MS · Forward model · Bear case

The Bear case, 20 quarters out

Each segment is projected from its own operating driver, rolled up into consolidated cash flow, and discounted back to a fair value per share. The assumptions below are editable — change them and every number on this page moves with them.

READ THE FIELD NAMES WITH CARE: this is an equity-level DCF wearing an enterprise-value engine's labels, the fourth in the r40 bank series after the BAC, JPM and GS pages and the first with FOUR verticals. The page says 'Enterprise value' and 'Net cash'. What it means is EQUITY value and a deliberate zero: $436,497M of deposits and $392,556M of borrowings are FUNDING at a bank, not financing to be netted, so netCash is 0, the engine's 'Equity' IS the equity value and fair value is a true per-share number. Eight substitutions make that reading correct, all checked against the June quarter. (1) FOUR VERTICALS, NOT THREE. Morgan Stanley's segment table has four lines - Institutional Securities, Wealth Management, Investment Management AND Intersegment Eliminations - and all four are carried. They sum EXACTLY to reported total net revenues in all fourteen quarters ($11,040 + $8,856 + $1,646 - $194 = $21,348M in the basis quarter) and their pre-tax income sums EXACTLY to firm income before provision for income taxes in all fourteen ($4,262 + $2,697 + $404 - $15 = $7,348M). Without the elimination line the verticals would overstate revenue by 0.91%. Nothing is split below a reportable segment: the firm publishes revenue sub-lines inside every segment but pre-tax income ONLY at segment level, so Banking cannot be separated from Markets and the Wealth fee business cannot be separated from the Wealth spread business. (2) 'EBITDA margin' is PRE-TAX INCOME margin, already net of segment non-interest expenses AND of each segment's own provision for credit losses. Verified line by line: 11,040 - 6,707 - 71 = 4,262; 8,856 - 6,132 - 27 = 2,697; 1,646 - 1,242 - 0 = 404; -194 - (-179) = -15. It is NOT pre-provision pre-tax income, which would lift Institutional Securities from 38.61% to 39.25%. (3) capexIntensity is 0.2% of revenue as a premises proxy on the three operating verticals and EXACTLY 0 on Intersegment Eliminations, whose revenue is negative in all fourteen quarters and where a positive intensity would produce a negative capex, a phantom cash inflow. Morgan Stanley discloses no capital-expenditure line at all - unlike Goldman, which discloses $502M against $509M of depreciation - so the 0.2% cannot be checked against anything and is a comparability convention, not an estimate. (4) corporate.taxRate is 23.07%, the basis quarter's OWN GAAP effective rate ($1,695M over $7,348M, which the release rounds to 23.1%). THERE IS NO FULLY-TAXABLE-EQUIVALENT BASIS AT MORGAN STANLEY and none is imported from the BAC or JPM precedents: the consolidated income statement runs total non-interest revenues plus net interest (income $15,902M less expense $13,122M) to 'Net revenues' with no taxable-equivalent adjustment anywhere in the 8-K, the supplement or the 10-Q, exactly as the GS page found for Goldman. The 21.4% first-half rate is depressed by a 19.6% March quarter carrying share-based-payment benefits; applying it here would overstate net income by $121M. The 22-23% full-year figure exists only on a transcript the capture marks non-authoritative and is not used. (5) overheadPctRevenue of 1.32% carries PREFERRED DIVIDENDS AND NONCONTROLLING INTERESTS and nothing else. This is a step beyond the Goldman page, which needed preferred dividends only. $145M of preferred dividends plus $72M of net income applicable to noncontrolling interests is $217M of AFTER-TAX leakage, deducted below tax on the income statement; grossed up at 23.07% that is $282.07M of pre-tax equivalent, 1.3213% of $21,348M, entered as 1.32. Segment operating expenses are already inside each vertical's margin, so a second overhead would double-count them. Without this line free cash flow would land on net income of $5,653M rather than on earnings applicable to COMMON of $5,436M, and dividing the former by common shares would overstate value by 4.0%. (6) discountRate is a COST OF EQUITY of 9.5%, derived by inverting P/TBV = (ROTCE - g)/(CoE - g) on Morgan Stanley's own numbers, not reused from another bank. Inverting at the basis quarter's 26.6% ROTCE gives 8.84% at g=3% and 9.59% at g=4%; at the first-half 26.8% it gives 8.89% and 9.64%. The band is 8.8-9.6% and 9.5% sits at the top of it. It is set BELOW BAC's 10.0% and GS's 10.5% on purpose: Wealth and Investment Management together are 49.2% of net revenues and 42.2% of pre-tax income, a fee annuity on $10.1 trillion of client assets, where Goldman is 76.3% Global Banking & Markets. THE INVERSION IS NOT CLEAN AND THIS PAGE SAYS SO: run it on the trailing-twelve-month ROTCE of 23.38% and it returns 8.04% at g=3%, which is not a credible cost of equity for a G-SIB; run it on the company's own published 