What the business is worth from its cash flows, growth and risk: the probability-weighted average (expected value) of 2,000 simulated outcomes, with the discount rate at its estimate. The median and the base case, the model run with every input at its estimate, sit beside it.
| Measure | Value |
|---|---|
| Expected value (headline) | ¥2,871 |
| Median of the simulated values | ¥2,801 |
| Base case (deterministic run) | ¥2,802 |
| Assumption | Value | Source |
|---|---|---|
| Risk-free rate | 2.49% | 3.05% latest 10Y govt bond (FRED IRLTLT01JPM156N, 2026-09) - 0.56% sovereign default spread |
| Equity risk premium | 4.07% | mature 3.25% (Damodaran implied ERP, Oct 2026, less the US large caps' country risk 0.67%) + country risk 0.82% weighted by FY2026 revenue (Japan 37%, United States 28%, Asia + Australia & New Zealand 18%, rest of world 10%, other; Damodaran country risk, Jul 2026) |
| Industry (own tables) | Banks - Diversified, Japan | own financial tables, Banks - Diversified: Japan n=1 > Developed Asia-Pacific n=6 > global n=46 [ours: beta; the rest Damodaran industry averages, Jan 2026, Bank (Money Center), Japan (unlevered_beta: Global, the Japan figure is implausible)] |
| own_beta | 135.08% | its own beta against its home market (2/3 two-year + 1/3 five-year regression, or the one that exists), standard error 0.13 |
| Beta | 1.15 | industry levered beta: own financial tables, Banks - Diversified: Japan n=1 > Developed Asia-Pacific n=6 > global n=46 [ours: beta; the rest Damodaran industry averages, Jan 2026, Bank (Money Center), Japan (unlevered_beta: Global, the Japan figure is implausible)]; own-beta mix: 47% of the way from 0.98 to its own 1.35 |
| Cost of equity | 7.17% | Rf + beta x ERP (CAPM) |
| Cost of equity (mature) | 6.56% | Rf + ERP (beta 1: a mature company); from year 5 a straight line to it by year 10, and the terminal's |
| Terminal growth | 2.49% | = risk-free rate (long-run nominal growth) |
| ROE (history) | 7.05% | median ROE, 5y |
| Years of EPS consensus | 3 | years of consensus net income / diluted shares |
| Share of year 1 still to come | 39.22% | valued 2026-11-08, 61% of year 1 (the fiscal year after 2026-03-31) past: year t discounted over t - 1 + 0.39 years; 75% of year 1's cash flow counted, the rest is in the 2026-06-30 balance sheet's cash and debt |
| ROE (stable) | 6.95% | mature cost of equity + 80% of the own median ROE's excess over it (7.0%), the excess at most 3%; nothing at or below it |
| ROE (year 10) | 6.95% | last ROE (analyst, else median) moved linearly to the stable ROE |
| Payout ratio | 74.14% | median (dividends + buybacks - share issuance) / net income, 5y (74%), within 0-100%, held through the 3 consensus years, then moving linearly to the stable payout by year 10; book grows by earnings x (1 - payout) |
| Payout ratio (stable) | 64.18% | 1 - g / stable ROE: retained earnings fund growth at g |
| Spread: stable ROE (simulation) | 1.00% | own ROE behind the stable ROE (the terminal rule applied to each draw with the draw's Ke: nothing at or below it, the excess at most 3%): floor 1% (standard deviation of own ROE, 5-year rolling averages (10y): 0.7%) |
| Spread: EPS per share (simulation) | ¥25.20 | per share, year 3: analyst EPS high/low range / d2(analysts) x sqrt(1 + 1/analysts); same driver as the stable ROE |
| Spread: discount rate (simulation) | 1.00% | discount rate: one shift of the whole schedule (today's and the mature rate together), standard deviation of the estimate's own error (beta, equity risk premium); terminal growth stays MIN_SPREAD below it |
| Year | Book/share | ROE | Cost of capital | Excess/share | Present value |
|---|---|---|---|---|---|
| 1 | ¥2,003 | 12.66% | 7.17% | ¥109.86 | ¥80.28 |
| 2 | ¥2,052 | 13.47% | 7.17% | ¥129.28 | ¥117.39 |
| 3 | ¥2,124 | 13.86% | 7.17% | ¥141.90 | ¥120.23 |
| 4 | ¥2,200 | 12.87% | 7.17% | ¥125.29 | ¥99.05 |
| 5 | ¥2,277 | 11.88% | 7.17% | ¥107.22 | ¥79.09 |
| 6 | ¥2,355 | 10.90% | 7.05% | ¥90.53 | ¥62.39 |
| 7 | ¥2,432 | 9.91% | 6.93% | ¥72.49 | ¥46.72 |
| 8 | ¥2,508 | 8.92% | 6.81% | ¥53.09 | ¥32.03 |
| 9 | ¥2,582 | 7.94% | 6.68% | ¥32.35 | ¥18.30 |
| 10 | ¥2,652 | 6.95% | 6.56% | ¥10.32 | ¥5.48 |
Terminal value: 4.92% of the DCF value.
