The Companies That Do Not Exist in English

A page on the Tokyo Stock Exchange's website used to list the companies that had asked for more time to exist in English, together with the dates by which each said it would arrive. As of 26 July 2026 the page returns 404 in both Japanese and English, and the link to it is gone from the contents of the exchange's English disclosure portal. The deferral it documented expired on 1 April 2026, which is consistent with the deletion, though no announcement of the removal could be located. The reason for the removal is inference, not disclosure.

The obligation being deferred took effect on 1 April 2025. Under Article 436-4(1) of the listing rules, amended and published on 9 May 2024, a domestic company listed on the Prime Market must disclose earnings information and timely disclosure material in English at the same time as it does in Japanese. This sits in the corporate code of conduct as a matter to be observed, not encouraged. A companion provision, Article 445-8, supplies the encouragement: companies are asked to make efforts to publish identical English versions of company information generally, simultaneously. Foreign issuers are outside the article by its terms. So are the Standard and Growth markets.

A Part or a Summary

The scope is broader than earnings. It covers the earnings summary and the quarterly summary; the supplementary explanatory materials, where a Japanese version exists; and every piece of company information disclosed through TDnet, including decided facts, occurred facts, forecast revisions, items a company chooses to disclose voluntarily through the system, and items falling under the materiality thresholds that would otherwise excuse disclosure. Investor-relations press material and inspection documents such as general meeting notices are voluntary.

The English text may be, in the words of Article 436-4(2), a part or a summary of the Japanese. No standard defines how much of a part, and the exchange's own guidance says each company decides. It does draw one line: where a single Japanese document carries several disclosure items, the English must carry all of them, and translating one of the items and stopping is not permitted.

The enforcement shape is the more instructive feature. English disclosure is positioned as a reference translation, and its accuracy is expressly outside the exchange's enforcement measures. Publishing nothing at all, without having filed for deferral, is inside them, though not automatically: the guidance says publication measures may follow depending on the content, circumstances and cause. The enforceable object is the existence of an English document rather than its correctness.

Simultaneity has a written exception as well. Where a matter requires urgent handling, or the Japanese wording is unsettled until shortly before release, and waiting for English would delay the Japanese, the Japanese goes first. English is then expected the same day: for evening releases, from 17:00, the exchange accepts the English filing until 19:00, and where 19:00 is not feasible, by the 9:00 open of the next business day. Descriptions elsewhere of a two-business-day allowance during an initial phase after the rule took effect do not appear in the primary materials.

The deferral itself was a form. A company already listed on Prime before the effective date could file a document between 6 January and 14 March 2025 stating when it intended to begin. Doing so bought a year. The stated date did not matter: every deferring company became subject to the obligation for disclosures made on or after 1 April 2026, uniformly. The form asks when you plan to comply and then disregards the answer.

How many filed it is not settled. A follow-up council document dated 2 September 2025 says about 7 percent, or 114 companies, took the maximum one-year deferral. A survey report published on 26 January 2026, on a 31 December 2025 basis, says 96 of the 1,598 Prime companies applied. The two figures come from different documents with different as-of dates and different sources within the exchange, and the difference is not explained. Delistings, market-segment changes and withdrawn applications would all produce movement in that direction; none of them is confirmed. The numbers should not be subtracted from each other.

The Series Behind the Jump

What the rule produced is not in doubt. As of 31 December 2025, on the exchange's self-reported survey, 99.6 percent of Prime companies published an English version of their earnings summaries, in whole or in part, the rule permitting either, up 5.8 points. For timely disclosure other than earnings, the figure is 96.8 percent, up 37.6 points in twelve months. The series behind that jump runs 47.5 percent at the end of 2022, 52.1 percent in 2023, 59.2 percent in 2024, then 96.8. Three years of incremental drift, followed by a step change that lands on a date printed in a rulebook eleven months earlier. Translation capacity did not appear in 2025. The requirement did.

