Thailand: Why Property Market Reports Give Different Numbers
Let’s explore this page together
Published · Updated
When two valid numbers appear to contradict each other
CBRE's Bangkok update put new condominium launches up 265% year on year in the second quarter of 2026. REIC subsequently reported a 19.5% increase in new condominium sales for that same quarter. Seen in two headlines, those numbers might look like competing accounts of a single market rebound. Read the labels, though, and the apparent disagreement starts to unravel: one series tracks supply entering the market, the other tracks units newly sold.
The publication dates also tell a useful story. CBRE issued its Bangkok Overall Figures on 18 August 2026; REIC's regional housing release followed on 22 September. Neither percentage describes what was happening on its publication date. Both refer to April–June and compare it with April–June 2025. Moving a figure from a quarterly report into a September news article does not turn it into a September observation.
A launch is the introduction of new condominium supply, while a new sale is a buyer transaction recorded in a developer-market series. Developers can launch substantially more homes while buyers continue purchasing units from projects that were already on sale. Alternatively, sales may improve while developers restrict launches. The two measures may inform the same market discussion, but they do not have to rise at the same pace or even move in the same direction. The public CBRE summary does not provide the full unit-level classification required to reconcile its launches with REIC's newly sold units.
Geography makes a direct comparison harder still. REIC specifies Bangkok plus five surrounding provinces: Nonthaburi, Pathum Thani, Samut Prakan, Samut Sakhon and Nakhon Pathom. CBRE labels its series Bangkok, without providing an equivalent province-by-province boundary for this headline in the public summary. Its 265% should therefore not be treated as a six-province result. This matters if, for example, a buyer is trying to understand a condominium in a peripheral province, not downtown Bangkok.
The striking CBRE increase also has an identifiable comparison-base effect. CBRE attributes much of it to unusually low launches in Q2 2025 following the earthquake. A 265% increase is 3.65 times the previous period's level, whatever that level was. Without the underlying Q2 2025 count, the percentage alone cannot tell a reader whether the recovered volume was historically large. It certainly does not establish that actual purchases rose 265%.
Eight headlines, eight different measures
These are not eight estimates of the same thing. The rows combine changes in activity, a price index, an outlook and a listing-site statistic, each with a different observation window or population.
| Source and metric | Published figure | Coverage |
|---|---|---|
| CBRE: new condo launches | +265% YoY; Q2 2026 | Bangkok; new supply |
| REIC: new condo sales | +19.5% YoY; Q2 2026 | Bangkok + 5 provinces |
| REIC: all housing transfers | 167,665; +17.6%; H1 2026 | Nationwide; houses + condos |
| REIC: foreign condo transfers | 6,533; −8.8%; H1 2026 | Thailand; foreign recipients |
| REIC: same series in Q2 | 3,292; +1.4%; Q2 2026 | Thailand; foreign recipients |
| REIC: new-condo price index | 161.8; +2.2%; Q3 2026 | Bangkok–vicinity index area |
| CBRE: asking-price outlook | Up to +15%; 2026 forecast | Downtown Bangkok; average |
| FazWaz: listing median | No live value confirmed | Bangkok; changing listings |
Why listing prices and indices point in different directions
Price headlines are particularly tempting because they seem ready to plug into a purchase budget. In Thailand, three figures that sound like price growth can refer to very different evidence: REIC's index of new condominiums offered for sale, CBRE's forecast for average downtown asking prices, and FazWaz's median of advertised Bangkok condos. Each describes a different statistical object. None automatically gives the final amount paid to register a particular transfer.
Take the REIC figure published on 7 October 2026. The Q3 new-condominium asking-price index stood at 161.8, up 2.2% from Q3 2025 and 0.1% from Q2 2026. An index level is not a baht-per-square-metre quote. Nor does a 2.2% annual rise imply that every individual unit appreciated by 2.2%. The series is about prices offered for new condos within REIC's index coverage, not resale properties of every age or achieved prices on completed transactions.
There is a further boundary worth preserving. REIC labels the index's territory Bangkok and its vicinity, but the accessible Q3 summary does not enumerate every province and weighting used in the composite. Earlier releases of this index separately discussed Bangkok and selected neighbouring provinces. That is insufficient grounds to assign to the index the full Bangkok-plus-five-province boundary used in REIC's Q2 new-sales release. Before comparing two statistical areas, the underlying maps or exact coverage definitions need to agree.
