Bank of America

IThe business

A bank takes deposits, lends them out, and keeps the difference. Bank of America does that for about 38.9 million consumer checking accounts and for companies, and it does three other things with the same balance sheet: it manages money for the wealthy through Merrill and the Private Bank, it trades stocks, bonds and currencies for clients, and it advises companies on deals. At 30 September 2026 it had $3.43T of assets, $2.04T of deposits and $1.25T of loans.⁴ It had about 213,000 employees at the end of 2025.² Common shareholders own $271.8B of equity, of which $201.8B is tangible, meaning what is left after $69.0B of goodwill and $1.8B of other intangibles.¹¹

The firm reports four businesses and a fifth bucket for what does not fit. Here is what each earned in 2025, on the firm’s taxable-equivalent basis.²

Three facts in that table matter more than the rest. First, consumer banking is the best business: 40% of net income on 27% of the assigned capital, and with wealth management, 55% of profit C (calculated).² Second, Global Banking and Global Markets use 61% of the assigned capital and earn 15% and 13% on it C (calculated), and Global Markets earned only 9% in the third quarter.⁴ Third, the unassigned $111B. If it sits in bonds yielding less than the cost of equity, it drags the return and no business owns the drag. I cannot see what it holds, so I do not build on it.

What it sells

Net interest income, the difference between what the firm earns on loans and securities and what it pays for deposits and debt, was $60.1B in 2025, 53% of revenue C (calculated).² Fees and commissions were $39.4B: $20.0B of investment and brokerage fees, $6.6B of investment banking fees, $6.5B of service charges and $6.4B of card income. Trading, which the firm calls market making, added $12.0B.² Net interest income rose 9% in the first nine months of 2026 to $48.2B, and the net interest yield was 2.16% in the third quarter, up 15 basis points in a year.³⁴ Management attributes the rise to loan and deposit growth, more interest income in the trading business, and fixed-rate assets repricing.⁴

Fees grew faster. In the first nine months of 2026 fees and commissions were $32.4B, up 11%, and asset management fees rose 17% in the third quarter to $4.5B, helped by $74B of net client flows over a year.⁴ Trading revenue was $10.2B in nine months, up 2%, and investment banking fees fell 11% in the third quarter to $1.8B.³⁴ The steady parts, interest and wealth fees, are growing. Trading and deal fees did not help in the third quarter, when revenue grew only 3% and net income fell 17%, or 2% before the settlement charge.⁴

Revenue grew 3.4% a year from 2016 to 2025, from $83.5B to $113.1B as reported, and about 3.0% on a consistent basis once the roughly $3.8B that an accounting change added to 2025 is taken out C (calculated). That is slow, because for most of the decade rates were low: net interest income was $41.5B in 2016 and $42.9B in 2021.¹⁴¹³ The jump since is the rate cycle, up 40% from 2021 to 2025 C (calculated).¹³² The accounting change, in the fourth quarter of 2025, moved about $4B a year of tax-related equity investment results from noninterest income into tax. The 2023 to 2025 figures here are restated and earlier ones are not, so I flag the break where it shows.¹⁶²

Where the customers are

The firm says it has the largest share of US consumer deposits, and the largest share of small business lending, according to bank regulatory filings for the second quarter of 2026.⁴ It had 38.9 million consumer checking accounts, 92% of them primary, with 59% of consumer deposits sitting in checking, and net growth in checking accounts for 31 quarters in a row.⁴ It spent over $100B on technology across the last decade, and its AI assistant, Erica, had 24.2 million users in the third quarter.⁵⁴ The card book had $105.8B of average balances, up 5%, and card spending was up 8%.⁴ The firm says it is first in US commercial loans and top five in nine of ten global capital markets products.⁵

The cost of being wrong is in that chart. In 2020 the firm set aside $11.3B for credit losses against a pandemic. In 2021 it released $4.6B back into profit. In 2025 it provisioned $5.7B.¹³² Over ten years the provision averaged $3.9B a year, 0.39% of year-end loans C (calculated). At 30 September 2026 the allowance for loan losses was 1.04% of loans, down from 1.14% a year earlier, and nonperforming loans were 0.44% of loans.⁴

The balance sheet

This is where a bank differs from every other company on this site, and where Bank of America differs from JPMorgan. Deposits were 163% of loans, and the firm holds $842B of debt securities, 25% of its assets.³ Of that, $496B are held to maturity, carried at cost, at an average yield of 1.91%. They were bought when rates were near zero. The ten-year Treasury yielded 4.91% on 2 November 2026, and the portfolio is worth $392.5B, $103.9B less than its cost.¹⁷³ Across all debt securities the unrealized losses net of gains are $108.8B before tax, $4.8B of it on bonds held for sale C (calculated).³ That is 54% of tangible common equity C (calculated). It is not a loss in the accounts, and not a cash loss unless the bonds are sold, but it is the largest thing on this page I cannot value precisely.

