Valuation Multiples for Banks and Insurers
Banks and insurers live in a different valuation universe than most stocks. Their balance sheets are their business—deposits and premiums collected upfront, claims and withdrawals paid later. Their leverage is structural, not a capital structure choice. Their earnings are cyclical and subject to regulatory constraints. Traditional multiples like P/E and P/B require heavy adjustment, and wholly new metrics like price-to-tangible-book-value and embedded-value ratios become essential. This chapter equips you to navigate the unique valuation language of financial institutions.
Quick definition: Price-to-tangible-book-value (P/TBV) is a bank's stock price divided by its tangible equity (book value minus intangible assets)—a more conservative view of equity than standard book value, since goodwill from acquisitions evaporates in a crisis.
Key Takeaways
- Price-to-book is misleading for banks; price-to-tangible-book is more honest, because tangible equity is what survives a crisis.
- P/E ratios for banks and insurers depend critically on loan-loss provisions and underwriting cycles, which swing more wildly than industrial companies.
- Return on equity (ROE) is the key comparator, but adjust for leverage: a bank with 10x leverage earning 12% ROE may be riskier than a 5x-levered bank earning 10% ROE.
- Float value in insurance is the equity cushion, not accounting equity—it's the underwriting profit available to invest long-term.
- Regulatory capital ratios determine dividend capacity and growth, not just accounting earnings.
- Tangible book value per share grows only if ROE exceeds the dividend payout—a crucial insight for assessing future shareholder value.
Why Standard Metrics Fail for Financials
Traditional valuation multiples break down for banks and insurers for three reasons:
First, balance-sheet leverage is structural. A bank operates 10–12x levered by design; an insurer, 2–3x. This leverage is not a capital structure choice but a requirement of the business. Comparing a bank's P/E to an industrial company's P/E is nonsensical—the leverage inflates bank earnings per share artificially.
Second, accounting earnings swing with reserves and provisions. Banks set aside loan-loss provisions when the economy softens; insurers adjust loss reserves when claims emerge. These are non-cash charges that depress reported earnings but protect future solvency. A bank reporting 20% earnings growth while its loan-loss provisions spike may be taking on risk, not creating value.
Third, regulatory capital constraints bind. A bank cannot pay out all its earnings as dividends if doing so would breach minimum capital ratios (typically 10–13% Tier 1 capital). An insurer cannot return excess capital without regulatory approval. These constraints mean accounting earnings are not fully distributable to shareholders.
Price-to-Book and Price-to-Tangible-Book
The most-used valuation metric for banks is price-to-book (P/B): stock price divided by book value per share. But book value includes intangible assets—goodwill from acquisitions, core-deposit intangibles—that evaporate in a crisis. The more prudent metric is price-to-tangible-book (P/TBV): stock price divided by tangible equity (book value minus all intangibles).
Example: JPMorgan in late 2023 had:
- Book value per share: $95
- Tangible book value per share: $72
- Stock price: $165
The P/B ratio was 1.74x; the P/TBV ratio was 2.29x. The tangible metric is higher because it strips out ~$28B in goodwill and intangibles (mostly from past acquisitions). Investors typically prefer to see P/TBV between 0.8x and 1.5x in normal times; much above 1.5x suggests the market is pricing in significant earnings growth or ROE expansion.
What P/TBV really means: It's the equity cushion—the amount of pure capital available to absorb losses before equity holders lose money. A bank trading at 1.2x tangible book is asking the market to pay $1.20 for every $1.00 of loss-absorbing capital. If that bank's ROE is 15%, the multiple is justified; if ROE is 8%, it's stretched.
ROE Adjustment for Leverage
Bank earnings per share depend heavily on leverage. A bank earning 10% ROE on $150B equity (using $1.5 trillion in assets, 10x leverage) generates $15B net income. Divide by 1B shares and you get $15 EPS. Lower the leverage to 8x, and the same 10% ROE on $187.5B equity generates $18.75B net income, or $18.75 EPS—20% higher earnings per share.
This is why adjusting ROE for leverage is essential.
A robust framework compares banks on risk-adjusted ROE: ROE scaled by the leverage ratio. If Bank A earns 12% ROE on 10x leverage and Bank B earns 11% ROE on 8x leverage, Bank A's risk-adjusted performance (12% ÷ 10 = 1.2% ROE per unit of leverage) is only marginally better, and Bank A carries more tail risk.
A better approach is to compare spread economics: the net interest margin (NIM) and non-interest income relative to assets. A bank with a 3% NIM on $1 trillion assets earns $30B in pre-provision operating income. If costs are $10B, pre-provision net revenue is $20B, and loan-loss provisions run 0.5–1.0% of loans (depending on cycle), the sustainable net income is clear. This avoids the ROE distortion from leverage and reveals the underlying economics.
Earnings Quality and Loan Losses
Bank earnings are only as good as the loan portfolio. P/E ratios for banks mean little without adjusting for the loan-loss cycle.
When the economy is strong and unemployment low, loan losses are minimal (0.1–0.3% of loans). Banks report high earnings. P/E ratios compress, and multiples look cheap. But as the cycle turns and unemployment rises, loan-loss provisions spike. Earnings collapse, and P/E ratios expand, catching unwary investors.
