A record MOP wave meets a thin launch pipeline: reading 2H 2026

Most half-year outlooks are a mood. This one had numbers, which makes it worth arguing with. The picture Marcus Chu and Eugene Lim put up on 23 July was of a market that has stopped sprinting: new-sale prices roughly flat since 2023, HDB resale down two quarters in a row, take-up rates so strong that several projects cleared almost everything on the opening weekend. Their advice followed naturally. If you are financially ready and you have found the right property, buy it. Or as Chu put it, time in the market beats timing the market.

I have very little to argue with there. What I want to add is the part the deck touched but did not join up. Two supply clocks are running at different speeds in the second half of 2026, and they point in opposite directions.

What is happening in Singapore's property market going into 2H 2026?

Going into the second half of 2026, Singapore's private new-sale market is supply-constrained and demand-strong, with roughly 3,620 new homes scheduled to launch between July and December against 1H 2026 take-up rates that reached 99% at Tengah Garden Residences and 100% at Rivelle Tampines (EC). At the same time the public housing market is doing the opposite: the HDB Resale Price Index has slipped for two consecutive quarters, about 0.5% quarter on quarter, while a record 13,480 flats reach the end of their five-year Minimum Occupation Period this year.

Those figures come from URA REALIS, HDB and MAS data as compiled for the 23 July briefing. Everything below is my reading of them, plus one cut of PropertyInsider.sg's own dataset that I think changes the conclusion.

Why is the price gap between new and resale condos narrowing?

Because the new-sale side stopped moving and the resale side did not. Median new-sale prices sat at $2,486 psf in 2023 and $2,561 psf in 1H 2026, which after three years is a rounding error. Median resale prices went from $1,578 psf to $1,777 psf across the same window. The gap between them fell from $908 psf to $784 psf.

This matters more than it looks. A buyer choosing between a new launch and a resale unit is not only comparing price per square foot; they are comparing a fresh 99-year lease, a current-generation layout and a developer warranty against whatever is left of an older lease. When the premium for all of that was over $900 psf, plenty of buyers took the resale unit and lived with the age. At $784 psf, and with 3-month compounded SORA at 1.08% in 2Q 2026 against a 3.7% peak in 1Q 2024, the maths tilts. That, not sentiment, is why the take-up rates looked the way they did.

My caveat is the one I give every client who repeats this argument back to me: a narrowing gap is not a discount. It closed because resale went up, not because new launches came down. If you are buying new because the gap narrowed, you are still paying $2,561 psf, and you should stress-test the repayment at 3% rather than 1.08% before you sign anything. Our affordability and stamp-duty calculators exist for exactly that exercise.

Are HDB resale prices falling in 2026?

They are flattening, not falling. Two quarters of roughly 0.5% quarterly decline after years of uninterrupted increases is what the top of a plateau looks like. Resale applications held at 6,285 in 1Q 2026 and about 6,268 in the 2Q flash estimate, so this is not a buyers' strike either. Sellers panicking over the word "decline" are reading a 0.5% move as a trend.

The million-dollar flat headlines confuse this further. There were 903 million-dollar resale transactions in 1H 2026, and the annual count has climbed steeply since 2021. But mapped against total resale volume it stays well under 10%, and around three quarters of resale flats still change hands below $1 million, mostly in the $500,000 to $800,000 band. The flats crossing the line are a recognisable type: mature estate, short walk to an MRT station or a good hawker centre and mall, near-new condition with a long lease left. If your flat is none of those things, the headlines are not about you.

The more useful number in that chart is the MOP curve itself. Roughly 13,480 flats reach the five-year mark in 2026, 18,939 in 2027 and 21,393 in 2028, with Punggol (3,222), Queenstown (2,405), Tampines (2,133), Toa Payoh (1,594) and Bedok (1,440) carrying the largest 2026 batches. That is a three-year wave of near-new stock arriving into a market whose price index has just stopped rising.

The two clocks: resale supply is expanding while launch supply contracts

Here is the join the briefing left implicit. Public resale supply is expanding fast, and private launch supply is unusually tight. Those are not independent facts, because the same household often sits on both sides: they sell the MOP flat and buy the launch.

On the launch side, ERA Project Marketing counted about 3,620 homes for 2H 2026 as of 8 July: roughly 3,189 condominium units, 420 EC units at Wynwood Grand under the pre-8 May rules, and 11 landed homes. Our own GLS pipeline tracker carries a slightly larger second-half set, mainly because it includes Chuan Grove GLS at around 1,055 units in Q4 2026, which the ERA schedule did not list. I flag the difference rather than reconcile it silently; preview dates on unlaunched GLS sites move, and neither list is wrong so much as differently cut off.

Set 3,620 against the 1H 2026 evidence that individual launches drew three to four interested buyers per available unit, and the practical implication is the one Chu drew: if you find the unit, do not spend three weekends thinking about it. My version is slightly less urgent. Thin supply in one half-year is a queueing problem, not a pricing law. The 2027 pipeline on our tracker is materially larger, and it includes sites bought at record land rates that will have to be priced accordingly. Waiting is a real strategy; it is just not a free one.

Will Prime and Plus flats reduce future upgrader liquidity?

