Basis Report/Resources/Investor Foundations
4 sections20 entries

Financials Stock Valuation: Pricing Banks, Insurers, and Asset Managers Correctly

The multiple for a financial stock should come from return on equity relative to cost of equity — not from a revenue screen or a sector average. Before you anchor on 10x forward earnings or 1.5x book, you need a clear view on normalized ROE, credit quality trajectory, and what the rate cycle is doing to the earnings power the market is pricing.

✓Derive the justified P/TBV multiple from the bank's normalized ROE relative to its cost of equity before calling it cheap or expensive.
✓Separate provision-adjusted earnings from reported earnings — reserve releases can overstate normalized profitability by 20-30% in credit-favorable periods.
✓Model NIM at normalized deposit costs, not at current peak-cycle spreads, before extending credit on the earnings power.
✓For insurers, build the investment income contribution separately from underwriting profitability — they have different durability and different risk profiles.
Go to frameworkSee a sample report
When to use this

Use this before initiating or sizing any bank, regional bank, insurance company, asset manager, broker-dealer, or specialty finance name. The framework is most useful when the stock looks optically cheap on a trailing earnings basis — which often reflects credit cycle normalization or NIM compression ahead — or when a bank looks expensive relative to book value and you need to decide whether the ROE justifies the premium.

Why it matters now

The rate cycle created the largest swing in bank net interest margins in two decades. Insurers are repricing property risk into an environment where catastrophe frequency and severity have permanently shifted. Asset managers are navigating fee compression and passive share gains that are structurally reshaping AUM economics. Applying a static multiple to any of these businesses without understanding the specific dynamic it is facing produces valuations that are wrong in both directions.

Where theses break

The playbook breaks when investors use trailing P/E without adjusting for provision timing, anchor on book value without testing whether the ROE trajectory supports a premium to book, and model net interest income using current NIM rather than the NIM the business will earn when the deposit repricing cycle completes and fixed-rate assets roll over.

Full framework

4 sections · 20 entries — work through each before you size a position.

Most financials valuation mistakes come from treating banks, insurers, and asset managers as interchangeable and applying the same EV/EBITDA or P/E framework across all three. Each sub-sector has a different economic engine, a different primary valuation anchor, and a different set of risks that can compress the multiple without appearing in the headline numbers. The hard work is decomposing the earnings into their durable and cyclical components, stress-testing the credit assumptions, and understanding whether the current ROE is structurally earned or borrowed from a favorable rate environment.

20 entries in view

Match the valuation lens to the sub-sector economics before building any model

The word 'financials' covers banks, insurers, asset managers, brokers, and specialty finance companies that have almost nothing in common economically. Choosing the wrong primary metric before you start produces conclusions that look disciplined and are fundamentally wrong.

Derive the justified P/TBV multiple before comparing it to where the stock trades

The justified price-to-tangible-book ratio for a bank is determined by one formula: (ROE − g) / (COE − g). A bank earning 14% normalized ROE with a 10% cost of equity and 4% sustainable growth should trade at 2.5x tangible book. A bank earning 9% normalized ROE with the same cost of equity and growth should trade at 0.83x — below book, because it is not covering its equity cost. Every P/TBV comparison across banks is misleading without running this formula for each company in the peer set. The market's current P/TBV pricing is not a fair value estimate; it is the starting point for evaluating whether the implied ROE expectation is too optimistic or too pessimistic.

Why it matters

P/TBV comparisons without adjusting for the ROE differential consistently mislead investors into buying the bank that looks statistically cheapest rather than the one that offers the best return per unit of equity risk. A bank at 0.9x TBV earning 9% ROE is not cheap — it is fairly priced relative to its cost of equity.

When it matters

Before initiating any bank position, and whenever comparing banks across the sector to identify relative value. Also critical before adding to a position after a sharp drawdown — the stock that has fallen 30% to 0.7x TBV is not automatically attractive if the ROE trajectory is deteriorating.

Investor take

Run the justified P/TBV formula for each bank in your comparison set using normalized ROE, not trailing ROE. Rank the peer group by the discount or premium to justified value, not by the raw P/TBV ratio. The most interesting positions are usually banks trading at a discount to their justified multiple because the market is pricing in a credit or rate scenario that your analysis does not support.

Build separate valuation models for underwriting profitability and investment income in insurance

An insurance company's earnings come from two independent engines: the underwriting book (premiums minus claims and expenses) and the investment portfolio (income on the float). These two income streams have different risk profiles, different sensitivity to rate and credit cycles, and different durability. A combined ratio of 98 on an investment yield of 4.5% looks similar to a combined ratio of 104 on a yield of 6%, but the second company is earning through above-market investment returns that will not repeat as the portfolio rolls over. Separate the two contributions before assigning any P/E or P/Book multiple — the combined ratio determines the quality of the franchise; the investment income determines how much credit you should give to the current rate environment.

Why it matters

Insurance investors who anchor on reported P/E without decomposing the underwriting and investment contributions consistently overpay when investment yields are above their long-run average and underpay when the underlying underwriting franchise is better than the current investment climate makes it appear.

When it matters

Before initiating any P&C or life insurance position, and whenever the investment portfolio yield has moved significantly from its historical average — which happens after prolonged low-rate or high-rate periods when the portfolio is rolling over at materially different yields.