Reference Range, central 50% (model-implied, not yet calibrated): ¥2,533 to ¥3,320. Half of 2,000 simulated values fall inside it (median ¥2,801); 29% are above today's price. The central 80% range, the 10th to 90th percentile, is ¥2,445 to ¥3,936.
Calibration: none. We have not yet tested how often real outcomes land inside ranges like this, so it is not a confidence interval and not a hit rate. It shows how far our inputs could move the value.
Simulated: analyst EPS and stable ROE (one driver), varied within the analysts' estimate range or the company's own history. The discount rate is drawn too, one standard deviation of 1 percentage point: it prices the cash flows' risk, but our estimate of it (beta, equity risk premium) is uncertain in its own right. Terminal growth stays below it. The base case is the run at the estimated rate, so it sits below the median of the draws (a lower rate adds more value than a higher one takes). The range width (narrow, moderate or wide) and the SE Grade's reading of uncertainty use the 80% range with the rate at its estimate, so they compare companies on their own cash-flow uncertainty (the rate's error is common to all).
| Method | Value | Range | How |
|---|---|---|---|
| industry P/B | ¥2,151 | ¥1,836 – ¥3,008 | BVPS x 1.1x: P/B from ROE 9.5% (ln P/B = -0.73 + 8.45 x ROE, R² 0.54, 42 peers (3 outliers left out), Financials sector banks, Japan), SA screener, 2026-11-10 |
| own P/B history | ¥1,230 | ¥828.57 – ¥2,454 | BVPS x 10Y median P/B 0.61x |
| industry P/E | ¥2,647 | ¥2,156 – ¥3,505 | TTM EPS x trailing 13.3x: median of 54 peers (Financials sector banks, Japan), SA screener, 2026-11-10; FY1 EPS x forward 11.6x: median of 32 peers (Financials sector banks, Japan), SA screener, 2026-11-10; mean of the two |
| own P/E history | ¥3,875 | ¥1,644 – ¥8,813 | FY1 EPS (reported basis) x 10Y median P/E 15.2x |
What similar companies (and this one, historically) sell for. A different question from intrinsic value, so not weighted into it. With three or more methods, one more than 3× off their median is left out. Range: the middle 80% of each method's simulated values, with the peer multiple read across the peer group's spread, widened so it holds 80% of peers' actual multiples (own multiples: across own history), and FY1 EPS across the analysts' range. For banks and capital markets firms, industry P/B is read off their return on equity, along the line their peers' P/B follows, when that line fits well enough.
Equity risk premiums, industry betas and returns: Damodaran; bond yields: OECD (snapshot 2026-10-25). Peer betas and multiples, analyst consensus: stockanalysis.com. An estimate, not advice.
How we got the stars: the business (Quality and Moat), the price's move of at most one notch, and the rank cap.