Japan publishes its capital’s consumer price index roughly a month before the national figure: the Tokyo ward-area CPI for a given month lands at 8:30 a.m. JST in that same month’s final week, while the nationwide print for the same month follows about three weeks into the next month. Because the two series have historically moved within a few tenths of a percentage point of each other, the Tokyo release is where Bank of Japan expectations and JGB yields reprice first. If you only have time to track one Japanese inflation release, track this one — the national print mostly confirms what Tokyo already told you.
The topic is timely as of October 2026: with the yen trading near 158 per dollar as of early October — keeping imported-cost pressure alive — every inflation data point on the calendar gets read as a Bank of Japan policy input. And given the release calendar, the Tokyo figure for the current month is often the freshest hard inflation observation available ahead of a BoJ meeting. The rest of this guide is the timeless mechanics.
What the Tokyo CPI actually is
Japan’s CPI is compiled by the Statistics Bureau, part of the Ministry of Internal Affairs and Communications. Alongside the monthly nationwide index, the Bureau publishes a separate series for the Tōkyō-to kubu (東京都区部) — the 23 special wards of Tokyo, the dense core of the capital. The Tokyo figure is released as a sokuhōchi (速報値, preliminary estimate) for the current month, based on prices collected around mid-month.
Two structural points matter. First, the Tokyo ward area is not a separate survey design — it uses the same item basket and methodology as the national index. Second, Tokyo is a subset of the national index, carrying roughly 7% of its weight. The Tokyo release is therefore best understood as an early, partial sample of the national number, not an independent forecast of it.
Why it lands a month early
The gap is purely operational. Compiling a nationwide CPI requires collecting and processing prices from cities across all of Japan’s prefectures, so the national figure for month M is published roughly three weeks into month M+1. Compiling a single metropolitan area is much faster: the Statistics Bureau can process the 23-ward data quickly enough to publish month M’s Tokyo reading before month M even ends. The Bureau explicitly positions the Tokyo series as a leading indicator of the national index for exactly this reason.
The practical consequence: for every month of Japanese inflation, you get two reads, three to four weeks apart. Tokyo arrives first and sets expectations; the national print arrives later and usually lands close to where Tokyo pointed. By the time the nationwide figure is published, rate markets have typically already traded the information.
Release timing and the three headline cuts
The Tokyo CPI is released at 8:30 a.m. Japan Standard Time, before the Tokyo Stock Exchange’s 9:00 a.m. open, typically in the final week of the month and usually on a Friday. That half-hour window before the cash open is when JGB futures, yen rates, and currency markets do their first repricing.
Three cuts of the index matter, and Japanese conventions differ from US ones, so learn the local labels:
| Series | Japanese term | What it excludes | Why it matters |
|---|---|---|---|
| All items | sōgō (総合) | Nothing | Headline number; the measure in which the BoJ’s 2% price-stability target is formally defined; noisiest due to fresh food swings |
| Core | seisen shokuhin o nozoku sōgō (生鮮食品を除く総合) | Fresh food only | The line consensus forecasts and market pricing key off; closely analyzed by the BoJ |
| Core-core | seisen shokuhin oyobi enerugī o nozoku sōgō | Fresh food and energy | Cleanest read on underlying, domestically driven inflation |
Note the trap for readers used to US data: Japan’s core excludes only fresh food, not energy. The ex-fresh-food, ex-energy measure — what a US reader would call core — is Japan’s core-core. Note also that while the BoJ’s 2% price-stability target is formally set on the all-items (headline) index, markets and the BoJ closely analyze core and underlying measures in practice. On release morning, the market reaction keys off core first; BoJ-watchers then check core-core to judge whether any surprise was energy-driven or genuinely underlying. (Our separate national CPI guide covers the core versus core-core concepts and the 2% target in depth; this guide is about the Tokyo release as a mechanism.)
How closely Tokyo tracks the national print
Historically, Tokyo core CPI and national core CPI move in the same direction in the large majority of months, and the year-on-year rates usually sit within a few tenths of a percentage point of each other. That is why the early read is tradable at all. But the tracking is good, not perfect, and the divergences have identifiable causes:
- Housing weight. Rent carries a heavier weight in the Tokyo basket than nationally, and Japanese rents move slowly — this tends to damp Tokyo readings relative to the national index when goods prices are running hot.