Investor take

Build a two-line model: line one is the underwriting contribution (1 − combined ratio) × net premiums earned; line two is investment income on the float. Normalize the investment line to the current portfolio yield plus a three-year rollover assumption. If the total earnings power at normalized yields is significantly different from reported earnings, use the normalized figure as the anchor for the P/E multiple.

Value asset managers on net flow trajectory, not on AUM at a point in time

An asset manager with $800 billion in AUM but experiencing 5% annual net outflows has a different forward earnings trajectory than one with $800 billion and 3% net inflows, even if they look identical on a trailing revenue multiple. AUM is a stock; net flows are the increment or decrement that determines where AUM — and therefore revenues — will be in three years. A manager with negative net flows requires above-average market returns just to hold its earnings flat. A manager with positive net flows has a revenue tailwind that compounds with market appreciation. The correct multiple sits on the endpoint of where AUM will be, not on where it is today.

Why it matters

Asset manager valuations consistently mislead when the AUM figure includes large institutional mandates that are close to redemption, or when recent AUM growth was driven entirely by market returns during a bull cycle with no organic flow contribution. Both situations make the business look stronger than it is going into the next market cycle.

When it matters

Whenever an asset manager's stock has moved more than 20% from its prior range, and before initiating any position in a manager that is transitioning its product mix — for example, from active to alternatives — because the fee rate and flow dynamics change significantly with the product shift.

Investor take

Track three metrics together: net new money as a percentage of beginning AUM, realized fee rate trend over the past eight quarters, and the share of AUM in products with long-term redemption restrictions versus liquid vehicles. A manager that scores well on all three has a more durable earnings engine than a simple multiple comparison reveals.

Apply EV/EBITDA only to financial sub-sectors where it actually works

EV/EBITDA is appropriate for brokerage firms, certain specialty finance companies, and alternative asset managers with defined management fee and carry structures because their debt is genuine leverage rather than operational funding. It is not appropriate for banks, insurance companies, or traditional asset managers where the 'debt' is either the product (deposits) or the long-duration liability that funds the investment portfolio. Before choosing any EV-based metric for a financials company, ask: is the debt on this balance sheet leverage the company uses to amplify returns on equity, or is it the input material for the business? If it is the latter, EV-based metrics will produce numbers that look precise but carry no economic meaning.

Why it matters

The most common valuation category error in financials is applying EV/EBITDA to a bank because it looks like a fast way to compare it across sectors. The resulting number is not comparable to anything else and will not predict how the stock behaves relative to peers.

When it matters

Whenever a financials stock appears in a cross-sector screen that uses EV/EBITDA, before accepting the implied conclusion, and when a broker or research analyst presents a financials comparison using enterprise value-based multiples.

Investor take

If someone hands you a P/E screen for banks alongside a P/E screen for industrial companies, the numbers exist in the same format but measure different things because the industrial company's P/E reflects equity ownership in an enterprise that uses debt to enhance returns, while the bank's P/E reflects equity ownership in an entity where debt is the product itself. Keep the valuation frameworks separate and compare financials only within their own sub-sector.

Check whether the current rate environment is inflating or compressing normalized earnings before assigning a multiple

Net interest margin for banks — and investment yield for insurers — is more sensitive to the rate cycle than any other input in the earnings model. At the peak of a rate-rising cycle, a bank's NIM may be 50–100 basis points above its normalized level because the deposit repricing lag has allowed it to temporarily earn more on assets than it is paying on funding. That tailwind reverses as deposits reprice competitively. Modeling NIM at current peak levels and then applying a normal P/E multiple produces a valuation that is simultaneously wrong on the earnings estimate and wrong on the multiple — the earnings will fall and the multiple will compress at the same time.

Why it matters

The rate cycle's impact on bank earnings has been responsible for more large-cap financial sector valuation errors in the past two decades than any other single factor. The NIM that looks stable today is often already building in the seeds of a significant earnings decline over the next four to eight quarters as the portfolio reprices.

When it matters

Before initiating any bank or insurance position in a rising or falling rate environment, and after any significant change in the Federal Funds rate or in the yield curve slope. The NIM sensitivity analysis should be done on any bank that has been growing NIM faster than peers — above-average NIM expansion usually carries above-average repricing risk.

Investor take

Build a NIM sensitivity table with three scenarios: current NIM held constant, NIM normalizing to the five-year historical average over two years, and NIM compressing 30 basis points below the five-year average. The earnings output from each scenario defines the range of outcomes the P/E multiple must accommodate. If the stock only looks attractive in the flat-NIM scenario, the thesis requires a specific argument about why deposit costs will not compete away the current spread — not just an assumption that they will not.

Normalize earnings before assigning any multiple

Reported financials earnings are among the most cyclically distorted numbers in any sector. Provision timing, reserve releases, trading gains, and investment portfolio mark-to-market all create gaps between what the income statement reports and what the business actually earns on a normalized basis.

Replace actual provisions with a through-the-cycle provision rate before building any earnings model

Bank loan loss provisions move with the credit cycle in ways that can overstate or understate true earnings power by 20–35%. In a credit-favorable environment, actual provisions are below the through-the-cycle average because credit losses are low. In a downturn, they spike above average as management builds reserves against future losses. The reported EPS figure in any given year is therefore a function of where the credit cycle sits, not of the underlying earnings power of the business. Estimate the through-the-cycle provision rate for the specific loan portfolio — typically 40–70 basis points of average loans for a diversified commercial bank, higher for consumer finance, lower for prime residential mortgage — and apply it consistently to normalize the earnings series.

