Essays

This is the page that spends nothing and sells nothing. Every other door on this site — buy, sell, borrow, invest, landlord, estate — ends in a decision, and behind every one of those doors is a person who is paid if you act on it, starting with me. This door is different on purpose. It is where a position gets written down before anyone asks you for a transaction: what the market actually did last quarter, what a rule change actually says versus what three blogs say it says, where a number everyone repeats turns out to trace back to a guide rather than a regulator. Nothing here is dated forward. A position is written, dated, sourced, and then revised in public when the facts under it move — not quietly edited, not deleted. The test for anything published on this page is not whether it is exciting. It is whether it is still standing next to its own date and its own source six months later. Reading this page will not tell you whether to buy a flat. It will tell you whether the person who might eventually help you buy one keeps his own score where you can check it.

The mechanics: how a position gets made here

The flagship of this practice is the monthly room-read, revised on a fixed date and built the same way every time: read the primary source first — the regulator's own release, not a summary of it — and say plainly when the primary source would not render and a secondary one had to stand in for it. A figure carries its date because policy and price both drift, and a reader returning in six months should be able to tell at a glance whether the ground has moved under a position written earlier. The discipline is to separate what changed this period from what is merely still true, because most months for most readers the honest answer is that nothing they own or plan needs to change — and that answer is worth stating plainly, not skipped past to get to something more interesting.

A worked example of what "dated, no call" looks like in practice. URA's own flash estimate for the second quarter of 2026 put the private residential price index down 0.1% overall — but that single headline number hides four different markets moving in different directions inside it: the Core Central Region up 2.0%, the Rest of Central Region down 1.4%, the Outside Central Region down 0.2%, and landed homes up 2.6% on the same quarter. A position that reports only the headline -0.1% and stops there is not wrong, but it is not describing the room either — it is describing one number pulled from a room with at least four distinct weather systems in it. On land, URA's own media-release page recorded the Government Land Sales tender for the Lorong Puntong / Sin Ming Avenue site closing 15 September 2026 and awarded 18 September 2026 — the newest confirmed award as at this run — with the award price itself not stated on the page. That gap is left visible rather than filled with a guess, because a land-bid price is exactly the kind of number that tells you where launch prices go next, and inventing one would be worse than admitting it is missing.

Supply is the slower-moving number underneath both of those, and it is the one worth watching longest because a price index reports what already happened while a pipeline reports what is coming. About 60,600 private homes are slated for completion through 2029 on one market aggregator's count, with 9,700 due in 2027 and 11,200 in 2028 — a wave that lands two to three years after the land is awarded, which is the lag between the GLS tender above and the rent or price pressure it eventually produces. Nothing on this page forecasts where that pressure lands. It records the pipeline, dates it, and leaves the forecasting to the reader.

Underneath all of it sits one rule that is not visible in the prose but is enforced on every figure: a fact is cited by an identifier tied to a source and a date, and a figure that later turns out to differ from an earlier reading is never silently overwritten. It is flagged as contested, both values are kept, and the page carries a visible marker until the disagreement is resolved. That is the difference between an opinion column and a ledger — an opinion can quietly change its mind; a ledger has to show its work.

Eligibility: who this door is for

Every transaction door on this site gates on something — a citizenship line, an income ceiling, a minimum occupation period. This one gates on nothing. A consumer of thought does not need to be a Singapore Citizen, does not need a household income under any ceiling, and does not need to be within five years of a lease running out. What differs, reader to reader, is which of the transaction doors downstream eventually applies — and the sections below on citizenship, tax and borrowing exist so that a reader can see, before going anywhere near a transaction, which room they are actually standing in once they do.

Citizenship status: the room reads differently depending who is in it

The same headline price move means a different thing to a different buyer before a single dollar of price is even discussed, because the duty schedule sitting on top of that price is not one schedule. A Singapore Citizen pays Additional Buyer's Stamp Duty of 0% on a first residential property, 20% on a second, 30% on a third and beyond. A Singapore Permanent Resident pays 5%, 30% and 35%. A foreigner pays 60% from the first purchase, unchanged. An entity or a trust pays 65%. None of these rates moved in Budget 2026, and none have moved since 27 April 2023 — which is itself a dated position worth stating: the longest-standing fact in this section is that nothing in it has changed for over three years.

