Japan's September 2026 Tankan points to still-positive corporate conditions, but it does not say that every part of the economy is equally strong or that the Bank of Japan has reached a policy conclusion. The useful reading is a split one: large-company manufacturing and services remained positive, medium-sized companies improved in manufacturing, and small-company readings stayed positive but lower. The same respondents expected a softer reading three months ahead.
The Tankan is the Bank of Japan's quarterly Short-Term Economic Survey of Enterprises in Japan. It is one of the most useful recurring primary sources for international readers, provided that its diffusion indexes, company-size categories and forecasts are kept in their proper lanes. The September survey covered 9,104 enterprises and had a 99.4% response rate; responses were collected from August 26 through September 30, 2026.
Key takeaways
- The all-enterprise business-conditions index was +21 in September 2026: +19 for manufacturing and +20 for nonmanufacturing. A positive figure means more respondents called conditions favorable than unfavorable; it is not a percentage growth rate.
- Large companies remained the strongest group in level terms, at +24 for manufacturing and +35 for nonmanufacturing. Medium-sized manufacturing improved to +23; small-company indexes were +14 and +15, respectively.
- Respondents expected lower readings three months ahead: +16 for all-enterprise manufacturing, +16 for nonmanufacturing and +15 for all industries. Those are company responses, not Bank of Japan forecasts.
- All enterprises planned FY2026 fixed investment, including land purchases, up 7.6% year on year, revised 0.8 percentage point higher than in June. It is a plan reported in a survey, not an actual national-accounts result.
- The practical read is to compare current conditions, the next-quarter expectation, investment plans and sector/size differences—not to treat one headline DI as a verdict on Japanese equities or interest rates.
What Japan's Tankan survey measures
The Bank of Japan's Tankan FAQ describes a nationwide quarterly survey that asks companies about business conditions, supply and demand, capacity, employment, prices, investment and annual projections. It pairs judgement questions with quantitative responses on items such as sales and fixed investment.
The widely quoted business-conditions number is a diffusion index, or DI. For this item, the calculation is the share of respondents saying conditions are favorable minus the share saying they are unfavorable. A DI of +21 therefore means the favorable share exceeded the unfavorable share by 21 percentage points. It does not mean 21% growth, a 21% probability, or that 21% of companies are doing well.
The Bank groups non-financial enterprises by paid-in capital: large enterprises have at least ¥1 billion, medium-sized enterprises have ¥100 million to less than ¥1 billion, and small enterprises have ¥20 million to less than ¥100 million. Those labels are survey classifications, not Tokyo Stock Exchange market segments or rankings of business quality. For the latter distinction, see ATG Press's TSE Prime, Standard and Growth guide.
The FAQ also makes an important boundary explicit: the published figures aggregate respondents' views and do not constitute the Bank of Japan's assessment of economic conditions or its projections. The Tankan is highly useful evidence, but it is not a policy statement.
September 2026: positive conditions, softer expectations
The September 2026 summary shows why the headline needs more than one number. Large-manufacturer sentiment improved from +22 in June to +24, while large nonmanufacturers slipped from +37 to +35. The changes were more pronounced below the largest companies: medium-sized manufacturers rose from +17 to +23, while small manufacturers rose from +9 to +14.
Swipe horizontally to view all columns.
| Enterprise group | Manufacturing DI: June actual | Manufacturing DI: September actual | Manufacturing DI: three-month outlook | Nonmanufacturing DI: June actual | Nonmanufacturing DI: September actual | Nonmanufacturing DI: three-month outlook | Source |
|---|---|---|---|---|---|---|---|
| Large enterprises | +22 | +24 | +21 | +37 | +35 | +30 | BOJ Tankan Summary, September 2026 |
| Medium-sized enterprises | +17 | +23 | +18 | +26 | +24 | +19 | BOJ Tankan Summary, September 2026 |
| Small enterprises | +9 | +14 | +12 | +15 | +15 | +10 | BOJ Tankan Summary, September 2026 |
| All enterprises | +14 | +19 | +16 | +21 | +20 | +16 | BOJ Tankan Summary, September 2026 |
The table shows two different comparisons. First, September conditions were generally positive and manufacturing improved across the size groups. Second, the next-quarter outlook was below the September actual reading in every displayed all-enterprise aggregate. That is a signal to investigate the detail, not proof that activity will weaken by a known amount.