Why it matters

Banks that appear cheap on a trailing P/E during credit-favorable periods are usually appropriately priced on normalized earnings. The apparent discount exists because investors are anchoring on earnings that include below-average provision expense and potentially above-average reserve releases — both of which will reverse when the cycle turns.

When it matters

Before initiating any bank or credit company position, and every time a bank reports quarterly results that include a significant provision beat — which is usually driven by better-than-expected credit conditions rather than by operating outperformance.

Investor take

Build a provision normalization bridge: start with reported net income, add back the reported provision, subtract the through-the-cycle provision you have estimated for this loan portfolio, then tax-effect the difference. The result is normalized after-tax earnings. Divide by shares to get normalized EPS, and use that as the denominator for the P/E multiple.

Track reserve coverage ratios to understand whether the current provision run-rate is sustainable

The allowance for loan losses (or CECL reserve) as a percentage of total loans tells you how much cushion the bank has built against future credit deterioration. A coverage ratio of 1.5% on a diversified commercial portfolio suggests the bank has reserved for roughly 2.5–3 years of normal charge-offs. When coverage ratios fall below historical norms — either through years of low provisioning or through reserve releases — the bank has less buffer and is more exposed to an earnings shock if credit deteriorates. A rising coverage ratio, by contrast, often signals that management sees stress building in the book before it shows up in reported charge-offs.

Why it matters

Reserve coverage ratios provide a leading indicator of provision pressure that the income statement does not yet show. Banks with below-average coverage ratios heading into a credit softening will need to rebuild reserves through above-normal provisioning that compresses earnings, while peers with adequate coverage can provision at normalized rates.

When it matters

Every quarter when reviewing bank earnings, and especially before initiating a position in any bank whose provision expense has been below historical averages for multiple consecutive years.

Investor take

Compare the current allowance-to-loans ratio against the bank's own five-year average and against peers with similar loan mixes. If the current coverage is more than 20 basis points below the five-year average, estimate the provisioning required to rebuild it over two years — then subtract that from normalized earnings and re-run the P/E.

Separate recurring fee revenue from market-sensitive or transaction-dependent income

Bank fee revenue contains multiple components with very different durability: wealth management fees tied to AUM levels (market-sensitive but recurring), service charges and card interchange (behavioral and relatively stable), capital markets revenue from underwriting and M&A advisory (highly cyclical), and mortgage banking gains (rate-cycle dependent). Investors who apply a bank's headline P/E to blended revenue that includes a large trading or capital markets contribution are valuing a volatile revenue stream at the same multiple as a recurring deposit franchise. Separate the stable recurring fee income from the market-sensitive components, then decide which deserves the higher multiple in the blended earnings figure.

Why it matters

Fee revenue mix affects both the level of earnings and the quality multiple those earnings deserve. A bank with 35% of revenue from stable wealth management and card fees is a different investment than one with 35% from trading and investment banking, even if the reported earnings are identical.

When it matters

Whenever a bank has a significant capital markets or trading franchise, and when evaluating regional banks that have been growing mortgage banking or other rate-sensitive fee businesses as a NIM substitute.

Investor take

Build a fee revenue bridge that separates stable recurring fees (trust, card, treasury management) from market-sensitive fees (trading, capital markets, mortgage banking). Apply a higher multiple to the stable component and a lower multiple to the market-sensitive component, then weight by their current contribution. This produces a blended earnings quality multiple that reflects the actual composition of the fee engine.

Adjust insurance earnings for catastrophe normalization before assigning a P/E or P/Book multiple

Property and casualty insurance results vary dramatically with catastrophe frequency and severity. A year with below-average hurricane, wildfire, or severe convective storm losses will produce a combined ratio that makes the underwriting franchise look stronger than it is on a through-the-cycle basis. The correct approach is to normalize the catastrophe load to a multi-year average — typically using a five-to-ten year rolling catastrophe loss history, adjusted for inflation and exposure growth — before assigning any multiple. An insurer trading at 15x earnings in a low-catastrophe year may be trading at 20x normalized earnings when you apply a realistic catastrophe assumption.

Why it matters

Insurance catastrophe normalization is the single most common source of overvaluation in P&C insurance. Investors who extend a premium multiple after a quiet catastrophe year are pricing in a permanence to favorable conditions that the historical frequency of major loss events does not support.

When it matters

Before any P&C insurance initiation, and whenever a P&C insurer has reported two or more consecutive years of below-average combined ratios without a clear structural reason — better underwriting terms, exposure reduction, or reserve strengthening — that explains why the improvement is sustainable.

Investor take

Add three lines to every P&C insurance model: the reported combined ratio, the normalized catastrophe load using the five-year rolling average, and the resulting normalized combined ratio. If the difference is more than three points, the earnings normalization exercise will produce a materially different intrinsic value than the one implied by reported results.

Model the unrealized loss position in the securities portfolio before calculating tangible book value

Banks hold large investment securities portfolios — often 20–30% of total assets — classified as either available-for-sale (AFS, marked to market through other comprehensive income) or held-to-maturity (HTM, not marked to market but still bearing economic risk). When rates rise sharply, AFS securities generate unrealized losses that flow directly to tangible equity and reduce tangible book value per share, even though reported earnings are unaffected. Tangible book value comparisons that use the balance sheet figure without checking the unrealized loss position routinely overstate the equity cushion available to shareholders. A bank reporting 1.3x P/TBV may effectively be trading at 1.7x P/TBV once the unrealized HTM losses are included — and both multiples may be on top of tangible equity that is partially impaired by the rate environment.