Two features make the loss less alarming than it looks. First, the deposits that funded those bonds cost less than the market rate: $548.7B of them, 27%, pay no interest, and the interest-bearing deposits cost an average of 2.00% in the first nine months of 2026, down from 2.40% a year earlier.³ That gap offsets some of the bond loss, but nobody can measure it exactly. Second, the bonds run off. If the $496B were reinvested at 4.5% instead of 1.91%, pretax income would rise by about $13B a year E (estimate), about a third of today’s pretax profit. It would take many years: $375B of it is due after ten years, almost all mortgage-backed securities, though they are repaid sooner in practice.³ That is the strongest reason to think the return can rise, and it is already in the revenue growth I use. It is not extra.

The cost is that the same bonds hold the firm’s capital down. Book value per share fell from $39.34 to $39.26 in the third quarter, and tangible book from $29.37 to $29.14, even while the firm earned $6.9B, because the firm paid out $8.3B and accumulated other comprehensive loss grew.⁴ The accumulated loss was $15.3B after tax at 30 September 2026 against $10.5B at the end of 2025.³

A bank’s operating cash flow moves with the size of its trading assets, so I do not use a free-cash-flow model here. Section VI explains the method I use.

IIThe moat

A durable advantage is a cost or a habit that a rival with money and talent cannot copy in a reasonable time. A bank’s edge is often narrower than it looks, because loans and trades are commodities and the customer can usually go elsewhere. Where Bank of America is hard to copy is in deposits, scale and breadth. I can test each against numbers.

The test of an advantage is whether it has widened. Here is the evidence both ways.

Evidence it is widening

• Tangible book value per share grew 6.1% a year from 2016 to 2025, from $16.89 to $28.73, while the firm returned 81% of its earnings and cut the share count by 28% C (calculated).¹⁴²¹⁵
• The overhead ratio, expenses over revenue, fell from 64% in 2023 to 62% in 2025, both on the restated basis, and was 62% in the third quarter of 2026 before the settlement charge C (calculated).²⁴
• Return on tangible equity rose from 9.48% in 2020 to 14.22% in 2025.¹³²
• Asset quality improved: commercial criticized exposure fell from $26.3B to $19.9B in a year.⁴

Evidence it is narrowing

• Expenses grew faster than revenue in the third quarter, excluding the settlement: up 7% against 3%, so the adjusted overhead ratio went from 60% to 62%.⁴
• The bond book is a drag. Held-to-maturity bonds yield 1.91% and the unrealized loss on them grew by $23.7B in nine months C (calculated).³
• The capital requirement is rising. The surcharge adds 0.5 point to required common equity on 1 January 2027, and two proposed rules would change it again.³
• Compliance. The Comptroller, the Treasury’s financial crimes unit and its sanctions office announced orders and a settlement on the firm’s anti-money-laundering and sanctions programs. The firm has been working on them since late 2023 and has now accrued the penalty.³

On balance the advantage is wide in deposits and breadth, and narrower than the best rival’s in returns. The question for the valuation is how much of today’s return it will keep.

IIIThe competitive advantage

The last section was about why a rival cannot copy the firm. The advantage is what the firm earns because of it. For a bank, the measure that matters most is how much profit it makes on the tangible equity it holds. The cleanest test is to put it beside the others that are as big or bigger in what they do.

The multiple tracks the return. Bank of America earns 12.6%, or 14.9% before the settlement charge, and trades at 1.7 times tangible book. Wells Fargo earns 16.3% at 1.9 times, JPMorgan 20% at 3.0, Morgan Stanley 22.6% at 3.4 and Citigroup 11.5% at 1.2.⁴⁶⁷ Bank of America sits close to the line. Compared with JPMorgan it earns about six points less and the market pays about 1.3 turns of book less, roughly what the gap in returns predicts C (calculated). I cannot call the shares cheap because they are lower. I can only ask whether the return is likely to rise.

Why it earns less than JPMorgan

Three things stand out when I set the two firms side by side. The first is cost. The overhead ratio, expenses over revenue, was 62% in the third quarter of 2026 before the settlement charge, against 54% at JPMorgan.⁴⁶ On the 2026 revenue I use, $122B, each point of overhead is about $1.2B of pretax profit C (calculated). The gap of eight points is worth about $9.8B before tax, or about 3.8 points of return on tangible equity C (calculated). Part of it is mix, since consumer and wealth banking cost more to run than a trading desk, and part is spending that the firm says will continue.⁴⁵

The second is the bond book. At the end of 2025, securities were 27% of assets, and $522.7B of them were held to maturity.² At 30 September 2026 the held-to-maturity bonds yielded 1.91%. Those yields were fixed when rates were near zero, and they are the main reason the net interest yield is only 2.16%.³⁴ Net interest yield excluding trading was 2.65% in the third quarter, up 17 basis points in a year.²⁰ The repricing is a one-way tailwind for earnings, and it is the main reason I think the return can rise from here.

