protects you on the charge, and it also caps your upside. A developer buying an over-built site cannot build bigger, so the deal has to work on the existing bulk alone. Over-built is protection, not a windfall. Both things are true at once.

CPF: where leasehold bites long before the endgame

Everything above is about the endgame. But most buyers will never see the endgame, and leasehold affects them anyway — through financing.

The amount of CPF you can use for a property depends on whether the remaining lease covers the youngest buyer to age 95. That single rule is one of the most consequential constraints in Singapore residential property and it is wildly under-discussed in tenure debates. It is also the mechanism that explains the shape of the Bala curve better than any discounting argument.

Think it through. A lease short enough to restrict CPF use and shorten loan tenure does not just become slightly less attractive. It loses entire categories of buyer. The young couple who would have used CPF cannot. The buyer who needed a long loan tenure to make the monthly payment work cannot get one. The bank gets conservative on valuation. What is left is a smaller pool of cash-rich buyers who know exactly how short the lease is and will price accordingly.

Prices are set at the margin by whoever can bid. Strip out the financed buyers and you have not reduced demand by a bit — you have removed most of the bidding. That is why the bottom of the curve is so steep, and why the steepness is not really about discounting cash flows at all. It is about liquidity. Bala’s Table happens to describe the outcome reasonably well, but the mechanism underneath it is a financing cliff, not a mathematical one.

The practical implication for a buyer in the 60-to-40 band is this. Do not only ask what the tenure cost is over your holding period. Ask what the lease will look like to your buyer on the day you sell, and whether that buyer will be able to use CPF and get a normal loan. If you buy at 55 years remaining and hold fifteen years, you are selling at 40 years remaining — and you are selling into a thinner pool than you bought from. That dynamic is not in the Bala figures; it sits on top of them.

The pro-ration rules are detailed and I will not re-explain them here, because we have a page that does it properly. Read our page on CPF usage where the balance lease is shorter and how to make sure the youngest buyer can stay in the property to age 95 before you commit to anything with a shortened lease. It is the single most useful piece of homework a leasehold buyer can do.

If you want to put actual numbers on your own situation rather than reading in the abstract, run the calculations on the Listings.sg tools. This article is money math, and money math is better done with your figures than mine.

Freehold vs leasehold condo: who should buy which

I want to write this section fairly in both directions, because the tenure debate in Singapore is unusually tribal and a lot of it is snobbery dressed as analysis. A leasehold buyer is not a lesser buyer. Plenty of the smartest property decisions I have been involved in over seventeen years were leasehold purchases made by people who understood exactly where on the curve they were standing.

Who should lean freehold and who should lean leasehold. My own judgement from seventeen years of advising on both, informed by the tenure-cost arithmetic in this article.
Lean freehold if you are Lean leasehold if you are
Buying to hold across generations, with a genuine intention to pass the asset down rather than a vague sentiment about it. Buying a fresh or near-fresh 99-year lease with a holding horizon of roughly ten to twenty-five years. Tenure cost over that window is small; the freehold premium usually is not.
Unwilling to carry any terminal risk at all, and prepared to pay a measurable premium for that peace of mind as a consumption choice rather than an investment one. Price-sensitive and want more square footage, a better location or newer facilities for the same budget. Leasehold routinely buys more house per dollar and that is a real benefit, not a consolation.
Buying in a submarket where the freehold premium is genuinely thin — it happens, and where it does, freehold is close to free optionality. Yield-focused, where rental income is the point. Tenants do not pay more for freehold. Rent is set by the flat, the location and the commute, so leasehold frequently yields better on a lower capital base.
A business owner or family office treating property as a long-duration store of value rather than a traded asset. Buying with a clear redevelopment thesis on an ageing site, having actually run the over-built test and thought about the lease premium.
Buying an older freehold property with redevelopment or rebuild potential on a decent plot, where the land is the asset and the building is incidental. A younger buyer who needs CPF and a long loan tenure to make the numbers work, and is buying a long lease where neither is constrained.
Genuinely unable to sleep while owning a depreciating clock. This is not irrational. Know yourself. Comfortable managing an exit on a timetable, and disciplined enough to actually sell when the plan says to rather than drifting into the steep part of the curve.

Notice what is not in either column: anything about status. If the only reason a freehold purchase appeals is that it sounds better at a reunion dinner, that is a very expensive dinner. And if the only reason a leasehold purchase appeals is that the headline price is lower, you have not done the analysis — you have read the first line of it.