Why it matters

The held-to-maturity portfolio concealment problem became visibly consequential during the 2022–2023 rate cycle, when several regional banks held HTM portfolios with embedded losses equal to 40–80% of tangible equity. The losses were not visible in P/TBV calculations until the portfolios were sold under pressure, crystallizing the economic reality.

When it matters

When evaluating any bank after a significant rate move up or down, and before concluding that a bank trading near tangible book is cheap — the tangible book figure may not reflect the true economic equity position once the securities portfolio is marked at current rates.

Investor take

Request the securities portfolio unrealized gain/loss disclosure from the 10-Q or 10-K (typically in the investment securities note). Calculate the after-tax impact of the unrealized position on tangible equity. Then recompute P/TBV using the adjusted tangible book and compare that to the justified multiple from the ROE formula. A bank can look cheap on unadjusted P/TBV and expensive on adjusted P/TBV simultaneously.

Model the credit cycle rather than extrapolating the current environment

Credit cycles in financials are the most reliable source of surprise in any direction. The best time to prepare for a credit downturn is when credit metrics are benign — because that is when the forward-looking signals are most available and the price of protection is lowest.

Monitor net charge-off trends to identify where the bank is in the credit cycle

Net charge-offs (NCOs) — the loans a bank writes off as uncollectible after recoveries — are the clearest lagging indicator of credit quality and the most useful input for estimating future provision requirements. When NCOs are below their five-year average, the bank is likely in a credit-favorable period where reserves are being released and provisioning is below through-the-cycle norms. When NCOs are rising — even from a low base — it signals that the credit cycle is inflecting. The pace of NCO increase matters as much as the absolute level: a bank moving from 20 basis points to 35 basis points of average loans is early-cycle; one moving from 80 basis points to 120 basis points is mid-stress.

Why it matters

NCO trends are available in every quarterly supplement and are among the most actionable credit metrics because they lag delinquencies and non-performing loans — meaning by the time NCOs are clearly rising, the leading indicators have already been available for one to three quarters.

When it matters

Every quarter when reviewing bank financials, and specifically before initiating any position in a bank whose NCO rate is currently running below its five-year historical average — which is the base condition for most banks in credit-favorable periods.

Investor take

Build an NCO trend table for each bank with the trailing eight quarters plus the five-year and ten-year averages. Flag any bank where current NCOs are more than 15 basis points below the five-year average — the mean-reversion math in a provisioning normalization year can be significant.

Evaluate commercial real estate concentration as the primary structural credit risk for regional banks

Commercial real estate (CRE) loans — including office, multifamily, retail, and construction lending — represent the largest single source of credit stress in regional banking cycles. Regulatory guidelines flag banks with CRE concentrations above 300% of risk-based capital as elevated-risk; banks above 400% are subject to heightened examiner scrutiny. Office CRE is the current focal point of concern because remote and hybrid work has permanently impaired office utilization and consequently collateral values across major markets. Before investing in any regional bank, calculate the CRE concentration ratio, segment it by property type, and stress the collateral values on the office portion at a 30–40% mark from peak to understand the potential impairment to book value.

Why it matters

Regional bank CRE concentration is typically invisible in the headline metrics — a bank with a 1.2% NCO rate and 150% CRE-to-capital concentration looks fine until the office book reprices at maturity and the underlying collateral does not support the outstanding loan balance. The stress is absorbed over a cycle, not in a single quarter.

When it matters

Before initiating any regional bank position, and whenever a regional bank has reported sequential loan growth in CRE categories without a corresponding increase in reserve coverage. Also essential after any major commercial real estate market event — office vacancies crossing a threshold, a large institutional landlord defaulting — that changes the mark-to-market framework for collateral values.

Investor take

Request the bank's loan portfolio disclosure from the 10-Q. Calculate CRE as a percentage of risk-based capital. Then segment: what share of CRE is office? What share of office is in gateway markets with above-average vacancy? Stress those balances at a 30% collateral decline. If the loss estimate exceeds 15–20% of current tangible equity, the credit risk is non-trivial and deserves explicit representation in the downside scenario.

Check deposit mix and repricing velocity to model NIM at different points in the rate cycle

A bank's deposit base is its primary funding source, and the rate sensitivity of that funding determines how quickly NIM compresses when the Fed cuts rates or when competition forces deposit cost increases. Non-interest-bearing demand deposits (DDA) — checking accounts that pay zero interest — are the highest-quality funding; they are free now and stay free as rates rise. Certificates of deposit and money market accounts reprice immediately with market rates. The higher the DDA mix, the more stable the NIM through a rate cycle. Banks with DDA comprising 35%+ of total deposits have structurally lower funding cost sensitivity than those with 15–20% DDA and a higher share of rate-sensitive CDs.

Why it matters

Deposit mix analysis is the most direct way to understand NIM trajectory in a rate-changing environment — more reliable than asking management, because the deposit composition is auditable in the balance sheet footnotes. Banks that grew aggressively using brokered deposits or high-rate CDs to fund loan growth during a rising rate cycle face the sharpest NIM compression when the rate environment softens.

When it matters

Before any bank initiation, and whenever a bank has reported NIM expansion that looks materially faster than peers — above-average NIM expansion often means the bank accepted funding concentration risk to generate it, and the reversal can be sharper than the expansion was.

