Buffett’s Japan Stocks Portfolio (2026): The Complete Picture After Tokio Marine

Reading Kabutan’s segment-by-segment breakdown of Mitsubishi Corp’s FY2026 決算短信 on my phone at Shinjuku station — hours before Bloomberg ran a paragraph summary — I realized the full Berkshire Japan story still hasn’t been told in English: this is no longer just a trading-house bet, it’s a coherent, float-driven, governance-anchored portfolio that dividend investors outside Japan can actually replicate.

Berkshire Hathaway’s Japan position has evolved from a headline-grabbing 2020 anniversary trade into one of the most carefully constructed international portfolios in the company’s history. Six names, two asset classes, one funding mechanism, and a unifying philosophy rooted in capital discipline and long-duration compounding. This article assembles the complete picture — from the original sogo shosha stakes through the March 2026 Tokio Marine move — and translates it into a concrete, actionable framework for US dividend investors who want to mirror the trade without Berkshire’s balance sheet. Read our Megabanks & Insurance Sector Overview for broader context on Japan’s financial sector before diving in.

Disclosure: The author may or may not hold positions in securities mentioned in this article. Nothing here constitutes investment advice. Please read our full Disclaimer before acting on any information. This article complies with FTC 16 CFR Part 255.

How Berkshire Built a $25B+ Beachhead in Japan’s Trading Houses (2020–2025)

On August 30, 2020 — Berkshire Hathaway’s 80th anniversary — Warren Buffett disclosed simultaneous 5% stakes in all five of Japan’s major sogo shosha (総合商社): Mitsubishi Corp (8058), Mitsui & Co (8031), Itochu (8001), Sumitomo Corp (8053), and Marubeni (8002). The initial outlay was approximately ¥600 billion. The announcement was made in Japanese, filed in Japan, and covered in detail by Nikkei Shimbun and Toyo Keizai before most US investors had finished their morning coffee. That information asymmetry — Japanese-language primary sources preceding English-language summaries by hours — has characterized the entire Berkshire Japan story ever since.

The Yen-Bond Carry: Why Berkshire Borrowed in Japan to Buy Japan

The funding mechanism is the trade’s most elegant feature and the detail most frequently omitted from English-language coverage. Berkshire did not convert dollars to yen to buy these stocks. Instead, it issued yen-denominated bonds in the Japanese market — borrowing at near-zero Japanese interest rates — and used those yen proceeds to purchase yen-denominated equity. The result is a structurally currency-neutral position: yen dividend income services yen bond coupons; yen equity appreciation repays yen principal at maturity. Berkshire is not speculating on yen direction. It is locking in a structural yield spread between the dividend yield on the basket (approximately 3–5% per name at the time of purchase) and its yen borrowing cost (sub-1%). That spread, compounded over a decade or more, is the return engine — regardless of what USD/JPY does.

This structure also signals something important about time horizon. Issuing long-duration yen bonds is not a trade you put on for 18 months. It is a statement that Berkshire expects to hold these positions for as long as the bonds are outstanding — potentially 10 to 30 years.

From 5% to 10%: The Quiet Accumulation Timeline

After the initial 5% disclosure, Berkshire steadily accumulated additional shares across all five names. By 2024–2025, stakes had been ramped to the 7.5%–10% range, with Berkshire publicly stating it had received permission from each company to exceed the standard 10% disclosure threshold if it chose to do so — an extraordinary signal of trust from management teams that are culturally conservative about large foreign shareholders. Total exposure across the five trading houses now exceeds $25 billion, making this one of the largest single-country equity concentrations in Berkshire’s history outside the United States.

Mitsubishi Corp (8058) offers the clearest window into what that accumulation has delivered. The company completed a ¥1 trillion share repurchase program on March 25, 2026, buying back 318,397,611 shares — reducing the share count meaningfully and enhancing per-share dividend and earnings metrics for all remaining shareholders, including Berkshire. Then, on May 1, 2026, Mitsubishi announced FY2026 consolidated results and FY2027 guidance: net profit forecast up 37% to ¥1.1 trillion, annual dividend raised ¥15 to ¥125 per share. Kabutan (株探) published the segment-by-segment breakdown — including granular data on Natural Gas, Mineral Resources, Food Industry, and Urban Development contributions — within hours of the 決算短信 (kessan tanshin) filing on EDINET, well before Bloomberg or Reuters ran comprehensive English summaries. The stock surged 9.7% on the day.

