New Launch or Resale? Whoever Picked the Dates Picked the Winner

If you have spent an evening watching Singapore property videos, you have heard both of these said with total confidence.

“Good resale beats new launch. Always.”

“New launch is the sure win. That is where the money is made.”

They cannot both be right. And yet both sides can point at real transaction data, from the same public source. Neither is making anything up.

The same study, three answers

The most thorough public attempt at this question — Stacked Homes, “Are New Launches A ‘Sure Win’?” — matched 152,948 pairs of transactions, the same unit bought and later sold, identified by address including the unit number, out of more than 450,000 URA records going back to 1995. That is far better than the two-projects-side-by-side comparisons the genre usually runs on.

From 1995From 2011From 2015

As published — total gain per transaction

Bought at launch

—

Bought resale

—

0%20%40%

Per year, on the same holding basis our calculation

Bought at launch

—

Bought resale

—

0%3%6% a year

As published—

Try the other two windows.

Window	Launch, total	Resale, total	Launch, per yr	Resale, per yr
1 Jan 1995 – 31 Mar 2021	27.1%	34.7%	2.83%	5.18%
14 Jan 2011 – 31 Mar 2021	7.9%	6.9%	1.20%	1.16%
1 Jan 2015 – 31 Mar 2021	11.8%	10.6%	not published	not published

Quote the first window and resale wins. Quote either of the others and new launch wins.

The per-year columns matter more than they look. Launch buyers in that study held for about 8.6 years; resale buyers held 5.9. So the published totals are not like-for-like — one side had years longer to earn its gain.

Put both on the same annual basis and the second window’s famous “new launch wins” nearly vanishes. The lead works out at four hundredths of a percentage point a year, which is a number nobody can feel, so here it is in money instead. On a S$1.5 million home held six years, the entire advantage is about S$3,800. The buyer’s stamp duty on that same purchase is S$44,600.

The tax is twelve times the whole thing. That is the margin half the internet is arguing about.

The third window cannot be annualised at all, because no holding period was published for it.

Why the window decides it

Here is the part that is easy to miss. Those three windows are not three different periods. They all end on the same day — 31 March 2021.

from 199526.2 years

Resale clearly aheadby 2.35 points a year

from 201110.2 years

New launch aheadby about S$3,800 on S$1.5m

from 20156.2 years

Cannot be comparedno holding period published

← further backall end 31 Mar 2021

Reach back to 1995 and resale wins clearly. Start in 2011 and new launch “wins” by a margin worth less than a tenth of the stamp duty on the purchase. Start in 2015 and nobody can say.

Nobody has to be lying. Two Singapore rules put the two groups on different parts of the market cycle: Seller’s Stamp Duty penalises an early exit, and a new launch is not a building yet.

In plain English: this is not a race. It is two runners who started at different times, on different parts of the track, and an argument about who is faster.

And the rule that sets the minimum holding period changed twice. For purchases between 14 January 2011 and 10 March 2017 the lock ran four years, topping out at 16%. From 11 March 2017 to 3 July 2025 it was three years and 12%. On and after 4 July 2025 it is back to four years and 16% (IRAS).

Look at what that does to the middle row. The window labelled “since SSD began” runs from January 2011 to March 2021 — so it contains two different SSDs, switching about six years in. Buyers who faced a four-year lock are averaged with buyers who faced three, as though they made the same decision.

We got this wrong ourselves

So we ran it on our own data. The first result came back beautifully: launch buyers underperformed by 4.04% a year across twenty-four projects, only 8.3% of them positive. Consistent sign, tight spread, segments in a sensible order. It agreed with what we already half-believed.

It was wrong.

Our data started in late 2021, and the code treated each project’s first visible sale as its launch. But several of those projects launched in 2018 and 2019 and were merely still selling their last units. We had recorded late-phase prices — higher, because developers raise prices as a project sells through — as launch prices, then measured how badly those buyers did. The window had manufactured the entire result.

So we tightened the rule: count a project only if we had actually seen it launch, and if somebody had since resold a unit in it. Of 184 projects, not one qualified. Almost none has had a resale yet — a launch takes years to build, then four more before the stamp-duty lock lifts, and five years of data cannot span that.

We could not answer the question either. We are telling you because that first result — clean, confident and wrong — is exactly the sort of thing that gets screenshotted and repeated for years.

We are not debunking anyone. Over the longest window published, resale did come out ahead. But the gap is small, it reverses with the window, and it excludes the costs that would decide it. Anyone who says this is settled has skipped a step — including anyone who says it is settled in resale’s favour.

