Basis Report/Resources/Investor Foundations
4 sections20 entries

Financial stock analysis: credit cycles, NIM, and capital adequacy

Financial stock analysis fails when investors apply industrial earnings frameworks to businesses where leverage is a feature, not a bug, and where credit quality determines whether the earnings are real. The right order is: classify the subsector, map the earnings drivers, then build a valuation that survives a credit cycle.

✓Write where you believe the bank's NIM sits relative to its normalized range before you model earnings — not just the current rate environment.
✓Check the reserve-to-NPL coverage ratio and ask whether it reflects the vintage composition of the loan book, not just the stated reserve-to-loans ratio.
✓Classify the business as bank, insurer, asset manager, or payment processor before touching a valuation multiple — each requires a different primary lens.
✓For any position where commercial real estate is above 200% of Tier 1 capital, run a collateral stress scenario before sizing.
Go to frameworkSee a sample report
When to use this

Use this before initiating on any commercial bank, regional bank, investment bank, life insurer, P&C insurer, asset manager, or payment processor. Start with the subsector classification — it determines which analytical framework applies and which risks deserve the most weight.

Why it matters now

The rate cycle from 2022 to 2024 produced the fastest Fed Funds increase since the 1980s, pushed net interest margins to multi-decade highs for some banks while creating unrealized losses that impaired tangible book value, and triggered the 2023 regional bank failures when rate-exposed fixed-income portfolios met deposit outflows. The credit cycle is now operating through a commercial real estate correction that will take years to resolve. Investors who can model normalized NIM, reserve adequacy, and capital at risk have a structural edge over those reading reported earnings as if the cycle were frozen.

Where theses break

The blueprint breaks when NIM is modeled at current levels without stress-testing the rate sensitivity of both assets and liabilities; when reserve adequacy is judged by the reserve-to-loans ratio without examining the vintage composition and collateral values of the underlying book; and when a low P/TBV ratio is treated as a margin of safety without asking whether tangible book value is accurately stated after marking HTM portfolios and commercial real estate collateral.

Full framework

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

Most weak financial research mistakes accounting earnings for economic reality, ignores reserve adequacy until charge-offs confirm what provisioning data already signaled, and values banks on P/E multiples that embed no view on whether the current net interest margin is normalized, peak, or trough. Strong financial research classifies the business model first — bank, insurer, asset manager, payment processor — then maps valuation to the specific risk architecture of that subsector.

20 entries in view

Classify the financial business model before you price the earnings

Financial sector stocks look like a single category on a screen but operate under fundamentally different economic models. The valuation framework must match the earnings architecture — a bank's NIM math does not apply to an asset manager, and an insurer's combined ratio framework does not apply to a payment processor.

Distinguish commercial banking from investment banking from asset management before touching a multiple

Commercial banking earns NIM on a portfolio of loans funded by deposits — the economics are driven by credit quality, rate sensitivity, and deposit franchise value. Investment banking earns episodic fee income on transactions and volatile trading revenue tied to capital markets activity. Asset management earns AUM-linked fees that grow with market performance and organic inflows. These are different financial models with different leverage profiles, different cyclicality, and different risk architectures. A single P/E multiple applied across all three produces a number that means something different in each case.

Why it matters

The most common valuation error in financials is applying the same multiple to businesses that earn money in structurally different ways. The multiple must reflect the earnings architecture, not just the sector label.

When it matters

Before initiating any position in a diversified financial company and whenever a business announces a strategic shift toward banking, wealth management, or capital markets.

Investor take

Write one sentence on how the majority of the company's pre-tax earnings are generated — spread income, fee income, or trading income — then choose the valuation framework appropriate for the primary earnings source.

Separate interest-rate sensitive revenue from fee-based revenue in the earnings bridge

For a bank, NIM-driven spread income is highly sensitive to interest rate moves and deposit repricing lags. Fee income from card processing, wealth management, and transaction services is less rate-sensitive and generally more recurring. An investor who values the entire earnings stream at a single multiple is assigning the same quality premium to rate-cycle income and structural franchise income. The interest-rate sensitive portion should carry a lower multiple because it mean-reverts as the cycle turns.

Why it matters

Banks with high fee-income-to-revenue ratios (above 40%) are more valuable than their spread-focused peers at the same P/E because their earnings are more cycle-stable. That quality difference deserves explicit modeling, not a single blended multiple.

When it matters

When building a sum-of-parts valuation for any large bank with material wealth management, card, or payments businesses, and whenever the rate cycle turns and NIM compression begins to affect reported earnings.