What Sogo Shosha Are (and Aren’t)

US investors often reach for the nearest mental model — “commodity trader” — and miss the point entirely. Sogo shosha are diversified industrial conglomerates with global supply chain management capabilities, financial services arms, real estate portfolios, food distribution networks, and infrastructure investments layered on top of their resource and energy businesses. Mitsubishi Corp operates across ten segments including Natural Gas, Mineral Resources, Petroleum & Chemicals, Industrial Infrastructure, Automotive & Mobility, Food Industry, Consumer Industry, Power Solutions, and Urban Development. Itochu has deliberately shifted toward downstream consumer businesses and consistently delivers superior ROE relative to peers. Marubeni has strengthened its balance sheet and forecasts 7% net profit growth. These are not pure-play commodity bets — they are diversified cash flow machines with explicit shareholder return targets published in their 中期経営計画 (medium-term management plans).

The Tokio Marine Move That Changes the Thesis (March 2026)

Five trading houses was already an extraordinary Japan position. But in March 2026, Berkshire’s National Indemnity Company (NICO) acquired a 2.49% stake in Tokio Marine Holdings (8766) for approximately ¥287.4 billion (~$1.8 billion), and the entire Japan thesis shifted from “capital-disciplined conglomerates” to something more structurally Berkshire-native: a float-driven insurance play layered on top of the existing industrial foundation.

Float Meets Float — Why Berkshire Wants a Piece of Japan’s Largest P&C Insurer

Tokio Marine Holdings is Japan’s largest property and casualty insurer, with a consolidated investment portfolio approaching ¥10 trillion. For readers familiar with Berkshire’s model, that number is the key: Berkshire’s entire competitive advantage is built on insurance float — premiums collected before claims are paid, invested in equities and bonds in the interim. Tokio Marine runs the same playbook at scale in Japan and across Asia. NICO acquiring a stake in Tokio Marine is not a passive equity investment. It is Berkshire recognizing a kindred institution — a company that thinks about capital allocation the way Berkshire does — and buying a seat at the table.

The reinsurance partnership announced alongside the equity stake deepens this reading. Berkshire and Tokio Marine are not merely co-investors; they are now business partners, with NICO providing reinsurance capacity to Tokio Marine’s global book. Japanese financial press covered the strategic rationale from Tokio Marine’s side — including commentary from Tokio Marine’s IR team on how a Berkshire partnership enhances their global reinsurance capacity — in detail that was largely absent from English-language wire coverage at the time. Tokio Marine’s 有価証券報告書 (yukashoken hokokusho) filed on EDINET provides the full ¥10 trillion investment portfolio composition that underpins the float-parallel argument for readers who want to verify the thesis directly.

The 9.9% Authorization: Optionality Built Into the Deal Structure

The initial 2.49% stake is almost certainly not the endpoint. NICO received authorization to scale up to 9.9% — a structure that precisely mirrors the sogo shosha playbook: initial disclosure at a modest percentage, management approval to exceed standard thresholds, and quiet accumulation over time. Investors who watched Berkshire’s trading-house stakes grow from 5% to near 10% over four years should interpret the 9.9% authorization on Tokio Marine as a roadmap, not a ceiling. The yen-bond funding mechanism is again in use, preserving the currency-neutral structure that has defined the entire Japan program.

Reinsurance Partnership — Strategic Glue Beyond the Equity Stake

Equity stakes can be sold. Reinsurance partnerships create operational interdependence that outlasts any single investment decision. By weaving a reinsurance relationship into the Tokio Marine deal, Berkshire has created a structural reason to maintain and grow the equity position — the two sides of the relationship reinforce each other. For long-term investors, this is the most underappreciated aspect of the March 2026 announcement, and it is the detail that most clearly signals this is not a trade but a relationship.