Investor take

Request the deposit detail table from the bank's 10-Q. Calculate DDA as a percentage of total deposits, then compare against the peer group. Estimate the blended deposit beta — the percentage of a rate move that passes through to deposit costs — for the specific mix. A bank with 40% DDA has a deposit beta of roughly 30%; one with 20% DDA has a beta closer to 50%. Use those betas to model NIM at three rate scenarios: current, flat for two years, and down 100 basis points.

Stress the capital stack before extending credit on capital return capacity

Capital return — buybacks and dividends — is a key driver of bank stock returns and a major determinant of P/TBV premiums. But capital return is only sustainable when the bank maintains its CET1 ratio above regulatory minimums with a sufficient buffer for stress. Most large banks target 50–100 basis points above their regulatory minimum; a bank that is returning 80–90% of earnings as capital while hovering near the minimum buffer is one credit event away from suspending buybacks and re-pricing the stock. Before including capital return in the bull case for any bank, stress the CET1 ratio through a moderate recession scenario and verify that the current return program survives.

Why it matters

Capital return expectations are frequently the most important driver of premium P/TBV multiples for banks — the market pays up for the cash distribution that a well-capitalized bank can sustain. When those expectations are disrupted by a credit event or a regulatory capital change, the multiple compresses and the loss of the buyback accretion often makes the downside larger than the initial capital event itself.

When it matters

Before initiating any bank position where the bull case includes above-average capital return, and whenever a bank has recently changed its CET1 target — either as a strategic choice or in response to examiner guidance — because the change in target changes the implied capital buffer available for returns.

Investor take

Run the following stress test: start with current CET1, assume a 150 basis point increase in the loan loss rate on the total portfolio over two years (a mild-to-moderate recession scenario), add back the resulting provision impact (after-tax) as capital consumption, and compare the resulting CET1 to the stated minimum target. If the bank falls below its target in the stress scenario, the buyback program is at risk and should be removed from the base case.

Build the credit cycle timeline explicitly rather than assuming mean reversion is linear

Credit cycles do not revert linearly. Delinquencies lead charge-offs by two to four quarters. Charge-offs lead peak provisioning by one to two quarters. Peak provisioning leads reserve stabilization by two to four quarters. Investors who wait for the headline NCO rate to peak before buying bank stocks typically arrive late — the forward-looking indicators (delinquency rates, criticized loan migration, watch list expansion) have usually been visible for six to twelve months before the NCO peak. Conversely, investors who sell too early in a credit recovery because NCOs are still elevated miss the period when earnings are most leveraged to provisioning normalization.

Why it matters

Most investors think of credit stress as an event that happens and then resolves. It is actually a process with a predictable internal logic: leading indicators deteriorate first, then lagging indicators catch up, then provisioning peaks, then the earnings trough is established, then reserves are rebuilt, then the earnings recovery begins. Each stage has observable markers that allow an investor to estimate where in the process the bank sits.

When it matters

During any period of credit cycle stress or emerging stress — rising commercial real estate vacancies, softening consumer delinquencies, widening investment-grade credit spreads — to position the bank exposure appropriately across the expected timeline of earnings pressure.

Investor take

Build a credit cycle roadmap for each bank: where is the NCO rate relative to the five-year average? Are delinquencies rising or stable? Is the watch list expanding? Is management guiding provision higher for the next two quarters? Assign the bank to a credit cycle stage — early-stress, mid-stress, late-stress, or early-recovery — and position sizing accordingly. Banks in late-stress or early-recovery stages have a different risk-reward than ones entering stress.

Use peer comparison to calibrate, not to set, the multiple

Peer comparison in financials tells you where the market is pricing a bank or insurer relative to its competitors. It cannot tell you whether the entire peer group is fairly valued — that requires the fundamental work done in the prior sections.

Rank banks by return on tangible equity before comparing P/TBV ratios

Return on tangible equity (ROTE) is the correct profitability metric for comparing banks because it strips goodwill and acquired intangibles from the equity base — which is the book that generates the banking returns. A bank with $20 billion in goodwill from past acquisitions has a higher tangible equity base than its GAAP equity implies, and comparing it on P/TBV without the tangible adjustment will mislead. Build the ROTE comparison across the peer group: rank by ROTE, then overlay the P/TBV multiples. The market's willingness to pay a multiple premium for ROTE leaders is a calibration check — outliers that trade at a significant premium or discount to what the ROTE ranking predicts are where the most interesting relative value lies.

Why it matters

The most common financials relative valuation error is comparing P/TBV without adjusting for the ROTE differential, which produces a ranking that looks like relative cheapness and is actually a reflection of structural quality differences the market is pricing correctly.

When it matters

When building any bank or insurance comparable analysis, and whenever a bank that ranks near the bottom on ROTE appears to look attractive on a simple P/TBV screen.

Investor take

Plot ROTE versus P/TBV for the peer group. Draw the regression line. Banks above the line are trading at a premium relative to their ROTE; banks below are trading at a discount. The banks below the line are the starting point for investigating whether the discount is explained by credit risk, concentration, or management quality — or whether it is a genuine valuation opportunity.

Adjust insurance P/B comparisons for float quality and investment portfolio duration

Insurance companies with similar reported combined ratios and similar P/Book ratios can have very different investment income trajectories based on float quality and portfolio duration. A short-duration portfolio in a rising rate environment reprices quickly, boosting investment income. A long-duration portfolio in the same environment is locked into below-market yields for years. Float quality — measured by the consistency and stickiness of policy renewal rates — determines how long the premium cash is available to invest before it is paid out as claims. Before concluding that two insurers are comparably valued at similar P/Book ratios, verify that their float duration and investment portfolio duration are aligned.