Investor take

Decompose total revenue into spread income and non-interest income. Apply separate quality multiples: a lower multiple to rate-sensitive NIM, a higher multiple to recurring fee revenue. Blending these correctly produces a more accurate intrinsic value than any screened P/E comparison.

Understand the role of leverage in financial business models before assigning a multiple

A bank's balance sheet operates with 8-12x leverage by design. An asset manager's balance sheet is typically unlevered. An insurance company's leverage is embedded in its float — premiums collected before claims are paid. These are not the same risk profiles, and the appropriate return on equity targets and valuation multiples reflect that difference. A bank earning 12% ROE on 10x leverage is generating 1.2% return on assets — a reasonable but not exceptional return given the credit and rate risks being carried.

Why it matters

Financial sector ROE comparisons across subsectors are meaningless unless they account for the underlying leverage ratio and the quality of the assets being levered. Asset managers earning 25% ROE on unlevered equity are more valuable per dollar of book value than banks earning 12% ROE on 10x leverage — even though the bank's headline ROE looks higher.

When it matters

When comparing valuation metrics across financial subsectors and when deciding whether a financial company's ROE justifies a premium to book value.

Investor take

Decompose ROE into ROA times leverage ratio. That decomposition separates the business quality signal (ROA) from the structural leverage decision. A bank with 1.5% ROA is a significantly better business than one with 0.9% ROA, even if both post similar ROE because they use different leverage ratios.

Identify the funding structure — deposits, wholesale funding, or AUM — before modeling earnings stability

A bank funded primarily by core checking and savings deposits has the most stable, lowest-cost funding base in financial services — deposit inertia is a competitive moat. A bank dependent on brokered deposits or wholesale funding markets faces higher funding costs and funding risk that increases during stress. An asset manager's 'funding' is the clients' voluntary decision to keep assets under management — a structurally different stability profile. These funding architectures determine cost of funds, NIM stability, and the speed at which earnings deteriorate in a stress scenario.

Why it matters

The 2023 regional bank failures were primarily funding failures, not credit failures. SVB and Signature had high uninsured deposit concentrations in a narrow customer base that withdrew simultaneously when confidence broke. A bank with 85% insured deposits in retail accounts is a fundamentally different business from one with 40% insured deposits concentrated in institutional clients.

When it matters

Before initiating any bank or thrift position, and every time a bank reports a significant change in deposit composition, deposit costs, or wholesale borrowing levels.

Investor take

Calculate the uninsured deposit percentage and the deposit concentration in the top 10 client relationships. A bank where the top 20 clients represent more than 30% of total deposits has a structural liquidity risk that deserves explicit modeling in the bear case, not just a qualitative risk factor mention.

Map regulatory capital requirements to earnings potential before sizing any financial position

Large U.S. banks operate under stress capital buffer requirements that set a floor on CET1 ratios based on annual stress test results. Banks near their stress capital buffer floor have minimal buyback capacity and limited room to absorb unexpected losses without cutting dividends or issuing equity. Banks operating well above their target CET1 have excess capital that is either a return-of-capital opportunity or a sign of management's inability to deploy it productively. The gap between current CET1 and the target operating range is the first-order determinant of near-term capital returns.

Why it matters

Capital adequacy is not just a risk metric — it is a return-of-capital metric. A bank with 50 basis points of CET1 above its target generates roughly $500M of annual buyback capacity for every $100B of risk-weighted assets. A bank below target cannot buy back shares and may need to build capital through earnings retention.

When it matters

Before initiating any large bank position and whenever a bank reports quarterly results with CET1 commentary, announces a capital plan update, or receives stress test results from the Federal Reserve.

Investor take

Calculate the dollar value of excess CET1 above the stated operating target. Divide by shares outstanding to estimate per-share capital available for distribution. Compare that to the current buyback pace to assess whether the bank is returning capital faster or slower than it accumulates.

Read credit quality signals before charge-offs land in earnings

Credit deterioration in financial institutions follows a predictable sequence: stress first shows up in NPL trends and criticized loan balances, then in reserve builds, then in charge-offs. Investors who wait for charge-offs are typically 2-4 quarters behind the economics.

Track NPL inflows and criticized loan balances as leading credit cycle indicators

Non-performing loans measure the stock of credits already in trouble. Criticized loans (Special Mention and Substandard under bank regulatory classifications) measure the flow of credits moving toward trouble. A bank where criticized balances are rising while NPLs and charge-offs are still low is in the early phase of a credit cycle — provisions will follow, earnings will compress, and the question is whether reserves are adequate to cover the losses that are now visible in the pipeline.

