Japan Cross-Shareholding Unwind 2026: What the Record Sell-Off Means for U.S. Investors

I’ve been watching the 持ち合い解消 (cross-shareholding unwind) story build for three years now, and what’s happening in 2025-2026 finally feels different — the Japanese-language FSA supervisory guidelines and TSE naming-and-shaming data I’ve been reading make it clear this isn’t another false start, and that gap between what’s in the Japanese IR filings and what English-language media is reporting is exactly where the opportunity lives.

Indicator	Value / Estimate
Strategic shareholdings on insurer/bank B/S (FY2024 est.)	¥50+ trillion
Annual unwind pace (FY2024-2026 est.)	¥5-8 trillion/year
FY2024 TSE-wide buyback authorizations (record)	¥17+ trillion
TSE Prime companies with PBR < 1 (early 2023 → end-2024)	~50% → ~43%
Foreign ownership of TSE Prime (2020 → 2024)	~29% → ~32%
Tokio Marine disclosed cross-holding book value (FY2023)	¥3.5 trillion

The structural story behind Japan’s cross-shareholding unwind is one of the most consequential — and least understood — capital-reallocation events in global equity markets right now. For U.S. dividend and value investors, it translates directly into rising buybacks, expanding payout ratios, and improving return-on-equity at some of Japan’s largest and most stable companies. This article sizes the opportunity, maps the sector winners and losers, and gives you a practical playbook grounded in Japanese-language primary sources.

Disclaimer: This article is for informational purposes only and does not constitute investment advice. I may or may not hold positions in the securities discussed. See full Disclaimer for details.

Why Japan’s “Stable Shareholder” System Lasted 60 Years — and Why It’s Finally Breaking

The Keiretsu Logic — How Cross-Holdings Became a Corporate Governance Moat

Japan’s 持ち合い (cross-shareholding) system emerged from the post-war reconstruction era. Banks, insurers, and industrial companies purchased each other’s shares to cement business relationships, ensure stable financing, and — critically — create a bloc of “friendly” shareholders that made hostile takeovers nearly impossible.

The logic was coherent for its time. A Toyota supplier holding Toyota shares, and Toyota holding the supplier’s shares, aligned incentives and reduced transaction costs across the keiretsu network. Insurers like Tokio Marine held shares in virtually every major corporate client as a relationship anchor.

By the 1980s bubble peak, cross-held shares accounted for an estimated 30-40% of all listed Japanese equities by market value. The system was not just cultural — it was structural, embedded in financing relationships, board interlocks, and insurance contract renewals.

Three Prior “Unwind” Moments That Fizzled (1998, 2002, 2015) and What’s Different Now

The 1990s bubble burst triggered the first serious unwind pressure. Banks sold some stakes to repair capital ratios, but non-life insurers largely held firm, arguing that cross-holdings were strategic, not speculative.

The 2002-2003 banking crisis produced another partial unwind, and the 2015 Corporate Governance Code introduced a requirement for boards to explain the rationale for each strategic holding. Progress was slow — companies filed boilerplate justifications and retained most stakes.

What makes 2023-2026 structurally different is the combination of three simultaneous pressures that did not co-exist in prior cycles:

• TSE naming-and-shaming: The TSE’s January 2023 “PBR below 1” disclosure request, followed by its March 2024 follow-up progress report (東証「資本コストや株価を意識した経営の実現に向けた対応」), publicly identifies non-compliant companies by name — a reputational lever that prior governance reforms lacked.
• FSA supervisory teeth: The FSA’s FY2023-24 金融行政方針 (Financial Administration Policy) explicitly targets 政策保有株式 reduction as a capital-efficiency priority for insurers, with supervisory follow-up — not just a recommendation.
• Corporate Governance Code revisions (2021, 2023): Boards must now justify each strategic holding individually and annually, with a presumption toward reduction unless a specific business rationale is documented.

The key information arbitrage here: TSE’s Japanese-language progress report contains company-level compliance data — which firms disclosed a plan, which are executing, and which are lagging — that has not been summarized in any English-language financial publication I’ve found. U.S. investors relying on Bloomberg or Reuters are working with a significant information lag.