The third is capital. Bank of America holds 10.9% in common equity against a requirement of 10.0%, a thin cushion, and has to raise it in January. JPMorgan holds 13.9% against a requirement of 11.5%.³⁴⁶ A thin cushion helps the return, since less equity is needed to earn the same profit, but it limits what the firm can do with buybacks and growth.

The ROTCE chart is the picture of the advantage, and of its limits. The bank earned more than my cost of equity in seven of the ten years, but by a margin of 2.3 to 6.4 points, and below it in 2016, 2017 and 2020 C (calculated).^1413122 The ten-year average was 13.2% and the five-year average to 2025 was 14.6%. The 2025 figure was 14.22% and the adjusted third-quarter figure 14.9%. JPMorgan’s ten-year average is 17.9% on its page, and its cost of equity is 9.5%.²¹ So here the picture is of a good bank that earns a bit more than it costs, rather than a great one that earns far more. It sits where the price is.

The advantage is therefore real but modest, and it is in the price. The valuation asks which return to expect for the next ten years: the 14 to 16% of today, the 17% to 18% that the best case would need, or the 9 to 10% of 2016 and 2017.

IVManagement

Brian Moynihan has been chief executive since January 2010 and chair since October 2014. He was also president until September 2025, when the board named Dean Athanasia and Jim DeMare Co-Presidents with responsibility for the lines of business.⁹²² The chief financial officer is Alastair Borthwick. The board named no successor to Moynihan, and in October 2026 it grew from 12 to 14 directors, adding James Fitterling, the former chief executive of Dow, and William Thomas, the former global chairman and chief executive of KPMG International.²³

The bench is stable, and nobody on the team has run the firm through a transition: Moynihan has been in charge for almost seventeen years. I did not find a named successor in the filings.

Pay, and what it pays them to do

The board set Moynihan’s pay for 2025 at $41.0M, up from $35.0M. That was a $1.5M salary and $39.5M of equity awards, no cash bonus.⁹¹⁰ The equity is 50% performance units, 30% units settled in cash after a year, and 20% units settled in shares over four years.¹⁰ The performance units have to be earned again over 2026 to 2028 against two tests. One is the three-year average return on assets, tax-normalized: 90 basis points for a full payout and 110 for the maximum of 150%. The other is the three-year average growth of adjusted tangible book value: 10.5% for a full payout and 12.5% for the maximum.¹⁰ The board said the 110 basis point test needs about $37B of profit a year on today’s assets, more than the firm has ever reported.¹⁰

Follow that incentive through. Return on assets is profit over the whole balance sheet, whatever equity supports it, so it does not care whether the firm holds $200B of tangible equity or $250B. The proxy defines adjusted tangible book value as common equity less goodwill and intangibles, with the effect of board-approved capital actions taken out.⁹ Dividends and buybacks are therefore added back, and the test is close to a plain return on tangible equity: earning 10.5% and paying nothing out would grow it 10.5% C (calculated). The target pays in full at about my cost of equity, and the firm earned 14.2% in 2025.⁹² So I read the targets as achievable on today’s earnings, not stretching. Neither test is per share, so neither is hurt by a buyback at a high price. They do not tell the chief executive what to pay for the firm’s own shares.

Other facts on pay. Half of the net shares Moynihan receives have to be held until a year after he retires, and all of the awards have clawbacks.⁹

Their own money

Moynihan owned 2,803,195 shares in the proxy, plus 1,724,434 unvested units, 4,527,629 in all. At $49.57 the shares are worth about $139M, and with the units $224M C (calculated).⁹¹ Directors and executive officers as a group, 25 people, owned 9,523,597 shares, about $472M C (calculated), and under 0.2% of the company.⁹ I did not check insider trades by Form 4 for this page.

The large outside holder is Berkshire Hathaway. It reported 517,295,934 shares at 31 December 2025, 513,624,165 at 31 March 2026 and 483,394,015 at 30 June 2026, worth $27.5B on that date, a 6.6% cut in the first half C (calculated).²⁴ The proxy cites a Schedule 13G/A filed on 14 November 2025 that shows 568,070,012 shares, so if the two filings count the same shares, the position is down about 15% since then C (calculated).⁹ Himalaya Capital went from 10,431,387 shares at the end of 2025 to 2,997,987 at 31 March and none at 30 June.²⁵ I do not read much into it, but two well-regarded holders were selling in the first half while the bank was buying its own stock at higher prices.