So ask a question that has an answer

Knowing the evidence cannot settle it is worth something. It is not much help on a Saturday afternoon in a showflat.

The useful move is to notice that for most buyers this was never mainly a returns question. Four things genuinely separate a new launch from a resale, and every one is knowable today.

You cannot move in. A new launch is a floor plan and a showflat. A resale is a home you can stand in, in the light it actually gets, next to the neighbours you will actually have. And you do not need an industry average for the wait: your sale and purchase agreement names a Vacant Possession Date, and clause 13.4 makes the developer liable for liquidated damages if it slips. Use yours, not an average of other people’s contracts.

You pay for two homes while you wait. This is left out of every comparison, including the studies — and the schedule is not the developer’s choice. It is set in law, in the sale and purchase agreement prescribed by the Housing Developers Rules (First Schedule, clause 5.1):

• 20% on signing, within eight weeks of the option — inclusive of the booking fee, not on top of it
• 10% when the foundation and pile caps are done
• 10% at the reinforced concrete framework
• 5% each at partition walls, roofing, wiring and plumbing, and the car park, roads and drains

That is 60% of the price before the building is finished. The 25% at Temporary Occupation Permit takes you to 85% paid on the day you get the keys — and every instalment before it is owed on somewhere you cannot live, while you are also paying to live somewhere else.

The lock starts before the home does. Seller’s Stamp Duty runs four years from the day you buy, not from the day you collect keys. On a new launch, most of that lock is served while you do not yet have the property. Your exit clock and your occupation clock are not the same clock.

And the gap buys floor area. Which is the one number we can put on it.

What the gap costs you, on your own budget

Your budget

S$

S$300kS$1.5mS$5m

Outside Centralnew costs +52% per sq ft

Existing

801

New

527

274sq ft less

Rest of Centralnew costs +47% per sq ft

Existing

635

New

432

203sq ft less

Core Centralnew costs +42% per sq ft

Existing

544

New

383

162sq ft less

Square feet, at each segment’s own resale median. The gap is not a measure of overpaying — it is measured against every existing home in the district, including stock built thirty years ago, so part of it buys a genuinely newer building. What it does tell you is what the same money reaches.

Segment	New costs more per sq ft	Resale median $psf	Comparisons
Outside Central	+52.0%	$1,872	405
Rest of Central	+46.9%	$2,362	483
Core Central	+42.2%	$2,756	293

Sources, both ours, and not the same measure: the gap is the median of new-sale $psf against the resale median in the same district and quarter, within size band and tenure class — 1,373 quarterly comparisons, 2021Q3–2026Q3. The resale median $psf is each segment’s own 12-month median across all sizes and tenures. Applying one to the other shows the scale of the difference; it is not a valuation of any particular unit. Both from URA caveat data.

Whether that trade is worth it depends on something this article cannot know: whether you are buying somewhere to live or somewhere to let. For an owner-occupier those square feet are the point. For an investor they are inventory.

Worth noticing too that the core-central gap is more than double the outside-central one. Land cost, scarce new supply and where the state has been selling sites are all plausible explanations. We have not tested any of them.

We have tested whether the gap is moving, though, and it is. Outside central it has widened sharply, from about 39% in 2021 to about 56% in 2025. The core central region has barely moved at all — 43% then, 43% now. The rest of central sits between them, drifting from 42% to 47%.

So the premium for buying new is diverging, not converging. In 2021 the three segments sat within about four percentage points of one another; by 2024 they were twenty-six apart, and in 2025 still thirteen. And the spread is being opened from the suburbs upward rather than from the centre downward, which is the reverse of the story usually told about Singapore’s prime districts.

What to take from this

Two things, and neither requires you to become a data analyst.

First, when someone shows you a property comparison, ask three questions. Over what window — if they cannot give you start and end dates, the number means nothing, and this one question retires most of what you will be shown. How many transactions — two projects side by side is an anecdote, however confident the delivery. And were the holding periods matched — if one side held eight years and the other six, the comparison is measuring time as much as strategy.

Second, and more usefully: stop trying to settle the returns argument. It has swallowed the entire conversation and it cannot be won on the evidence that exists. Work the four differences instead — when you need to move in, what you would pay to live somewhere else while you wait, when your exit clock actually starts, and what the entry gap costs you in floor area. Those have answers, and the answers are yours rather than the market’s.

We cannot tell you whether new or resale wins, and on the published evidence nobody else can either. But that was probably never the question worth asking.