- Local policy measures. The Tokyo Metropolitan Government periodically runs its own subsidies — tuition relief, utility-charge measures — that mechanically push the Tokyo index away from the national trend for as long as they are in effect. These are the single biggest source of misreads.
- Fresh food and travel noise. Weather-driven produce swings and hotel-rate seasonality can hit the two indices with different force in any single month.
The practical technique: before reacting to a Tokyo surprise, check the recent spread between Tokyo core and national core over the prior several months. If the spread has been stable, project the surprise onto the national print roughly one-for-one. If a known Tokyo-only subsidy is in effect or expiring, adjust for it — the national index will not inherit that distortion at full strength.
Why rate markets react to Tokyo first: a worked example
The BoJ’s policy path is priced most directly in short-maturity Japanese government bonds, so that is where a Tokyo CPI surprise shows up first. Here is how to read an actual curve, using Ministry of Finance official yields as of October 8, 2026: 2-year 1.936%, 5-year 2.414%, 10-year 3.089%, 30-year 4.136%.
- Start with the 2-year. A 2-year JGB yield approximates the average expected policy rate over the next two years. At 1.936% (as of October 2026), the market was pricing substantial further normalization. This is the maturity most sensitive to a CPI surprise.
- Watch the change, not the level. If Tokyo core prints above consensus at 8:30 a.m., a move of even a few basis points in the 2-year that morning means the market is pulling rate hikes forward. A flat 2-year after a hot print means the market judged the surprise to be noise — often because core-core did not confirm it.
- Use the 10-year as the cross-check. The 10-year (3.089% on that date) embeds term premium and fiscal factors on top of the policy path, so it is a noisier inflation read. If the 2-year moves and the 10-year barely does, the curve is flattening: the market expects earlier hikes but no higher destination.
- Glance at the yen. A hawkish Tokyo surprise that lifts short-end yields typically supports the yen at the margin; if USD/JPY does not budge, the rates move may reflect something other than the CPI.
This same four-step read works on every Tokyo CPI morning, regardless of where the level of yields happens to be.
Common mistakes on release day
- Mixing up reference months. In a typical calendar month, the national CPI for the previous month is released around the third week, and the Tokyo CPI for the current month lands about a week later. Two inflation headlines a week apart refer to different months — always check the reference period before comparing.
- Trading the headline instead of core. The all-items figure swings with fresh food. Consensus forecasts, BoJ commentary, and market pricing all key off the ex-fresh-food core series.
- Treating Tokyo as a forecast rather than a sample. Tokyo is folded into the national index at roughly 7% weight. A Tokyo surprise shifts the expected national print, but regional divergence means it does not determine it.
- Ignoring known local distortions. When a Tokyo-specific subsidy starts or expires, the month-on-month step in the Tokyo index is policy mechanics, not an inflation signal. The national print will show a much smaller echo.
FAQ
Is the Tokyo CPI revised later?
The initial release is a preliminary estimate (sokuhōchi), finalized when the corresponding national CPI is published about three weeks later. Revisions are typically small, which is precisely why markets treat the preliminary figure as actionable.
Which line should I watch at 8:30 a.m.?
Core — the ex-fresh-food index — because it is the series consensus forecasts are built on and the line markets trade; the formal 2% price-stability target is defined on the all-items index, but core is the operational reference on release morning. Check core-core immediately afterward to see whether any surprise survives once energy is stripped out.
Why doesn’t the BoJ just wait for the national number?
It effectively gets both: given the release calendar, the Tokyo figure for the current month is often the freshest inflation observation available ahead of a policy meeting, while the latest national print covers the month before. BoJ communication regularly references underlying trend measures, and the Tokyo series feeds the market’s real-time estimate of that trend.
How large is Tokyo’s weight in the national CPI?
The 23-ward area carries roughly 7% of the nationwide index weight. Its predictive value comes less from the weight itself than from the fact that national price trends — energy, processed food, services — tend to show up in the capital’s data on essentially the same schedule as everywhere else.
Sources
- Bank of Japan — monetary policy and price-stability target materials: www.boj.or.jp/en
- Ministry of Finance Japan — official JGB yield data: www.mof.go.jp/english
- JPX — market statistics: www.jpx.co.jp/english
- Nikkei Indexes — index data: indexes.nikkei.co.jp/en
Disclaimer: This is an information and analysis publication, not investment advice. See our Methodology for data sources, standards, and our corrections policy.