CBRE's 15% figure is not another Q3 index reading. Its Thailand market outlook, issued on 24 February 2026, anticipated that average asking prices in downtown Bangkok could increase by up to 15% year on year during 2026. The words 'expected' and 'up to' are doing substantial work. As at the 11 October review of these publications, this remains a statement about the outlook published in February, not a verified full-year outcome. CBRE also expected more luxury and super-luxury launches; its location, property mix and averaging method are not the same as the published REIC index.
Even two asking-price statistics can diverge without one being defective. Consider a purely illustrative market in which a number of high-end projects start advertising large, expensive apartments. The average asking price of the advertised new stock can rise because the composition of available homes has shifted. A price index using a defined methodology might behave differently, while the advertised price of a particular older one-bedroom flat stays where it was. This example explains a possibility, not a measured explanation of the difference between REIC's 2.2% and CBRE's forecast. The public summaries do not contain enough detailed observations and weights to prove that particular causal story.
Listing portals add another complication. FazWaz's Bangkok condos-for-sale page displays both a 'median sales price' and a median price per square metre. Despite that wording, the figures appear in the setting of homes being marketed for sale. The accessible page does not establish that the median consists of registered, closed transactions. When listings arrive, expire, are repriced or move under different filters, a portal median may move even though no like-for-like property has changed value. It can also conceal enormous variation between districts, building ages and unit sizes.
We have not included a numerical FazWaz median as a fresh figure for 11 October 2026. A currently reproducible, timestamped capture of its live filtered inventory was not established; an older accessible page copy is not evidence of what was displayed at today's precise time. A defensible snapshot would need the observation time, currency, geography, listing filters and price-per-square-metre field. Even then, it would be a snapshot of advertised inventory, not a completed-sales database, and duplicate or stale advertisements could affect the statistic.
For a seller deciding an asking price, the evidence is often much narrower than any citywide median. A 45-square-metre unit should be compared with genuinely similar units in its building or immediate area, distinguishing floor, outlook, condition, tenure and furnishing. The total ticket and the price per square metre tell different stories; advertised incentives and seller circumstances further complicate the amount eventually agreed. REIC provides a market index, CBRE provides a dated expectation and FazWaz provides asking-price context. Treating any one of them as a guaranteed achievable price creates false precision exactly where a buyer needs it least.
What is worth verifying
A metric is more useful when its boundaries can be checked. These are encouraging signals, not proof that a conclusion is correct.
Observation period and release date stated
Why it mattersCBRE published its Q2 figures on 18 August 2026; REIC released its Q3 price index on 7 October. Release dates are not the months or quarters measured.
Geography and property type defined
Why it mattersREIC explicitly covers Bangkok and five provinces for new sales. The geographic shorthand on an index or CBRE summary does not establish that both cover the same six areas.
Denominator and event timing disclosed
Why it mattersREIC labels new condominium sales separately from registered foreign ownership transfers. Both can use year-on-year comparisons, yet their underlying events take place at different stages.
Revisions and missing data identified
Why it mattersPublished revision notes and exclusions help readers assess changes over time. CBRE’s short public launch summary leaves some classification details open, and an unarchived listing median cannot be recreated retrospectively.
Launches, developer sales and registered transfers are different events
In a property development, commercial interest and legal ownership can be separated by a long stretch of time. Three expressions in Thai market reports mark different points along that path: a new launch, a new sale and an ownership transfer. They are related, but a dwelling does not necessarily move through all three within a reporting quarter. A resale home, for example, can change owners without its building appearing in any new-launch series. A launch signals new marketed supply, not completed buyer transactions.
REIC's new-sales measure sits further along the commercial process. For condominiums across Bangkok and its five surrounding provinces, it reported 19.5% year-on-year growth in Q2 2026. This is a defined developer-market sales indicator, and it is useful for understanding how new housing stock is being absorbed within that survey area. It is not a nationwide count of all home transactions and should not be silently extended to second-hand properties. Most importantly, its date of sale need not match the date on which a legal title will be registered.
Ownership-transfer statistics capture that later event. In a release dated 27 August 2026, REIC reported 167,665 residential ownership transfers across Thailand during the first six months of the year, an increase of 17.6% on H1 2025. Of those, 111,624 involved low-rise housing and 56,041 involved condominiums. Condo transfers were up 24.1% year on year. The figures are nested categories: the 56,041 condo units are part of the 167,665 housing total. Adding them together as if they were separate markets would double-count the condos.
