01 · The questionWhat does a cooling measure actually cool?
The assumption behind every headline, tested against the record.
Four tightenings in five years. The transaction record disagrees with the standard story about what they did.
Every announcement produces the same reaction: prices will fall. It rarely happens, and the reason is visible in the data. A measure removes a category of buyer. It does not remove the seller’s reservation price, their mortgage, or their unwillingness to sell at a loss.
So the first thing that moves is volume, and volume moves hard.
Measures remove buyers, not prices. Volume takes the hit while sellers simply wait.
02 · The claimVolume is the signal, price is the lag
Narrow, testable, and occasionally wrong — which we show.
Two of these look like textbook cooling. Two look like the textbook is wrong. Both readings are in the same dataset, and section 06 deals with the awkward pair honestly.
03 · MethodHow we measured this
A deliberately crude test, because crude tests are harder to rig.
- The dataset. URA caveats, 2021-07 to 2026-07, private non-landed, ECs excluded.
- The test. Average monthly transactions for the three months before each measure against the three months after. No modelling, no seasonal adjustment.
- Why crude. Every adjustment is a place to hide an assumption. A blunt before-and-after can be checked by anyone with the same data.
- The consequence. Our figures include seasonality and any concurrent events — a limitation we state rather than smooth away (section 07).
Three months before against three months after, unadjusted — so the test can be checked rather than trusted.
04 · The evidenceVolume around each measure
Transactions per month, with every measure marked.
The dashed lines mark the measures. The drops after the first two are immediate and steep.
Now the same window for price. The line barely acknowledges the events that halved volume.
Where volume moved 36%, price moved a fraction of that. Sellers withdraw rather than discount.
05 · The mechanismWhy prices are so stubborn
Four reasons the seller does not blink.
- No forced sellers. Most owners live in the asset and have no deadline. Absent distress, they wait.
- The anchor is the neighbour. Sellers price against the last transaction in their block, which is public and recent.
- Financing is already conservative. TDSR limits leverage before a downturn arrives, so fewer owners are forced out.
- Measures target specific buyers. Removing foreign demand barely touches a suburban market that never had it.
06 · The counter-caseThe two that went the other way
The findings that contradict the headline — presented, not buried.
Two measures in our window were followed by volume rising: the April 2023 foreign-buyer duty (+47.1%) and the July 2025 SSD extension (+82.6%).
We can think of three explanations and cannot definitively separate them: the market was recovering from a low base and would have risen anyway; the measures targeted narrow buyer groups whose absence was swamped by domestic demand; or announcement timing coincided with seasonal strength. What we will not do is quietly drop these two because they spoil a tidy story.
The defensible conclusion is narrower than the headline: measures aimed at broad demand or financing move volume sharply; measures aimed at narrow groups may not move it at all.
Two of four measures were followed by higher volume. A measure only cools what it actually targets.
07 · LimitsWhat this test cannot see
Three honest weaknesses.
- No control group. We cannot observe the Singapore that had no measure, so “after” is not the same as “because”.
- Seasonality is not removed. Deliberately — but it means a measure landing before a strong quarter looks better than it was.
- Caveats lag. A transaction agreed before a measure can be lodged after it, smearing the boundary by weeks.
Correlation, three-month windows, no control group. Directionally useful; not a causal claim.
08 · Your moveWhat this means for you
Three situations, three rules.
If you are waiting for measures to lower prices
The evidence says you will wait a long time. What arrives is thinner choice, not cheaper stock.
Do not time an entry on a measure. Volume falls up to 36%; price barely moves.
If you are selling into a measure
Fewer buyers are looking, so presentation and pricing precision matter more than in a busy market.
In a post-measure quarter, price to the last comparable transaction in your block, not to the last asking price.
If a measure targets you directly
Then it is not a market event, it is a personal one — and the arithmetic changes for you specifically.
Recalculate your own duty position before reacting to a market-wide headline.
How do the current rules apply to you?
We’ll map the measures that actually touch your situation — and the ones that don’t.
- Which duties apply to your specific purchase
- The SSD clock on anything you already hold
- What the last measure did to your district’s volume
One WhatsApp message back — usually same day.
We’ll WhatsApp it shortly — usually within the day.
09 · The recordEvery measure, before and after
Average monthly transactions either side of each event.
10 · MethodSources and limits
Data: URA caveats, 2021-07 to 2026-07, private non-landed, ECs excluded. Test: mean monthly transactions, 3 months before vs 3 months after, unadjusted. Policy dates from our Singapore property policy knowledge base. Limits: no control group, no seasonal adjustment, caveat lag of several weeks — correlation only. Reproducibility: guides.py. Not investment advice.
Policy is general. Your position isn’t.
Duties, timelines and the SSD clock apply to your specific purchase. The Scenario Planner runs them on your numbers.
Map my move →