UBS Bubble Index 2026: Reading the Risk Map

What Does the UBS Bubble Index Actually Measure?#

The UBS Global Real Estate Bubble Index is UBS's annual attempt to grade housing markets for overvaluation, not to call a top. Now in its twelfth year of publication, the 2026 edition tracks residential prices across 23 major cities, with Lisbon and Seoul added to the sample for the first time (UBS, Global Real Estate Bubble Index 2026, 22 September 2026). UBS defines a bubble as "a significant and sustained overvaluation of property prices," with warning signs it lists as "a disconnect between prices, incomes, and rents, alongside excessive credit growth and construction activity" (same source).

The output sorts the 23 cities into four bands: high risk, elevated risk, moderate risk, and low risk. Zurich and Tokyo carry the high-risk label this year. Miami, Dubai, Seoul, Geneva and Lisbon sit in elevated risk. Milan and Madrid are moderate, UBS notes, with imbalances increasing. London and New York are rated low risk (UBS, Global Real Estate Bubble Index 2026, 22 September 2026). None of the bands is a timeline. UBS states that the index does not predict the timing of price declines (same source).

Reading a city-level score is a screening step, not a verdict on a segment. The index prices an average home, sold to an average buyer, financed the way that market usually finances a purchase. Prime addresses, resort second homes and off-plan pre-completion stock are a different transaction, with a different buyer, different leverage and often a different currency of ambition. UBS's own list of 23 cities makes the boundary explicit: markets outside the sample carry no UBS score at all, not a hidden low one.

The chart below ranks all 23 cities by their 2026 score. Dubai, fourth at 1.16, sits in the elevated band. Milan, at 0.50, sits exactly on the lower edge of the moderate band. Both are marked because they are the sampled cities nearest to Ras Al Khaimah and Lake Como, not because their scores describe either market.

Zurich and Tokyo Head the High-Risk List. What Does That Signal?#

Zurich and Tokyo are the only two cities UBS places in the high-risk band for 2026, at scores of 1.69 and 1.54. UBS describes Zurich as the city with the strongest price increase in the study over 20 years, driven by low interest rates and tight supply, with vacancy close to zero and an exceptionally high and still-rising price-to-rent ratio that leaves the market increasingly dependent on low financing costs. In Tokyo prices have outpaced incomes after a long run of gains, and UBS warns that further increases in financing costs could weaken the investment appeal of ownership (UBS, Global Real Estate Bubble Index 2026, 22 September 2026). High risk, in UBS's language, describes the gap between prices and the fundamentals meant to anchor them, not the presence of foreign capital or the absence of construction.

A high-risk score has coexisted with a strong prime segment before, and it does again this year. Tokyo's mass-market housing stock, the one UBS prices, is not the same asset as a branded penthouse near Ginza. Knight Frank's Prime Global Cities Index, tracking high-end values across a 47-city basket, put Tokyo's luxury segment up 50.7% in the twelve months to June 2026, the sharpest move in the basket, with a 12.6% rise in the second quarter reversing an 8.6% drop in the first (Knight Frank Prime Global Cities Index, reported by World Property Journal, 22 September 2026). One city, two different signals, depending on which slice is measured.

The lesson for an allocator is about the label, not the city. "High risk" on a bubble index is a statement about a broad, owner-occupied, mortgage-financed market. It travels badly when applied whole-cloth to a segment the index was never built to price. A branded residence or a resort villa is financed differently, sold to a different buyer pool, and priced against a different set of comparables than the median owner-occupied flat UBS is actually measuring. Zurich's and Tokyo's mass housing markets and their prime pockets can move in opposite directions in the same quarter, and both readings can be true at once. Treating a single city-level score as a verdict on every asset class within that city collapses a distinction the index itself is built to preserve.

Is Dubai's Bubble Risk Rising or Easing?#

Dubai's answer is both, depending on the window. UBS's 2026 score for Dubai is 1.16, up from 1.09 a year earlier, placing the emirate fourth among the 23 sampled cities and inside the elevated-risk band (UBS Global Real Estate Bubble Index 2026, reported by The National, 24 September 2026). Read year over year, the direction is up. Read against the index's own commentary, UBS also says bubble risk in Dubai "has eased since March," a signal that the trajectory within 2026 has been cooling even as the annual comparison still shows a higher score than 2025.