A second REIC release, dated 11 September, narrows the same general type of event to condos transferred to foreign recipients. The H1 total was 6,533 units nationwide, down 8.8% from H1 2025 and equivalent to about 11.7% of all condo transfers recorded in the broader series. That subgroup can contract while overall transfers rise, because local and foreign recipients, as well as low-rise homes and condos, make up different portions of the overall activity. Foreign condo transfers also should not be treated as a census of everything overseas buyers considered, booked or intended to buy.
Quantity and transaction value introduce a different denominator again. REIC recorded 3,292 foreign condo transfers during Q2 2026, up 1.4% from Q2 2025. The aggregate value associated with those transfers was THB 14,803 million, up 20.2% over the same interval. More money in total does not establish that the price of the same apartment, or the price per square metre, rose by 20.2%. Unit sizes and the composition of transferred homes can change the total. REIC also reports floor area, illustrating why counting units, baht and square metres are distinct ways to describe a period.
There is a real-world reason for the timing gap. An off-plan buyer might sign long before handover and registration, while a resale purchase may move to registration on a different schedule. A quarterly registration count therefore describes transactions completed at the Land Department stage, not necessarily the moment buyers first committed. For investors assessing the stock developers have put on the market, ownership-transfer data answer a different question. For an owner checking how active completed transactions have been, launches provide only distant context. The choice should follow the decision, not whichever headline sounds strongest.
Which option fits your situation
Launches indicate marketed supply, not a guaranteed pipeline of completed homes. REIC permit data relate to a separate administrative event and should not be added to launches.
Use matched quarters and the same REIC survey area. A developer’s stock figures for one building can add detail but are not an independent measure of the wider market.
These are registered changes of ownership, not the date of booking. Separate condos from houses and foreign transferees from the total.
Use dated listings for similar floor areas, locations and condition, including price per square metre. Asking prices are negotiation evidence, not confirmed closing prices.
How a quarterly rise can coexist with a half-year decline
One REIC series gives a clean demonstration of how apparently opposite headlines can both be right. Nationwide transfers of condominiums to foreign recipients rose 1.4% year on year in Q2 2026, yet the total for January–June fell 8.8% year on year. Unlike CBRE's launches and REIC's new sales, these figures measure the same kind of registered event in the same country. Only the observation window changes.
The quarterly counts tell the story. Q1 2026 recorded 3,241 transfers, a 17.3% drop against Q1 2025. Q2 recorded 3,292, an increase of 1.4% against Q2 2025. The first-half unit count is simply 3,241 + 3,292 = 6,533. A reporting period can accumulate units from its constituent quarters; it cannot accumulate percentage changes in the same way, because the percentages use separate denominators.
The 2025 numbers are the missing piece in many short news reports. REIC's earlier figures show 3,919 foreign condo transfers in Q1 2025 and 3,248 in Q2 2025, totalling 7,167 for that half-year. Q1 2026 therefore fell by 678 units, while Q2 gained only 44. Across both quarters, 634 fewer units changed hands than a year earlier. Dividing that decline by the 7,167-unit H1 2025 base produces approximately −8.8%: (6,533 − 7,167) / 7,167.
That calculation also explains why a simple average of −17.3% and +1.4% would be wrong. A year-on-year quarterly growth rate uses the matching quarter of the previous year as its base, and those bases differ. These are not quarter-on-quarter changes, either: Q2's +1.4% compares April–June with the same months of 2025, not with January–March 2026. Replacing the underlying counts with an average of rounded percentages discards exactly the information needed to calculate the half-year outcome.
The conclusion is more restrained than either headline alone. Q2 performed a little better than its year-earlier counterpart, while the considerable Q1 decline still outweighed that gain in the half-year result. Neither number predicts July–December, and neither proves that every part of the foreign-buyer market improved or weakened together. Here REIC is not contradicting itself. It is reporting distinct time windows on a consistent registered-transfer measure.
Common mistakes and how to fix them
What it costsBuyer demand may look stronger than recorded, distorting expectations for sell-through and legal completion.
What to do insteadDistinguish CBRE launches from REIC new sales and transfers, and align the observation area and quarter.
What it costsA forecast can become an unjustified claim that an individual home has already gained value.
What to do insteadKeep CBRE’s up-to-15% as a 2026 expectation; observed performance requires a later matching measure.
What it costsIt can falsely suggest official inconsistency or hide the weaker six-month total.
What to do insteadAdd the quarterly unit counts, then compare that total with the matching combined 2025 base.
