Shanghai’s new and resale homes tell different stories

Which housing market belongs in the brief?

A consumer business investigating Shanghai could start from the positive new-home print and write a brief around a property recovery. The resale series complicates that premise: it remained below its level a year earlier. Neither series measures household confidence or discretionary spending. Read together, they identify a question for research that one headline would conceal. NBS July tables I and II.

The difference is a gap between growth rates. It does not measure a purchase-price premium between a new home and an existing home. The indices cover separate markets, with their own reference prices; they cannot be subtracted to estimate a cash premium or the value of someone’s home.

The split is narrowing, with resale still below last year

Calculating the difference between the rates reported by NBS, Shanghai’s gap narrowed from 7.5 percentage points in May to 6.3 in June and 5.0 in July. Over those reference months, new-home annual growth edged down from 3.2% to 3.0%; the resale annual decline eased from 4.3% to 2.0%. May tables, June tables, July tables.

The monthly resale indices were above 100 in each of those months: 100.6, 100.4 and 100.3. That records successive monthly increases alongside prices still below a year earlier. A brief can acknowledge improving recent prices while keeping the annual comparison visible. The data does not identify the cause of the change.

Turn the comparison into a better research question

For a home-furnishings brief, distinguish people purchasing newly built homes from existing owners considering renovation. Ask each group about timing, budgets and postponed purchases. This table gives a reason to test those differences; it does not predict either group’s spending.

For a broader consumer brief, establish housing tenure and purchase timing before using “the property market” to explain confidence. A renter, a recent buyer and a long-standing owner have different exposures that a city-level price index cannot resolve. Compare their accounts with transaction volumes, income and category sales where those data are available.

Beijing, Guangzhou and Shenzhen also had weaker annual resale readings than new-home readings in July, but the gaps ranged from 0.7 to 2.5 points. Copying Shanghai’s five-point split into a four-city study would erase part of what the comparison is meant to test.

Method and limits

The latest property period in China Snapshot’s archive when this investigation was prepared was July. We checked its four city-level new-home readings against the NBS release, then added the same cities’ resale readings and the two preceding months directly from NBS tables I and II. The downloadable extract contains 24 observations: four cities, two market types and three months.

NBS publishes the year-on-year indices with the same month a year earlier equal to 100. Subtracting 100 yields the percentage change shown here. We compare new and resale observations only within the same city and reference month. We do not average cities, mix monthly and annual rates, or infer a national trend from this selection.

The reference month, publisher’s release date and our retrieval date are separate fields in the export. All three source pages were retrieved on 14 September 2026. The NBS notes explain that new-home data comes from online transaction registrations; resale collection uses reporting and survey methods. These are city-level indices, not valuations of a constant set of individual homes.

The reproduction instructions explain the four-city extraction and include a script that calculates the comparison from the downloaded rows. Reproduce the displayed figures with yoy_index − 100 and pair the rows by city and period.

• NBS May 2026 tables / published 16 June 2026
• NBS June 2026 tables / published 16 July 2026
• NBS July 2026 tables / published 17 August 2026