The components behind the score cut against a simple headline. Inflation-adjusted home prices in Dubai rose only 0.4% in the year to the second quarter of 2026, UBS reports, while real rents fell 4% over the same period. A market where rents are softening faster than prices are rising is not a market accelerating into a classic price-income-credit spiral; it looks more like one where supply is catching up with demand, at the specific moment UBS takes its snapshot.

Affordability, on UBS's own comparison, remains a genuine structural feature. Dubai still requires around five years of income to buy a 60-square-metre apartment, among the more attainable ownership costs of the cities in the sample (UBS Global Real Estate Bubble Index 2026, reported by The National, 24 September 2026). Elevated risk and comparative affordability are not a contradiction; they describe a market that has run up from a lower base and is still, on this one metric, less costly to enter than most of the cities ranked above it.

Miami, Seoul, Geneva, Lisbon: The Elevated Cohort#

Four cities share Dubai's elevated-risk band, for four different reasons. Miami ranked highest in the index in the 2024 and 2025 editions and has now fallen back into the elevated band, at 1.41. Seoul and Lisbon are new to the 23-city sample this year, both in the elevated band, and UBS reports that they recorded the largest increase in bubble risk this year (UBS, Global Real Estate Bubble Index 2026, 22 September 2026). Geneva shares with Zurich the effect UBS attributes to low Swiss financing costs, which continued to add to risk in both cities, without crossing into the high-risk band.

Elevated is UBS's second-highest classification, one notch below Zurich and Tokyo. It is not a forecast that any of the four corrects, and UBS's own release attaches no timeline or probability to the label. What the band does say is that the gap between prices and the incomes, rents and credit that should support them has widened enough, on UBS's methodology, to warrant attention. Four different housing markets, four different demand drivers, and one shared statistical flag. UBS's own city notes show how different the drivers are. Miami, after the strongest real price growth in the sample over 15 years, is cooling under high borrowing costs and rising insurance premiums. Lisbon's real prices rose by nearly 7% a year over the past decade, the fastest in the sample, and by another 10% since mid-2025, while rental growth has stalled (UBS, Global Real Estate Bubble Index 2026, 22 September 2026). Reading the four together as one cohort risks flattening exactly the differences that would matter to an allocator underwriting a specific transaction in any one of them.

Seoul and Lisbon entering the sample for the first time also changes the shape of the index itself. A 23-city sample with two new members is not directly comparable, city for city, with the 21-city 2025 edition. UBS's own note on the addition is a reminder that even a well-established index has a moving perimeter, and that a city's absence from an earlier edition never meant it was risk-free, only that it had not yet been measured. The same logic applies beyond UBS's own perimeter. A market that has never appeared in any bubble index, anywhere, is not thereby validated as sound; it is simply a market no comparable methodology has yet tested against its own prices, incomes and rents.

What Does the UBS Index Not Sample?#

Ras Al Khaimah, Lake Como, Zanzibar and Bali carry no UBS bubble score, positive or negative, because none of them are in the sample. The 2026 edition covers 23 cities, a list of large metropolitan housing markets (UBS, Global Real Estate Bubble Index 2026, 22 September 2026). A resort market financed largely by cash and foreign buyers, with a small owner-occupier base, does not fit the index's own inputs: price-to-income for a local skilled worker, or mortgage growth against a domestic economy, both assume a market this index was not designed to describe.

Absence from the sample is not a clean bill of health, and it is not a warning either. It means the specific statistical test UBS runs, a gap between prices and the incomes, rents and credit of the resident population, has not been run on that market at all. An allocator underwriting a resort position in Ras Al Khaimah, on Lake Como, in Zanzibar or on Bali needs a different diligence stack: absorption, buyer concentration, currency of the price list, and the credibility of the completion timeline, not a city-level score built for owner-occupied Zurich apartments.

The prime segment inside the sampled cities faces the same gap. UBS's headline score for Dubai is a mass-market number; Knight Frank's tracked prime segment for the same city moved far faster over the same period, in nominal terms. Multiply that gap by a market UBS never samples at all, and the case for treating "bubble risk" as a single, citywide verdict gets weaker, not stronger. The practical implication is not that bubble indices are unreliable; it is that they answer a narrower question than headlines usually credit them with, and an allocator who reads past the headline number gets a materially different, and more useful, picture of the same city.

