Methodology¶
日本語 (full version) · All areas
What the market phase means¶
Each area is placed in one of five phases: Overheating / still rising → Plateau (high prices, easing demand) → Early correction → Correction under way → Declining.
When a housing market starts to cool, volumes usually move first and prices follow later: unsold inventory builds up, fewer homes sell, and asking prices are cut before closing prices start to fall. The score therefore puts most weight on volume indicators. An area can be in a correction phase while closing prices are still above last year.
Adjustment pressure (0-100)¶
Each indicator is turned into a stress score from 0 (none) to 3 (strong), and the weighted average is scaled to 0-100.
| Indicator | Weight | What raises the score |
|---|---|---|
| Gap between asking (new listings) and closing prices per m² | 2.0 | A wider gap (above 12%, 18%, 25%) |
| Closing price per m², year on year | 2.0 | Slower growth or falls (below +2%, -2%, -10%) |
| Inventory, year on year | 1.5 | Growth (above 0%, +5%, +20%) |
| Closed sales, year on year | 1.5 | Declines (below 0%, -5%, -15%) |
| Consecutive months of falling closing prices | 1.0 | 1, 2-3, 4 or more months |
| Asking price (70 m² basis), year on year | 1.0 | As for closing prices |
| Share of listings with price cuts; new-build first-month contract rate | 1.0 each | Higher cuts; lower contract rates |
| MLIT quarterly closing price and closed sales (only where no monthly data) | 1.0 and 0.75 | As above, year on year |
- Bands: 0-20 overheating, 20-40 plateau, 40-60 early correction, 60-80 correction under way, 80-100 declining.
- Confidence: high with 5 or more indicators, mid with 3-4, low with 2 or fewer; fewer than 2 is shown as insufficient data.
- Areas without monthly statistics get a simplified call from MLIT quarterly closing data, marked "(simplified)".
- Monthly market statistics come from the regional REINS networks and other private sources. Under their terms, their figures are shown only to members; public pages show the resulting phase and the strength of each sign, not the figures.
From the centre outwards¶
To see how a correction spreads within a city, we split cities the same way the REINS monthly area tables do:
- Tokyo 23 wards: Central 3 Wards (Chiyoda, Chuo, Minato) → Josai (Shinjuku, Shibuya, Nakano, Suginami) → Jonan (Shinagawa, Meguro, Ota, Setagaya) → Johoku (Bunkyo, Toshima, Kita, Itabashi, Nerima) → Joto (Taito, Sumida, Koto, Arakawa, Adachi, Katsushika, Edogawa).
- Osaka City: central six wards → the other 18 wards. Kyoto City: central six wards → whole city. Kobe City: central three wards → whole city.
For each area we show the first month of the current run at 40 (early correction) or above. If the score dropped below 40 on the way, the count restarts. Past values are recomputed with the current method using only data up to each month, so they are not the values we published at the time. No date means the area has not reached 40 yet. Like the phase itself, this describes overall supply and demand, not the price of any single property.
For Nagoya and Fukuoka there are no monthly in-city tables, so their pages only compare MLIT quarterly year-on-year changes for the city centre against the whole city.
Market-phase calls are our own interpretation, not investment advice.