What they did with the money, over a full cycle

Over the ten years to 2025, through a pandemic, a rate surge and the 2023 bank failures, the firm earned $253B, paid $61B in common dividends and bought back $143B of stock C (calculated).¹⁵ That is 81% of earnings returned. The share count fell 28%, from 10,053M to 7,213M, and tangible book per share rose from $16.89 to $28.73, 6.1% a year, after those payouts C (calculated).¹⁴²

Three things stand out. First, the firm made no large acquisition in the period; the two it made in 2008 are the origin of today’s scale.⁵ Second, it slowed buybacks when it had to: $4.6B in 2023, after the bank failures, against $28.1B in 2019.¹⁵ Third, and this is the part I weigh most, the 2026 buybacks are large and were made at high prices. In the first nine months the firm paid $6.3B in common dividends and bought back $19.3B of stock, together 109% of the $23.4B it earned for common holders C (calculated).³ The average price was $55.21 a share, and $62.07 in the third quarter, about 2.1 times tangible book C (calculated).³ The price is $49.57 now. On my base value of $46, each dollar spent at $55.21 bought about 83 cents of value for the holders who stayed E (estimate).

The cost shows in the capital ratio. CET1 fell from 11.4% at the end of 2025 to 10.9%, because the firm returned more than it earned and risk-weighted assets grew.³ The $2.0B preferred issue in October, at 7.2%, is a sign that the firm needs Tier 1 capital to keep returning it.²⁶³ If buybacks slow in 2027, my payout in the early years is too high, and the value depends on what the capital earns instead.

The record on capital is mixed: good over ten years, with the dividend raised 14% in July to $0.32 a quarter and the share count down 28%.²⁷ But the 2026 pace is above what earnings support, and the compliance record is the other mark against management.³

VFinancials

I use calendar years, since the firm reports on them. The ten-year table has the figures the rest of this page leans on, with nine months of 2026 added in the last column. Everything is as the firm reports it, with no adjustment of mine, and there is one break in the series that I want to flag before you read it.

Revenue grew 3.4% a year and earnings per share 11.0% a year over 2016 to 2025, helped by a 28% fall in the share count C (calculated).¹⁴² Tangible book per share grew 6.1% a year C (calculated). That last figure is the number to anchor on for a bank. It is what the business adds to the owners’ capital per share after paying out most of what it earns. Profit per share grew three times as fast as revenue: the share count fell, and the overhead ratio fell from 64% in 2023 to 62% in 2025, both on the restated basis, while credit costs stayed low.

Reading 2026

The first nine months of 2026 brought net income of $24.6B on revenue of $91.8B, both up 8%.³ The third quarter included a $1.2B regulatory settlement charge that is not deductible for tax and took $0.17 off earnings per share.⁴ I take it out, but not fully. Charges outside the ordinary run have recurred: the industry-wide FDIC special assessment in the fourth quarter of 2023, a consent order in December 2024, and now this penalty.³¹² I put back the $1.2B and charge $0.6B a year to expenses, a guess at what legal and regulatory costs run at in an ordinary year E (estimate).

Two things in the normalization are judgement calls. I leave in an ordinary year for trading, and I leave in credit costs that are low: the net charge-off rate was 0.44% in the third quarter, against 0.50% for 2025 and a ten-year average of 0.40% C (calculated).⁴² A return of 15.5% on that base is what the firm earns in a good year. It is a point above the 14.2% of 2025 and just above the adjusted third quarter.

Capital, and the capital the owner can reach

This is the part of the bank that most limits what it can do. The firm’s CET1 ratio was 10.9% at 30 September 2026 against a requirement of 10.0%.³ The requirement has three parts: a 4.5% minimum, a 3.0% surcharge for being a global systemically important bank, and a 2.5% stress buffer. The cushion of 0.9 points is about $16B C (calculated). But the ratio that binds is not CET1. It is the Total capital ratio, which counts preferred stock and bonds that count as capital. That was 14.2% against a minimum of 13.5%, a cushion of 0.7 points, about $13B C (calculated).³

On 1 January 2027 the surcharge rises by 0.5 point, to 3.5%, under the rules as they stand. The required CET1 ratio becomes 10.5% and the Total capital minimum 14.0%.³ At today’s risk-weighted assets of $1,821B, that is about $9.1B of extra capital to hold, and the cushion falls to about 0.4 points on CET1, $7B, and about 0.2 points on Total capital, $4B C (calculated). In the third quarter the firm returned $8.3B and its CET1 ratio fell from 11.2% to 10.9% and its Total capital ratio from 14.7% to 14.2%.⁴¹¹ At that pace the cushion would be used in about a quarter. This is the reason I model slower payouts, and why the Series WW preferred issue came when it did.