Building Your Own Mini-Berkshire Japan Basket

With six positions now defined, the natural question for a US dividend investor is: how do I actually own this? The answer requires working through four practical layers: portfolio construction, valuation context, tax treatment, and broker access.

The Portfolio Table — Six Names, Three Weighting Schemes

The table below presents the six-name basket across three weighting approaches. Note that Mitsubishi Corp’s PBR of 2.00 and PER in the high-20s are outliers — the other five names pull the weighted average down to approximately PBR 1.2x and PER 10x, which is where the “deep value” characterization of the basket as a whole still holds.

TSE Ticker	Company	Equal-Weight	Approx. Yield	Approx. PBR	Approx. PER	US ADR?
8058	Mitsubishi Corp	16.7%	~2.2%	2.00	~27–28x	OTC (MSBHF)
8031	Mitsui & Co	16.7%	~3.8%	~1.3x	~9–10x	OTC (MITSY)
8001	Itochu Corp	16.7%	~3.2%	~2.1x	~10–11x	OTC (ITOCY)
8053	Sumitomo Corp	16.7%	~4.5%	~1.0x	~8–9x	OTC (SSUMY)
8002	Marubeni	16.7%	~4.0%	~1.0x	~8–9x	OTC (MARUY)
8766	Tokio Marine	16.7%	~3.5%	~2.2x	~16–17x	OTC (TKOMY)

A market-cap-weighted approach will naturally overweight Mitsubishi Corp and Itochu — the two largest by market capitalization — and underweight Marubeni and Sumitomo. A Berkshire-mirror approach would require knowing Berkshire’s exact position sizes across all six names, which is not fully disclosed; the equal-weight approach is the most practical starting point for most retail investors. The Mitsubishi Corporation Investor Relations page and equivalent IR pages for each name provide current dividend and payout ratio data directly from the source.

One governance data point worth flagging: the Japan Exchange Group’s Corporate Governance Code revision announcement (April 2026) confirms that TSE is continuing to press Prime Market companies on capital efficiency disclosures. All six basket names are Prime Market listed and have submitted improvement plans ahead of peers — a governance quality signal not captured in any standard English-language ETF screener. The JPX website maintains a Japanese-language disclosure list tracking which companies have submitted capital efficiency plans; checking each basket name against this list is a due-diligence step that takes ten minutes and is invisible to investors relying solely on English-language research.

Tax Math for US Investors — The 15% Withholding and Form 1116 Walkthrough

Japan imposes a 20.315% withholding tax on dividends paid to foreign investors as a default. However, under the US-Japan tax treaty, this rate is reduced to 15% for US residents holding shares directly (either TSE-direct or via OTC ADR). The practical implication: on a ¥125 Mitsubishi dividend, you receive approximately ¥106.25 after Japanese withholding. The withheld ¥18.75 is not lost — it is recoverable as a foreign tax credit on your US federal return via Form 1116, subject to the passive income basket limitation. For investors in taxable accounts, the net after-credit cost of the withholding is typically zero or near-zero, assuming sufficient US tax liability to absorb the credit. For investors holding in tax-advantaged accounts (IRA, 401k), the withholding is generally not recoverable — a meaningful drag that makes taxable accounts the preferred wrapper for this basket.

Broker Access and Execution — TSE Direct vs. OTC ADR Trade-offs

All six names have OTC ADRs tradeable through standard US brokers, but liquidity varies significantly. The trading-house ADRs (MSBHF, MITSY, ITOCY, SSUMY, MARUY) trade with wider spreads and lower volume than their TSE-listed parent shares. For investors comfortable with international trading, Interactive Brokers, Saxo Bank, and Webull all support direct TSE execution with competitive commission structures. TSE-direct trading gives you the tightest spreads, access to the full order book, and the ability to participate in shareholder meetings — but requires a yen-denominated account and attention to TSE trading hours (09:00–15:30 JST, which is overnight for US East Coast investors). For most retail investors starting out, OTC ADRs are the path of least resistance; for those building meaningful positions, TSE-direct is worth the setup cost. There is no single ETF that cleanly mirrors this six-name basket — WisdomTree Japan SmallCap Dividend (DFJ) and iShares MSCI Japan Value (EWJV) are partial proxies at best, with materially different holdings and no specific sogo shosha concentration.