The broader measure, the share of Prime companies doing any English disclosure of any of the seven surveyed document types, reached 99.8 percent, up 0.8 points, having climbed from 85.8 percent at the end of 2021 through 97.1, 98.2 and 99.0. The population is not constant across those points: Prime held 1,837 companies at the end of 2022 and 1,598 at the end of 2025. A rate that rises while its denominator falls by 13 percent is measuring a different set of companies each year.

Documents the rule does not reach moved less, and in a pattern. Prime general meeting notices reached 94.9 percent, up 2.5 points, and IR presentation materials 87.1 percent, up 4.2. The statutory annual securities report sat at 27.5 percent, up 0.4 points, business reports at 32.1 percent and corporate governance reports at 40.6 percent. The annual report is not comparable to the rest: it is a filing under the financial instruments law, where the Japanese text is the operative original, and translating it is a separate exercise from producing an English summary. Still, the sorting is clean. English exists where the exchange requires it, and where companies meet investors in person.

There is a second definition problem inside the headline. Implementation means an English document exists, not that the Japanese document exists in English. Restricted to full translation, the same survey reports 57.9 percent for earnings summaries, 81.3 percent for timely disclosure, 78.4 percent for IR materials, 60.8 percent for meeting notices and 10.4 percent for annual reports. A separate count as of August 2025 put full translation of earnings summaries at 56 percent and of timely disclosures at 76 percent. So a 99.6 percent implementation rate sits on top of a full-translation rate in the high fifties, and both are accurate statements about the same population.

The survey is a survey. The exchange has run it annually since November 2019 by questionnaire, sent for the 2025 round to 3,777 listed companies excluding the TOKYO PRO Market. Companies report on themselves; the survey is not a machine aggregation of TDnet data, and the exchange does not present the responses as checked against actual filings. The overall response rate was 90.2 percent, with Prime at 100 percent, and non-respondents are counted as non-disclosers.

The IR-materials series carries a seam: the same date, 31 December 2024, is 76.4 percent using all companies as the denominator and 82.9 percent once companies that publish no Japanese version are excluded, and the 2025 report standardised on the latter. A composite timely-disclosure column present in earlier appendices is not in the current one. The exchange's own listing statistics, meanwhile, count 3,782 companies excluding the PRO Market at the same date, five more than the survey's universe.

The Other Side of the Line

Outside the boundary, the picture holds still. Standard-listed companies published English earnings summaries at 25.9 percent and English timely disclosures at 7.9 percent. Growth was at 24.0 percent and 14.1 percent. On the any-English measure, Standard reached 34.1 percent and Growth 31.4 percent, against an all-market figure of 61.5 percent. A few Growth line items fell year on year, by up to seven tenths of a point, which in a segment with constant new listings likely reflects turnover in its constituents rather than individual companies retreating. Response rates were 88.0 percent for Standard and 69.9 percent for Growth, and non-response counts as no, so these are floors.

The only rule reaching them is the governance code's comply-or-explain expectation. Overseas investors have asked for the mandate to be extended; the exchange says extension and a wider document scope remain under continuing consideration, with Standard listed as under consideration in a January 2026 council document and no date attached.

The two-tier structure is clearest when the same question is asked twice with different weights. Across all markets, 56.8 percent of companies publish an English earnings summary and 46.5 percent an English timely disclosure. Weighted by market capitalisation, the same items are 98.0 percent and 97.0 percent, and even English annual reports reach 66.3 percent by value against 12.7 percent by company count. The two series are built differently and cannot be mixed, the cap-weighted figures being more inclusive in what they count.

What survives the caveat is the shape: English exists in proportion to market value. It did so before the mandate as well. At the end of 2024, 93.8 percent of Prime companies published an English earnings summary, against 54.1 percent across all markets. Which of those two facts explains the other is not in the survey.