The all-industries DI gives the same pattern: +21 in September, up from +18 in June, with an outlook of +15. The September reading is a snapshot of responses during the survey window; the outlook is the same respondents' view of conditions three months later. Neither should be rewritten as a Bank forecast.
Investment, profits and exchange-rate assumptions: separate the definitions
The Tankan's quantitative tables can add useful context, but their labels matter as much as their direction. All enterprises planned FY2026 fixed investment including land-purchase expenses to increase 7.6% year on year, an upward revision of 0.8 percentage point from the June survey. The Bank's own documentation notes that Tankan investment plans are surveyed repeatedly through a fiscal year and later finalized; an early plan is not the same object as an official expenditure estimate.
Large companies' manufacturing investment plan was +11.6%, while large nonmanufacturing was +11.2%. Smaller-company plans were much less even: medium-sized manufacturing was +1.9%, medium-sized nonmanufacturing was +11.4%, small manufacturing was +4.0%, and small nonmanufacturing was -8.4%. That spread is more informative than a single aggregate number when the reader is assessing where spending momentum is expected.
The summary also reports the average exchange-rate assumption supplied by respondents. For FY2026, all enterprises expected ¥154.00 per U.S. dollar in the second half, compared with ¥152.51 in the June survey's second-half assumption. That is an input reported by companies, not an exchange-rate prediction by the Bank of Japan or a trading signal.
For context on another distinct price indicator, see ATG Press's Japan Core CPI vs Core-Core CPI explainer. CPI tracks consumer-price changes; the Tankan captures company responses and plans. They can inform the same macro discussion without being interchangeable.
A five-step reading routine for each quarterly release
1. Identify the exact series. Confirm whether the number is a business-conditions DI, a price DI, a capacity/employment measure, or a quantitative annual projection. 2. Separate actual and outlook. The current result and the three-month forecast are different survey answers. Do not report the outlook as if it were the current reading. 3. Check size and sector. Manufacturing, nonmanufacturing, large, medium-sized and small enterprises can move differently. The aggregate can hide that dispersion. 4. Label plans as plans. Sales, profit, investment and exchange-rate figures may be annual projections or revisions. Check the fiscal year and whether the table says actual, forecast or revision rate. 5. Cross-check with company disclosures. The Tankan is aggregate evidence. Use earnings releases and guidance to test which companies or sectors are driving a result; ATG Press's Japan earnings season guide explains when those company disclosures usually arrive.
What to watch next
The next useful checks are whether the expected three-month softening appears in the December survey, whether fixed-investment plans hold or are revised again, and whether the difference between larger and smaller companies narrows or widens. Readers should also track price, capacity and employment measures in the underlying release rather than relying only on the business-conditions headline.
The Bank publishes a September 2026 release hub with the summary, outline, comprehensive data set and long-term time series. That is the best starting point for a specific industry, company-size group or historical comparison. It is preferable to treating a media headline as the entire survey.
Sources
- Bank of Japan: Tankan Summary (September 2026) (PDF) — official sample, response rate, diffusion indexes, outlooks, annual projections and exchange-rate assumptions; released 2026-10-01; accessed 2026-10-11.
- Bank of Japan: Tankan (September 2026 Survey) — official release hub for the summary, outline, comprehensive data set and long-term series; dated 2026-10-02; accessed 2026-10-11.
- Bank of Japan: FAQs on the Tankan — official purpose, coverage, classifications, methodology boundaries and release materials; dated 2026-03; accessed 2026-10-11.
Disclaimer
This article is for informational purposes only and does not constitute investment advice. Survey responses, company projections, exchange rates and policy conditions can change; verify current primary disclosures before making an investment decision.