This was the last of the four long-run trends in the briefing, and I think it was the most important one on the slide. It also came with the biggest hedge, correctly, because nothing here happens before the mid-2030s.

The mechanism is simple enough. Prime and Plus flats carry a 10-year MOP rather than five, so the upgrade clock starts twice as late. A subsidy clawback takes a slice of the resale proceeds, so the equity that comes out is smaller. And a $14,000 monthly household income ceiling on eligible resale buyers caps the pool bidding for the flat, which caps the price. Delay the sale, shrink the proceeds, cap the buyer pool: each one on its own is manageable, all three together change what an upgrader can afford next.

The reason this matters beyond public housing is that HDB upgraders have quietly underwritten the private market for decades. Here is the cut I ran on our own data, which I have not seen published anywhere else. Across the 67 selling projects PropertyInsider.sg tracks with buyer-profile data, HDB upgraders are a median 42.5% of buyers in Outside Central Region launches, 28.3% in the Rest of Central Region and 18.7% in the Core Central Region.

Read those two things together. A policy that thins the upgrader flow does not hit the market evenly; it hits hardest exactly where upgraders make up two buyers in five, which is the suburban mass market, not Orchard. If you own an OCR unit and you are modelling your exit in the mid-2030s, that is the number to keep on the wall. It is also, to be fair to the briefing, a decade of policy adjustments away from being certain, and HDB has changed income ceilings before.

The trends I would weight differently

Three of the four long-run trends in the deck were demographic. On the ageing population and the shrinking household, I think the presenters were, if anything, too gentle. One in five citizens is already 65 or over and that becomes one in four after 2030, and the household-formation arithmetic is the quieter half of it: one-person and two-person households have been rising steadily since 2015 while larger households flatten. Split a six-person household into three two-person households and the population has not changed at all, but the housing requirement has tripled. That is the strongest structural demand argument available in Singapore, and it does not depend on immigration policy holding.

The right-sizing point is the corollary and it gets underweighted. As retirees move out of five-room and executive flats, those flats become the resale supply that today's growing families are short of. If you are hunting a large resale flat, the next few years should be kinder than the last few.

The fourth trend, the future of work and AI, I would drop from a property outlook. The argument, that AI reshapes jobs, jobs drive income, income drives housing decisions, is true at a level of generality that makes it unfalsifiable. It cannot tell you anything about a purchase in the next 18 months. The ERA and Ngee Ann Polytechnic My Dream Home survey finding in the same section was more useful and more concrete: flexible layouts were the most requested feature in every age band, at 54% among 21 to 35s and 63% among 60 to 69s. That is an actionable brief for developers and a real thing to check on a floor plan.

Where this reading could be wrong

Three ways, and I would revise on any of them.

The first is rates. Every affordability argument in the second half of 2026 rests on 3-month compounded SORA near 1.08%. Rates that fell 2.62 percentage points in two years can retrace faster than anyone's holding plan. The briefing acknowledged this, and I would put it higher on the risk list than it sat on the slide, because the loan quantum a buyer qualifies for is computed off prevailing rates.

The second is policy. The May 2026 EC changes, which lifted first-timer allocation from 70% to 90% and applied a 10-year MOP to sites sold from 8 May, are a live demonstration that the rules can move against a segment overnight. I wrote about that reshaping in The New Shape of Singapore Housing Demand. Any 2H 2026 plan that depends on an EC being available to a second-timer is planning on thin ice.

The third is my own upgrader-liquidity thesis. It rests on tracker buyer-profile data for 67 projects, several of them small enough that a handful of transactions moves the percentage, and on a Prime and Plus resale market that does not exist yet and will not for another 10 to 15 years. If HDB lifts the $14,000 ceiling in step with incomes, most of the price-capping effect disappears. I hold the direction with more confidence than the magnitude.

What I would actually do in the second half of 2026

For first-timers, the October BTO exercise is the last of 2026 and the Bedok and Bayshore projects sit next to a station in an area whose private launches have already sold well. The system is built to favour you; use it while that is true, and remember your first flat does not need to be your last one. Eligibility questions belong with HDB, not with me.

For upgraders, the honest advice is to stop waiting for a BTO or an EC ballot that the new allocation rules have made a long shot. The 13,480 MOP flats arriving this year are near-new, they come without a five-year wait, and the resale index has been flat for two quarters. That combination has not been available for a while. Our buyer guides cover the sequencing mechanics, and the calculators handle the cash-and-CPF side.

For investors, the transformative-corridor argument the briefing made is sound but it needs a filter, because everywhere in Singapore is on some version of the Master Plan. The one I use is whether infrastructure is funded and dated rather than merely drawn. Beyond that, the land market is the forward price curve: our land cost dataset and the 1H 2026 GLS review show how tender prices have been translating into launch prices, and I keep my own site-by-site GLS notes for clients weighing today's launches against tomorrow's costlier ones.

Everyone else should note the shape of the thing. Public housing supply is loosening while private launch supply tightens, and the policy bridge between them is being narrowed for the generation after this one. Prices are the noisy part. The bridge is the part I would watch.

Frequently asked questions

← Back to the Research desk