Private Residences
Search
ENRUARFRESCNJP
Stories/What the index measures
Reading another house’s measurement

What the bubble index measures

A 60-square-metre flat and a skilled worker’s wage — and why that says nothing about the buildings in this register.

The index measures a different building from the one this register holds, and says so in its own methodology. The UBS Global Real Estate Bubble Index gauges “fundamental valuations in owner-occupied housing markets” from five sub-indices — price-to-income, price-to-rent, changes in mortgage-to-GDP and construction-to-GDP, and the city-to-country price ratio. Its two published benchmarks are explicit about the unit: the number of years a skilled service worker must work to afford a 60-square-metre apartment near the city centre, and the number of years that apartment must be rented out to pay for itself. Its price series are national housing indices — Nationwide for London, the FHFA and S&P for the American cities, the Bank of Korea for Seoul.

And UBS itself says the segment this register covers has come apart from the one it measures. Its 2026 editorial: “prime residential segments are outperforming broader housing markets across many cities”, and “market overheating is becoming more localized, with prime districts often diverging significantly from broader citywide trends.” It goes further, and the sentence is almost a description of this register's subject: “established markets characterized by political stability and trophy assets are well positioned to benefit from these trends.” The authors are telling a reader that the index and the trophy asset are moving apart.

So the register did the only test it is entitled to do: it counted its own records against the index's cities. UBS covers twenty-three. This register holds published records in 14 of them, 138 records in all. Nine hold nothing — and the three emptiest are not the cheap ones. Zurich, the index's highest-risk city at 1.69, holds no record here. Nor does Seoul at 1.13, nor Lisbon at 1.04. Three of the seven cities the index calls elevated or high are, to a register of branded and landmark residences, blank.

The relationship between the two is nothing at all, and that is the finding. Across the twenty-three cities the correlation between bubble-risk score and the number of records this register holds is Pearson +0.12, Spearman -0.08. Removing Dubai — which alone carries 56 of the 138 and distorts any line drawn through the set — gives -0.11 and -0.21. On twenty-three points every one of those is indistinguishable from zero. The register states them to show it looked, not to claim a result: a correlation computed on twenty-three cities with one outlier holding more than a third of the mass is not evidence of anything, including of absence.

What can honestly be said is the weaker and more useful thing. The index's score does not tell a reader of this register whether it holds anything in that city, and the register's coverage does not tell a reader of the index anything about bubble risk. Two measurements of the same city that do not inform each other are measuring two different cities — which is what UBS's own editorial says is happening, and what a reader of either document should carry away from the other.

The register publishes this page because the temptation ran the other way. A bubble index naming Dubai, Miami and Hong Kong, in a week when this register holds 56, 12 and 19 records in those three, invites a page attaching the index's figures to the register's records. That would be the same error the register spent October correcting — a citizenship threshold printed as a residence price, a resort's room count printed as a villa's, a group's thirteen-year-old figure printed as current. A number true of one object is not a fact about another, however close the two sit on a map.

The twenty-three cities, against this register’s own count

UBS bubble-risk score, 2026 edition, against published records held by this register in the same city. Counts are computed on every build and are not typed.

Source: UBS Global Real Estate Bubble Index 2026, Chief Investment Office GWM, authors Matthias Holzhey, Maciej Skoczek and Claudio Saputelli; data collected to 26 August 2026, editorial deadline 21 September 2026. Scores and methodology quoted from that report. Record counts are this register’s own.