Why it matters

The gap between criticized balance growth and NPL/charge-off recognition is where credit cycle positioning lives for investors who read regulatory disclosures rather than just the headline earnings. Most investors only see the charge-off line, which is the last signal in the sequence.

When it matters

Every quarter when evaluating any bank or finance company with material loan exposure to cyclically sensitive sectors — commercial real estate, leveraged lending, consumer unsecured credit.

Investor take

Build a loan quality waterfall: criticized balances → NPLs → net charge-offs, tracked quarterly for 8-12 quarters. If criticized balances are growing faster than the loan book and management's provision guidance is not accelerating, the reserves are likely inadequate.

Evaluate CECL reserve adequacy by loan category, not by the blended reserve-to-loans ratio

The Current Expected Credit Loss (CECL) standard requires banks to reserve for lifetime expected losses at origination. The blended reserve-to-loans ratio across the entire portfolio conceals where credit risk is concentrated. A portfolio where commercial real estate reserves sit at 1.5% alongside unsecured consumer credit at 6% looks healthier when blended than the CRE exposure actually warrants. Reserve adequacy must be assessed category by category, with assumptions stress-tested against vacancy rates, cap rate trends, and borrower debt service coverage for CRE, and unemployment rate assumptions for consumer credit.

Why it matters

A bank can report a stable 1.2% reserve-to-loans ratio while systematically under-reserving against a deteriorating commercial real estate book because the headline ratio blends a shrinking CRE reserve percentage with adequately-reserved consumer loans.

When it matters

Whenever a bank has material exposure to office commercial real estate (above 100% of Tier 1 capital), construction lending, or leveraged loan portfolios, and whenever credit metrics are diverging across loan categories even when the blended ratio looks stable.

Investor take

Extract reserve percentages by loan category from the allowance footnote in the 10-Q. Compare the CRE reserve percentage to the reported delinquency rate in that category and to industry survey data on collateral values. If the reserve percentage is below the delinquency rate in any major category, provisions are likely to increase.

Analyze loan growth versus underwriting standard changes to assess credit risk accumulation

Loan growth above GDP or above the bank's historical average is not inherently a positive signal. It requires the question: what had to change in underwriting standards to generate that growth? Banks that grow loans 15% in a slow-demand environment are typically lowering rates, extending terms, relaxing collateral standards, or entering unfamiliar markets. Each of those concessions increases future credit loss probability. The question is not whether the bank is growing loans but whether the marginal loans being booked are of the same quality as the existing portfolio.

Why it matters

The classic pre-crisis lending pattern: banks that grow fastest in the expansion phase tend to have the highest charge-off rates in the correction phase. Loan growth alone is not underwriting quality — rapid loan growth combined with declining reserve builds is a contradiction that deserves scrutiny.

When it matters

When any bank reports loan growth materially above peers or above the prior year's pace, and when disclosure of underwriting metrics (average LTV, DSCR, credit score distribution) shows tightening collateral coverage or looser income verification.

Investor take

Normalize loan growth by end market. If commercial loan growth is concentrated in construction or bridge lending to unfamiliar geographies, ask management about average LTV at origination and borrower DSCR for the new vintages. If those disclosures are vague, assume the conservatism is not there.

Monitor commercial real estate concentration relative to Tier 1 capital and by vintage

Regulators use 300% of Tier 1 capital as an informal threshold for commercial real estate concentration requiring enhanced scrutiny. Banks near or above that threshold in office, retail, or multifamily carry idiosyncratic credit risk that is correlated with the broader CRE market cycle. More important than total concentration is vintage composition: loans originated at 2019-2022 cap rates and LTVs look significantly impaired at 2024 market values, while recent originations may reflect current market clearing levels. The vintage that deserves the most scrutiny is whatever was originated at the peak of the real estate cycle for the relevant property type.

Why it matters

A bank with $3B of Tier 1 capital and $6B of office CRE loans underwritten at 4% cap rates when the market is now clearing at 6.5-7% cap rates has loan-to-value ratios far above what the reported book suggests. Refinancing risk and maturity wall timing determine when the impairment becomes a realized loss.

When it matters

When evaluating any regional bank with CRE concentration above 200% of Tier 1 capital, and whenever office, multifamily, or retail property value indices show sustained declines that impair the underwriting assumptions used at origination.

Investor take

Request or calculate the loan-to-value distribution of the CRE portfolio at current market values, not at appraisal values from origination. A portfolio where more than 20% of loans have current LTV above 90% requires explicit reserve stress-testing, not a historical loss rate applied to current balances.