One nationality question sits outside that table entirely. Nationals of the United States, Switzerland, Liechtenstein, Norway and Iceland — plus the EFTA-linked permanent residents the relief extends to — are taxed as Singapore Citizens under free-trade-agreement parity, but that treatment is not automatic. It runs through a remission application on the IRAS myTax Portal, filed after the ABSD is paid in full, not a discount applied at the point of purchase. A US green-card holder who is not a US national does not qualify — the remission tracks nationality, not residence. A US citizen who also holds Singapore PR status is charged at the Citizen rate outright, and a trust with a US citizen beneficiary can have the full 65% remitted if that is the beneficiary's only residential property. Each of those is a distinct rule, not a restatement of the same one, and collapsing them into "Americans pay no ABSD" would be exactly the kind of flattening this page exists to avoid.

Taxes: the steadiest input, and the one whose stillness is itself information

Buyer's Stamp Duty is the one duty every buyer pays regardless of citizenship, and it has been the same progressive schedule since 15 February 2023: 1% on the first $180,000, 2% on the next $180,000, 3% on the next $640,000, 4% on the next $500,000, 5% on the next $1,500,000, and 6% on anything above $3,000,000. Seller's Stamp Duty, by contrast, moved as recently as 4 July 2025, when the holding period lengthened to four years and the rates rose: 16% inside the first year, 12% in the second, 8% in the third, 4% in the fourth, nothing after. A position on this page states which of those two duties is currently stable and which one moved within the last eighteen months, because treating both as equally settled would misdescribe the room.

Property tax runs on two entirely different schedules depending on one fact about the property, not the owner: is it owner-occupied. The owner-occupier band starts at 0% on the first $12,000 of Annual Value and rises through several steps to 32% above $140,000; a rented or vacant home is taxed on the non-owner-occupier schedule instead, starting at 12% on the first $30,000 of Annual Value and rising to 36% above $60,000. The owner-occupier rate applies to exactly one property per owner — a second home the owner also occupies is still charged at the higher rate. Rental income itself is taxable, with the CPF Board's own guidance on what CPF Ordinary Account money may or may not be used for drawing a sharp line: OA can pay stamp duty and legal fees, but never the monthly service and conservancy charges, property tax or utilities that keep accruing after the purchase closes.

Borrowing: reading the financing weather, not just the price index

MAS's own explainer, last revised 16 December 2021 and unchanged on every re-read since, sets the Total Debt Servicing Ratio ceiling at 55% of gross monthly income for any property loan, and the Mortgage Servicing Ratio at 30% for an HDB flat or executive condominium during its minimum occupation period. Underneath that ceiling sits a floor a reader rarely hears about: TDSR is assessed at the higher of the loan's actual rate or a MAS-set floor — 4% for residential property, 5% for non-residential — so a headline rate of 2% on a bank's rate card is not the number a bank actually stress-tests your income against.

Loan-to-value follows the same layered logic. A first housing loan can reach 75% of the purchase price, dropping to 55% if the loan tenure runs past 30 years or past the borrower's 65th birthday — a rule that quietly re-prices a purchase for an older borrower or a longer tenure without changing the headline price of the flat at all. HDB and bank financing then split further: HDB's concessionary rate has sat at 2.5% to 2.6% for years with no lock-in and no prepayment penalty, while a floating bank rate tracks compounded SORA plus a spread — 1.19% for three-month compounded SORA as quoted in early September 2026, against spreads typically 0.8 to 1.2 percentage points on top. One market study covering July 2005 to May 2026 found a bank loan at SORA plus 0.5% was actually cheaper than the HDB rate in roughly three of every four months across that entire twenty-one-year run — a dated position that cuts directly against the folk wisdom that an HDB loan is always the safer, cheaper choice.

Bridging finance is where this weather changes fastest, because it is priced weekly and quoted differently by every guide that reports it. Bank bridging rates were reported at 4.88% to 5.75% a year in April 2026, typically three-month SORA plus 3.5 to 4.5 percentage points, with an admin fee around 1% of the loan and an early-repayment fee of a further 1 to 1.5%. A worked illustration on this site's own bridging page puts a S$235,000 bridge at 5.5% for two and a half months on a S$1.65 million condominium purchase at about S$2,698 in interest alone — small next to the ABSD due on the same purchase, but real money that a position stated once and never revisited would understate the moment rates move.