20% goal and it returns 7.21%. A 4.04x tangible book cannot be justified by capitalising anything but a record half-year. (7) exitEvEbitda of 10.0x is a 13x P/E GROSSED DOWN BY TAX, because the engine applies the multiple to terminal PRE-TAX earnings: 13 x (1 - 0.2307) = 10.0009. Entering 13 would have inflated the terminal value by 30.0%. The 13x target is itself a de-rate - the tape is 17.37x trailing-twelve-month diluted EPS of $12.37 and 15.52x annualised basis-quarter EPS of $13.84 - and it is set one turn above the 12x used for Goldman for the same reason the cost of equity is set a point below. (8) SEASONALITY ON INVESTMENT MANAGEMENT ONLY, factors [0.9924, 0.9569, 0.9909, 1.0598] on calendar quarters, mean exactly 1.0000, derived by ratio to a CENTRED four-quarter moving average and NOT ratio-to-year-mean; the basis quarter is deseasonalised to $1,646M / 0.9569 = $1,720.1M before the driver runs. Institutional Securities was TESTED for the Q1-strong markets pattern that earned BAC's Global Markets its factors and REJECTED at 1.86:1 signal-to-spread, below the 2.34:1 at which Goldman's GBM was rejected. Wealth Management fails the gate outright (spread 0.0488 against signal 0.0262, 0.54:1). Intersegment Eliminations carries no factors by rule, because its revenue is negative. CALIBRATION, WHICH IS THE TEST OF ALL EIGHT: the first projected quarter's free cash flow is $5,415.8M against actual 2026 Q2 earnings applicable to Morgan Stanley common shareholders of $5,436M - a -0.37% error, against -0.05% on the BAC page and -0.43% on the GS page. The static version with no growth and no glide - ($11,040 x 38.61%) + ($8,856 x 30.45%) + ($1,646 x 24.54%) + (-$194 x 7.732%) = $7,348.0M of segment pre-tax, less 1.32% x $21,348M of overhead ($281.79M), less 0.2% x $21,542M of capex ($43.08M), all times (1 - 0.2307) - gives $5,402.9M, -0.61%. THE ENTIRE STATIC RESIDUAL IS THE CAPEX PROXY: at capexIntensity 0 the static calibration is $5,436.0M, +0.001%. If a future revision breaks this, the tax or the margin basis has drifted. BANK ANCHORS, all 2026 Q2, because a bank valued on a DCF alone is not credible. Tangible book value per share $53.18 (on $83,602M of tangible common equity) and book value per share $67.80. ROTCE 26.6% for the quarter, 26.8% for the half and 23.38% trailing twelve months on ending tangible common equity, against a 20.7% ROE for the quarter, 21.6% ROTCE for the whole of 2025 and the 17.6% 2021-2025 average Morgan Stanley published itself on a ROTCE series running 9.3 / 9.2 / 13.5 / 13.4 / 15.2 / 19.8 / 15.3 / 12.8 / 18.8 / 21.6 for 2016-2025. CAPITAL: Standardized CET1 14.9% ($87,568M of CET1 over $589,397M of RWA) against an 11.8% requirement - 4.5% minimum plus a 7.3% buffer that is a 4.3% stress capital buffer plus a 3.0% G-SIB surcharge - so 306 basis points and $18.0B of headroom. STANDARDIZED IS BINDING, and the 10-Q says so directly: headroom is 3.1 points Standardized against 6.2 Advanced (16.2% versus 10.0%), 2.0 on Tier 1 leverage (6.0% versus 4.0%) and 1.4 on the SLR (4.9% versus 3.5%, the requirement cut by the eSLR reform adopted 1 January 2026). The 11.8% requirement holds until 1 October 2027 on the Federal Reserve's own confirmation. BUYBACK AND DIVIDEND: $3,250M repurchased in the first half ($1,750M in March at an average $169.15, $1,500M in June at an average $197.64), against a $20B multi-year authorisation reauthorised without an expiry from the third quarter of 2026 - about 6.2 half-years, or 3.1 years, at that pace. The quarterly dividend was raised 15 cents to $1.15 payable 14 August 2026, a 2.14% yield on the $214.86 tape. THE PAYOUT PREMISE, WHICH IS NOT BAC'S AND NOT GOLDMAN'S. Morgan Stanley returned $6,334M to common in the first half - $3,250M of buybacks and $3,084M of dividends, the $3,385M cash-dividend line less $301M of preferred - against $10,847M of first-half earnings applicable to common. That is a 58.4% payout, NOT the roughly 100% of BAC or the 99.5% of GS. And unlike Goldman's, the gap is not a policy choice being spent down: holding the 31 December 2025 Standardized CET1 ratio of 15.05% against the $36,882M of Standardized RWA growth in the half required retaining $5,551M, 51.2% of what was earned, which implies a 48.8% SUSTAINABLE payout. Morgan Stanley paid 58.4% and the ratio duly slipped from 15.05% to 14.86%. So 'earnings are distributable' is HONEST HERE ONLY AT ABOUT HALF: the engine's implicit full distribution overstates distributable cash by roughly $11.1B a year while RWAs grow at the first-half pace. That is stated in prose rather than plugged into capexIntensity, because programme spend is deducted pre-tax and would corrupt the calibration above. LIMITATION ONE, THE LABELS: 'Enterprise value' is equity value and 'Net cash' is a chosen zero, not a measured balance. Read them any other way and the page misleads. LIMITATION TWO, THE SHARE COUNT: the engine divides by a flat 1,571,931,108 period-end BASIC shares, the same convention