Understand net interest margin sensitivity to the full rate cycle, not just the current rate environment

NIM is a dynamic spread between asset yields and funding costs that changes with a lag when rates move. On the asset side, fixed-rate loans and bonds reprice only at maturity; floating-rate loans reprice immediately. On the liability side, demand deposits may not reprice at all in a rising rate environment until deposit competition forces banks to compete; CDs and brokered deposits reprice quickly when they mature. A bank with a 2-year fixed-rate loan book and a short-duration deposit base earns expanding NIM in rising rate environments — until deposits reprice and the benefit reverses. Mapping the repricing schedule of both sides of the balance sheet tells you when the current NIM peak will begin to mean-revert.

Why it matters

NIM peaked for most banks in 2023-2024 after asset yields repriced upward while deposit costs lagged. The reversal is now underway: as CD maturities roll and competitive deposit pricing increases, funding costs are rising toward the rate level while assets begin to mature into a lower-rate environment. Investors who underwrite current NIM as normalized will overestimate earnings power as the cycle turns.

When it matters

Before initiating any rate-sensitive bank position, after any Federal Reserve rate decision, and when a bank reports a material change in its deposit composition, funding mix, or asset duration.

Investor take

Build a NIM sensitivity table: at +200, +100, flat, -100, and -200 basis point rate scenarios, what does NIM look like in year 1 and year 2, accounting for repricing schedules? That two-period sensitivity is more useful than any single-scenario NIM assumption.

Stress the capital position and management's allocation discipline

Capital adequacy is both a risk floor and a return-of-capital enabler. The quality of management's capital allocation decisions — when they buy back stock, when they build or release reserves, how they grow the loan book — separates banks that compound per-share value from those that do not.

Use CET1 relative to the stress capital buffer as the actual capital flexibility metric

Regulatory minimum CET1 requirements are not the binding constraint for large banks — the stress capital buffer (SCB) determined annually by Federal Reserve stress tests is the binding floor. The SCB reflects the hypothetical losses in the Fed's severe adverse scenario and determines how much additional equity the bank must hold above the minimum. A bank with a 4.5% minimum CET1 and a 3.5% SCB must maintain 8% CET1 to maintain dividend and buyback capacity. Every basis point above that floor is deployable capital — every basis point below means restrictions on capital returns.

Why it matters

Investors who use 4.5% as the capital floor for large U.S. banks are using the wrong number. The effective floor is the regulatory minimum plus the stress capital buffer, which can be 2-4 percentage points higher. That gap determines actual buyback capacity and risk absorption capacity.

When it matters

When the Fed publishes annual stress test results in June, and whenever a bank reports quarterly CET1 that is approaching or moving away from its stated target operating range.

Investor take

Calculate: current CET1 minus (minimum CET1 plus stress capital buffer plus management's stated operating buffer above SCB floor). That residual is the actual excess capital available for accelerated buybacks. Express it as a percentage of the stock's current market cap to understand the buyback optionality.

Evaluate PPNR growth as the core operating quality metric — not reported EPS

Pre-provision net revenue isolates the bank's core spread-and-fee earnings power before provisioning decisions, loan loss reserve builds, and credit cycle noise. A bank that grows PPNR 5% annually for four years while provisioning for credit cycle normalization is demonstrating structural earnings improvement. A bank that reports growing EPS driven by reserve releases while PPNR is flat or declining is reporting an accounting gain, not an economic improvement. PPNR growth is the metric that persists through the credit cycle.

Why it matters

The distinction between PPNR growth and EPS growth matters most at credit cycle inflections: when reserves are released in a benign environment, EPS overstates earnings quality; when reserves are built in a deteriorating environment, EPS understates the underlying spread-and-fee business quality.

When it matters

Every quarter when evaluating any bank's reported results, and whenever the bank's EPS trajectory diverges from its PPNR trajectory by more than 10% over two or more quarters.

Investor take

Build a PPNR bridge: start with total revenue, subtract non-interest expense, arrive at PPNR before provisions. Track this series quarterly and compare to EPS. The gap between PPNR growth and EPS growth is entirely driven by provisioning decisions and below-the-line items.

Score buyback timing against tangible book value to test management's capital allocation discipline

A management team that buys back shares at 1.8x tangible book during peak earnings is spending permanently scarce equity capital at a price that requires sustained premium ROTE to justify. The same team buying at 0.9x TBV in a temporary earnings trough is creating per-share value arithmetically. The timing of buybacks relative to TBV and ROTE is one of the clearest signals of whether management is allocating capital for shareholders or managing optics by returning capital when the stock looks expensive.