The mistake I see most often on the freehold side is overpaying for permanence that will never be collected. Somebody pays a thumping premium for freehold, holds eleven years, sells, and the tenure they paid so much for was never tested. Under our arithmetic, eleven years off a fresh 99-year lease costs around three per cent of value. If the freehold premium was fifteen per cent, that buyer paid five times over for a benefit they did not use.

The mistake I see most often on the leasehold side is buying the bargain without pricing the clock. Forty-something years remaining looks cheap against freehold in the same district, and sometimes it genuinely is. But at 1.76% a year compounding, plus a shrinking financed buyer pool on exit, a twenty-year hold has to produce a lot of capital growth before you are ahead. Sometimes it does. You should know in advance whether you are relying on it.

Five expensive mistakes I watch people make

Seventeen years, five hundred deals and more tenure arguments than I can count. These are the ones that cost real money.

One: treating lease decay as a flat annual rate. This is the big one and it is why the stage table above is the centrepiece of this article. If you apply a single average decay rate across the life of a lease, you will simultaneously overestimate the cost at the top and badly underestimate it at the bottom. The curve is not a line. Averages lie about curves.

Two: comparing a new leasehold launch to an old freehold property on tenure alone. I hear this constantly — somebody rejects a new 99-year launch on principle and buys a forty-year-old freehold apartment instead. Under the table, the new lease is at 96.0% of freehold. The old freehold property is at 100% of freehold, but it is also forty years old, with forty-year-old plumbing, a forty-year-old layout and an owners’ committee that has never agreed on anything. You traded four percentage points of tenure for four decades of building age. That is rarely the trade people think they are making.

Three: assuming en bloc will rescue a short lease. Collective sales depend on a developer being able to make the numbers work after paying the lease premium, after construction cost, after holding cost, and with a market willing to buy the new units. Several of those variables have gone the wrong way for developers in recent years. En bloc is a possibility to be assessed, not a parachute to be assumed. If your entire thesis on a 40-year-remaining asset is “it will go en bloc”, your thesis is a hope.

Four: ignoring the buyer on your exit. Covered above but worth restating because it is so frequently missed. You do not sell into the market you bought in. You sell into the market your lease length creates fifteen years later, with CPF and loan tenure constraints that will be tighter than yours were. Model your buyer, not just your price.

Five: confusing the land betterment charge with the lease premium. They are different instruments. One prices intensity, one prices tenure. Mixing them up leads owners to worry about the wrong number and to set reserve prices that no developer can meet. At Lakeside Towers the charge was zero and the premium was nine figures, on our inference. If you take one structural fact away from this article, take that one.

What we are deliberately not claiming

I think publishing the boundary of your own work is more valuable than publishing the work, so here is ours, stated clearly.

We have not measured the freehold price premium in transaction data in this piece. Not once. There is no regression here, no matched-pair analysis, no price-per-square-foot comparison between freehold and leasehold condos. If you came looking for “freehold condos trade at X per cent above comparable leasehold”, this article does not contain that number and does not pretend to.

Stacked Homes has done that work, on both sides of the question, and their ten-year study is the right place to start for a reader who wants the return comparison. I would rather send you to good work than produce a weaker version of it for the sake of keyword coverage.

What we have done is answer a different question. What is the tenure difference worth at the end of a lease, and what does the only real cheque anyone writes for tenure imply about that value? That question has two original answers in this piece: the 37.82x shape of the decay curve, which is our arithmetic from the industry’s own table, and the Lakeside Towers reverse-engineering, which is our inference from one site with caveats attached in writing.

We also are not claiming to know the Chief Valuer’s method. It is not published. We assumed Bala’s Table and said so. If that assumption is wrong, the second finding collapses and the first one stands, because the first one depends only on the published table and arithmetic.

And none of this is a valuation of any specific property. I am a licensed agent with seventeen years of transaction experience. I am not a valuer and I am not a solicitor. If you need a valuation for a decision, engage a valuer. If you need advice on a collective sale agreement, engage a solicitor. If you need somebody who has sat on both sides of five hundred of these conversations and will tell you which number actually matters, that part I can do.

My verdict on freehold vs leasehold condo

Seventeen years in, five hundred deals in, here is where I have landed. The tenure question is real, it is worth money, and almost everybody frames it wrongly. It is not a quality grade. It is a position on a curve, and the curve is published, and the curve is not a straight line.