Milan and Madrid: Moderate Risk, With Imbalances Building#

Milan and Madrid sit in UBS's moderate-risk band for 2026, one step above London and New York and two below Zurich and Tokyo. They reached it from different directions. Milan was rated low risk in the 2024 and 2025 editions and has moved up into moderate at 0.50, exactly on the line UBS draws between the two bands. Madrid was already moderate in 2025 and now scores 0.86, closer to the elevated threshold. UBS flags imbalances as increasing in both cities (UBS, Global Real Estate Bubble Index 2026, 22 September 2026; UBS, Global Real Estate Bubble Index 2025, 23 September 2025; UBS, Global Real Estate Bubble Index 2024, 24 September 2024). Moderate is the band UBS assigns to markets it does not yet consider elevated, not a statement that the trend line is flat. A city can spend several consecutive editions in the moderate band while its underlying imbalance measures keep widening, right up to the point where UBS reclassifies it as elevated; Miami's own history in this index shows that the move between bands can happen within a small number of years.

Milan's housing market has been reshaped by external demand and infrastructure spending this cycle, a dynamic covered in more detail in Victaura Research's reading of Milan's prime segment after the Winter Games. A moderate UBS score for the city's broad housing stock and a hot prime segment around the same event calendar are two separate readings, on two separate populations of buyers, and the gap between them is the same pattern already visible in Zurich, Tokyo and Dubai. The pattern recurs often enough across this index that it is worth treating as a standing feature of how mass and prime markets relate, rather than a coincidence specific to any one host city or one event calendar.

Madrid's moderate score comes with one of the strongest price runs in the sample. UBS reports real house prices up by around 10% over the past year in Seoul, Lisbon, Madrid and Hong Kong (UBS, Global Real Estate Bubble Index 2026, 22 September 2026). UBS's release does not break down which driver, price-to-income, price-to-rent, credit growth or construction, is doing the most work in Madrid's number. What the release does say plainly is that the direction of travel is toward more imbalance, not less, even from a base UBS still calls moderate rather than elevated. Spain's broader housing debate, over supply shortages and the pace of new construction relative to household formation, is consistent with a construction-side pressure on the score, though UBS's release does not isolate Madrid's specific mix of causes from the aggregate number it publishes.

Why Are London and New York Rated Low Risk?#

London and New York are two of UBS's five low-risk cities in the 2026 index, alongside Paris, San Francisco and São Paulo. For London specifically, UBS attaches a number to the story: inflation-adjusted home prices are now more than 15% below their 2021 level (UBS, Global Real Estate Bubble Index 2026, 22 September 2026). A market can be rated low risk on a bubble index precisely because it has already worked through years of real price decline, not because prices are cheap in any absolute sense.

Low risk, on this index, describes the absence of a statistical gap, not the presence of a good entry point. London's real prices falling for several years is the mechanism that closed the gap UBS measures between prices, incomes and rents. It is the opposite of a market accelerating into a bubble, and also the opposite of a market handing out discounts, since the nominal price a buyer pays has not moved anywhere near as much as the real, inflation-adjusted figure UBS quotes.

New York's low-risk score coexists with a soft prime segment, a slice UBS's methodology does not track directly. Knight Frank's prime segment for New York was down 0.4% in the twelve months to June 2026, and London's was down 3.6% over the same window, while Beijing, Toronto and Wellington posted the weakest annual results in the basket (Knight Frank Prime Global Cities Index, reported by World Property Journal, 22 September 2026). Low bubble risk and a soft prime market are two readings pointing in a similar direction for once, worth noting precisely because it is the exception, not the rule, elsewhere in this index.

Bubble Risk Is Not a Crash Forecast#

UBS defines a bubble as a sustained overvaluation of prices relative to fundamentals, and is explicit that the index flags a gap, not a date. The warning signs UBS lists are "a disconnect between prices, incomes, and rents, alongside excessive credit growth and construction activity" (UBS, Global Real Estate Bubble Index 2026, 22 September 2026). None of those variables carries a forward-looking probability. A city can sit in the high-risk band for several consecutive editions without a correction, and a city can move from moderate to elevated to high without the index ever specifying a trigger. The record is not neutral either. UBS reports that the markets it identified as high risk in 2021 subsequently recorded the sharpest price declines, averaging roughly 15% since then, and that in over half of the cities analysed housing did not protect against inflation over the last five years (UBS, Global Real Estate Bubble Index 2026, 22 September 2026).