Three further rules are not final: a March 2026 proposal on risk-based capital for large banks, a proposed change to the surcharge calculation, and a 30 September 2026 proposal on how fee income is projected under stress.³ On that date the Federal Reserve also finalized two rules: results of the two most recent stress tests will be averaged from 2029, and the buffer’s start date moves from 1 October to 1 January from 2028. The firm’s stress buffer stays at 2.5% through 31 December 2027.³ One final rule helps: since 1 January 2026 the minimum leverage ratio is 3.75%, not 5.0%, against the firm’s 5.5%.³ The firm says final rules may differ from the proposals, so I show a range in the valuation, not one number.

Reported earnings and owner cash

Reported earnings are not what the owner can take out. A bank has to keep its capital ratio as its balance sheet grows, and now has to meet a higher requirement, so some of every year’s profit is spoken for before it is free. I call what is left owner cash. Stock pay is already inside net income as an expense, so it is deducted and not added back: it was $4.0B in 2025.¹⁵ The capital to stand still is the equity the firm must add so its ratio does not slip when the balance sheet grows only with inflation, which I take as 3% a year, plus the one-time step to the new requirement.

So the owner can take out about 52% of reported earnings in the first year, about 4.9% of the market value. That is a long way below the 109% the firm paid out in 2026. The gap is why I think payouts have to come down, or the ratio has to. The cash I put in the valuation is the owner cash after growth, not this standstill figure.

VIValuation

Start with what would go wrong

Before any numbers, I ask what would most likely make this a bad investment from here, and work back from that. The bank is sound. These are the three ways it goes badly wrong, in the order I think most likely.

1. The capital runs out before the payouts do. The firm paid out 109% of its earnings to common holders in the first nine months of 2026, and its CET1 ratio fell from 11.4% to 10.9%.³ From 1 January the requirement rises 0.5 point, which leaves about $7B of cushion over CET1 and about $4B over the Total capital minimum that binds C (calculated).³ If the firm has to cut buybacks hard to rebuild, owner cash in the early years is lower than the price assumes. Repurchases cut the share count by 4.0% in the first nine months of 2026 C (calculated), and the stock has been carrying that support.³² Nothing in the business has to break for that to happen.
2. The return stays where it has been and the multiple falls back. The ten-year average return on tangible equity is 13.2%, and the stock traded at between 1.3 and 2.1 times tangible book at year ends, 1.6 on average C (calculated).^14131221 If the bank earns what it did and the market pays 1.25 times tangible book in 2036, the return from $49.57 is about 8.8% a year with the cash paid out on the way, which is 3.9 points above the 4.91% Treasury E (estimate).¹⁷ That is a poor trade for taking bank risk.
3. A credit cycle and a markets slump arrive together while the bond losses limit what the firm can do. The Federal Reserve’s severely adverse scenario, run in 2026, gives Bank of America loan losses of $64.9B over nine quarters, including 16.2% of its credit cards and 9.3% of its domestic commercial real estate, against pre-provision revenue of $66.2B. The firm loses $12.5B before tax, and its CET1 ratio falls from 11.4% to a low of 9.9%.²⁸ My bear case is milder than that for one year: revenue down 4% in 2028 and credit costs at 1.5% of loans, a return of 8% E (estimate). And it never gets back above 13%. Under those terms the value is $33.

I can show what the second does, and I do below. The third is my bear case. The first I model through a higher capital target and a lower payout. The bond book I flag and do not model.

What the price implies

Over the ten years the price rose 3.01 times, from $16.48 in November 2016, while tangible book per share rose 1.73 times to September 2026 C (calculated).¹¹⁴⁴ The multiple supplied the rest, about 1.74 times of it, from 0.98 times book at the start. The price peaked at $64.81 on 12 August 2026 and is 24% below that, and it fell 4.8% on 14 October, the day the firm reported.¹ The market is working through a debate between a bank that is repricing its bonds and one that is running out of capital room.

How I value a bank

For most businesses on this site I discount free cash flow. A bank cannot be valued that way. Loans and deposits are its raw material, so the cash flow statement moves with the size of the balance sheet. The right question is how much a shareholder can take out and still leave the bank as strong. So I work on tangible equity, as I did for JPMorgan Chase: an excess-return model, which values the shares as tangible book plus the present value of what the bank earns above its cost of equity.

The idea is simple. A dollar of tangible equity that earns exactly the cost of equity is worth a dollar. One that earns more is worth more, and the more it earns above the cost, and the longer it can, the more it is worth. That gives the same number as discounting the cash the owner can take out each year, after keeping the capital ratio at a target. I do both, and they agree to the cent, which is a check on my arithmetic.

About 50% of the value sits beyond 2036, so it is sensitive to what the firm earns in ten years; I show that sensitivity below. Note also how little of the value is above tangible book: of $46 a share, $29 is the book itself. A bank earning a point or two over its cost of equity does not earn a big premium.