I have given you the table the industry uses, two independent sources for it, the stage-by-stage cost of a year of tenure that nobody else prints, the ratio between the gentlest stretch and the brutal one, and a nine-figure reverse-engineering of the only cheque anybody in Singapore writes purely for years of lease. I have also told you, in writing and in a bordered box, exactly which of those rests on an assumption I cannot verify. That boundary is not a weakness in the work. It is the work.

The verdict

Stop asking whether to buy freehold or leasehold and start asking where on the lease curve you are buying and where on it you expect to sell. Under Bala’s Table, the first nineteen years of a 99-year lease cost about 0.2811% a year in tenure value and the last five cost about 10.6335% a year — a ratio of 37.82x by our arithmetic. For a buyer holding a fresh 99-year lease for fifteen to twenty years, tenure is a rounding error against almost any realistic freehold premium, and paying up heavily for permanence you will never collect is simply an expensive habit. For a buyer of a forty-year-remaining leasehold property, tenure is the dominant term and needs a thesis, not a hope. And at the bottom of the lease, where the only real cheque for tenure gets written, our reverse-engineering of the Lakeside Towers collective sale suggests the market paid 92.7646% of implied freehold value for freshly restored 99-year land against the table’s 96.0% — one site, our inference, caveats published. Two points on a curve. That is the whole analysis.

Want this run on your actual property?

If you own a leasehold condo and want to know where you are on this curve, what a lease premium might do to a collective sale case, or whether your block is over-built, message me directly. I will do the arithmetic with your real numbers, and I will tell you plainly if the answer is that you should do nothing.

WhatsApp Gary Lim on +65 8986 1688

You can also request a free property valuation at buycondo.sg, watch the walkthroughs and market breakdowns on our video hub at buyers.sg, or run the numbers yourself with the calculators at Listings.sg. We Serve with Heart.

Frequently asked questions

Is a freehold condo always a better investment than a leasehold condo

No, and the data says otherwise more often than people expect. Stacked Homes has published a study finding that leasehold condos can deliver better ten-year returns than freehold ones, and I have no quarrel with that finding. The reason is straightforward: leasehold typically costs less to enter, and under Bala’s Table a fresh 99-year lease is worth 96.0% of freehold, so the tenure you give up over a normal holding period is small while the price discount you receive is often large. Freehold wins when the holding period is genuinely multi-generational or when the premium you pay for it is thin. It loses when you pay fifteen per cent extra for permanence and sell after twelve years.

What is Bala Table and who actually wrote it

Bala’s Table is the reference table Singapore valuers use to express a leasehold interest as a percentage of freehold value for the same land. It has been in use for roughly three quarters of a century and was published online by the Singapore Land Authority from 31 July 2000. Nobody knows who wrote it. The Centre for Liveable Cities commentary by Kwek Sian Choo and Dionne Hoh from 2017 says it is widely believed to have been prepared by a Land Office employee named Bala, and that the exact time and basis of its adoption are not known, with evidence pointing to around 1948. The same commentary is candid that the table and its implicit discount rate are not an exact science, but have served for decades.

How fast does a leasehold condo actually lose value each year

It depends entirely on where the lease sits, and the variation is enormous. Our arithmetic from Bala’s Table gives about 0.2811% a year compounded as a lease falls from 99 to 80 years, about 0.7730% a year from 60 to 40, about 1.7624% a year from 40 to 20, and about 10.6335% a year over the last five years before expiry. The last five years destroy value 37.82 times faster per year than the first nineteen. Anyone quoting a single flat annual decay rate for leasehold property is giving you an average of a curve, which will mislead you at both ends.

What is a lease top-up premium and why does it matter to me

When a developer buys a leasehold collective sale site, it must pay the State to restore the lease to a fresh 99 years. That payment is the lease upgrading premium or lease top-up premium, and it is the one place in Singapore where somebody writes a cheque purely for tenure with nothing else changing. It matters to an owner because it comes out of the same pot as your sale proceeds from the developer point of view. At Lakeside Towers, our inference puts the premium at about $111,813,412 against a $350,000,000 sale, which is 31.9467% of what the owners received. If you are in an en bloc discussion and have not modelled the premium, your reserve price may be a number no developer can justify.

Is the land betterment charge the same thing as the lease premium

No, and this is the most common mix-up in the whole subject. The land betterment charge prices an uplift in intensity or use, meaning what you are allowed to do with the land. The lease premium prices an uplift in tenure, meaning how long you are allowed to do it for. They are independent, and a site can owe a very large lease premium and zero land betterment charge at the same time. Lakeside Towers is exactly that case: no land betterment charge payable because the site is already over-built, and roughly $111.8 million of lease premium on our inference. Being over-built saved them the charge, and being leasehold cost them far more than the charge would have.

