Key takeaways
- CAPM gives the cost of equity — the discount rate equity investors require for taking on the risk of holding a stock.
- Formula: Re = Rf + β × ERP, where Rf is the risk-free rate, β is beta (sensitivity to the market), and ERP is the equity risk premium.
- Inputs to source carefully: 10-year government bond for Rf; regressed beta over 2–5 years; Damodaran's implied ERP for ERP.
- Plugs directly into WACC as the cost-of-equity component.
- Critiques are real (single-factor, doesn't capture all risk premia), but CAPM is the practitioner default — use it, sense-check the inputs, sensitise the WACC.
The formula, walked through
where (Rm − Rf) is the equity risk premium (ERP)
The risk-free rate (Rf)
The yield on a riskless investment of the same currency and rough duration as the cash flows you're discounting. Standard choice: 10-year government bond yield in the company's reporting currency.
- USD: 10-year US Treasury yield (~4.0–4.5% in 2025–2026).
- EUR: 10-year German Bund yield.
- GBP: 10-year UK Gilt.
- JPY: 10-year JGB yield.
For emerging markets, use the country's 10-year sovereign yield or a mature-market rate plus a country-risk premium (Damodaran publishes country risk premia).
Beta (β)
Beta measures the stock's sensitivity to broad market moves. β = 1 means the stock moves 1% for every 1% market move. β = 1.5 means 1.5%. β below 1 means the stock is less volatile than the market.
Source: regress the stock's monthly returns against a broad market index (S&P 500 for US, MSCI World for global) over 2–5 years. Bloomberg, Yahoo Finance, and Damodaran's data tables all publish beta. For private companies or recent IPOs, use a "bottom-up beta" — average the betas of public peers, unlever to remove capital-structure effects, then re-lever to your target's structure.
Common defaults: 1.0–1.2 for mature large-caps, 1.3–1.6 for growth, 0.6–0.9 for utilities and consumer staples, 1.6+ for highly cyclical or early-stage companies.
Equity risk premium (ERP)
The extra return investors require for holding equities over the risk-free asset. The most contested input in CAPM.
- Historical ERP: long-run realised return on equities minus realised risk-free rate. US data suggests ~6.5–7.5% over 100+ years.
- Implied ERP: solve backward from current S&P 500 prices and consensus dividend / earnings forecasts. Damodaran publishes monthly. Typically ~4.5–6.0% in normal times.
- Survey ERP: surveys of CFOs, academics, and practitioners. Currently ~5–6%.
Practitioner default: 5.0–6.0% in developed markets. Damodaran's implied ERP is the most defensible single anchor — it's market-consistent (uses current prices) and updated regularly.
Worked example
A US large-cap with beta 1.2:
- Rf = 4.25% (10-year Treasury)
- β = 1.2
- ERP = 5.5% (Damodaran implied)
- Re = 4.25% + 1.2 × 5.5% = 10.85%
That's the cost of equity to plug into WACC. If the company is 70% equity / 30% debt at market values, with after-tax cost of debt of 4%, then WACC = 0.70 × 10.85% + 0.30 × 4% = 8.8%.
Sense checks on the inputs
| Input | Healthy range | If outside |
|---|---|---|
| Rf | ~3.5–5.0% USD (2025–2026 environment) | Below 1%: extreme rate environment, may need to normalise. Above 6%: stress regime. |
| β | 0.6–1.6 for most listed companies | Above 2.0: very speculative, double-check the regression. Below 0.5: utility-like, defensible if industry confirms. |
| ERP | 4.5–6.5% developed markets | Above 7%: sourcing historical realised premium (overstated for forward-looking valuations). Below 4%: optimistic — would imply cost of equity below most banks' lending rates. |
| Re (output) | 7–14% for most listed companies | Below 7% or above 15%: trace inputs — likely an error in beta or ERP. |
Critiques of CAPM
CAPM has known limitations. It's a single-factor model — only market sensitivity matters — when in reality returns are also driven by size (small-caps outperform), value (high book-to-market outperforms), momentum, profitability, and other factors documented in the academic literature (Fama-French, Carhart).
The Fama-French three-factor and five-factor models add these as explicit risk premia. They're more accurate empirically but more cumbersome to use. For practical valuation work, CAPM is the default — it's transparent, auditable, and good enough when the inputs are sense-checked. Where CAPM produces a clearly wrong answer (a stable utility with cost of equity of 18%, say), use the sense check to find the input error rather than abandoning the model.
Common errors
- Wrong Rf tenor. Using the 3-month T-bill for a 10-year DCF mismatches the duration. Match Rf to cash-flow duration — 10-year is standard for general DCF work.
- Stale beta. Beta from 5 years ago may not reflect current business mix (e.g., post-acquisition, post-divestiture). Re-regress on recent data.
- Historical ERP for forward-looking work. Realised premia are usually higher than implied — using historical inflates cost of equity. Implied ERP is the better forward-looking anchor.
- Bottom-up beta without re-levering. If you average peer betas, you've averaged across capital structures. Unlever to the asset beta, then re-lever to your target's leverage.
- Country risk premium missed. Emerging-market companies need a country-risk premium added on top of the mature-market ERP, or use the local sovereign as Rf.
How Smalt AI builds it
Every DCF Smalt AI builds includes a dedicated WACC tab with a CAPM build for cost of equity:
- Rf sourced from the current 10-year sovereign yield in the company's reporting currency, with the source date in the cell comment.
- Beta regressed on 5-year monthly returns vs the relevant market index. For private targets or recent IPOs, an unlever / re-lever bottom-up beta from public peers.
- ERP defaulted to Damodaran's implied premium (most recent month available), with country-risk premium added if the company is emerging-market based.
- Cost of equity output sense-checked against the 7–14% band. Outside that range triggers a flag for review.
- Sensitivity on Re via the WACC × terminal growth table — the indirect way to stress-test CAPM inputs that flow through to the implied share price.
Further reading
- Damodaran, Aswath — Investment Valuation, Chapter 7 on cost of capital.
- Damodaran online data — implied ERP updated monthly, country risk premia, beta tables. pages.stern.nyu.edu/~adamodar.
- Fama & French (1992) — original three-factor critique of CAPM.