Why it matters

Banks with the best long-run per-share value creation tend to be counter-cyclical buyers: building capital during downturns when capital is scarce and buying back stock when prices are low, rather than accelerating buybacks during peak earnings when the stock is at a premium multiple.

When it matters

After each quarterly earnings release with a buyback disclosure, and when evaluating any bank that has announced a large accelerated share repurchase program during or after a period of peak NIM or peak earnings.

Investor take

Compare the five-year history of quarterly buyback volume and price paid per share against tangible book value per share in the same quarter. A management team that consistently buys above 1.5x TBV while the stock trades at peak P/E is an inferior capital allocator relative to one that buys countercyclically near or below TBV.

Decompose return on equity into ROA and leverage to separate business quality from capital structure

A bank with 1.4% ROA on 9x leverage earns 12.6% ROE. A bank with 0.9% ROA on 12x leverage earns approximately 10.8% ROE. These look similar on a headline basis but represent fundamentally different businesses. The first has a structurally superior asset-level return that does not require aggressive leverage to reach a reasonable ROE. The second depends on leverage to compensate for mediocre asset returns. In a credit downturn, the high-leverage, low-ROA bank deteriorates much faster — because the credit losses flow through the same thin ROA, and the leverage amplifies the impairment to equity.

Why it matters

ROE is a composite signal that conflates business quality (ROA) with a balance sheet decision (leverage). Improving ROE by increasing leverage is not the same as improving ROE by generating better spread income or lower operating costs. The two paths to the same ROE have completely different risk profiles.

When it matters

When comparing banks across different size categories or business models, and whenever a bank reports ROE improvement driven by leverage expansion rather than ROA improvement.

Investor take

Build a ROA-leverage decomposition for the bank and its three closest peers over 5 years. Sort by ROA, not by ROE. The highest-ROA banks in any peer group are almost always the ones that compound per-share value most reliably over a full cycle.

Use the efficiency ratio trend to separate structural cost discipline from temporary revenue tailwinds

The efficiency ratio — non-interest expense divided by net revenue — measures how much it costs to generate a dollar of income. A declining efficiency ratio signals operating leverage: revenue is growing faster than costs. But efficiency ratios can decline purely because NIM-driven revenue is rising in a favorable rate environment, not because management has actually controlled expenses. The test is: what does the efficiency ratio look like when NIM normalizes? A bank with an 82% efficiency ratio on peak NIM may be a 90%+ bank at normalized NIM — revealing that the apparent cost discipline is mostly rate leverage.

Why it matters

Efficiency ratio improvement driven by rate leverage is the accounting equivalent of confusing a tailwind for skill. The genuine cost discipline signal is an efficiency ratio that improves even when revenue growth is modest — because that requires actual expense control rather than revenue expansion.

When it matters

When evaluating any bank's multi-year financial performance, and when comparing efficiency ratios between a rate-sensitive bank and a fee-income-heavy bank — the latter's efficiency ratio is more comparable across rate environments.

Investor take

Build the efficiency ratio on a NIM-normalized basis: replace current NIM with the bank's 10-year average NIM in the revenue denominator and calculate what the efficiency ratio would look like at that revenue base. If the normalized efficiency ratio is above 80%, the current improvement is rate leverage, not structural progress.

Build a valuation that prices the credit cycle and rate cycle separately

Financial sector valuations that hold through cycles are built on normalized earnings, accurately stated tangible book value, and a clear view of what ROTE looks like when NIM mean-reverts and credit losses normalize — not on current reported earnings.

Anchor on price-to-tangible-book-value relative to through-cycle ROTE as the primary valuation framework

The correct valuation anchor for a bank is P/TBV relative to normalized ROTE (return on tangible equity, excluding goodwill and intangibles from the equity base). A bank that earns 14% ROTE sustainably through a full credit cycle deserves to trade at a significant premium to tangible book. A bank earning 8% ROTE should trade below TBV. The key discipline is the 'through-cycle' modifier: peak ROTE overstates the sustainable return, and the multiple must reflect the normalized level. Paying 1.8x TBV for a bank earning 14% ROTE at peak when normalized ROTE is 10% overpays by the gap between peak and normalized.

Why it matters

The P/TBV-vs.-ROTE relationship is the most reliable valuation framework in bank analysis because it anchors on the equity base (TBV) and the return on that base (ROTE), which together determine whether the stock is value-creating or value-destroying at the current price.

When it matters

Before initiating any bank position and when the stock re-rates significantly in either direction, prompting a reassessment of whether the implied ROTE in the current price is achievable through the cycle.