Reverse the question: what return does the price need?

I take my base growth and my base capital ratios and ask a different question. What return on tangible equity, held for ten years and for ever after, makes the shares worth $49.57? I solve it.

The answer is a return on tangible equity of 14.6% every year C (calculated). With the faster growth of my bull case it is 14.2%, and with the slower growth of my bear case, 15.6% C (calculated). There is a shorter route. If the firm holds a return R for ever and grows at g, the multiple of tangible book is (R minus g) over (cost of equity minus g). At today’s 1.70 times, R has to be 15.6% if growth is 3.5%, 15.9% if it is 3% and 14.5% if it is 5% C (calculated). Either way the price needs the firm to keep earning what it earns today, for a decade or more.

That is the honest headline for this page. The price does not pay for a better bank. It pays for the bank as it is now, with an adjusted third-quarter return of 14.9% and a five-year average of 14.6%. JPMorgan is the opposite: its price needs about 19% for ten years, against 17% as the firm’s own target.²¹

A flat 14.6% is a little above the ten-year average of 13.2% and in line with the five-year average. The price is not unreasonable. It prices the recent run as the norm. At a flat 10%, near the 2016 and 2017 returns, the shares would be worth about $28; at 16%, $56 E (estimate).

The discount rate

I discount at 10.6%. I use the same method as on my JPMorgan Chase page. The capital asset pricing model with the 4.91% ten-year Treasury and Damodaran’s 4.20% implied equity risk premium gives 9.9% on Bank of America’s beta against the market over 36 months of 1.19, 10.0% over 60 months at 1.21 and 10.6% over 119 months at 1.37 C (calculated).^17293031 I take the highest of the three, because a bank’s equity is the thin top layer of a very leveraged balance sheet and its beta tends to read low in calm years. The betas use price returns, without dividends. JPMorgan’s three betas on the same method were 0.97, 1.01 and 1.08, which is why its rate is 9.5%.²¹

That 1.1 point difference is the biggest reason this page values the bank so much closer to its book than the JPMorgan page does. The data show a higher beta than JPMorgan’s over every window I tried, and I would not argue the other way hard. I show the value at 9.5% below.

At 9.5%, the rate on my JPMorgan page, the base is $54 and the price is 8% below it E (estimate). At 9.0% it is $59. So the price is within reach of a lower required return. I use 10.6% because a bank’s earnings depend on credit and on a bond book in ways I cannot see from outside.

Three cases

My base case is the bank keeping what it has and giving a little back. Revenue grows 5.5% in 2027 and eases to 3.5% by 2032, about 4.0% a year on average, a little above the 3.0% of the last decade on a consistent basis, because bond repricing helps net interest income for several years E (estimate).¹⁴² The overhead ratio drifts up from 61.4% to 62.0%, since the firm keeps investing and the settlement shows compliance costs do not disappear. Credit costs run at 0.52% of loans, above the 0.43% I estimate for 2026 and the 0.40% ten-year average of the charge-off rate, and close to the 0.54% and 0.50% of provision over average loans in 2024 and 2025 C (calculated).² Risk-weighted assets grow half a point faster than revenue. The result is a return on tangible equity of 15.6% in 2027, 14.6% in 2031 and 14.2% in 2036, an average of 14.6%, above the ten-year average of 13.2% and about the same as the last five years C (calculated). I let the fade run toward 13% because every large bank’s return drifts toward the cost of capital over time, and because capital rules are rising.

The bull case has the bond book reprice for a decade and the firm close part of its cost gap: revenue growth of 6.5% easing to 4%, an overhead ratio falling to 56.5%, credit costs at 0.42%, and the capital ratio falling to 10.9% as the proposed rules turn out kinder than today’s E (estimate). That gives an average return of 17.5%, above the 17.0% of 2021, the best year in my ten-year table C (calculated). The bear case has revenue down 4% in 2028 as markets fall and rates drop, credit costs at 1.5% of loans, a higher capital target of 11.8%, and afterwards an overhead ratio stuck near 63% and a return that settles near 13% and then 10% for ever E (estimate). Even the bear is a profitable bank, worth about 1.1 times tangible book.

What I can’t judge

Some inputs I can’t know, and I would rather say so than pretend. The first is the deposit franchise against the bond loss. The $103.9B loss on held-to-maturity bonds exists because the bank holds a pile of 1.9% bonds when new ones pay about 4.9%. The same low-rate world gave it $548.7B of deposits that cost nothing. How much of the loss that is worth, and how much of the deposits would leave if rates rose further, I cannot see from outside. If I took the loss off tangible book at a 25% tax rate, the book would be $17.89 a share and the stock 2.8 times it E (estimate). I do not do that in the value, because it would double count the low yield that is already in my revenue. But it is a number worth keeping in view.