Sizing the Unwind — How Many Trillions Are Actually on the Table

Insurer Balance Sheets — The Largest Single Source of Supply

The Bank of Japan’s 資金循環統計 (Flow of Funds statistics) tracks equity held by domestic financial institutions. As of end-2023, domestic insurers held approximately ¥50 trillion in listed equities, with a significant portion classified as 政策保有株式 (strategic shareholdings).

Tokio Marine Holdings is the clearest case study. Its FY2023 annual report disclosed ¥3.5 trillion in book value of cross-held shares, and management has pledged to reduce strategic holdings to effectively zero by FY2026. That is not a gradual wind-down — it is a near-complete liquidation of a position that took 60 years to accumulate.

MS&AD Insurance Group and Sompo Holdings made comparable pledges. Combined, the three major non-life insurers are estimated to sell ¥10+ trillion in cross-held equities over FY2024-2026 — roughly ¥3-4 trillion per year from this cohort alone.

Megabank Pledges — Smaller in Scale, Faster in Execution

Japan’s three megabanks — MUFG, SMFG, and Mizuho — each disclosed 政策保有株式 reduction plans in their FY2024 integrated reports. MUFG alone disclosed a target reduction of approximately ¥2 trillion in strategic shareholdings over a multi-year horizon.

The megabanks’ unwind is smaller in absolute yen terms than the insurers’, but execution has been faster. Banks face tighter Basel III capital constraints that make holding low-yielding equity stakes increasingly expensive from a regulatory capital perspective.

Combining insurers and megabanks, the estimated annual sell-off pace is ¥5-8 trillion per year in 2024-2026, compared to approximately ¥2 trillion per year in 2019-2022. This is a 2.5-4x acceleration in supply hitting the market.

How to Read a 政策保有株式一覧 on EDINET (Step-by-Step for U.S. Investors)

Every listed Japanese insurer and bank files a 政策保有株式一覧 (list of strategic shareholdings) as part of their 有価証券報告書 (annual securities report) on EDINET (edinet.fsa.go.jp). This filing discloses issuer name, share count, book value, and the stated business rationale for each holding.

To find it: search for the company name on EDINET, open the most recent 有価証券報告書, and navigate to the section titled「政策保有株式」(typically in the corporate governance section, around pages 40-60 of the PDF). The table format is standardized, so once you’ve read one, the others follow the same structure.

Cross-referencing EDINET filings with TSE’s 適時開示 (timely disclosure) feed reveals which specific stocks are being sold before they appear in foreign financial media — a genuine information edge for investors willing to read Japanese-language filings.

The Capital Return Flywheel — Buybacks, Dividends, and PBR Repair

The Buyback Math — How ¥1 of Cross-Holding Sold Becomes ¥1.2 of Shareholder Value

The mechanism is straightforward but powerful. When an insurer sells ¥100 billion in cross-held shares, it receives cash. That cash, deployed as a share buyback at a PBR below 1.0x, is mathematically accretive to book value per share — every ¥1 spent buying back shares worth more than ¥1 in book value creates surplus value for remaining shareholders.

Tokio Marine used unwind proceeds to fund ¥300 billion+ in buybacks in FY2023 alone. MS&AD and Sompo followed with comparable programs. The result: EPS growth that is partially balance-sheet-driven rather than purely earnings-driven — a dynamic U.S. value investors familiar with financial-sector book-value investing will recognize immediately.

FY2024 TSE-wide buyback authorizations hit a record ¥17+ trillion across all listed companies, a figure that would have been unthinkable five years ago. The unwind is a significant driver of this surge.

Dividend Payout Expansion — Which Sectors Are Raising Fastest

Beyond buybacks, companies with completed or near-complete unwinds have been raising dividends materially. The logic: once the cross-holding portfolio is liquidated, the drag on ROE disappears, earnings per share improves via buybacks, and management faces pressure to demonstrate ongoing capital discipline through progressive dividends.