Renovation, A&A and rebuild: the cost side nobody indexes

Every price index this page cites measures what a property sold for. None of them measure what it cost to make that property worth selling at that price, and that gap is where a position most often goes stale fastest, because contractor pricing moves on its own schedule, disconnected from URA's quarterly release. Condominium renovation is reported at roughly S$24,800 to S$71,100 for a two-bedroom unit and S$32,200 to S$84,100 for a three-bedroom, rising to S$122,700 to S$181,100 for a penthouse — a range wide enough that quoting a single midpoint figure as "the cost of renovation" would itself misdescribe the room. For landed housing specifically, an addition-and-alteration under URA's own 50% envelope tests is reported at S$180 to S$300 per square foot of construction against S$300 to S$500 per square foot for a full rebuild — figures from contractor advisory pages, not from BCA or URA directly, and stated as such rather than borrowed as if they carried government weight.

Conveyancing: the room's clock

A price index reports a moment; a transaction takes weeks to actually close, and that lag is itself part of reading the room correctly. Where a mortgage is involved, a private purchase typically runs 10 to 12 weeks from signing the sale and purchase agreement to completion, with mortgage duty fixed at 0.4% of the loan amount, capped at S$500. One guide, updated 23 September 2026, makes a point worth repeating on a page about dated positions generally: a headline conveyancing fee is only comparable across providers when it covers the same defined scope of work — transfer, mortgage, registration and searches quoted separately — because a lower headline total that excludes work another quote includes is not actually a lower price.

Estate: a position that has to outlive the person holding it

The clearest example on the shelf right now of a dated position that refuses to guess at an outcome is the pending change to collective-sale consent thresholds. Under the current framework, an en-bloc sale needs consent from owners holding 80% of share value and strata area for developments over ten years old, and 90% for those under ten years. The Ministry of Law's own second-reading speech for the Land Titles (Strata) (Amendment) Bill 2026, delivered 8 September 2026, confirms new thresholds of 70% for developments 40 to 59 years old and 65% for those 60 years and older — but the speech itself does not state a commencement date, and secondary legal reporting states the amendments take effect only once the Minister appoints a date by Gazette notification, which had not happened as at this run. The honest position is not "the threshold is now 65%" and not "the threshold is still 80%." It is: a lower threshold has been legislated, and it is not yet law, and anyone timing a collective sale around it is timing around a date that does not exist yet.

On the personal side of estate, the same discipline applies to smaller facts that matter enormously to one family and not at all to the market: a Grant of Probate can only be applied for where a valid will exists and names the applicant as executor, a straightforward will from a professional runs S$200 to S$400, and a will's formalities — signed at its foot, witnessed by two people present at the same time, with any gift to a witness or a witness's spouse rendered void — have not changed. None of that is exciting. All of it is the kind of thing that is only ever checked after it is too late to matter, which is exactly why it belongs on a page that exists to be checked in advance. [⚠ estate.probate.eligibility_2026_06_15: sources differ; awaiting a ruling]

The twelve questions: the frame under every position published here

Every transaction page on this site runs its numbers through the same twelve-question frame, and on this page the frame does something slightly different: it is not applied to one purchase, it is the discipline that keeps a piece of market commentary honest. The two questions that actually sort a reader are not whether you are buying to invest or to live in it — nobody buys anything believing it will fall, so that question sorts nobody. They are: are you borrowing to buy, and once you hold it, are you lending it out. Cash and own-use is pure appreciation with no carry at all. Cash and let-out is yield, positive carry by construction. Borrowing and own-use is negative carry by construction — most owner-occupiers sit here and few describe it that way. Borrowing and let-out is the spread play, where the gap between rent and interest is the entire game. Every position below on this page reads differently depending which of those four households is reading it, and a position that does not say which household it is speaking to has usually smuggled in an assumption instead of stating one.

C0. The play: balance sheet or cash flow?

A price-index move is a balance-sheet fact: it changes what your asset is worth on paper and changes nothing else by itself. A rate-quote move is a cash-flow fact: it changes what leaves your account every month regardless of what the index does that quarter. A position that reports one and implies the other — "prices rose, so you're better off" said to someone with a floating-rate loan whose monthly instalment just rose too — has quietly swapped the play without saying so.

