The economics of sake becomes much clearer when every number arrives with its label attached. Production, shipment, consumption and exports are different measures. A licensed site isn’t necessarily a site that produced sake during the period, and an industry average isn’t the size of a typical brewery. The arithmetic is usually manageable; choosing the correct denominator takes more care.
This lesson uses explicitly dated snapshots rather than presenting statistics as timeless facts. The production example comes from the NTA survey for the 2024 brewing year, published in January 2026. Exports use the full 2025 calendar year in JSS data. Their different periods and measurement bases matter whenever you compare them.
Economics of sake starts with the unit
A koku is traditionally a volume measure. For study calculations, use approximately 180 liters per koku. One kiloliter is 1,000 liters, so one kiloliter corresponds to about 5.56 koku on that rounded basis. State the approximation instead of pretending the conversion has unlimited precision.
The size is easier to picture through familiar containers. At the 180-liter teaching approximation, one koku corresponds to 100 bottles of 1.8 liters or 250 bottles of 720 milliliters. These are volume equivalents, not a claim that the liquid will all be packaged in one format without losses.
Keep units consistent before dividing. A number in kiloliters divided directly by 180 liters gives the wrong result unless the kiloliters are first converted. This is a common mistake because the written digits look plausible even when the quantities don’t match.
A dated production snapshot
The NTA production survey covers July 1, 2024 to June 30, 2025. It reports 285,636 kiloliters of sake production on a 20%-alcohol-equivalent basis. This is a standardized quantity, not the final volume of bottles all sold at their actual individual strengths.
The same survey records 1,104 responding sites that produced sake. There were 1,501 responses from 1,707 surveyed sites, including storage sites in the target population. Therefore, 1,104 is the producing-respondent count in this survey, not a complete count of every licensed location or corporate company in Japan.
That distinction should travel with the number whenever you use it. Saying there are exactly 1,104 sake companies would change both the population and the unit being counted. A company can have multiple sites, and a license or storage location need not correspond to an active producing respondent.
Convert production into koku carefully
Using the rounded 180-liter conversion, 285,636 kiloliters becomes 285,636,000 liters. Divide by 180 to obtain approximately 1,586,867 koku, or about 1.59 million koku. The alcohol-equivalent basis remains attached after the unit conversion; changing liters into koku doesn’t remove it.
This is a good habit for any statistical transformation. Units can change while the underlying definition stays the same. A standardized production figure doesn’t become retail consumption merely because it has been expressed as bottle equivalents or a traditional measure.
For revision, write the calculation in words as well as numbers: standardized reported production, converted into liters, divided by liters per approximate koku. That short chain makes an error easier to spot than a single unexplained answer copied onto a flashcard.
Calculate an average without inventing a typical brewery
Dividing the survey’s production quantity by its 1,104 producing respondents gives about 258.7 kiloliters per producing site, or approximately 1,437 koku using the same rounded conversion. This is a calculated mean for that dataset, not a separately reported median or a full-industry census average.
The calculation uses matching numerator and denominator definitions from the same survey. Mixing this quantity with a different year’s licensed-site count would create a different statistic, even if both inputs were official. Official numbers can still be combined incorrectly.
A mean can also conceal a very uneven distribution. A few large producers can raise the average far above the output of many smaller sites. Don’t use the calculated mean as evidence that most breweries produce around that amount or as a minimum commercial scale.
Distribution explains what the mean cannot
A JSS chart for 2020 illustrates this issue: in its stated population of 1,020 breweries, 599 were in the below-100-kiloliter band. The historical chart is useful for understanding structure, but it must not be presented as a current count.
The lesson is mathematical as much as commercial. If many sites are small while a few produce substantial volume, the number of breweries and the share of output can tell different stories. Counting sites equally doesn’t measure the amount each contributes.
Imagine an original simplified market with nine small producers and one much larger producer. The mean output may describe none of them particularly well. Before using it to plan a business or describe a typical brewery, ask for the distribution, median and relevant segment information.
Exports have both volume and value
The JSS export table includes the complete 2025 year on its third page. It records 33,549 kiloliters and 45,879 million yen, approximately 45.88 billion yen. The first page is partial 2026 data, so don’t accidentally label those partial totals as a full year.
Volume measures liquid crossing the statistical boundary, while value measures its recorded monetary amount. Growth in one doesn’t require equal growth in the other. Price changes, product mix, exchange-rate effects on trading behavior and destination composition can produce different patterns.
An aggregate value-per-liter calculation can be informative, but it isn’t a retail price. Dividing the 2025 totals gives roughly 1,368 yen per liter. That ratio doesn’t include every later distribution cost or tell you what a customer paid for a particular bottle in a restaurant.
