An average is nobody's property
The latest published real estate price index is for December 2025, and publication has been postponed since. But even without the delay, the index is not the price of your own property. We trace back three things: the population, the conditions held equal, and the timing.

BENERU writes for readers who live in Japan. Taxes, loans and the other rules described here are those of Japan, and they assume a resident of Japan.
This page is a translation of the Japanese original. If the two differ, the Japanese original prevails.
Good day.
A notice had appeared on the Ministry of Land, Infrastructure, Transport and Tourism's real estate price index page. It was dated October 29, 2026 (Reiwa 8). A malfunction had occurred in the program used for the calculations, and the data for July 2026 (Reiwa 8) and the second quarter, scheduled for publication on October 30, was postponed. The data for January to June 2026 (Reiwa 8) and for the fourth quarter of 2025 (Reiwa 7) and the first quarter of 2026 (Reiwa 8), which had been scheduled from April onward, was also postponed. No new publication date has been set.
The latest index you can read today is the one for December 2025 (Reiwa 7). It was published on March 31, 2026, and the nationwide residential composite was 148.0 (up 0.5% from the previous month, seasonally adjusted). In terms of the month the transactions took place, that is December 2025, 11 months before today.
"The price index is rising, so the property I own is rising too." I often see it put this way.
But this is a story about a malfunction. Even without the delay, the index is not the price of your own property. The word I want to take apart this time is "average."
Where was that average counted?
Every published number has a population.
The Land Price Publication gives the prices of standard sites as of January 1, and in 2026 (Reiwa 8) it covered 25,565 of the 26,000 points nationwide. The prefectural land price survey is as of July 1, with 21,466 benchmark sites (of which 10 are suspended). The real estate price index is built from about 300,000 transaction price records a year.
Is your own property in any of these? The points for the Land Price Publication and the land price survey are fixed, and it is normal not to be among them. The population of the real estate price index is transactions, not points, so if you sell, your property may be included (limited to transactions collected through questionnaire surveys of the parties to the transaction and similar means). While you keep holding it, it is not included. What a number represents is that population only.
An average is a number after the conditions have been held equal
The real estate price index does not simply average transaction prices. It uses a hedonic regression method and, as the methodology document states, splits housing price changes into those caused by changes in attributes and "pure price changes." The movement is what remains after adjusting for floor area, distance from the nearest station, building age, municipality, zoning, and so on.
Your property's building age went up by 1 year in 1 year. The index, for detached houses and condominiums, includes building age as an explanatory variable and removes its effect (it is not used for residential land alone). To read a 5% rise in the index as a 5% rise in your own property, you would have to assume that your building age has not increased.
The box it sits in may differ from the box you think
The real estate price index (residential) covers 4 uses: residential composite, residential land, detached houses, and condominiums (sectional ownership, mainly pre-owned).
A whole apartment building whose "building type" in the building registration is "apartment house" is not included here. It is classified under "apartments (whole building)" in commercial real estate. However, properties not matched against the registration of the transfer of ownership (mainly new construction) with a total floor area of 300% or less of the site area go into "detached houses."
If you apply "the residential price index rose" to a whole apartment building, you are looking at a different box.
Your property moves by a different formula
The following is a fictional example for illustration.
An apartment building with annual rent assuming full occupancy of ¥6 million, operating expenses of ¥1.5 million, and net income after subtraction of ¥4.5 million. If the yield buyers require is 8.0%, then ¥4.5 million ÷ 8.0% gives ¥56.25 million.
One year later. Suppose net income stays at ¥4.5 million and the required yield moves to 8.4%. ¥4.5 million ÷ 8.4% gives about ¥53.57 million. That is a drop of about 4.8%. Even if the index rose 5% in the same year, this one goes down.
Conversely, if the vacancies fill and net income becomes ¥4.7 million while the required yield stays at 8.0%, the result is ¥58.75 million, a rise of about 4.4%. To get the same 5% as the index with net income staying at ¥4.5 million, the required yield would have to fall to 7.62%.
If you think in terms of the income capitalization approach, the price is net income divided by the yield the market requires. There are two numbers, the one divided and the one dividing, and both move somewhere apart from the index. The basic appraisal methods are three: the cost approach, the sales comparison approach, and the income capitalization approach, and it is said that more than one should be applied.
Then what is an average good for?
It may have sounded as if there is no point in looking at statistics. Not at all.
A broad population is a strength. A rushed sale, an inheritance, there happening to be one buyer who was just right. Each transaction's circumstances cancel each other out when gathered in numbers, and what remains is the direction of the market. Since the average exists precisely to erase individual circumstances, it makes sense that it cannot be applied back to an individual property.
It cannot be used to pin down a level. It can be used to see direction and range. If the assumptions in your own estimate run opposite to the direction of the market, that is a reason to question the assumptions.
Read the timing along with it. The real estate price index is published about 3 months after the transaction month, and is revised for 3 months after its first publication. The first release is a preliminary figure. Building starts statistics record a building at the point when the construction work notification under Article 15, Paragraph 1 of the Building Standards Act is accepted. The construction cost actually incurred after completion comes from a separate sample survey that looks at the gap from the planned amount at the start of construction, and is estimated nationwide from the sample.
I walk through a shopping street in Shinagawa Ward. Even if the number of vacant shops in the ward as a whole has not changed, on one street I often pass along, 3 shutters came down during October 2026. The more streets there are, the more the circumstances of one street are diluted in the average. Because they are diluted, the direction of the whole town becomes visible. But if you keep a shop open on that street, what you should be looking at is not the ward's average but the shutters of the 2 shops next door.
On the estimate sheet in front of you, try replacing the sale price field not with "index growth" but with two separate items: net income and the yield buyers require. Which of them would hurt you if it moved will show up there.