What it costsA property valuation or return model may use an unsupported entry or exit price.
What to do insteadLabel FazWaz as a dated asking-price reference; use verified comparable transactions where legitimately available.
Which discrepancies public data cannot resolve
Clearer definitions do not necessarily make a study reproducible from public data. The accessible CBRE Q2 Bangkok summary identifies the increase in condo launches and explicitly links it to the low Q2 2025 base following the earthquake. It does not supply the complete underlying launch count, project list, submarket coverage or treatment of separate phases in the short text. More detailed material may exist elsewhere, but it cannot be inferred from a headline. Nor can we decompose the gap with REIC's sales growth into a numerical 'geography effect' and 'event effect' without matching observations.
The Q3 REIC price-index summary provides a clear period and headline readings: 161.8, up 2.2% year on year and 0.1% quarter on quarter. It does not, on its own, furnish enough detail to reconstruct the contribution from each locality, the complete index boundary or the treatment of every developer incentive. What can be supported is that this is an index for new condos being offered for sale in the area named by REIC. What cannot be supported is an assumption that it shares the exact six-province perimeter of the separate new-sales survey, or that it identifies an achieved sale price for a particular unit.
A property-listing site presents a different transparency problem. FazWaz may display the median asking-price context visible to users along with live condo advertisements. Without a reproducible capture of the relevant listings as they stood on 11 October 2026, changes in the median cannot be attributed to any specific market event. Properties can disappear, prices can be edited, and the set of units included by a filter can change. The accessible page copy does not establish a fresh live numerical value for that date, which is why this comparison deliberately leaves the number blank instead of dressing an older observation as current.
Published statistics can also be revised. The date a REIC release was issued is distinct from the dates on which its recorded ownership transfers occurred; two versions of a series may differ if later corrections were made. No particular such revision has been verified here as the cause of the figures discussed. A media story repeating a CBRE or REIC release is also not a second independent dataset. These are limitations to retain in the explanation, not evidence that either publisher is deceptive or systematically less accurate.
The comparison of these specific publications did not establish a genuine conflict between two measures with identical events, territories, periods, units and definitions. That is a narrow conclusion, not a verdict on every Thai property report. For someone negotiating a purchase, the remaining question is usually more concrete: how does this apartment compare with genuinely similar homes available on the same date, and is there lawful, reliable evidence of completed prices? A market index supplies context, an outlook sets out a scenario, and advertising reveals sellers' initial expectations. Where the underlying completed-sale information is unavailable, an honest estimate should retain that uncertainty instead of pretending to have discovered the one true market price.
Expert view

A 265% rise in condominium launches is the kind of headline that can change a buyer’s expectations before it changes the evidence. It means that much more supply was launched than in the comparable quarter, not that purchases rose by the same amount. CBRE explicitly links the jump to an unusually weak second quarter of 2025 following the earthquake. Before treating that percentage as proof of stronger demand, I would compare new sales within a matching area and period. Registered ownership transfers would answer a later question, since title can change hands well after the initial sale agreement. Without the underlying launch count and precise market boundary, the number should remain a limited observation, not a verdict on Bangkok property.
Sources and check dates
Sources4
- REIC — official market statistics
Release of 22 September 2026: new condo sales +19.5% YoY in Bangkok and five surrounding provinces; not a launch count.
Original title: REIC — Bangkok and five adjacent provinces, Q2 2026 housing market (22 September 2026) - REIC — official market statistics
Release of 27 August 2026: 167,665 housing transfers (+17.6%), including 56,041 condos (+24.1%) and 111,624 low-rise units.
Original title: REIC — Nationwide housing ownership transfers, H1 2026 (27 August 2026) - REIC — official market statistics
Dated 11 September 2026: foreign condo transfers were 3,292 in Q2 (+1.4% YoY), 6,533 in H1 (−8.8%); Q2 value THB 14,803m (+20.2%).
Original title: REIC — Foreign condominium ownership transfers, Q2 and H1 2026 (11 September 2026) - REIC — official market statistics
Q3 2026 new-condo asking-price index 161.8 (+2.2% YoY, +0.1% QoQ); the accessible summary does not fully specify geographic weights and perimeter.
Original title: REIC — New condominium asking-price index, Q3 2026 (7 October 2026)
Useful next reading
A practical second opinion
Need help assessing a specific property?
Send the property link, price and your goal. We will separate the documented facts from the points that still need a property-specific check.
What to send: link, budget and target date