Investor take

Calculate normalized ROTE by averaging ROTE across the last two full credit cycles — including at least one trough year. Compare P/TBV at the current price to the P/TBV implied by the normalized ROTE relative to peers. If the current P/TBV is more than 20% above the peer-implied level given the normalized ROTE, the stock is pricing in above-average earnings quality or above-average growth that needs explicit justification.

Model earnings under at least three NIM scenarios before deciding the stock is cheap

Net interest margin for most banks in 2024-2025 sits above 10-year historical averages because of the rate cycle. Modeling earnings on current NIM is not a conservative assumption — it is modeling the cycle peak as the steady state. The three scenarios to underwrite: a base case with NIM at the 10-year average, a bull case where rate cuts are gradual and the deposit repricing lag extends the current NIM environment for 2-3 more quarters, and a bear case where aggressive rate cuts and competitive deposit pricing compress NIM 40-60 basis points from peak over 18 months. The current stock price should be evaluated against all three.

Why it matters

NIM is the most important variable in most bank earnings models and the one that changes the fastest when rate cycles turn. Investors who do not run NIM scenarios are essentially making a rate bet without acknowledging it.

When it matters

After any Federal Reserve rate decision, when building any 2-year earnings model for a bank with material NIM sensitivity, and whenever the bank discloses a change in deposit repricing dynamics or asset duration.

Investor take

Build the NIM bridge: starting NIM → asset repricing impact → deposit cost change → NIM for each quarter in the 8-quarter forward model. Run this for all three rate scenarios. The range of 8-quarter cumulative EPS across the three scenarios is the honest uncertainty band for the earnings estimate.

Calculate tangible book value at fair value — not at amortized cost — before calling any financial stock cheap on P/TBV

Banks hold significant portions of their investment portfolios in held-to-maturity (HTM) securities that are carried at amortized cost rather than fair value. After 2022, rising rates created large unrealized losses in these portfolios that do not appear in stated tangible book value. The fair-value-adjusted tangible book value — disclosed in the footnotes as 'fair value of held-to-maturity securities' minus the amortized cost — can be materially lower than stated TBV. A bank trading at 0.8x stated TBV with $2B of unrealized HTM losses on $10B of TBV is actually trading at 1.0x fair-value-adjusted TBV — not the discount the screen suggests.

Why it matters

The post-2022 period created a systematic overstatement of tangible book value for banks with large HTM securities portfolios. Investors who compare P/TBV ratios without adjusting for HTM unrealized losses are comparing stated TBV to fair-value TBV across a peer group with very different HTM portfolio compositions.

When it matters

Whenever evaluating any bank with a material HTM portfolio that was built prior to 2022 rate increases, and whenever building a downside scenario that depends on whether the bank can survive a liquidity stress without selling HTM securities.

Investor take

Extract the fair value of HTM securities from the footnotes (required disclosure under ASC 820). Calculate fair-value-adjusted TBV = stated TBV minus (HTM amortized cost minus HTM fair value) after tax. Use this as your denominator in the P/TBV calculation rather than stated TBV.

Stress tangible book value through a credit cycle scenario before sizing any position

The downside anchor in bank valuation is not the current stock price — it is the credit-stressed tangible book value: what TBV would be after absorbing the expected loss in a credit downturn scenario. For a bank with material commercial real estate exposure, this means applying collateral haircuts to CRE loans above certain LTV thresholds, estimating additional provisions, netting the after-tax loss against equity, and calculating the resulting TBV. If the credit-stressed TBV is below the current stock price, the stock is priced as if the credit cycle cannot stress the equity base — which is only appropriate if the loan book quality is genuinely above cycle-average.

Why it matters

The bear case for a bank is not that earnings disappoint — it is that tangible book value gets impaired through credit losses, which then compresses the stock well below the current multiple because investors re-anchor on the new (lower) TBV. That sequence can produce 30-50% drawdowns for banks with concentrated problem loan exposure.

When it matters

Before initiating any bank position with CRE concentration above 200% of Tier 1 capital, leveraged loan exposure above 100% of Tier 1, or reported NPL coverage ratios below 80%.

Investor take

Build a credit stress model: take the most concentrated risky loan category → apply a loss rate consistent with the 2008-2010 or equivalent sector-specific historical loss rates → calculate after-tax losses → subtract from TBV → compute the resulting stressed P/TBV. If the stressed P/TBV is still below 1.0x at the current price, the stock may be genuinely cheap. If not, the current P/TBV discount is not a margin of safety.