C1. Balance sheet: what a headline number does and does not change

The URA flash by segment above moved the asset side of a Core Central Region owner's balance sheet up 2.0% and a Rest of Central Region owner's down 1.4% in the same quarter. It moved nobody's liability side at all — no loan balance changed because of it. What did move for a specific household is the loan-to-value ceiling available on their next purchase, which falls with every existing housing loan already on the books, quietly changing how much of the next balance sheet they can build with debt versus cash.

C2. Cash flow: what actually moves, and when

This is a separate thing from the balance sheet and from profit, and conflating the three is precisely how a household reading only the price pages gets surprised by a cash squeeze. Stamp duty is due within 14 days of signing regardless of what the index did that month. A bridging loan's interest accrues daily whether or not the old sale completes on schedule. A position that reports the market move but not the cash clock underneath it has described only half the room.

C3. P&L: what it does to profit, which is not the same as cash

Rental income minus the expenses IRAS allows is profit; the principal portion of a mortgage instalment is cash out and is not an expense at all; stamp duty paid on day one shows up as a loss only against a future sale, not against this year's rent. A dated position that quotes a yield figure as if it were a profit figure has already made an error the frame is built to catch.

C4. Carry: income minus financing cost

This is where the four households actually separate. Cash and own-use carries nothing, because there is no rent and no interest. Cash and let-out carries the whole rent, positive by construction. Borrowing and own-use carries negative by construction — rent is zero, interest is not. Borrowing and let-out carries the gap between an average gross yield reported around 3.36% for non-landed condominiums in 2026 and whatever the prevailing financing rate is that quarter — and that gap, not the price index, is the number that actually tells a leveraged landlord whether the room is currently working for or against them.

C5. Give and get: the trade, stated plainly

You give cash upfront of at least 25% of price with a cash floor inside that, duty within 14 days, and years of exposure to Seller's Stamp Duty if you leave early. You get, depending on which household you are, a home, a rent stream, a stake in whatever the index does next, or some mix of the three. Stating the give plainly, before the get, is the part most sales conversations skip — this page does not skip it because it is not a sales conversation.

C6. Exposed to: what can move against you

Interest rates, if the loan floats. Price, but not uniformly — the same quarter that showed landed property up 2.6% showed the Rest of Central Region down 1.4%, so exposure to one segment is not exposure to the other in the same direction. Policy, because a consent threshold can be legislated and still not be law, as the en-bloc example above shows directly — a household timing an exit around a threshold that has not yet commenced is exposed to a date nobody can currently name.

C7. For how long: the horizon you are committing to

At minimum four years, because of Seller's Stamp Duty, for anyone who might sell early. Longer for an HDB flat inside its minimum occupation period. Two years or so of lock-in on a fixed bank-rate package, with a prepayment penalty typically around 1.5% for leaving early. The shortest horizon that actually applies to a given reader is the earliest exit that carries no penalty — and it is almost never the horizon the marketing around a purchase implies.

C8. Failure cascade: if it goes wrong, in what order

For either household carrying debt, it goes wrong in one order. Negative carry first — rent, if there is one, stops covering interest, or an owner-occupier's income stops covering the instalment. That hits cash flow, because a monthly shortfall now needs funding from savings or from selling something else. It accumulates into profit and loss over successive months, alongside whatever duty was already sunk on day one. If price falls at the same time the cash-flow strain becomes negative equity, where the loan exceeds what the property would fetch. Negative equity becomes illiquidity, because a forced sale at that point crystallises a loss instead of realising a gain. Illiquidity becomes insolvency only if there is no other asset to bridge the gap. The two cash-only households on the grid above cannot reach this cascade at all, because there is no financing cost to turn negative on in the first place — which is the entire reason the frame sorts on borrowing and letting-out first, and not on "investment or own use".

C9. Liquidity: how fast this becomes money again, and at what haircut

A private resale runs roughly two weeks on option, two weeks on exercise, then completion in ten to twelve weeks where a mortgage is involved. The haircut on the way out is agent commission, legal fees, Seller's Stamp Duty if inside the holding period, and — for anyone who used CPF — a refund of principal plus 2.5% accrued interest before any cash actually reaches the seller. That haircut is the standing cost of turning any position in this room back into cash, and it does not appear on a price index at all.

C10. Control: can you move the outcome, or are you a price-taker?