The Five Disciplines Behind Buffett’s Japan Selection Framework

The six positions are not random. They share a common investment logic that, once extracted, becomes a reusable framework for evaluating any Japanese equity — not just these six names.

Capital Allocation as the Moat — Why Explicit ROE Targets Matter More Than Sector

Discipline 1: Explicit capital allocation targets. Every name in the basket has published ROE and dividend payout targets in its 中期経営計画 (medium-term management plan). Mitsubishi’s “Corporate Strategy 2027” is the most detailed example — specifying segment-level ROE hurdles, a minimum dividend payout ratio, and a buyback authorization framework. This is not boilerplate. These targets create accountability: management compensation is linked to them, and TSE governance reforms now require public disclosure of progress. Berkshire is not buying sectors; it is buying managements that have committed, in writing, to treating shareholders as partners.

Discipline 2: Yen-bond funding as a long-duration signal. The yen-bond carry is not merely a clever financing tactic — it is a public commitment to a holding period measured in decades. You do not issue 10-year yen bonds to fund a position you plan to exit in three years. Every additional yen-bond issuance by Berkshire is a renewal of that commitment, visible to anyone watching the Japanese bond market. The FSA’s April 2026 Corporate Governance Code revision — which elevates executive remuneration disclosure and diversity policies to principle status — reinforces the long-term governance improvement trajectory that makes a decade-long holding period rational.

The Yen-Bond Carry as a Signal, Not Just a Tactic

Discipline 3: Quiet accumulation without activist pressure. Berkshire has never filed an activist letter in Japan, never demanded board seats, never issued public ultimatums. It accumulates via open market purchases and negotiated block trades, builds relationships with management over time, and lets the governance reforms and shareholder return programs play out on their own schedule. This approach is culturally calibrated to Japan’s corporate environment, where activist pressure from foreign shareholders triggers defensive reactions rather than constructive change. The results — ¥1 trillion buybacks, ¥15 dividend hikes, explicit ROE targets — speak for themselves.

Discipline 4: Deep-moat industries that do not require constant reinvention. Sogo shosha have been managing global supply chains for a century. Tokio Marine has been writing P&C insurance in Japan since 1879. These are not businesses that need to pivot to AI or reinvent their business model every five years. They generate durable cash flows from structural positions in resources, finance, industrial trade, and insurance — exactly the kind of businesses Buffett has always preferred. What he deliberately avoided in Japan is as instructive as what he chose: no megabank pure-plays (despite Japan’s banks trading at deep discounts to book), no technology conglomerates, no consumer cyclicals. The negative selection reveals the framework as clearly as the positive selection does.

What Buffett Deliberately Avoided — The Negative Selection

Discipline 5: Compounding shareholder return — buybacks plus dividend growth. Mitsubishi’s ¥1 trillion buyback combined with its ¥15 dividend hike is the clearest recent illustration of this discipline in action. The buyback reduces share count; the dividend hike increases per-share income; the combination compounds per-share value at a rate that exceeds the headline earnings growth rate. Marubeni’s 7% net profit forecast with accompanying dividend growth follows the same pattern. Itochu’s premium valuation relative to peers reflects the market’s recognition that superior ROE, sustained over time, justifies a higher multiple. Berkshire is not buying yield; it is buying yield growth, and the distinction matters enormously over a 10-year horizon.

The BOJ rate path adds a nuance worth tracking. With the policy rate held at 0.75% as of April 28, 2026 — in a 6-3 split vote signaling growing internal pressure — and Fitch Ratings forecasting an additional 75bp of tightening in 2026 to reach 1.5%, Berkshire’s yen borrowing cost will rise modestly. At 1.5%, the carry spread narrows but does not disappear — a 3–4% dividend yield against a 1.5% funding cost still generates a meaningful positive spread. The thesis survives moderate rate normalization; it only breaks under a scenario of sustained, aggressive BOJ tightening well beyond current forecasts.