Use sum-of-parts valuation for complex financials with multiple distinct business models

A large financial holding company with commercial banking, investment banking, wealth management, and capital markets businesses is not a single-multiple stock. The commercial bank deserves a P/TBV-based valuation anchored on its ROTE. The wealth management franchise deserves a P/E multiple on recurring fee revenue, similar to an independent asset manager. The investment banking and trading businesses deserve a P/TBV or P/E multiple on through-cycle revenue, explicitly discounted for the volatility and capital intensity of those businesses. Summing these parts and comparing to where the company trades on a blended basis reveals whether the market is ascribing the right value to each business.

Why it matters

Conglomerate discounts are real in financial services. A company with a high-multiple wealth management business inside a lower-multiple commercial bank may trade at a blended multiple below the stand-alone value of either business. That gap is either an opportunity or a structural premium problem that requires a catalyst to resolve.

When it matters

When initiating on any large, diversified financial company and when a strategic review, spin-off discussion, or activist investor raises the question of whether the parts are worth more than the whole.

Investor take

Segment the business by revenue source and earnings contribution. Assign a standalone comparable multiple to each segment based on the most relevant single-line peer group. Sum the parts to a theoretical fair value and compare to market cap. If the gap exceeds 15%, investigate why the market is assigning a discount — usually it is justified by execution complexity, regulatory capital requirements, or management quality.

Evidence

Financial sector metrics scorecard

The metrics that separate real earnings quality from cycle-driven noise

These metrics vary by subsector. A managed care ratio does not belong in a bank model, and NIM does not belong in an insurance analysis. Match the metric to the business.

Net Interest Margin (NIM)
Net Interest Income ÷ Average Earning Assets
The core spread for banks and thrifts. NIM above the bank's 10-year average during a rate cycle peak is not normalized earnings — it is a temporary benefit from asset repricing that precedes liability repricing. Model NIM at mid-cycle levels before underwriting a P/E-based valuation. Watch for NIM expansion driven by unrealized security portfolio concentration that creates duration risk on the asset side.
Pre-Provision Net Revenue (PPNR)
Net Revenue − Non-interest Expense (before provisions)
The best measure of core bank earnings power, stripped of provisioning decisions. A rising PPNR trend signals genuine operating leverage in the spread-and-fee business. A declining PPNR masked by reserve releases is a quality deterioration signal most investors miss until charge-offs arrive.
CET1 Ratio
Common Equity Tier 1 ÷ Risk-Weighted Assets
The core capital adequacy metric. Large U.S. banks operate with CET1 targets 100-200 basis points above regulatory minimums to create stress buffers. A bank operating near its regulatory minimum has limited buyback capacity and less room to absorb credit cycle losses without diluting equity. When CET1 declines toward the lower bound of the bank's stated operating range, dividend and buyback capacity is the next thing to come under pressure.
NPL / Charge-off Divergence
NPLs rising; NCOs flat = lag, not resolution
Non-performing loans are the leading indicator; net charge-offs are the lagging confirmation. A book where NPLs are rising while NCOs are still low is a classic early-cycle deterioration signal. The bank is recognizing credit problems but has not yet taken the losses — reserves may be understated, and earnings may be overstated relative to economic reality.
Combined Ratio (Insurance)
(Losses + Expenses) ÷ Premiums Earned
The primary P&C underwriting quality metric. A combined ratio above 100% means the insurer is losing on underwriting before investment income. More important than the absolute level is the direction and whether reserve releases are masking deterioration in the current accident year. Favorable prior-year development is a non-recurring earnings item — strip it out before assessing underwriting profitability.
AUM Organic Flow Rate
Net New Flows ÷ Beginning AUM
For asset managers, total AUM growth is a misleading metric because market appreciation can obscure net outflows. Organic flow rate — net inflows divided by beginning AUM — is the true business health indicator. Sustained negative organic flows signal that the product lineup is losing competitive positioning to passive or lower-fee alternatives, regardless of what the AUM headline shows.

Subsector valuation map

Financials is not one sector — the valuation discipline must match the business model

Applying a bank earnings multiple to an asset manager or a P/E framework to an insurance company produces conclusions that look precise and are almost always wrong. Classify first.

