Tankan Large Manufacturers Index

It is one headline of the Short-Term Economic Survey of Enterprises in Japan (a quarterly business poll), nicknamed Tankan. The Bank of Japan runs it, while answers reflect firms only. The current survey started in 1974, following a 1957 predecessor. It covers private firms with capital of 20 million yen or more, excluding financial institutions. Large means capital of 1 billion yen or more across 17 manufacturing industries in 93 strata, with firms judging own conditions mainly on profits.

Surveys run each March, June, September, and December, with forms going out about 1 month before release. The return deadline falls about 2 weeks after dispatch, and collection runs until 1 business day before release day. Results appear at the start of April, July, and October, and in mid December, always at 8:50 am Japan Standard Time. Dates for the next 12 months are announced in advance.

Each release carries a summary, an outline, a full data set, and time series data. It prints actual results plus forecasts for 3 months ahead alongside annual sales, profit, and investment projections. Corrections appear as swiftly as possible.

The headline reacts to profit outlooks, demand balances, cost pressures, capacity strains, and credit ease. Zero is the only dividing line: above 0 optimists lead, below 0 pessimists lead.

• Domestic supply and demand: industry balance of excess demand minus excess supply.
• Overseas supply and demand: the same balance for foreign product markets.
• Inventory levels: own stocks of finished goods plus wholesalers stocks.
• Production capacity and employment: own equipment and staffing judged excessive minus insufficient.
• Financial position and lending: own cash position plus lending attitude of financial institutions as perceived.
• Prices and interest: changes in output prices, input prices, and loan interest rates as rises minus falls.

The survey released December 13, 2023, the 199th round, put large manufacturers at 12, up by 3 from 9 in September 2023. Firms had forecast 10, so the result beat expectations by 2, while they forecast only 8 ahead, down 4. Response shares were 22 favourable, 68 not so favourable, and 10 unfavourable. Motor vehicles read 28, up by 13, showing gains varied by industry.

1. Answer choice: Each firm picks Favourable, Not so favourable, or Unfavourable at the survey date and for 3 months hence, ignoring seasonal effects.
2. Formula: The DI equals the Favourable share minus the Unfavourable share, in percentage points. For example, 30 Favourable, 60 Not so favourable and 10 Unfavourable gives a DI of 20.
3. No size weighting: The Bank simply aggregates replies instead of weighting big firms more heavily.
4. Rounding: Shares round to whole numbers, so parts can look slightly inconsistent.
5. Seasonality: No statistical adjustment applies because firms must exclude seasonal factors themselves.