How do I tell if my development is over-built

Divide your development as-built gross floor area by the site area to get the as-built gross plot ratio, then compare it against the gross plot ratio allowed for your site under Master Plan 2025. Above it and you are over-built, which means no intensity uplift to capture and therefore no land betterment charge. Below it and you are exposed to the charge on the uplift. Equal and you are protected on intensity but get no bonus bulk. One counterintuitive point worth knowing: the cramped 1970s and early-1980s blocks with failing facades are frequently the over-built ones, because they predate tightened plot ratios. The uglier your block, the better your baseline may be.

Does a shorter lease affect how much CPF I can use

Yes, and this is where leasehold bites an ordinary buyer long before the endgame arrives. The usable CPF for a property depends on whether the remaining lease covers the youngest buyer to age 95. That rule puts a floor under how short a lease the market will realistically finance, and it is the mechanism that makes the bottom of the Bala curve so steep. A lease short enough to restrict CPF use and shorten loan tenure loses the buyers who would otherwise have bid for it, and prices are set at the margin by whoever can bid. Read our dedicated CPF page for the pro-ration detail before committing to anything with a shortened lease.

About the author

Gary Lim · ERA Senior Division Director · CEA Licence R009877B

17+ years in Singapore residential property and more than 500 completed transactions across HDB resale, private condominium and landed sales and leasing. Gary leads the BuyCondo Team, which specialises in condominium sales, leasing and owner-side advisory, including collective sale positioning for ageing developments. He publishes the arithmetic behind his advice, names his sources, and corrects himself in public when he gets something wrong — nine times so far and counting.

Reach him on WhatsApp at +65 8986 1688, request a free valuation at buycondo.sg, or watch the team’s market breakdowns at buyers.sg. 新加坡买房,就找对的团队.

Disclaimer. Gary Lim is a licensed real estate salesperson registered with the Council for Estate Agencies, CEA registration number R009877B, representing ERA Realty Network Pte Ltd, CEA estate agent licence number L3002382K. Nothing in this article is a valuation of any specific property, an offer, or financial, tax or legal advice. Gary Lim is a licensed real estate salesperson, not a licensed valuer and not a solicitor. Readers should engage a licensed valuer for any valuation and a solicitor for any matter concerning collective sale agreements, lease documentation or conveyancing.

Methodology and sources. Bala’s Table figures are sourced from the Centre for Liveable Cities commentary by Kwek Sian Choo and Dionne Hoh (2017), Appendix 1, and cross-checked against figures published by Lee Sze Teck, Senior Director of Data Analytics and Research at Huttons Asia, in EdgeProp on 12 June 2025; the two sources agreed at all three marks tested. The stage-by-stage tenure cost table, the 37.82x ratio, the interpolated 73.540% value at 48 years remaining, and the entire Lakeside Towers reverse-engineering are our own arithmetic derived from those published figures and from reported transaction facts. Lakeside Towers site facts, including the 153,237 sq ft site area, the 144 units, the leasehold commencement in 1975, the above-80% consent level, the 1 October 2026 tender close, the $350,000,000 reported sale price and the $1,277 per square foot per plot ratio land rate inclusive of lease top-up premium and bonus balcony gross floor area with no land betterment charge payable, are as reported in market coverage; the explicit as-built gross plot ratio of 2.36 against a base of 2.1 was published by JLL. The land betterment charge statement that it is not a 100% charge on the value increase is from the Singapore Land Authority. The ten-year leasehold-versus-freehold return finding referenced in this article is the work of Stacked Homes and is credited to them. The method by which the Chief Valuer assesses lease upgrading premiums is not published, and our inference assumes Bala’s Table as the basis — an assumption we cannot verify and have flagged in the article. Stacked Homes, EdgeProp, Huttons Asia, JLL and the Centre for Liveable Cities are credited by name in the text without links.

Corrections policy. We correct published errors in public, on the page where the error appeared, and we keep a running count. This article’s Bala’s Table figures follow our ninth published self-correction, in which four of six previously published figures were found to be wrong and all four were wrong in the same direction. If you find an error in this piece, message Gary on WhatsApp and we will fix it and say so.

 

Share

Can we get your Personal Insights...

Free Home Valuation