For U.S. dividend investors comparing Japan to domestic alternatives, this is the key frame: Japanese non-life insurers currently trade at PBR 1.0-1.5x versus global insurance peers at 1.5-2.5x, while offering dividend yields in the 2.5-4% range and buyback-enhanced total return profiles. Think of it as a Japanese version of a Travelers or Hartford Financial — but with a multi-year governance catalyst that has no U.S. equivalent.

You can track the dividend history and payout ratio trends for individual Japanese insurers on TradingView, which provides clean historical data going back to the pre-Abenomics era — useful for calibrating how much of the current payout expansion is structural versus cyclical.

Index Inclusion Tailwind — How Governance Scores Drive Passive Inflows

TSE’s “action plan” scoring correlates with inclusion in the JPX Prime 150 index and MSCI Japan ESG indices. Companies that disclose specific unwind timelines and demonstrate execution receive higher governance scores, which triggers passive fund inflows from ESG-mandated institutional investors.

Foreign investor ownership of TSE Prime has risen from approximately 29% in 2020 to approximately 32% in 2024, partly driven by this governance-reform narrative. TSE publishes a monthly 株主分布状況調査 (shareholder distribution survey) in Japanese that breaks down foreign vs. domestic institutional vs. cross-held ownership by sector — this data predicts which sectors have the most remaining unwind runway and is not translated into English.

For U.S. investors holding Japan exposure in a taxable brokerage or IRA via IBKR or Schwab, this passive inflow dynamic matters: it creates a price-support mechanism for governance-compliant names that partially offsets the selling pressure from the unwind itself.

Sector Map — Who Sells, Who Gets Sold, and Who Wins

Non-Life Insurers — The Cleanest Unwind Story

Non-life insurers (損保) are simultaneously the largest sellers of cross-held shares AND the primary beneficiaries of the capital return flywheel. They sit at the intersection of both dynamics: they sell, generate cash, buy back their own shares, improve ROE, and attract passive inflows. This double-benefit structure makes them the cleanest expression of the unwind theme.

Tokio Marine (8766), MS&AD (8725), and Sompo (8630) are the three names to focus on. All three trade on TSE Prime, all three have made near-zero cross-holding pledges with specific FY2026 timelines, and all three have demonstrated execution credibility through FY2023-2024 buyback programs. For a deeper dive on Tokio Marine specifically, see our analysis at Why U.S. Investors Need Japan’s 2.5% Yields (TSE 2026).

Auto and Parts Keiretsu — Selling Pressure vs. Buyback Offset

Auto manufacturers (Toyota, Honda) and auto parts suppliers (Denso, Aisin) are among the most cross-held sectors per BOJ flow-of-funds data. This means they face selling pressure as insurers and banks liquidate positions — but simultaneously, the Toyota Group and Hitachi Group are unwinding their own internal cross-holdings as part of group restructuring programs.

The net effect is ambiguous for auto stocks. Selling pressure from insurer exits is partially offset by buybacks funded by the parent companies’ own unwind proceeds. Toyota Industries (6201) and Denso (6902) both raised dividends materially in FY2024 as their unwind programs progressed — a nuance that requires reading the Japanese-language 統合報告書 (integrated reports) to fully appreciate.

METI’s グループ・ガバナンス・システムに関する実務指針 (Group Governance Guidelines) and the 2023 企業買収における行動指針 (M&A Guidelines) explicitly encourage keiretsu unwinds — providing regulatory tailwind for group restructuring that English-language coverage has largely missed.

Regional Banks — The Laggard Cohort and FSA Enforcement Risk

Regional banks (地銀) are the laggard cohort. They hold significant cross-shareholdings in local industrial companies, but their unwind pace has been slower than megabanks. The nuance missed by English-language coverage: some regional banks are selling cross-held shares not to fund buybacks, but to shore up loan-loss provisions as their regional economies face demographic headwinds.

For U.S. investors, regional banks are a lower-conviction play on the unwind theme. The FSA enforcement risk cuts both ways — pressure to unwind may accelerate, but the proceeds may not reach shareholders. Stick with megabanks and non-life insurers for the cleanest capital-return exposure.