Define top five before listing markets
For 2025, ranking the destinations in the JSS table by export value gives the following five. The values are in million yen and are rounded as displayed in the source.
| Rank by value | Market | Value, million yen |
|---|---|---|
| 1 | China | 13,304 |
| 2 | United States | 11,045 |
| 3 | Hong Kong | 4,824 |
| 4 | Korea | 4,395 |
| 5 | Taiwan | 2,698 |
The order differs from a volume ranking: the United States leads these listed markets by volume, while China leads by value. State which measure is being ranked. Hong Kong is separately reported in the trade table and is treated as a separate market here, rather than silently combined with another row.
Don’t calculate a false export share
It is tempting to divide the 2025 export volume by the brewing-year production quantity and announce a precise proportion exported. That would combine a calendar year with a different brewing-year period and actual export volume with standardized 20%-alcohol-equivalent production.
The result would look mathematical without being a clean like-for-like share. Stocks, timing and measurement differences also complicate the relationship. A responsible comparison can describe the scale of both datasets while explaining why they aren’t directly interchangeable.
If a question supplies compatible definitions for a share calculation, use them. If it doesn’t, identify what additional information is required. Refusing an invalid ratio isn’t avoiding arithmetic; it is choosing arithmetic that actually answers the question.
Commercial scale affects practical choices
A larger production operation may spread some equipment and distribution costs across more volume, while a smaller producer may pursue different product and sales strategies. These are economic possibilities, not automatic quality rankings. Size alone doesn’t establish balance, complexity or care.
More extensive polishing reduces the rice remaining for brewing and adds processing demands. Storage, refrigeration, packaging and distribution also use resources. The finished price reflects more than the amount of alcohol in the bottle, and a higher price isn’t a direct sensory score.
Export adds another chain of decisions: suitable products, destination requirements, logistics, storage and communication. A fragile nama expression has different handling needs from another product. An export total can’t show whether every bottle was transported under appropriate conditions or whether each producer earned the same margin.
Read trends with their dates intact
The JSS production and export resource pages make different series available. Check the period printed inside the file, not just the publication date or the year in its filename. A newly uploaded document can contain historical observations.
When comparing growth, use the same measure across the same periods. Value growth isn’t volume growth, and a decline in one product category doesn’t automatically describe all sake. Record the source, definition, period and unit together so later updates can replace the numbers without breaking the reasoning.
For a quick data audit, ask four questions before quoting any statistic: who collected it, what was counted, which period applies and which unit was used? Add a fifth when calculating an average: does the denominator describe the same population? These checks catch more errors than memorizing another decimal place.
The roles behind the data are explained in the trade-organization lesson. Compare the commercial scale with the category framework, then practice unit and denominator reasoning in the complete mock papers.
One-page revision summary
Use about 180 liters per koku for stated teaching calculations. The dated production survey gives about 1.59 million standardized koku and a calculated mean of about 1,437 koku per producing respondent. Keep that basis separate from actual 2025 exports and value-ranked markets. Define every denominator, distinguish mean from typical, and learn how to update statistics rather than memorizing unlabelled numbers.
Knowledge check: 10 original practice questions
Independent TFE questions, not official WSET examination items. Choose one answer per question; explanations follow separately.
1. What does aggregate export value per liter represent?
2. What does 1,104 describe in the cited survey?
3. Why can a mean misrepresent a typical brewery?
4. What basis accompanies the 285,636-kiloliter production figure?
5. Why not divide these export and production totals as a precise export share?
6. What is the approximate mean in this survey using 180 liters per koku?
7. Which market leads the cited 2025 export-value ranking?
8. Which list gives the top five by value in order?
9. What is the production period used in this lesson?
10. Using the teaching approximation, one koku is about how much?
Answers and explanations
1. C — A ratio of recorded totals, not a universal retail price.. Later costs, product mix and individual transactions are not resolved by the aggregate ratio. Mapping: Unit 1 LO3.
2. A — Responding sites that produced sake.. It is a defined survey population rather than an unrestricted count of companies or licenses. Mapping: Unit 1 LO3.
3. C — Uneven production sizes can pull it away from most sites.. Distribution matters when a few larger operations account for substantial volume. Mapping: Unit 1 LO3.
4. D — 20%-alcohol-equivalent quantity.. The standardized basis remains relevant after converting the volume into koku. Mapping: Unit 1 LO3.
5. B — The periods and volume bases differ.. A valid ratio requires compatible numerator and denominator definitions. Mapping: Unit 1 LO3.
6. B — 1,437 koku per producing respondent.. The calculation divides matching survey production by producing respondents and is a mean, not a median. Mapping: Unit 1 LO3.
7. A — China.. China leads by value in the table, while volume ordering differs. Mapping: Unit 1 LO3.
8. D — China, USA, Hong Kong, Korea, Taiwan.. The measure and year must accompany the ranking to make it meaningful. Mapping: Unit 1 LO3.
9. C — July 2024 through June 2025.. The NTA brewing-year survey period differs from the calendar-year export snapshot. Mapping: Unit 1 LO3.
10. B — 180 liters.. The rounded 180-liter basis must be stated and applied with consistent units. Mapping: Unit 1 LO3.