Why it matters

Float duration differences can cause two insurers with identical combined ratios and identical P/Book ratios to have very different investment income trajectories over the next five years. The one with a shorter-duration float and a longer-duration portfolio — collecting long-lived premiums and holding long-dated bonds — will look most attractive when rates fall and least attractive when rates rise.

When it matters

When comparing P&C or life insurance companies on any valuation metric that includes investment income, and whenever the interest rate environment has changed significantly since the comparison baseline was established.

Investor take

Add two columns to any insurance peer comparison: average duration of the investment portfolio and average float duration from premium collection to claim payment. Pair these against the current interest rate environment to assess which insurer has the most favorable reinvestment setup over the next two to four years.

Compare asset manager fee rates per dollar of AUM alongside the AUM multiple

Asset managers that appear to trade at a discount on an AUM multiple may deserve the discount because their fee rate per dollar of AUM is declining faster than peers — a function of product mix shifting toward passive or near-passive mandates, institutional fee renegotiations, or redemptions from high-fee products into lower-fee alternatives. The AUM multiple (price per dollar of managed assets) is a useful starting point but must be paired with the fee rate per dollar of AUM to understand revenue quality. A manager with $500 billion in AUM at a 0.30% average fee rate has $1.5 billion in fee revenue; the same AUM at 0.18% has $900 million — the difference is a 40% gap in revenue per unit of asset that will not appear in an AUM comparison alone.

Why it matters

Fee rate compression in asset management is structural, persistent, and historically underestimated by investors who rely on AUM multiples without modeling the fee rate trajectory. A manager at a 20% discount to peers on AUM is fairly priced if the fee rate is declining 4–5 basis points per year while peers hold their rates flat.

When it matters

Before initiating any position in an asset manager, and whenever an asset manager reports AUM growth while simultaneously reporting revenue growth that lags — which signals fee rate dilution from the product mix shift.

Investor take

Build a three-year fee rate trend for the asset manager: divide total management fee revenue by average AUM for each year. If the fee rate is declining and the rate of decline is accelerating, model the revenue under an extrapolation of that trend before anchoring on the current multiple.

Verify that the dividend yield is sustainable through a stress scenario before using it as a valuation anchor

Bank and insurance dividend yields are frequently used as valuation anchors — a stock at a 4% dividend yield looks attractive relative to Treasuries when the market's risk premium justifies it. But a bank dividend that represents 60–70% of through-the-cycle earnings while the bank is running near its minimum CET1 buffer is not a durable 4% yield; it is a 4% yield with significant cut risk in a credit event. Before anchoring any bank or insurance valuation on the dividend yield, verify that the payout ratio is sustainable under a credit-normalized earnings scenario and that the bank has the capital buffer to maintain the dividend if earnings temporarily decline.

Why it matters

Dividend cut expectations in bank stocks are a larger source of drawdown than the dividend income is a source of return. A bank that cuts its dividend from a $2.00 annual rate to $1.00 when credit deteriorates loses 10–15% of price from the repricing of the yield alone, plus additional multiple compression from the credit concern that caused the cut.

When it matters

Before using dividend yield as a primary valuation argument for any financial stock, and whenever a bank or insurer has increased its dividend significantly in the past two years without a corresponding improvement in through-the-cycle earnings power.

Investor take

Calculate the dividend payout ratio on through-the-cycle earnings (using the normalized provision rate from section two). If the payout ratio exceeds 50% on normalized earnings while the bank is operating within 75 basis points of its regulatory minimum CET1, model the dividend as at risk in the downside scenario and remove it from the base-case yield argument.

Avoid the cheap-financials-screen trap when the sector is cheap for a structural reason

Financials occasionally trade at what appears to be a historically significant discount to book value or to earnings power across the entire sector. Sometimes this reflects a genuine cyclical opportunity — credit fears are exaggerated, rate concerns are overstated, and the sector is recovering toward normalized earnings. Other times, the entire sector is cheap because a structural headwind is repricing the long-term earnings trajectory — rising regulatory capital requirements, secular fee compression, credit cycle normalization that has not yet flowed through to reported earnings. The screening exercise cannot tell you which condition you are in. The fundamental work — credit cycle position, NIM trajectory, capital adequacy, fee rate trends — is required to decide whether the sector-wide discount is an opportunity or a warning.

Why it matters

Sector-wide cheapness in financials is itself a signal that requires investigation rather than a conclusion. The market is usually right that something is creating uncertainty — the question is whether the uncertainty is temporary or structural. Getting that wrong is the most costly error a financials investor can make, because betting on cyclical recovery when the headwind is structural means averaging into a multi-year de-rating.

When it matters

Whenever the financial sector appears at a multi-year low on P/TBV or P/E screens on a sector-wide basis, which tends to happen when credit concerns are elevated or rate concerns are causing earnings estimate revisions.

Investor take

Before concluding that a cheap financials screen is an opportunity, identify which of the following is driving the discount: (1) cyclical credit stress that will normalize, (2) cyclical NIM compression from a rate cycle that will reverse, (3) structural earnings erosion from regulatory requirements or fee compression that will not reverse. If the answer is (1) or (2), price in the recovery timeline. If the answer is (3), the sector may be cheap on trailing metrics and still be correctly valued on forward normalized earnings.