U.S. Investor Playbook — ADRs, ETFs, and Direct Access to the Unwind Theme

Direct Insurer Plays — ADR Liquidity and Valuation Entry Points

All three major non-life insurers have OTC ADR programs accessible to U.S. investors: Tokio Marine (TKOMY), MS&AD (MSADY), and Sompo (SMPNY). Liquidity is thin relative to U.S. large-caps — typical daily volume in the hundreds of thousands of ADR shares — but sufficient for retail-sized positions up to approximately $50,000-100,000.

For larger positions, direct TSE access via Interactive Brokers (IBKR) is preferable. IBKR provides direct TSE execution with competitive JPY/USD spreads, and the bid-ask on TSE Prime shares is tighter than the OTC ADR market. U.S. investors can hold TSE-listed shares in a taxable brokerage account at IBKR; IRA eligibility for foreign equities varies by account type — verify with IBKR directly.

For screening and monitoring the insurer names alongside their cross-holding disclosure calendars, TradingView’s screener allows filtering TSE Prime stocks by yield and PBR — a useful starting point before diving into EDINET filings.

ETF Proxies — What You Get and What You Miss

iShares MSCI Japan ETF (EWJ) carries approximately 8% financials weight, providing broad Japan exposure but diluting the unwind theme significantly. WisdomTree Japan Hedged Equity (DXJ) overweights exporters, making it more of a yen-hedge play than a governance-reform play.

For a more targeted approach, the JPX-Nikkei 400 index overweights high-ROE companies — many of which are unwind beneficiaries — and is accessible via NEXT FUNDS ETFs on the Tokyo Stock Exchange or through a Japan brokerage account. No U.S.-listed ETF currently provides a pure-play on the cross-shareholding unwind theme; stock selection remains the most direct route.

For context on how Japan dividend yields compare to U.S. alternatives and why the governance reform narrative strengthens the yield case, see our pillar article How U.S. Investors Can Target Sub-1× PBR Japan Stocks (2026).

Timing the Trade Around Japan’s Fiscal Calendar

Japan’s fiscal year ends March 31. Cross-holding sales are often executed in Q4 (January-March) to book gains before fiscal year-end, and the updated 政策保有株式一覧 disclosures appear in 有価証券報告書 filings on EDINET in June-July (for March fiscal year-end companies).

The practical timing signal: monitor EDINET filings in February-March for 適時開示 notices of large-block equity sales, and in June-July for updated strategic shareholding lists that reveal next year’s unwind capacity. This two-step calendar gives U.S. investors a 3-6 month lead on buyback announcements that typically follow in the autumn earnings season.

On currency: the yen-unhedged vs. hedged decision matters here. Unwind proceeds are yen-denominated. If the BOJ’s 2024-2025 rate normalization cycle continues, yen appreciation could add 5-10% to USD-denominated returns. Japanese dividend withholding tax is 15.315% for U.S. treaty-eligible investors — claimable as a foreign tax credit on IRS Form 1116, making buyback-heavy names slightly more tax-efficient than pure dividend payers for U.S. taxable accounts.

Risks and Counter-View — Three Reasons the Unwind Could Disappoint

Risk 1 — The M&A Escape Valve: When Sellers Don’t Return Cash

The most substantive risk to the capital-return thesis is that unwind proceeds get recycled into overseas M&A rather than buybacks or dividends. This is not hypothetical — Tokio Marine’s acquisition of Pure Group and Sompo’s acquisition of Endurance Specialty Holdings are both examples of insurers deploying capital internationally rather than returning it domestically.

METI’s 2023 M&A Guidelines actually encourage outbound M&A as a growth strategy, creating a competing use of capital that directly undermines the shareholder-return thesis. Investors should monitor each insurer’s capital allocation commentary in earnings calls and 中期経営計画 (medium-term management plans) for signals of M&A appetite versus buyback commitment.

Risk 2 — Supply Overhang Math: Can the Market Absorb ¥8 Trillion/Year?

¥5-8 trillion per year in cross-held equity supply hitting a market where daily TSE Prime turnover averages approximately ¥3-4 trillion creates a structural overhang. The stocks being sold — auto manufacturers, electronics conglomerates, trading companies — may underperform the broader market even as the sellers’ own balance sheets improve.