What Could Come Next — Pattern-Matching Berkshire’s Japan Playbook

Important disclaimer: The following is speculative pattern-matching based on Berkshire’s known preferences and publicly observable criteria. This is NOT a prediction. I have NO insider knowledge of Berkshire’s intentions, and nothing below should be interpreted as a forecast of Berkshire’s actual investment decisions.

With six positions established and a clear selection framework visible, it is analytically reasonable to ask: what other Japanese companies fit the same pattern? The criteria derived from Berkshire’s existing six positions suggest the following filters: dividend yield above 3%, PBR below 1.5x at time of purchase, explicit shareholder return targets in a published medium-term plan, a global or diversified business mix rather than purely domestic exposure, a deep-moat competitive position, and a business model that generates yen cash flows sufficient to service yen-funded debt.

The Insurance Expansion Hypothesis — MS&AD and SOMPO as Pattern Fits

The most obvious pattern extension is within the P&C insurance sector. MS&AD Insurance Group (8725) and SOMPO Holdings (8630) are Japan’s second and third-largest P&C insurers, respectively, and both share the float-advantage logic that made Tokio Marine attractive. Both have historically traded at lower PBR multiples than Tokio Marine, both have published capital efficiency targets under TSE governance pressure, and both have global reinsurance operations that would be consistent with a Berkshire partnership structure. Whether Berkshire would find the same management chemistry and strategic alignment that it found with Tokio Marine is unknowable from the outside — but the financial profile would fit the pattern. This is pure speculation; investors should not position in MS&AD or SOMPO on the assumption that Berkshire will follow.

Beyond Insurance — Industrial and Utility Candidates

Beyond insurance, Japan’s post-deregulation power sector and capital goods manufacturers with explicit ROE targets and global operations could theoretically fit the Berkshire criteria — though Berkshire has no current Japan utility exposure and has not signaled interest in the sector. The more important takeaway for individual investors is not which specific names Berkshire might buy next, but rather that the five-discipline framework described in the previous section is itself a reusable stock-selection tool. Any Japanese equity that satisfies all five criteria — explicit capital allocation targets, durable moat, quiet management culture, deep-moat industry, and compounding shareholder return — deserves a place on a watchlist, regardless of whether Berkshire is involved.

Closing disclaimer on this section: This is pure pattern speculation. Berkshire’s actual decisions depend on price, availability, management relationships, regulatory approvals, and factors that are entirely unobservable from the outside. Do not buy any security on the assumption that Berkshire will invest in it.

Risks to the Buffett Japan Thesis

Intellectual honesty requires stress-testing the thesis against its most serious vulnerabilities. There are four that deserve substantive treatment, not generic disclaimers.

Risk 1 — Yen Reversal and the Retail Investor’s Missing Hedge

Berkshire’s yen-bond funding structure creates a natural hedge: yen liabilities offset yen assets, so currency moves largely cancel out at the portfolio level. A US retail investor holding TSE-direct positions or OTC ADRs has no such hedge. If USD/JPY reverts from the current ~150 range toward 130 — a 13% yen appreciation — the dollar value of the basket declines by 13% before any change in underlying equity prices. This is not a tail risk; the yen has moved 20–30 points in a single year multiple times in the past decade. Investors with meaningful Japan exposure should consider whether partial currency hedging via FX forwards or currency-hedged ETFs makes sense for their portfolio, accepting that hedging costs money and erodes the yield advantage. Japanese retail investors are themselves net sellers of domestic equities and net buyers of overseas assets, partly in response to yen weakness — a behavioral signal worth noting.