The Audience, by Value and Turnover

The demand side has been setting records. Foreign investors held 34.7 percent of the market by value as of 31 March 2026, up 2.3 points and the third consecutive high since the survey began. In money, that is 420.6 trillion yen ($2.57 trillion), against a total market value for the 3,975 surveyed companies of 1,213.9 trillion yen ($7.41 trillion). The holding rose by 113.9 trillion yen ($695 billion) in one year, but TOPIX rose 31.6 percent over the same period, so most of the increase is price rather than purchase. A second series, computed on trading units rather than value, puts foreign ownership at 28.3 percent. The two are not connectable, and the gap between them is itself a statement about which companies foreign investors own.

Trading is counted on a different calendar and answers a different question. In calendar 2025, foreign investors accounted for 64.9 percent of brokered trading value across the Tokyo and Nagoya markets, 66.2 percent on Prime alone, on 1,725.7 trillion yen ($10.5 trillion) of activity. That share was 66.0 percent in 2024. Measured against total trading value including proprietary trading by member firms, 2,978.8 trillion yen ($18.2 trillion), the foreign share is about 57.9 percent, a ratio the exchange does not publish and which is arithmetic on its own spreadsheet. Recent readings run 64.3 percent for June 2026 and 66.1 percent for the third week of July, which is a week and not a trend.

The declining share is a share of turnover, not a direction of money: over the fiscal year to 31 March 2026, foreign investors bought 10.3 trillion yen ($63 billion) more than they sold, after net selling of 4.9 trillion yen ($30 billion) the year before. A composition ratio in a high-turnover market went down while the holding went up. Those are two sentences about two things.

Coverage Estimates and Undefined Universes

How many companies no analyst is looking at is, remarkably, not a published number. QUICK computes consensus estimates for 1,548 Japanese listed companies as of 26 July 2026, which against 3,782 works out to about 41 percent covered and about 59 percent not, a ratio neither QUICK nor the exchange publishes and which is arithmetic done here; QUICK's own English materials say more than 1,600. The inclusion threshold is not disclosed, so companies with one analyst may sit outside the consensus set. The published estimates, such as they are, describe different universes.

A March 2025 newspaper report put the share of Japanese stocks with no analyst forecasts at about 70 percent, without a visible basis, date or definition in the free portion. An asset manager's 2024 white paper, using Bloomberg data, reported an average of 0.9 analysts per small-cap and 62 percent uncovered, against 1.7 and 56 percent in early 2010, with mid-cap coverage falling from eight analysts to five. Another manager wrote in July 2024 that over 70 percent of the TOPIX Small index is uncovered, which is a statement about an index and not about the market.

A fund manager's article from February 2025 put small-cap coverage of at least one analyst at 42 percent, from which the uncovered 58 percent is again arithmetic done here rather than a published figure. None of these is an official statistic, most of the universes are undefined, and each source's definition of small is its own. A sixth source, cap-weighted, puts MSCI Japan at 18 analysts per company against 45 for the S&P 500, which describes the largest companies twice over: the index leaves out most of the small-cap tail, and weighting by value would bury whatever remained of it.

The mandate covers roughly 1,600 Prime-listed domestic companies. Foreign issuers are outside it by the article's terms, and 1,598 is the number the survey wrote to rather than a published count of the obliged. On the other side of the line sit about 2,180 companies, 1,567 on Standard and 612 on Growth in the same survey, to whom the rule does not apply at all. Around a third of them publish something in English, which is to say Japanese is the only version for about two-thirds. Fewer than one in ten Standard-listed companies translates a timely disclosure.

Whether that changes what those companies are worth is not a question the disclosure data can answer; ownership share is a measurement, translation rate is a measurement, and the line between them is a hypothesis. What the data does show is the audience: 34.7 percent of the market by value, 64.9 percent of the brokered turnover. What it does not show is how much of either sits on the far side of the boundary. The surveys do not break it down that way.

All yen amounts here are converted at 163.8 yen to the dollar, the rate on 24 July 2026, and that single rate is applied to figures reported for earlier dates as well. Dollar amounts across years therefore move only when the yen amounts move; the exchange rate is held still and says nothing. The yen figures are the reported ones, and the dollars are a convenience.