The net market impact is genuinely ambiguous. Sellers improve; the stocks they sell face headwinds. U.S. investors should be selective: overweight the sellers (insurers, megabanks) and underweight the most cross-held names in auto and electronics unless those companies have independent buyback programs of their own.

Risk 3 — Regulatory Backsliding or Governance Fatigue

TSE’s “request” is non-binding. Companies can comply on paper — filing a disclosure plan — without executing. The 2015 Corporate Governance Code cycle demonstrated this pattern clearly: initial enthusiasm, boilerplate disclosures, slow implementation. Most disclosed unwind plans contain an explicit “market conditions” clause allowing boards to pause execution during market downturns.

Domestic Japanese sell-side analysts at Nomura, Daiwa, and SMBC Nikko publish Japanese-language 政策保有株式 reduction tracker reports quarterly, containing company-level execution rates versus pledged targets. These reports — unavailable in English — reveal which companies are behind schedule before it surfaces in foreign media coverage. This is the single most valuable monitoring tool for tracking execution risk on this theme.

Bottom Line — The Unwind Is Real, but Stock Selection Separates Alpha from Beta

The 持ち合い解消 unwind is the most structurally significant capital-reallocation event in Japanese equities since the post-bubble era. Unlike the false starts of 1998, 2002, and 2015, this cycle has three simultaneous enforcement mechanisms — TSE naming-and-shaming, FSA supervisory guidelines with real teeth, and a revised Corporate Governance Code requiring annual board-level justification of each holding.

The clearest beneficiaries remain the non-life insurers: Tokio Marine, MS&AD, and Sompo. These companies are simultaneously the largest sellers of cross-held shares and the primary recipients of the resulting buyback and dividend capacity. They trade at PBR 1.0-1.5x versus global insurance peers at 1.5-2.5x, with dividend yields in the 2.5-4% range and a multi-year governance catalyst that has no direct U.S. equivalent. For a U.S. dividend investor comparing Japan to a domestic financial-sector allocation, this combination of valuation discount, yield, and structural catalyst is difficult to replicate in the S&P 500 financials sector.

The practical monitoring routine: check EDINET 有価証券報告書 filings each June-July for updated 政策保有株式一覧 disclosures, and watch the autumn earnings season (October-November) for buyback authorizations funded by that year’s unwind proceeds. That two-step calendar is the single best leading indicator of shareholder return capacity for the following fiscal year — and it requires reading Japanese-language filings that the vast majority of English-language investors will never see.

The risks are real: M&A appetite, supply overhang in auto and electronics, and the ever-present possibility of governance fatigue. Size positions accordingly, diversify across the three major non-life insurers rather than concentrating in one name, and treat the megabanks as a secondary, faster-execution complement to the insurer core.

Frequently Asked Questions

Q: What exactly is 持ち合い (cross-shareholding) and why does its unwind matter for dividends?

持ち合い refers to the Japanese practice of companies mutually holding each other’s shares as a relationship anchor rather than for investment return. When these stakes are sold, the proceeds are increasingly being returned to shareholders via buybacks and dividends rather than reinvested in low-return assets. For dividend investors, this means higher payout ratios and improving EPS from share count reduction — a structural tailwind that is separate from, and additive to, underlying earnings growth.

Q: How does Japanese dividend withholding tax work for U.S. investors?

Under the U.S.-Japan tax treaty, Japanese dividend withholding tax is reduced to 15% for eligible U.S. investors (versus the standard 20.315% rate). This 15% withholding is claimable as a foreign tax credit on IRS Form 1116 for U.S. taxable accounts, effectively eliminating double taxation in most cases. IRA accounts cannot claim the foreign tax credit, making buyback-heavy names slightly more tax-efficient than pure dividend payers for tax-advantaged accounts. Always verify your specific situation with a tax advisor.

Q: Can I buy Tokio Marine, MS&AD, or Sompo through IBKR or Schwab?