Financials is not one sector — the valuation discipline must match the business model
SubsectorPrimary valuation lensKey metrics to anchorMost common mistake
Commercial / Regional BanksP/TBV relative to ROTENIM (normalized), PPNR/RWA, CET1 ratio, NPL coverage, reserve adequacyBuying at peak NIM on a P/E basis without modeling the liability repricing lag. The earnings that look cheap on current NIM become expensive when deposits re-price and spreads normalize.
Investment Banks / Capital MarketsP/TBV or P/E on through-cycle revenueTrading revenue variability, IB backlog, FICC vs. equities revenue mix, capital returns vs. RWA usageExtrapolating peak capital markets activity as a durable earnings base. IB and trading revenue is highly mean-reverting — applying cycle-high multiples to cycle-high revenue produces systematic overvaluation.
P&C InsurersP/B or P/E on normalized combined ratioCurrent accident year combined ratio (ex-reserve releases), reserve adequacy, catastrophe load vs. reinsurance coverage, premium rate trendTreating favorable reserve development as ongoing earnings quality. Prior-year development is non-recurring. The current-year accident year combined ratio is the durable signal.
Asset ManagersP/E on normalized fee revenueOrganic AUM flow rate, effective fee rate trend, operating margin trajectory, equity vs. fixed income mixValuing AUM growth without adjusting for fee rate compression. A manager growing AUM 10% annually while fee rates decline 5% annually is growing revenues at half the pace the AUM number implies.

Most common mistake

A low P/TBV ratio is not a margin of safety if tangible book value is not accurately stated

After 2022-2023, dozens of regional banks traded at 0.5-0.7x tangible book value while appearing optically cheap. The problem: tangible book value was stated at amortized cost on HTM bond portfolios that had lost 15-25% of fair value. When those unrealized losses were factored in, the actual P/adjusted-TBV ratios were at or above book. Separately, commercial real estate loan books carried at par may require significant write-downs as office and multifamily values reset below origination appraisals. Before calling any financial stock cheap on P/TBV, mark the bond portfolio to fair value and apply a stress haircut to the most concentrated loan categories. That adjusted tangible book value is the real floor.

Common questions

What investors ask about investor foundations for investor foundations stocks.

What is the most important metric for analyzing a bank stock?
Pre-provision net revenue (PPNR) divided by risk-weighted assets, tracked over several quarters, is the most informative bank earnings quality metric. PPNR strips out provisioning noise and shows whether the bank's core spread-and-fee business is generating real operating leverage before the credit cycle adds or subtracts from reported earnings. A bank with growing PPNR/RWA is improving its core economics. A bank with flat or declining PPNR/RWA is either under competitive pressure on spreads, losing operating leverage, or both — and provisioning is masking it.
How do you value a bank stock?
Price-to-tangible-book-value (P/TBV) relative to return on tangible equity (ROTE) is the primary valuation framework. A bank earning 12% ROTE at a 10% cost of equity should trade near or above tangible book. A bank earning 8% ROTE should trade below tangible book — the business is destroying per-share value. The key inputs are: how sustainable is the current ROTE, how accurate is the stated tangible book value after marking HTM portfolios and assessing loan book quality, and what does the ROTE look like in a normalized rate and credit environment rather than the current one.
What causes bank earnings to surprise to the downside?
The two most common causes are reserve build cycles that arrive with a lag after credit quality deteriorates, and NIM compression when liability repricing catches up to or exceeds asset repricing in a falling rate environment. Credit problems are typically signaled 2-4 quarters in advance by rising NPLs, declining coverage ratios, and specific portfolio disclosures — but investors who only track net charge-offs miss the early signals. NIM compression is predictable once you know the repricing schedule of the deposit book — banks with higher-rate CDs and brokered deposits face faster liability repricing and more NIM pressure when rates fall.
What does the combined ratio tell you about an insurance company?
The combined ratio — losses incurred plus expenses, divided by premiums earned — is the core underwriting profitability metric for P&C insurers. A ratio below 100% means the company is making money on underwriting before investment income. Above 100%, it is losing on underwriting and relying on the investment portfolio to generate overall profit. The direction and decomposition matter as much as the absolute level: a rising loss ratio driven by social inflation in litigation costs is harder to correct through repricing than a spike driven by a single large catastrophe. Sustained above-100% combined ratios with declining reserve releases signal that the underlying book of business is unprofitable and repricing is not keeping pace with loss trends.

Continue research

Keep building the foundation

How to Value a Stock Without Fooling YourselfInvestor FoundationsCapital allocation playbook for investorsInvestor FoundationsHow stock multiples actually workInvestor FoundationsEarnings Quality Checklist: Judge the Quarter, Not the HeadlineInvestor FoundationsReturn on invested capital: how to use ROIC without kidding yourselfInvestor FoundationsHow to analyze a stock: first pass to convictionInvestor 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 flowGS stock overviewJump to the live ticker page and report flowBRK.B stock overviewJump to the live ticker page and report flowV stock overviewJump to the live ticker page and report flow