Japanese megabank shares are, to a first approximation, a leveraged position on the Bank of Japan. The arithmetic is straightforward: the banks’ deposit costs have historically moved far less than policy rates when the BoJ hikes, while their floating-rate loan books reprice upward and maturing securities roll into higher-yielding paper — interest income and bond valuation move separately here, since existing fixed-rate holdings lose market value as yields rise, an effect treated further below — so, on the sensitivity estimates the banks themselves disclose, a meaningful share of each policy hike has tended to flow through into net interest income. That is why MUFG (8306), SMFG (8316) and Mizuho (8411) have tended to outperform the broader market on hawkish surprises and to sag when tightening expectations fade.
The topic is timely as this is written: Japan’s July national CPI is scheduled for release on August 21, 2026 (08:30 JST), and the yen traded around 159 per dollar in mid-August as markets weighed shifting Fed–BoJ rate expectations. But the mechanism below is structural. It will read the same at the next hike, and the one after that.
The deposit-beta machine: why hikes can fall largely to the bottom line
Start with the funding side. Japanese banks are financed overwhelmingly by retail and corporate deposits — the ordinary savings account, or futsu yokin, being the workhorse. The key concept is deposit beta: the fraction of a policy-rate increase that a bank passes on to depositors. Deposit betas typically rise as a hiking cycle matures. In Japan, after roughly three decades in which deposit rates were effectively zero, households have shown little inclination to shop for yield, deposit outflows have so far been limited, and the megabanks have to date passed on only a small fraction of each hike. But pass-through varies by bank, deposit type, competitive conditions and the stage of the cycle, and it can accelerate — the banks’ own published rate-sensitivity disclosures are the place to check current estimates rather than assuming the beta stays low. To the extent deposit betas remain low, funding costs stay pinned near the floor.
Now the asset side. Loans linked to short-term benchmark rates reprice within months of a policy change, and maturing securities are reinvested at whatever the JGB (Japanese Government Bond) curve now offers. When liabilities are sticky and assets float, the net interest margin (NIM) — interest earned on assets minus interest paid on funding, divided by earning assets — tends to widen with every hike. This is the single most important sentence in bank-sector analysis in Japan: the megabanks’ earnings sensitivity to BoJ policy has been high largely because their deposit franchises have, so far, demanded only a small share of each hike.
The corollary explains the sector’s long bear market. Under the negative-rate and yield-curve-control era (the BoJ held its policy rate below zero from 2016 until it exited negative rates in early 2024, and capped the 10-year yield near zero for years), the same machine ran in reverse: assets repriced down to nothing while deposit costs, already at zero, could not fall further. NIMs compressed for years. Every step of normalization since has been that compression unwinding.
Worked example: reading the JGB curve like a bank analyst
The Ministry of Finance publishes constant-maturity JGB yields daily. Here is the actual curve as of August 13, 2026, and how to read it in three steps:
| Maturity | Yield (2026-08-13, MOF) | What a bank analyst reads into it |
|---|---|---|
| 2-year | 1.651% | Market’s average expected policy rate over the horizon — the repricing anchor for floating-rate loans |
| 5-year | 2.134% | Reinvestment yield on the belly of a bank securities book |
| 10-year | 2.873% | Benchmark reinvestment yield; drives the carry story |
| 20-year | 3.741% | Super-long sector — matters more to life insurers than banks |
| 30-year | 4.002% | Same; watch for duration-supply stress, not bank NIM |
Step 1 — the level. A 2-year yield of 1.651% tells you the market is pricing a policy path well above the current rate; if you believe the BoJ will out-hike that path, bank earnings estimates are too low, and vice versa. The 2-year is the cleanest single gauge of what is already in the price of bank stocks.
Step 2 — the slope. The 2s10s spread — 2.873% minus 1.651% — is roughly 122 basis points. That slope is a rough indicator of the maturity-transformation environment: banks fund short (deposits priced off the short end) and invest long (loans and JGBs priced off the belly and 10-year), so a steepening curve is generally friendlier to future reinvestment margins and an inverting one is hostile. A bank’s actual NIM depends on its specific mix and repricing of deposits, loans, securities, hedges and wholesale funding, so treat the slope as a directional read rather than a margin forecast. Track the spread, not just the 10-year headline.
Step 3 — the reinvestment gap. Compare the 10-year at 2.873% with what the maturing bonds on a bank’s book were bought to yield. Securities purchased during the yield-curve-control years carry coupons near zero; every redemption rolled into the curve above is a step-up in interest income that recurs for the life of the new bond. This ‘back book versus front book’ gap is why interest income can keep improving for years after hikes stop — though whether that reaches earnings depends on where the curve sits when those bonds actually mature, and on loan demand, deposit pricing, hedging, funding mix and credit costs.
Revaluation: the bond book and the mochiai portfolio
Two balance-sheet effects sit alongside the margin story, and they cut in opposite directions.
- Bond revaluation is the short-term cost of higher rates. The megabanks hold JGB and foreign-bond portfolios measured in the tens of trillions of yen. When yields rise, the market value of existing holdings falls, producing unrealized losses. How much of that reaches earnings and capital depends on each bank’s portfolio classification and hedging: held-to-maturity holdings are largely insulated unless sold, while available-for-sale holdings are marked through other comprehensive income and can weigh on regulatory capital even before any sale crystallizes a loss. A violent, disorderly rise in yields spooks the market because it raises those capital questions acutely, along with the prospect of losses realized if bonds must be sold. This is why bank stocks prefer a gradual hiking cycle to a shock: same destination, very different journey.