Risk 2 — Commodity Cyclicality Across Multiple Names Simultaneously

Sogo shosha are diversified, but they are not immune to commodity cycles. Mitsubishi Corp’s FY2026 net profit fell 15.8% year-over-year — a meaningful decline — before the FY2027 recovery forecast of +37%. A synchronized commodity downturn affecting oil, LNG, and base metals simultaneously could compress earnings and, potentially, dividends across all five trading-house names at once. The diversification within each sogo shosha (food, consumer, infrastructure alongside resources) provides some buffer, but the resource and energy segments remain the primary earnings drivers in most years. Investors should size the basket with the understanding that a commodity bear market could produce two to three years of flat or declining dividends before the recovery cycle reasserts itself. The METI supply chain resilience initiatives announced in April 2026 provide some structural support for Japan’s energy procurement, but do not eliminate commodity price risk.

Risk 3 — Tokio Marine Catastrophe Concentration

Tokio Marine’s US homeowner insurance book, written through its Pure Group subsidiary, carries meaningful concentration risk in hurricane and wildfire-exposed geographies. A severe US catastrophe season — a repeat of 2017 or 2005 in terms of insured losses — could impair Tokio Marine’s earnings materially and create strain in the reinsurance partnership with NICO. This is precisely the kind of risk that makes insurance investing inherently lumpy: years of steady premium income interrupted by episodic large losses. Investors in Tokio Marine (8766) should treat it as a long-duration position where any single year’s earnings are less meaningful than the 10-year average combined ratio and investment return.

Risk 4 — Berkshire Successor Risk and the Post-Buffett Question

The Japan thesis is, at its core, a Buffett thesis. The yen-bond funding structure, the relationship-based accumulation approach, the patience to hold for decades without activist pressure — these reflect a specific investment philosophy and a specific set of management relationships built over years. Greg Abel, Berkshire’s designated successor, is a capable capital allocator, but his track record is in energy infrastructure, not international equity investing. There is a legitimate question about whether post-Buffett Berkshire will maintain the same Japan conviction, continue issuing yen bonds, and resist the temptation to trim positions when the next bear market creates pressure to deploy capital elsewhere. This is not a reason to avoid the basket — the underlying businesses are excellent regardless of what Berkshire does — but it is a reason not to treat Berkshire’s continued ownership as a permanent guarantee of the thesis.

Bottom Line

Recommendation: Buy (basket) — with eyes open on currency and commodity risks.

Berkshire’s Japan portfolio is the most carefully constructed international equity position the company has assembled in decades. Six names, two asset classes, one funding mechanism, and five repeatable selection disciplines that any investor can apply independently of Berkshire’s involvement. The blended basket yield of approximately 4.0%+ on a weighted PBR of roughly 1.2x, against a backdrop of accelerating TSE governance reform and explicit shareholder return commitments from every name in the portfolio, represents a compelling risk-adjusted entry point for long-duration dividend investors.

The Tokio Marine addition is the detail that most clearly reveals Berkshire’s intent: this is not a trade on cheap Japanese equities, it is a structural bet on Japan’s governance renaissance, funded in yen, designed to compound for decades. The reinsurance partnership transforms an equity stake into an operational relationship — the kind of durable connection that outlasts any single investment cycle.

For US investors, the practical implementation path is straightforward: six names, equal-weighted as a starting point, held in a taxable account to preserve the Form 1116 foreign tax credit benefit, accessed either via OTC ADRs for simplicity or TSE-direct for tighter spreads. Size the position to reflect the genuine risks — yen appreciation, commodity cyclicality, catastrophe exposure, and post-Buffett uncertainty — and treat it as a 10-year holding, not a 12-month trade. That is precisely how Berkshire is holding it, and there is no reason to believe a shorter time horizon will deliver the same results.

For deeper analysis of individual names within the basket, see our Megabanks & Insurance Sector Overview and the related coverage in the Japan Value & Dividend Investor series — both of which provide stock-level detail on the governance metrics and capital return programs that underpin the basket thesis.

Full Disclaimer: This article is for informational and educational purposes only and does not constitute investment advice, a solicitation, or a recommendation to buy or sell any security. The author may or may not hold positions in securities mentioned. All investments involve risk, including the possible loss of principal. Past performance is not indicative of future results. Japanese equity investing involves additional risks including currency risk, regulatory risk, and market liquidity risk. Please read our full Disclaimer before making any investment decision. This disclosure complies with FTC 16 CFR Part 255.