Yes. All three trade as OTC ADRs in the U.S. (TKOMY, MSADY, SMPNY) accessible through any U.S. brokerage including Schwab and Fidelity. For direct TSE access with tighter spreads, Interactive Brokers (IBKR) provides direct Tokyo Stock Exchange execution. ADR liquidity is sufficient for retail-sized positions; larger allocations benefit from direct TSE access via IBKR. Saxo Bank is the preferred option for Singapore and European-based investors seeking direct TSE coverage.

Q: How do I find the 政策保有株式 (strategic shareholding) disclosures on EDINET?

Go to EDINET (edinet.fsa.go.jp), search for the company by name or ticker, and open the most recent 有価証券報告書 (annual securities report). Navigate to the corporate governance section — typically pages 40-60 of the PDF — and look for the section titled「政策保有株式」. The table discloses each holding by issuer name, share count, book value, and stated business rationale. The format is standardized across all listed companies, making cross-company comparison straightforward once you’ve read one filing.

Q: Is the cross-shareholding unwind already priced into Japanese insurer stocks?

Partially, but not fully. Japanese non-life insurers have re-rated significantly since 2023, but still trade at a meaningful discount to global insurance peers on PBR (1.0-1.5x versus 1.5-2.5x for U.S. and European comparables). The market has priced in some unwind benefit, but the multi-year execution timeline — with ¥5-8 trillion per year in ongoing supply through FY2026 — means the buyback and dividend capacity is still being revealed quarter by quarter. The EDINET monitoring approach described in this article gives investors a 3-6 month lead on announced buyback programs, which is where the remaining alpha opportunity lies.

How to Access Japan’s Cross-Shareholding Unwind Theme as an International Investor

The primary vehicles for this theme are TSE Prime-listed non-life insurers — Tokio Marine Holdings (8766), MS&AD Insurance Group (8725), and Sompo Holdings (8630) — alongside megabanks MUFG (8306), SMFG (8316), and Mizuho (8411). OTC ADRs for the insurers (TKOMY, MSADY, SMPNY) are available in the U.S. market. No dedicated cross-shareholding unwind ETF exists; direct stock selection via TSE access provides the cleanest exposure.

International investors can access these names directly through:

• Saxo Bank — full TSE coverage, available in Singapore, Japan, Europe, and most countries. Strong platform for Japan equity access. Preferred broker for our Singapore/Asia-based readers.
• Interactive Brokers (IBKR) — direct TSE access, competitive JPY/USD spread, available in the U.S. and most countries. Strong choice for U.S.-based investors seeking tighter spreads than OTC ADRs.
• Webull — lower minimums, growing TSE coverage, good for smaller position sizes and ADR access (U.S. audience).

Tax notes by country:

• Singapore: No capital gains tax on Japan stocks. Japanese dividends are subject to 20.315% withholding (15% when the Japan-Singapore tax treaty applies via your broker). Net dividend received after withholding — check with your broker on treaty reclaim procedures.
• United States: Japanese dividend withholding is 15% under the U.S.-Japan tax treaty; claimable as a foreign tax credit on IRS Form 1116. ADR holders face similar 15% withholding plus a small custody fee. Buyback-driven EPS growth has no withholding event, making it tax-efficient for U.S. investors.
• Other countries: Withholding rates vary by treaty. Check Japan’s National Tax Agency treaty list or consult your broker.

Account opening eligibility varies by country of residence. I am not affiliated with these brokers; this is general information only. Always verify current terms directly with the broker.

This article is for informational and educational purposes only. It does not constitute investment advice or a solicitation to buy or sell any security. Opinions expressed are my own and do not represent those of any financial institution. I may or may not hold positions in Tokio Marine (8766 / TKOMY), MS&AD (8725 / MSADY), Sompo (8630 / SMPNY), MUFG (8306), or SMFG (8316) at the time of publication. All investment decisions carry risk, including the possible loss of principal. Past performance of Japanese equities, buyback programs, or dividend payments does not guarantee future results. Japanese-language source materials have been summarized to the best of my ability; readers are encouraged to verify all figures against primary EDINET and JPX filings. This disclosure is made in accordance with FTC 16 CFR Part 255 regarding endorsements and testimonials. As of May 2025. See full Disclaimer for complete terms.