the BAC, JPM and GS pages use. That is CORRECT arithmetic - the model discounts total firm earnings and divides by today's count, and buybacks at fair value are NPV-neutral, so shrinking the count would double-count them. What it misstates is the per-share EARNINGS path: the terminal four quarters produce $25.007B of free cash flow, $15.91 of implied EPS on today's count but $17.84 on the 1,401.5M shares the first-half buyback pace of 17.94M shares a half-year (1.14% of the count, 2.27% a year) would reach by 2031 - so the model's implied terminal EPS runs 10.8% below a Street figure built on a shrinking count. The buyback shows up here as price appreciation, not as EPS growth. THE DILUTED ALTERNATIVE: average diluted shares were 1,569 million in the basis quarter - UNUSUALLY, 2.9 million BELOW period-end basic, 0.19%, because the count is falling through the quarter. Substituting it would RAISE fair value from $190.44 to about $190.80, +0.19%. This is the opposite sign and a tenth of the magnitude of the Goldman case, where average diluted ran 2.2% higher and flipped the headline from +0.8% to -1.4%. THE THREE-WAY CROSS-CHECK, AND IT IS WORSE THAN GOLDMAN'S - which is the point of the page, not a footnote. The tape is 4.0402x tangible book. This DCF is $190.44, 3.581x. The Gordon identity at a 9.5% cost of equity and 4% growth justifies 4.145x ($220.46, +2.6% on the tape) on the record 26.8% first-half ROTCE, 3.524x ($187.39, -12.8%) on the trailing-twelve-month 23.38%, 2.909x ($154.71, -28.0%) on Morgan Stanley's own published 20% goal, and 2.473x ($131.50, -38.8%) on the 17.6% 2021-2025 average the company printed itself. The DCF lands on the trailing-twelve-month rung, not the record-half rung, and the three checks AGREE only at the top of the ladder and only because that rung capitalises one strong half-year. Goldman's equivalent ladder bottomed 40% below on a through-cycle 16%; Morgan Stanley's bottoms 38.8% below on a five-year average the company chose to publish. THE MOST SENSITIVE INPUT IS THE EXIT MULTIPLE and the second is the cost of equity, both measured on this model rather than asserted: 8.46x pre-tax (an 11x P/E) gives $170.07 and 11.54x (a 15x P/E) gives $210.81, plus or minus 10.7%, while 8.5% and 10.5% costs of equity give $198.01 and $183.26, +4.0% and -3.8%. Both inputs are doing the same job - deciding whether the 26.6% quarter is the new normal - and nothing about the DCF's construction settles that. TERMINAL SANITY CHECK: $25.007B of terminal-year earnings would be a 20% ROTCE - the company's own goal - only on $125.0B of tangible common equity, 49.5% above today's $83,602M, which is 8.4% a year. At the 58.4% payout actually run in the first half, retaining 41.6% of roughly $22B a year adds about $9.2B a year, 10.9% of today's tangible book. The terminal earnings are therefore consistent with a growing equity base at the company's published return goal, not with a permanently record return. SEGMENT AND SOURCE BASIS: U.S. GAAP throughout, fiscal quarters are calendar quarters, so 2026 Q2 is the June quarter and seasonal index 0 is January-March. From 1Q26 Morgan Stanley STOPPED publishing its ex-DCP non-GAAP net revenue and compensation measures - it now hedges deferred-cash-plan awards with derivatives - so 2026 has one presentation and every figure here is the GAAP one. All fifty-six segment data points are DISCLOSED, none derived: they are printed in the Consolidated Financial Summary of the quarterly financial supplements, and none is flagged estimated. THE LOCAL SERIES MUST NOT BE USED as a cross-check: data/companies/ms/series.json carries revenue on a net-revenues-LESS-provision basis from 2024 Q1 to 2025 Q4 and switches to gross net revenues from 2026 Q1, so it sits 0.1% to 1.2% below this history in eight of ten stored quarters and its 2025 Q4 figure of $17,872M is $18M below what Morgan Stanley reported. The history here comes from the filings. Reference price is the $214.86 close captured on 27 August 2026 in data/prices-close.json; data/companies/ms/profile.json still carries a stale $227.09 and a stale 1,589,309,311 share count, and using either would have overstated P/TBV by 5.7% and the count by 1.1%. NOT ASSERTED: no consensus block exists for MS in this repository, so nothing on this page says the quarter beat or missed. The only forward guides on the call - a modest sequential Q3 NII increase and a 22-23% full-year tax rate - exist solely on a transcript the capture marks NON-AUTHORITATIVE and are not model inputs.

Where each case comes from

From cash flow to fair value

The projected path

Every row is projected. A year-over-year change is shown only where the quarter it compares against exists — an em dash means there is no comparable quarter, not a flat year.

Model revisions

Assumptions are marked to reality as each quarter prints. Every change is appended here, with the fair value the model produced at the time, so the model's own history stays visible.