On the index, every household is a price-taker — nobody sets the quarter's -0.1% or +2.6%. What a household actually controls is narrower and specific to where they sit, and it is exactly the renovation, A&A and rebuild spoke above: a landed owner controls the reno-versus-rebuild decision, and that single decision wears different clothes at different points in the holding period — a tax question when it changes Seller's Stamp Duty exposure, a financing question when a land loan caps lower than a completed-home loan, a legal question the moment URA's envelope-control and BCA's approval process get involved. That is control the way this frame means it: not moving the market, but moving the asset within the market you cannot move.

C11. The exit: how you get out, and what it costs

Sell, and pay the haircut in C9. Let it out and stay invested, provided the carry in C4 actually holds. Draw against the equity through a mortgage equity withdrawal loan, itself capped by the same LTV and TDSR that governed the original purchase. Exit through a collective sale, on a timeline market guides report as two to four years from committee to completion, and currently sitting on a consent threshold that is legislated but not yet in force. Or hold to the estate, where no ABSD or BSD applies to a transfer on death, though the inherited property still counts against the beneficiary's own next purchase. None of those five exits changed this run. What can change under any of them, without warning, is the date a pending rule actually starts — which is the whole reason this page states "not yet" instead of guessing "soon".

The B2B spoke: what a company reads this page for

A company reader is not sorted by the Q1/Q2 household grid above at all — TDSR does not apply to a loan made to a company, though MAS's own look-through rule reaches a sole proprietor who incorporates purely to hold one property. What does reach a corporate buyer squarely is the 65% ABSD rate on any entity or trust acquiring residential property, a flat non-individual lending limit of 15% loan-to-value, and, on a share deal rather than an asset deal, Additional Conveyance Duty running up to 71% for the buyer and 16% for the seller on a property-holding entity, defined as one holding at least half its tangible assets in property. A corporate reader of this page is reading for a board decision, not a household decision, and the numbers that matter to them sit almost entirely outside the twelve-question frame above.

The realtor economics: what describing the room is worth, and what it costs him

This page earns nothing by itself. Nobody pays a fee for reading it, and it recommends no action for anyone to take. The value it produces shows up somewhere else entirely: the trust that a reader who checks a position here and finds it still standing six months later extends to the five doors that do end in a transaction — buy, sell, borrow, invest, landlord. This door earns the trust the other five spend. That is a real economic relationship, not a figure of speech, and it is worth stating plainly rather than leaving it implied.

CEA's own guidance is unambiguous and has been for years: there are no fixed commission rates for any property type, the amount is agreed between agent and client before work starts, and it is paid to the agency, not the individual agent. What is less comfortable to publish, and is published here anyway because the alternative is describing the room selectively: CEA reports that almost 8 in 10 consumers already know commission is negotiable, and of those who actually negotiated, 85% succeeded in lowering it. Every one of those readers is a potential client of mine. Publishing that figure does not help me collect a higher commission — if anything it invites the exact conversation that lowers one. It stays on the page because leaving it off, on a page whose entire premise is that a dated position does not get softened to suit the person publishing it, would be the one position on this page that actually failed its own test.

For scale: market-reported agent income runs to a median around S$72,000 a year, with a new agent's first year commonly S$10,000 to S$40,000 — a recruiter's figure, not CEA's, and stated as such. A typical resale commission runs about 2% on the seller's side, split with a co-broking buyer's agent; on a new launch the developer pays the buyer's agent instead, so the buyer sees no separate fee, which does not mean the advice was free — the price carries it. The plainest incentive worth naming: an agent is paid when a transaction completes, so an honest answer of wait, or do nothing — which is the correct answer on most pages on this site most months — pays nothing at all. That asymmetry does not go away because it is named here. Naming it is the only thing this page can actually do about it.

What is uncertain

• The Land Titles (Strata) (Amendment) Bill's new en-bloc consent thresholds (70% and 65%) have passed second reading but carry no confirmed commencement date; the 80%/90% thresholds remain the ones actually in force until a Gazette notification says otherwise.
• The Lorong Puntong / Sin Ming Avenue GLS award price was not stated on URA's own release, so land-cost direction from that tender cannot be read from this page.
• Renovation and A&A/rebuild cost figures throughout this page come from contractor and market guides, not from BCA or URA directly, and move on their own schedule disconnected from any government release.
• The split between an agency's cut and an individual agent's cut of a commission is not published anywhere I could find, government or commercial.
• Figures attributed to market aggregators and portals are dated as read and are not re-polled every time this page is revised; check the date on each one before relying on it.