- Cross-shareholdings are a slow-release capital return story. Japanese banks still hold large portfolios of seisaku hoyu kabu (policy-holding shares, colloquially mochiai — strategic cross-shareholdings in client companies). Under sustained corporate-governance pressure, the megabanks have been selling these down for years and recycling the gains into buybacks and dividends. A rising equity market inflates the value of what remains, adding a second, equity-linked kicker to the rate story.
The practical takeaway: when yields rise steadily, the margin effect has usually dominated and banks have tended to rally. When yields gap higher in a disorderly way — especially at the super-long end, where the curve was above 4% at 30 years as of August 2026 — the revaluation fear can temporarily overwhelm the margin story even though the long-run earnings news is good.
The eight meetings: why the surprise, not the hike, moves the stock
The BoJ’s Policy Board holds eight Monetary Policy Meetings (MPMs) a year, four of which are accompanied by the quarterly Outlook Report containing the Board’s inflation forecasts. Bank stocks do not simply rise when the BoJ hikes; they rise when the BoJ is more hawkish than the curve had priced. Before each meeting, check the 2-year JGB and overnight-index-swap pricing: if a hike is fully discounted, delivering it may produce no move — or a ‘sell the fact’ dip — while a surprise hold can hit the sector hard.
Three features make the bank-sector version of this trade distinctive:
- Banks are the natural long on hawkishness. A hawkish surprise typically strengthens the yen and pressures exporters, dragging on the broad indices — while banks tend to outperform. The sector is one of the few places in the Japanese market where higher rates are, on balance, supportive of earnings — though credit costs, bond revaluation and funding shifts can dilute the benefit — which is why relative performance around MPMs has often tracked the rate story.
- The between-meeting drift matters as much as the event. Speeches by the Governor and Board members, wage negotiation results in spring, and each monthly national CPI print reprice the path continuously. Bank stocks trade the path, not just the decision.
- The yen is the tell. Because both the yen and bank stocks key off the same rate differential, a firming yen alongside rising bank shares usually signals genuine domestic-rate repricing rather than a global risk move.
Reading the tape: data sources and common mistakes
A few practical habits, and the errors they prevent:
- Do not confuse ‘Financials’ with ‘Banks.’ In the TOPIX-17 sector framework, banks and other financials are separate categories. On August 14, 2026, for example, the ‘Financials ex-Banks’ grouping (insurers, brokers) fell 0.58% while MUFG closed at ¥3,686 (−0.62%) and Mizuho fell 1.77% — three different exposures. Insurers key off the super-long end of the curve; brokers key off market volumes; only banks are the clean NIM trade.
- Watch positioning, not just price. Japan Securities Finance publishes daily standardized margin balances per stock. As of August 13, 2026, MUFG carried a margin-loan (leveraged long) balance of roughly 3.55 million shares against a stock-lending (short-side) balance of about 508,000 shares — a heavily long retail book. Crowded longs into an MPM raise the cost of a dovish disappointment.
- Date-stamp every yield you use. The MOF publishes the official constant-maturity series daily; quoting ‘the 10-year is about 2.9%’ without a date is how stale analysis is born, because in a live hiking cycle the curve moves in basis points per day (the 10-year rose 1.7bp on August 13, 2026 alone).
- Remember the flow backdrop. Foreign flows can materially influence pricing of the megabanks; JPX publishes weekly investor-type flow data (foreigners were net sellers of about ¥0.49 trillion of TSE Prime equities in the week of August 3–7, 2026 — a market-wide figure, not bank-specific, but useful for the backdrop). A hawkish story with foreigners selling the market overall is a weaker trade than the same story with inflows.
FAQ
Why do bank stocks sometimes fall when JGB yields rise?
Two common cases. First, when the rise is fast and disorderly, the market prices the mark-to-market hit to bond portfolios before it credits the future margin gain. Second, when yields rise for global reasons (a US Treasury selloff) rather than domestic BoJ repricing, the yen may weaken and the domestic-rate story that drives bank NIMs is not actually improving. Check whether the 2-year JGB — the policy-expectation proxy — moved, not just the 10-year.
Which JGB maturity matters most for the megabanks?
The 2-year for what is priced into policy expectations, and the 2s10s slope for the margin outlook. The 20- to 40-year sector (above 3.7% across the board as of August 13, 2026) is primarily a life-insurer and fiscal-supply story; violent moves there affect bank stocks through sentiment more than through earnings.
Is a rate cut cycle simply the trade in reverse?
Broadly yes, but the symmetry is not exact, and how painful it is depends on where deposit rates sit when easing begins. The asymmetry sits on the funding side — margin gains in a hiking cycle rest on deposit costs lagging the policy rate, while in a cutting cycle the offset depends on how far deposit rates can still be reduced. If normalization has already lifted deposit rates meaningfully above zero, there is room to cut them back and cushion falling asset yields; only where funding costs are already close to the floor does easing compress NIMs with little offset. That conditionality is why the sector’s valuation is so tightly bound to the perceived durability of the BoJ’s normalization path.
Sources
- Ministry of Finance Japan — daily JGB interest rate data: www.mof.go.jp/english
- Bank of Japan — Monetary Policy Meeting schedule, statements and Outlook Report: www.boj.or.jp/en
- JPX (Japan Exchange Group) — investor-type flows, margin balances, short-selling statistics: www.jpx.co.jp/english
- Nikkei Indexes — index and volatility data: indexes.nikkei.co.jp/en
Disclaimer: This is an information and analysis publication, not investment advice. See our Methodology for data sources, standards, and our corrections policy.