Mizuho's Turnaround Is Real - The Question Is Whether You're Still Early

The headline writers say "buy Mizuho ahead of Q2 earnings." That framing is backward. The second-quarter (first-half) results for fiscal 2026 won't be announced until November - three months away. Q1 FY2026 dropped on July 30. The inflection this story is really about didn't start in the latest quarter; it started over a year ago when the market stopped looking at Mizuho as the broken bank of Japan's big three.

What matters now is whether the operating trajectory has earned the valuation it already trades at - or whether the rerating still has room to run.

The old story is stale

Until April 2025, Mizuho was the punchline among Japan's megabanks. Trump-era tariff shock drove its price-to-book ratio down to 0.703 on April 7 - well below the Tokyo Stock Exchange's own target of 1.0x. The market was pricing persistent legacy baggage: a bloated bond book, decades of cross-shareholdings propping up weak peers, and an investment banking franchise that trailed MUFG and Sumitomo Mitsui.

That story is dead. As of February 2026, Mizuho's P/B had climbed to above 1.5x, making it the highest of the three megabanks. The stock trades at roughly $10.32 (as of mid-July). The turnaround isn't a thesis anymore; it's a scoreboard.

What actually changed

Three things, in order of importance.

First, the interest rate environment finally worked in Mizuho's favor. Japan's slow climb out of near-zero rates has re-priced the loan book. Domestic loan yields rose to 1.08% in FY2025, up from 0.76% the year before. The domestic net interest margin expanded to 0.89%. Meanwhile, the cost of deposits moved to only 0.19%. That spread is the core engine: Mizuho has repriced assets faster than it had to reprice liabilities. Retail banking segment profit alone surged 86% year-over-year in the first three quarters of FY2025.

Second, capital is being returned and the share count is shrinking. FY2025 saw the company achieve a 60% total payout ratio - dividends plus buybacks - on normalized profit of ¥1,150 billion. Management has committed to maintaining a 50% or higher payout going forward and raising dividends by roughly ¥5 per share each year. Buybacks started at ¥100 billion in FY2025 and were expanded later that fiscal year. Mizuho is the only megabank to have increased its buyback program. Cross-shareholdings - the legacy holdings that trap capital - were reduced by ¥114.6 billion in FY2025, bringing total progress to 45% of the three-year target of ¥350 billion. Fewer shares, same or better earnings: that's the mechanical rerating path.

Third, operating leverage is real, not accounting. Normalized ROE reached 10.5% in FY2025. The expense ratio fell to 59.4% (57.3% through the first three quarters) as revenue growth outpaced cost increases. Return on risk-weighted assets jumped from 2.4% as of March 2019 to 3.6%. Asset quality held up - the NPL ratio improved to 0.75% as of December 2025, the end of the third quarter, despite tighter monetary policy. And the investment banking franchise, long Mizuho's weak link, is improving: domestic CIB gross profits grew 17%, IB income from large corporates grew at a 24% CAGR, and Mizuho ranked second in cross-border M&A involving Japan - a first, buoyed by the Greenhill acquisition.

Where the market is still wrong

The market's mistake isn't ignoring the turnaround. It's assuming the turnaround has finished. AInvest's aggregate signal labels Mizuho a Buy, and recent coverage acknowledges a 23% upside from valuation recalibration. But the consensus is still anchored to FY2025 results - it hasn't fully priced the compounding that happens between now and FY2027.

The bridge to the next inflection point is straightforward. Management is targeting 12% ROE by FY2028. At 10.5% in FY2025, that's a 150-basis-point climb in roughly three fiscal years. The ingredients are already in the machine: higher net interest margins (as BOJ rates continue normalizing), continued share count reduction through buybacks and cross-shareholding sales, and revenue growth in wealth management (individual AUM reached ¥68.5 trillion, up ¥4.3 trillion year-over-year) that isn't rate-dependent.

Even more direct: if normalized profit stays near ¥1,150 billion and the payout ratio holds at 50%+, returning roughly ¥575 billion annually to shareholders, the arithmetic on a shrinking share base pushes per-share metrics up mechanically. You don't need a DCF for that. Simple forward multiples on a shrinking denominator do the work.

The setup and the risk

At $10.32 (mid-July) and with a P/B that was above 1.5x in February, Mizuho isn't cheap by historical standards. But it's the highest-quality of the three megabanks right now - highest ROE trajectory, most aggressive share return program, and the only one with an in-house global CIB operation.

The thesis is: ROE continues climbing toward 12%, share count keeps shrinking, and the market eventually prices Mizuho at the same P/B as the best global banks, not the worst Japanese ones. That rerating path doesn't require earnings surprise. It requires the numbers to keep doing what they're already doing.

What could break it? A recession that reverses the rate normalization story and squeezes net interest margins. Or a credit event - NPLs are low now, but the BOJ is still tightening, and Japanese corporates haven't been stress-tested in a real rate environment since the 1990s. If credit costs spike and ROE stalls below 10%, the P/B expansion loses its anchor.

The tripwire is the ROE target. If FY2026 H1 results in November show ROE falling back toward 9% or below, or if management cuts the 12% target, the operating path has broken. Cut it.

If ROE stays above 10.5% and buybacks continue, the setup gets cleaner with each quarter. The market is still pricing the old Mizuho, even though the numbers already look like the new one.