The second is the shape of the final capital rules. The March 2026 proposal, the surcharge proposal and the fee-income proposal could each move the requirement, and a point of required CET1 is $18B of equity, or $2.63 a share C (calculated).³ The firm says it is evaluating them.³ The third is trading, which moved from $7.1B to $5.1B of revenue between two quarters of 2026.⁴ The fourth is what the consent orders do to growth; I found no restriction in the 10-Q but cannot see the orders. The fifth is AI, which the letter says will lift returns without a number.⁵ The sixth is how Moynihan hands over. The range between my bear and bull is wide because these are wide questions.

Multiply it out

The price needs several things together. A return on tangible equity averaging 14.6% or more for ten years: I would give that 40%. Payouts high enough to keep buying the stock without tougher capital rules: 55%. Credit losses inside a normal cycle: 70%. And a market in 2036 still paying 1.7 times tangible book or more: 45%. Multiplied, that is about 7% E (estimate). They overlap, so the true figure is higher, perhaps one in six. The point is that the price needs several things to go right at once.

Quality and price are two questions

How good is the business? Good, not great: the largest consumer deposit base in the country, a wealth business whose revenue rose 13% in the third quarter, a return that is above its cost of equity in seven of ten years but by a modest margin, a cost gap with its best rival, a thin capital cushion, and a compliance record that cost $1.2B this year. How good is the price? Fair. It is 8% above my base value and 21% below my bull value, and it is 24% below the August peak. A great business can be a poor purchase at the wrong price. This looks closer to a good business at about a fair price, and I say so without calling it a bargain or a trap.

It helps to compare it with the next best use of the same dollar. A ten-year Treasury pays 4.91% with no risk to the dollars promised.¹⁷ Bank of America returns about 10.1% a year on my base, five points more, against my required 10.6%. The closest comparison on this site is JPMorgan Chase, priced 15% above its base value, returning 8.7% a year on its base against a required 9.5%, and needing about 19% on tangible equity for ten years.²¹ The shortfall to the required return is 0.5 points here and 0.8 there. Bank of America is the better-priced of the two on my numbers, and JPMorgan is the better business. I have not valued Wells Fargo or Citigroup, which trade at 1.9 and 1.2 times tangible book.⁷⁸

The margin of safety

My margin-of-safety price is $30, 35% below my base value. At $30 my bear case returns about 11.3% a year, close to the 10.6% I require, and the base returns about 14.4% C (calculated). I use the same 35% cushion as on the JPMorgan page, for the same reason: I cannot see a bank’s credit, and here I also cannot value the bond book against the deposits. The price is 65% above it. The stock last closed below $30 in November 2023.¹

Where this could be in 5 and 10 years

These are my estimates of what a share is worth at later dates, from the model above. The value at a date is the present value, at that date, of the owner cash that comes after it, on today’s 6,922.7M shares and in that year’s dollars. It includes the value of the bank beyond 2036 and leaves out cash paid before the date, which the fourth column adds up from now to the end of 2036, including the fourth quarter of 2026.

For the base to stand at five years, revenue has to grow about 5% a year to 2031, the overhead ratio has to drift from 61.5% to 62.0%, and credit costs have to run at 0.52% of loans. That keeps the return on tangible equity near 15% in 2031 and pays out about $17 a share by then E (estimate). For it to stand at ten, the return has to settle near 13%, which is worth about 1.3 times tangible book. $63 is that multiple on $47 of tangible book per share E (estimate).

The bear and the bull differ mostly in the return on tangible equity. In the bear a 2028 recession cuts revenue 4% and the return falls below 8% before settling near 13%. In the bull it rises to 17 to 18%. The discount rate is 10.6% in all three.

The most likely way the base is wrong is the fade. If the return stayed at 16% for ten years and after, instead of easing to 13%, a share would be worth about $83 at the end of 2036 and $56 today, above the price E (estimate).

These are estimates of value, not forecasts of the share price. The market price can sit above or below value for years.

VIIThe case for the price

What most people miss

First, the bond book is also a source of future profit. The firm holds $496B of held-to-maturity bonds at a 1.91% average yield, while the ten-year Treasury pays 4.91%.³¹⁷ Reinvested at 4.5%, that gap is worth about $13B of pretax profit a year, about a third of today’s E (estimate). Second, consumer banking and wealth management made 55% of 2025 profit and earned 28% and 24% on their assigned capital C (calculated).² Third, in the Federal Reserve’s 2026 severely adverse test, pre-provision revenue of $66.2B paid for 95% of the $69.5B of loan-loss provisions, and CET1 fell 1.5 points to 9.9%, less than the 2.5-point stress buffer C (calculated).²⁸³

The bull case

The case is a high-quality deposit franchise whose earnings are still catching up with rates, with falling costs and a falling share count. The clearest version is the firm’s own. In the 2025 annual report, filed on 23 March 2026, Brian Moynihan wrote that the firm will keep growing tangible book value per share consistently and that AI will help improve returns.⁵ On 14 October 2026 chief financial officer Alastair Borthwick said in the third-quarter release that net interest income grew 8%, asset management fees rose 17%, and asset quality, liquidity and capital were strong.⁴ Of twenty managers whose latest 13F filings I checked, only Berkshire Hathaway held the stock at 30 June 2026.²⁴³² I could not confirm any named analyst target from a primary source.