Evidence

Financials valuation inputs scorecard

The six financials valuation inputs — what each tells you and when it misleads

No single metric works across banks, insurers, and asset managers. Each sub-sector has a primary lens and a set of conditions under which that lens breaks — know which you are in before drawing a conclusion.

Price / Tangible Book Value (P/TBV)
Market price ÷ tangible equity per share
The primary anchor for bank valuation. Justified P/TBV = (ROE − g) / (COE − g). A bank earning 15% normalized ROE with a 10% cost of equity should trade at approximately 2.0x TBV. Misleads when book value includes large unrealized securities losses held in accumulated other comprehensive income (AOCI) — during rate-rising cycles, these losses can reduce tangible book by 20–40% without appearing in reported earnings, making the multiple look higher than the economic reality.
Pre-Provision Net Revenue (PPNR) Yield
PPNR ÷ assets or ÷ risk-weighted assets
Strips credit cycle noise from earnings by separating the operating engine from provision decisions. A bank with high PPNR relative to assets has more cushion to absorb credit stress without equity impairment. Misleads when compared across banks with different loan mixes — a bank with 60% commercial real estate will generate higher NIM but lower PPNR quality than one with 40% CRE and more fee business.
Net Interest Margin (NIM)
Net interest income ÷ average earning assets
Measures the spread between what the bank earns on assets and pays on funding. The key variable is not the current NIM but the NIM trajectory: where is it going as fixed-rate loans reprice, deposits compete away rate benefits, and the funding mix shifts? Peak-cycle NIM figures routinely overstate normalized earnings power by 20–40 basis points. Always model NIM at fully-repriced deposit costs and rolled-over asset yields before anchoring any earnings estimate.
Combined Ratio (Insurance)
Loss ratio + expense ratio
The primary operating metric for property and casualty insurers. A combined ratio below 100% means the underwriting book is profitable before investment income. Above 100%, the insurer is generating underwriting losses that must be offset by investment returns. Misleads when investors accept management's reported combined ratio without checking for reserve development — favorable reserve development from prior accident years can make a structurally loss-making book look profitable in the current year. Always check the prior-year development line in the supplement.
Fee Revenue as % of Total Revenue (Banks)
Non-interest income ÷ total net revenue
Measures how much of the earnings stream is independent of NIM compression. Banks with 40%+ fee revenue — from wealth management, capital markets, card fees, treasury services — are structurally less sensitive to rate cycles than pure NIM businesses. Misleads when fee revenue includes trading revenue, which is volatile and not repeatable, or mortgage banking gains, which reverse sharply when rates rise. Separate recurring fee streams from market-sensitive ones before assigning a quality premium to fee-heavy revenue mix.
AUM Growth Rate and Fee Rate (Asset Managers)
AUM growth × realized fee rate
The two drivers of asset manager revenue: the size of the asset base and what the manager charges on it. Net flows (new AUM minus redemptions) and market appreciation determine the first; fee compression from passive competition and institutional negotiating power determines the second. Misleads when managers report gross AUM growth without disclosing net flows — a manager with 10% AUM growth driven entirely by market appreciation and flat-to-negative flows has a more fragile franchise than the headline implies.

Sub-sector to valuation lens

The financials sub-sector determines the right valuation method — P/E across all three is the most common mistake

Applying a bank P/E to an insurance company or an asset manager multiple to a broker produces conclusions that look reasonable and are almost certainly wrong. Map the lens to the economics.

The financials sub-sector determines the right valuation method — P/E across all three is the most common mistake
Sub-sectorPrimary valuation lensQuality gates requiredCommon mistake
Large-cap banks (JPM, BAC, WFC, C)P/TBV anchored on justified multiple from ROE vs. COE; P/E on through-the-cycle normalized earningsCET1 ratio above regulatory minimum with buffer for stress, through-the-cycle NCO rate confirmed, NIM modeled at normalized deposit costs not peak-cycle spreadsPaying a premium P/TBV on reported earnings that include reserve releases — the earnings look clean until the credit cycle turns and provisions normalize upward, compressing both the earnings estimate and the multiple at the same time.
Regional banks (USB, FITB, HBAN, KEY)P/TBV calibrated to franchise quality and loan concentration; PPNR yield as the operating quality gaugeCRE concentration below 300% of risk-based capital, core deposit franchise verified as rate-stable, fee revenue mix assessed for cycle sensitivityBuying at a discount to large-cap peers without verifying the CRE concentration or loan-to-deposit ratio — a regional with 400% CRE concentration and limited core deposits can lose a third of its earnings power when commercial real estate stress materializes.
P&C Insurance (PGR, TRV, ALL, CB)P/B calibrated to normalized combined ratio and return on equity; P/E on average combined ratio through a catastrophe cycleAccident-year combined ratio confirmed ex-reserve development, catastrophe load normalized over 5+ years rather than the current low-loss period, investment portfolio duration and credit quality assessedUsing a reported combined ratio that includes favorable prior-year reserve development as the operating baseline — the insurer looks underwriting-profitable until the development tailwind runs out and the underlying book's true cost reveals itself.
Life insurance (MET, PRU, AFL, LNC)P/Book or EV (embedded value) for policies with long-duration liabilities; P/E on normalized spread incomeLiability duration confirmed against asset duration, interest rate sensitivity stress-tested for 200+ basis point moves, actuarial assumptions reviewed for mortality and lapse rate stabilityValuing life insurance on reported GAAP earnings without understanding the sensitivity of long-duration policyholder liabilities to rate moves — a significant rate shift can impair statutory surplus and restrict capital return capacity even if GAAP earnings appear stable.
Asset managers (BLK, BEN, IVZ, AMG)P/E on normalized earnings with AUM-based cross-check; EV/EBITDA for diversified managers with significant alternative AUMNet flow trajectory assessed for organic vs. market-driven AUM growth, fee rate compression trend modeled explicitly, distribution leverage verified as durableAnchoring on current AUM as a revenue run-rate without modeling fee compression — a manager with 10% passive outflows per year and a fee rate declining 3 basis points annually has a revenue trajectory that looks stable today and erodes significantly over five years.