The index also compresses very different housing systems into one comparable number, which is its strength and its limit at the same time. Zurich's supply-starved, rate-sensitive market and Miami's migration-driven one, now squeezed by borrowing and insurance costs, are structurally unlike each other, yet both can register elevated or high scores through different combinations of the same inputs. Comparability across 23 cities is what makes the index useful for a first screen; it is also why the number alone cannot substitute for market-specific diligence once a specific address is on the table.

For the broader resilience case for prime property, the relevant point is narrower still. A UBS bubble score describes the mass, owner-occupied, mortgage-financed segment of a named city. It is one input among several an allocator should hold when weighing a specific transaction, alongside prime-segment data, currency exposure, and the liquidity of the asset class in question, never the only one. Treated as a single line in a wider diligence file, next to broker-reported prime comparables and a market's own transaction history, the UBS score does useful work. Treated as the entire file, it answers a question the allocator was not actually asking.

Reading Dubai's Premium-Segment Caveat#

UBS's own commentary on Dubai ties the market's next move to a variable outside real estate altogether. The bank notes that "uncertainty about a recovery of the inflow of high-income earners" is expected to weigh on Dubai's premium property segment, tying that inflow to the broader regional conflict, while stating that the emirate's structural advantages "remain firmly intact" (UBS, quoted in The National, 24 September 2026). That is UBS's own hedge on its own elevated-risk score, not a projection of when, or whether, high-income arrivals resume.

The conflict UBS references has no settlement as of 27 September 2026. No truce is in force. The June memorandum between the United States and Iran broke down on 8 July, the US blockade of Iranian ports remains in place, and on 27 September President Trump rejected Iran's seven-day plan to reopen the Strait of Hormuz (Al Jazeera, 27 September 2026). The Council on Foreign Relations' Global Conflict Tracker, updated 11 September 2026, rates the conflict as worsening, and the dated record is kept in our revised Ras Al Khaimah note. That status is a fact about the region, not a forecast Victaura Research is positioned to make. It is the swing factor UBS itself attaches to Dubai's premium segment, adding that an improvement in the geopolitical environment is likely to support a rapid recovery in market sentiment and price expectations (UBS, quoted in The National, 24 September 2026).

Tying a premium-segment recovery to migration flows rather than to the mass-market score is a more precise read than "Dubai is at risk" or "Dubai is safe." UBS's 1.16 score describes the broad market. The premium caveat describes a narrower, income-driven population the headline score does not isolate. Both statements can be true of the same city, in the same release, without contradicting each other. An allocator who collapses the two into a single headline, in either direction, discards the more precise of the two readings UBS actually published, and the caveat is the one that speaks directly to the segment where large-ticket resort and second-home transactions typically sit.

What the Allocator Takes From a City-Level Index#

A bubble index is a screening tool, and screening tools are meant to be read at the level they were built for. UBS's 23-city, mass-market, mortgage-financed sample says something true and specific about Zurich, Tokyo, Dubai and the other 20 cities on the list. It says nothing, directly, about a resort position, a prime address, or a market like Ras Al Khaimah, Lake Como, Zanzibar or Bali that was never sampled. The sample moves between editions, from 25 cities in 2024 to 21 in 2025 and 23 in 2026, and none of the 2026 cities is a resort market. Nothing in this release suggests that is about to change.

The discipline is in not stretching the label past its evidence. "Elevated risk" for Dubai is a statement about the gap between prices, incomes, rents and credit in the city's broad housing market, measured by UBS at a specific point in time, with a specific set of inputs. It is not a statement about a specific building, a specific developer, or a resort market outside the city, with a different buyer base and a different mix of cash and mortgage purchases. The same discipline applies in the other direction: a low-risk score for London or New York is not a statement that every asset in those cities is fairly priced, only that the city-wide gap UBS measures has narrowed or never opened in the first place.

Reading the index correctly means holding two facts at once. A high or elevated UBS score is real information about where a mass housing market sits relative to its own fundamentals. It is not a substitute for asset-level diligence on the specific transaction an allocator is actually underwriting, and it was never built to be one. The index earns its place in a diligence file precisely by staying in its lane: one dated, methodologically consistent read on 23 mass housing markets, no more and no less.

Skin in the game disclosure. Victaura, through its parent Greystone B.V. (Netherlands), holds an active operating position in prime resort property. Readers should assume commentary may be influenced by, or benefit, Greystone's position. This document is classified as marketing material under MiFID II Article 24(3). It is not investment advice.