Against reality

Much of the bull case is true today: net interest income is up, the yield is rising, fees are growing and credit is improving.⁴ If the bonds reprice for years, the overhead ratio falls to 56.5% and the capital rules come in softer, my bull case gives a return averaging 17.5% and is worth $63 E (estimate). If the bank does that and hits its targets, $49.57 is 21% below that value. The price needs less: about 14.6% held for ten years, against 14.2% in 2025, an adjusted 14.9% in the third quarter and a five-year average of 14.6% C (calculated). The bull case needs what the bank has not done: the best full year in ten was 17.0%, in 2021, and the overhead ratio has not been below 58%.²³

What a 10x would take. Ten times $49.57 is $496 a share, about $3.43T of market value C (calculated). At 12 times earnings that needs about $286B of net income to common, 9.0 times my normalized 2026 figure of $31.8B. At today’s 1.7 times tangible book it needs about $2.0T of tangible equity, ten times the $202B the bank has. At a net margin near 26%, $286B of profit means about $1.1T of revenue, nine times today’s, which takes about 29 years at 8% a year and about 74 years at the 3.0% the bank managed over the last decade on a consistent basis C (calculated).¹⁴² The stock rose 3.0 times from $16.48 in November 2016, when the company was worth about $166B, and 1.7 times of that was the multiple rising from 1.0 to 1.7 times book.¹⁴¹ A tenfold move is not plausible for a bank of this size, and I do not think it is a useful way to think about this stock.

Where I come down

The bond repricing and the consumer and wealth returns moved me: they are why my base holds a return near 15%, not the 13% average of the decade. The low multiple did not move me, since it reflects a return that barely clears the cost of equity. What would raise my values is a return rising toward 17% on an overhead ratio near 58%, and capital rules that let the firm hold less. What would lower them is a payout cut to meet the January requirement without earnings rising to make up for it.

VIIIRisks, and what would change my view

The three ways this goes badly wrong are at the start of Valuation. These are the rest.

• Capital and the rules. The surcharge adds 0.5 point on 1 January 2027, and the firm’s stress buffer is 2.5% through 31 December 2027, after which two years of stress tests will be averaged from 2029.³ Three proposals are open: the March 2026 capital proposal, the surcharge calculation and the fee-income model for the stress test.³ A point of CET1 is about $18B C (calculated). The proposals could help or hurt, and the 10-Q does not quantify them.³
• Bonds and deposits. Net unrealized losses on all debt securities are $108.8B, and about 42% of deposits were uninsured at the end of 2025, $857.9B.³² A 100 basis point rise in rates adds $1.4B to annual net interest income and a fall of the same size takes $1.9B away, in the firm’s estimate.²⁰ Higher long rates would widen the unrealized loss and cut tangible book.³ The firm held $914B of liquidity on average in the third quarter.⁴
• Credit. The card book is $107.3B with a loss rate of 3.22% in the third quarter, commercial real estate loans are $75.2B and commercial loans grew 8% in nine months, to $757B, while the allowance fell to 1.04% of loans from 1.14% a year earlier.³⁴
• Trading. Global Markets earned 9% on its capital in the third quarter against 20% in the second, and sales and trading revenue fell from $7.1B to $5.1B.⁴
• Law, compliance and control. A consent order from December 2024 is still in force, and the new orders and settlement announced on 13 October 2026 cost $1.2B. The firm says it has completed significant steps and does not expect the ongoing operating costs to be significant.³ A class action over unemployment prepaid cards can go to trial unless the court decertifies the class.³
• Succession. Moynihan has been chief executive for almost seventeen years. Two Co-Presidents were named in September 2025 and the board has not named a successor.²²⁹
• Technology and competition. AI is a source of efficiency for the firm and a risk if a rival uses it to cut cost faster, or a payment system routes around the firm’s cards and deposits.⁵

What would raise my base value: the fourth-quarter results in January showing an adjusted return of 16% or more on normal credit costs, an overhead ratio near 60%, and buybacks that fit inside the new requirement. What would lower it: an overhead ratio drifting up, charge-offs over 0.65%, or required capital moving up in the final rules while the payout has to be cut.⁴³