The reserve release trap

Reserve releases inflate bank earnings without creating economic value — and the reversal is non-linear

A bank that provisioned aggressively during a downturn and then releases reserves when the credit environment improves reports higher earnings per share without generating any additional interest income or fee revenue. The released reserve goes directly to the income statement as a negative provision expense, boosting reported earnings and reducing the apparent P/E ratio. Investors who anchor on this earnings figure are paying for value that was created in a prior period and cannot be repeated. When the credit cycle normalizes and the bank must provision at a through-the-cycle rate again, reported earnings decline without any deterioration in the underlying business — but the market reprices the stock as if the business is worse. Build a provision-normalized earnings figure before assigning any P/E multiple to a bank that has been releasing reserves for more than two consecutive quarters.

Common questions

What investors ask about investor foundations for investor foundations stocks.

Why can't you use EV/EBITDA to value a bank?
EV/EBITDA requires separating enterprise value from equity value by adding debt — but for a bank, deposits and wholesale borrowings are the raw material of the business, not leverage in the traditional sense. Removing them to compute an enterprise value produces a number that does not correspond to any economic reality. The bank uses the spread between its cost of funding and the yield on its loans and securities to generate earnings; stripping out that funding as 'debt' removes the input that creates the product. The correct frameworks are price-to-tangible-book (anchored on return on equity relative to cost of equity), price-to-earnings on normalized provisions, and dividend discount models for yield-oriented banks. For brokers, asset managers, and specialty finance companies, EV-based metrics become more applicable because their debt is genuine leverage rather than operational funding.
How do you derive the right price-to-book multiple for a bank?
The justified P/TBV ratio is a function of one equation: (ROE − g) / (COE − g), where ROE is normalized return on equity, g is sustainable growth, and COE is cost of equity. A bank earning 15% ROE with a 10% COE and 5% growth should trade at roughly 2.0x tangible book. A bank earning 9% ROE with the same COE and growth should trade at 0.8x — below book, because it is not covering its cost of equity. The practical implication is that every P/TBV comparison is meaningless without adjusting for the ROE differential. A bank at 1.2x TBV is not necessarily cheap relative to a bank at 1.8x TBV — if the first has a 9% normalized ROE and the second has a 16% normalized ROE, the second is cheaper on a value-per-unit-of-economic-return basis. Run the justified multiple formula before concluding that a bank is discounted to peers.
How do you normalize bank earnings for the credit cycle?
Three adjustments: first, replace actual loan loss provisions with a normalized through-the-cycle provision rate — typically 40–60 basis points of average loans for a diversified commercial bank, though this varies significantly by loan mix and market position. Second, add back any reserve releases from prior-year over-provisioning and remove the earnings benefit they created. Third, check whether current NCOs (net charge-offs) are below or above historical averages, and model the direction of normalization over the next two to four years. The resulting earnings figure — called pre-provision net revenue or PPNR, adjusted for through-the-cycle credit costs — is the number that should anchor the P/E multiple. Reported earnings in credit-favorable years routinely overstate normalized profitability by 15–30% once reserve releases are removed and a through-the-cycle provision rate is applied.
What drives the valuation difference between large-cap and regional banks?
Four factors: balance sheet complexity, funding cost stability, loan book concentration, and regulatory capital efficiency. Large-cap banks — JPMorgan, Bank of America, Wells Fargo — benefit from greater deposit franchise diversification, lower average funding costs across the cycle, diversified fee revenue streams that reduce dependence on NIM, and higher capital efficiency from GSIB surcharge management. Regional banks often have more concentrated loan books (commercial real estate is frequently 200–400% of risk-based capital in smaller regionals), less fee revenue to absorb NIM compression, and more deposit rate sensitivity as large-cap banks compete for their depositors in rate-rising cycles. The regional bank discount to large-cap is not simply a liquidity discount — it reflects structurally higher earnings volatility from concentration and less revenue diversification. Before concluding that a regional bank discount is unjustified, verify the CRE concentration, the loan-to-deposit ratio, and the deposit mix between retail core and rate-sensitive commercial.

Continue research

Keep building the foundation

Financial stocks: classify the model before the multipleInvestor FoundationsHow to Value a Stock Without Fooling YourselfInvestor FoundationsCapital allocation: grade management before you buyInvestor FoundationsEarnings Quality Checklist: Judge the Quarter, Not the HeadlineInvestor FoundationsReturn on invested capital: the incremental-returns testInvestor FoundationsStock multiples: read the argument inside the numberInvestor Foundations

Put it on a live stock

JPM stock overviewJump to the live ticker page and report flowBAC stock overviewJump to the live ticker page and report flowBRK.B stock overviewJump to the live ticker page and report flowPGR stock overviewJump to the live ticker page and report flow

Category hub

Valuation frameworks hubSee all guides in this category