Pro analysts compute "normalized" or "through-cycle" earnings by averaging net income across a full credit cycle (recession + recovery). A bank with $20B net income in boom years and $10B in recession might normalize to $15B. Valuing it on the boom number overstates sustainable earnings; valuing on the recession number understates the normacy case.
Example: During 2021–2022, US banks reported rock-bottom loan losses (0.1–0.2% of loans). In 2023, as Fed tightening took hold and commercial real estate softened, provisions climbed to 0.5–0.8%. Earnings fell 20–30%, though the fundamental economics hadn't collapsed. Investors who extrapolated 2021–2022 earnings got blindsided.
Adjust for this by asking: what is the loan-loss provision as a percent of loans, and is it rising or falling? If rising sharply from a multi-year low, assume provisions will normalize at mid-cycle rates.
Insurance Float and Book Value
For insurers, float is the lifeblood. Float is the premiums collected upfront, held in reserves, and invested until claims are paid. A $100B balance-sheet insurer might have $80B in float and $20B in shareholder equity.
Float is valuable if the insurer invests it profitably or underwrites insurance at a profit. If underwriting is profitable (premiums exceed claims and expenses), the float grows without the insurer having to raise more capital. If underwriting is unprofitable, the float shrinks and turns into a liability.
The key metric is underwriting profit as a percent of premium (the combined ratio). A combined ratio below 100% means the insurer makes money on the underwriting itself; a ratio above 100% means it loses money and relies on investment returns to offset losses.
Example: Berkshire Hathaway's insurance float generates underwriting profits in some years (combined ratio ~95%) and underwriting losses in others (combined ratio ~105%, during catastrophic years). The $180B+ float is valuable because Berkshire invests it in stocks and bonds at high returns, covering underwriting losses and generating shareholder value.
Standard P/E multiples for insurers miss this. A more relevant metric is price-to-embedded-value (PEV) or price-to-intrinsic-value (PIV), which estimate the present value of future float generation and underwriting profits. But these require detailed actuarial modeling, often available only from specialized insurers or analysts.
For quick screening, use price-to-book, but adjust for the quality of the float. A company with $20B float earning 5% annual returns (from smart investment) is worth more than a company with $20B float earning 2% returns.
A Mermaid View of Financial Institution Valuation
Regulatory Capital and Dividend Capacity
A key difference between banks and other businesses is that dividend capacity is constrained by regulatory capital rules, not just earnings.
A bank must maintain Tier 1 capital ratios of 10–13% of risk-weighted assets. If the bank earns $20B and its current Tier 1 ratio is at the minimum, it cannot pay out all $20B as a dividend; it must retain enough to keep the ratio above the minimum. This creates a hidden drag on shareholder returns.
Conversely, if a bank has a strong capital surplus and can push the Tier 1 ratio to 15%, it has flexibility to increase the dividend or buy back shares without impairing capital.
For valuation purposes, adjust earned ROE downward by the amount of earnings that must be retained for regulatory capital. A bank earning $20B with a 12% ROE might retain $5B for capital (25% of earnings) and distribute $15B to shareholders (via dividends and buybacks). The true return on retained capital is lower, boosting shareholder returns is lower too.
This is less transparent than for a regular company—hence the need to audit the capital position and regulatory guidance.
Tangible Book Value Growth as a Valuation Anchor
One of the most reliable long-term valuation anchors for banks is tangible book value (TBV) per share growth. Here's the math:
Tangible book value per share grows when (1) the bank earns a positive ROE and retains earnings, or (2) it buys back shares at less than TBV.
If a bank has $100B TBV, earns 12% ROE = $12B net income, and has 2B shares, then TBV per share = $50. If it retains 50% of earnings ($6B), TBV rises to $106B. But if it issued new shares (to maintain ratios), TBV per share might only grow to $52, a 4% increase.
The sustainable TBV growth rate is ROE × (1 − Payout Ratio). A bank with 12% ROE and a 40% payout ratio has 7.2% sustainable TBV growth. Over 10 years, TBV per share doubles. This is a powerful lens for long-term intrinsic value.
Real-World Examples
JPMorgan Chase, 2020–2023: JPM's tangible book value per share grew from ~$58 (2020) to ~$72 (2023), a 3-year CAGR of 7.5%, below the 12% ROE it earned. Why? Because JPM was constrained by regulatory capital rules from retaining all earnings; it returned significant capital via buybacks. But at 1.2x P/TBV, the market was pricing in the sustainable long-term TBV growth, making the stock fairly valued for a long-term holder.
Regional banks post-2023: Banks like SVB Financial (failed) and First Republic (failed) collapsed when asset mark-to-market exposure and deposit outflows exposed weak tangible book value. A bank trading at 0.6x tangible book often signals severe stress; here, it was warranted.
Berkshire Hathaway, float-driven value: Berkshire's float grew from $40B (2010) to $180B+ (2024) without raising new capital, because it underwritten insurance profitably and invested float wisely. The market paid up to 1.5x book value for Berkshire, partly because float was growing and earning high returns. The P/B multiple was justified by the float economics.
Arch Capital, underwriting cycle: Arch trades at 0.9–1.2x book, depending on the underwriting cycle. In 2021–2022, when property/casualty rates were rising and underwriting was profitable, Arch traded at 1.2x P/B. In 2023, as competition increased and underwriting margins compressed, it fell to 0.95x P/B. The earnings were steady, but the quality of future underwriting profit (and thus float growth) dimmed.
Common Mistakes
1. Comparing bank P/E to industrial P/E. A bank trading at 10x P/E and an industrial at 14x P/E doesn't mean the industrial is more expensive. The leverage embedded in bank earnings distorts the comparison. Always use P/TBV for banks.
2. Extrapolating a strong loan-loss environment into perpetuity. When loan losses are at multi-year lows, extrapolating them forward overstates sustainable earnings. Normalize through-cycle.
3. Ignoring deposit/funding costs. A bank's net interest margin depends on the deposit environment. Low rates meant cheap deposits (2015–2021); rising rates increased deposit costs (2022–2023). Compare NIMs to the rate environment and peer group, not in absolute terms.
4. Over-weighting capital management skill. A bank that shrinks its capital base and buys back shares may boost EPS even if ROE is declining—a value trap. Watch for organic TBV per share growth, not just reported EPS growth.
5. Assuming regulatory capital ratios are permanently loose. During crisis periods (2008–2009, 2020), regulators loosened capital rules. As the cycle normalizes, tighter capital rules return, constraining dividends and buybacks. Plan for normalized regulatory ratios.
6. Treating float like equity. Insurance float is liabilities. It's valuable only if invested profitably or underwritten profitably. Don't count it as equity cushion. A poorly managed insurer can lose float rapidly.
FAQ
Q: Is price-to-tangible-book the only metric I need for bank valuation?
A: It's a good starting point, but supplement it with ROE (adjusted for leverage), net interest margin trends, loan-loss provisions, and tangible book value per share growth. A bank at 0.9x P/TBV with declining ROE and rising loan losses is a value trap; one at 1.3x P/TBV with rising ROE and stable loan losses is fairly priced.
Q: How do I estimate normalized earnings for a bank?
A: Average net income over a full credit cycle (typically 7–10 years). Adjust loan-loss provisions to the through-cycle rate (often 0.5–0.8% of loans for most banks). If the bank is near-peak (unemployment very low), assume provisions will rise. If near-trough (unemployment high), assume provisions will fall.
Q: What's a good price-to-tangible-book range for banks?
A: In normal times, well-capitalized, profitable banks trade at 1.0x to 1.5x P/TBV. Below 0.9x often signals distress or market pessimism (sometimes a buy). Above 1.5x suggests the market expects significant ROE expansion. Tier 1 banks (JPM, BofA, Citigroup) trade 1.1–1.3x. Regional banks trade 0.9–1.2x.
Q: How do I value the float of an insurer?
A: Start with the stated float value on the balance sheet, then estimate annual underwriting profit or loss (premiums - claims - expenses). If the combined ratio is 95%, the insurer earns 5% on float. Invest that at the current cost of capital (5–7% for a mature insurer) to estimate the intrinsic value contribution from float. Add in the stated equity value, and you have a rough intrinsic value.
Q: Should I use ROE or ROIC for banks?
A: ROE is standard for banks because they don't report (and don't use) much invested capital in the traditional sense. But ROE is inflated by leverage; compare it to peers of similar leverage. If you want apples-to-apples comparisons across different leverage levels, adjust ROE downward by the leverage multiple, or compute return on assets (ROA) instead.
Q: Can I use dividend yield to value a bank?
A: Only with caution. Bank dividends are constrained by regulatory capital rules, not just earnings. A bank might have capacity to raise dividends, but regulators may forbid it during stress tests. Use dividend yield as a secondary anchor, not the primary one.
Related Concepts
- Leverage and financial risk – Understanding why 10x leverage is normal for a bank but dangerous for an industrial company.
- Credit cycles and loan losses – How loan-loss provisions swing with the economy and distort reported earnings.
- Net interest margin (NIM) and spread economics – The fundamental driver of bank profitability, separate from leverage effects.
- Insurance underwriting cycles – How combined ratios swing with competition and catastrophe patterns.
- Regulatory capital ratios – Why Tier 1 ratios matter more to dividends and growth than accounting earnings.
- Book value per share growth – The long-term anchor for intrinsic value in financial institutions.
Summary
Valuing banks and insurers requires stepping outside the standard P/E framework and mastering price-to-tangible-book, ROE (adjusted for leverage), and through-cycle earnings normalization. Regulatory capital rules constrain dividends more than earnings alone suggest. For insurers, float quality and underwriting profitability drive long-term value. Watch loan-loss provisions and credit-cycle dynamics for banks; underwriting combined ratios and investment returns for insurers. Use tangible book value per share growth as a long-term anchor for intrinsic value, and remember that most financial institution valuations compress when risk rises and expand when risk falls—more so than most sectors.