What the cost of equity is, and what it decides
The cost of equity is the annual return a shareholder must expect before committing money to your stock rather than to a government bond. It is an opportunity cost, not a cash expense: nothing in the accounts records it, no invoice arrives, and yet it sets the value of the business. Discount the same free cash flows at 9% instead of 11% and an equity value can move 25% or more.
Three decisions run on this number. Valuation: it is the discount rate for equity free cash flow and dividends, and the Re input to a weighted average cost of capital. Capital allocation: a project must clear the cost of the capital that funds it, so a hurdle rate built on equity that is too cheap approves value-destroying investments. Performance measurement: economic profit and economic value added both charge the business for equity at this rate.
CAPM — the Sharpe–Lintner capital asset pricing model, after the two economists who derived it independently in the mid-1960s — answers the question with one insight: investors hold diversified portfolios, so they will not pay for risk that diversification removes. Only the part of a stock's volatility that moves with the market — its beta — earns a premium. Idiosyncratic risk, however painful to a concentrated owner, is free to a diversified one and so commands nothing.
The CAPM formula, term by term
The model prices risk as a straight line. Start at the risk-free rate, then add a slice of the market's risk premium proportional to how much market risk you carry.
The risk-free rate (Rf) is the return available with no default and no reinvestment surprise. In practice you use a government bond yield whose maturity matches the horizon of your cash flows: for a going-concern valuation with a terminal value, the 10-year or 20-year Treasury, not a three-month bill. Take today's yield, not a historical average — the current yield already embeds the market's inflation expectation, which is exactly the expectation your nominal cash flows should share.
Beta (β) is the slope of a regression of the stock's excess returns on the market's excess returns. A beta of 1.15 says that when the index rises 10%, this stock has historically risen about 11.5%. Beta is the only company-specific term in pure CAPM, and it is where most of the error lives. The standard error on a five-year monthly regression is large enough that the point estimate should be read as the centre of a range rather than as a fact, which is why practitioners frequently discard the raw regression and build a bottom-up beta from the median of unlevered peer betas instead. The stock beta calculator runs the regression itself if you have the return series.
The equity risk premium (E[Rm] − Rf) is the extra return the whole equity market must offer over bonds. It is a forward-looking expectation nobody observes directly, so it is estimated — from long-run realised excess returns, from surveys, or by solving for the discount rate that reconciles current index prices with forecast cash flows, the implied ERP. Estimates cluster between 4% and 6% for the US. This single input drives the answer as hard as beta does: moving from 5% to 6% with a beta of 1.15 adds 1.15 points to the cost of equity.
The build-up premiums are extensions, not CAPM. Small companies have historically earned more than their betas predict, and appraisers add a size premium read from a published decile table. A company-specific premium covers risks a diversified investor genuinely cannot escape by holding your stock — a single customer at 60% of revenue, an owner whose departure ends the business. Both are legitimate in private-company work and both are where valuations get padded, so document each one.
Worked example: a beta-1.15 company with a 5% premium
You are valuing a listed industrial with a five-year regression beta of 1.15. The 20-year Treasury yields 4.25%, and your firm's house view on the equity risk premium is 5.0%.
- Fix the risk-free rate. Rf = 4.25%. This is the floor: an investor can have it without owning any equity.
- Compute the beta-adjusted premium. β × ERP = 1.15 × 5.00% = 5.75 percentage points. That is the compensation for bearing 1.15 times the market's systematic risk.
- Add them. Re = 4.25% + 5.75% = 10.00%. Pure CAPM is finished in two lines.
- Sanity-check the implied market return. Rf + ERP = 4.25% + 5.00% = 9.25%. If your model implies a return on the whole index you would not defend out loud, the inputs are wrong.
Now suppose the same business is private, with revenue of $40 million, one customer at 45% of sales, and a founder who signs every major contract. A published size-premium table for that market-cap decile gives 1.00 point, and you conclude customer concentration justifies 0.50 more:
- Size premium. +1.00 point, running total 11.00%.
- Company-specific premium. +0.50 point, giving 11.50%.
The two versions differ by 150 basis points. On a business generating $3 million of equity free cash flow growing at 3%, capitalising at 10.00% minus 3% gives $42.9 million; at 11.50% minus 3% it gives $35.3 million — an 18% haircut from two judgement calls. That is why the premiums belong in the report with their sources, not buried in a spreadsheet cell.
How to read the number you get
Judge your answer against the range the market supports, not against a feeling. With a risk-free rate near 4% and a 5% premium, large-cap costs of equity land roughly between 7% and 12%: a regulated utility with a 0.55 beta near 7%, a consumer staple near 8.5%, a broad industrial near 10%, a high-beta semiconductor or a levered cyclical between 12% and 14%. Private-company build-ups commonly land between 12% and 20% once size and specific premiums are stacked. A cost of equity below the risk-free rate requires a negative beta, which is almost always a regression artefact. Above 25% you are no longer using a cost of capital — you are using a venture-style target return, and you should say so.
Check three consistency conditions before you use the number. First, currency and inflation: a nominal dollar discount rate belongs with nominal dollar cash flows, so if you forecast in real terms, strip the inflation expectation out of the rate too. Second, leverage: a regression beta reflects the company's current capital structure, so if your plan levers up, relever the beta and the cost of equity rises. Third, country risk: cash flows earned in a market with meaningful sovereign risk need an explicit adjustment, which the country risk premium calculator handles.
Finally, the cost of equity always exceeds the after-tax cost of debt, and by a wide margin — equity is residual, unsecured, and its return is not tax-deductible. If your build-up produces a cost of equity within a point or two of your borrowing rate, something is wrong.
Cost of equity at a 4.25% risk-free rate
| Beta | Typical profile | ERP 4.0% | ERP 5.0% | ERP 6.0% | ERP 7.0% |
|---|---|---|---|---|---|
| 0.60 | Regulated utility, water | 6.65% | 7.25% | 7.85% | 8.45% |
| 0.80 | Consumer staples, telecom | 7.45% | 8.25% | 9.05% | 9.85% |
| 1.00 | The index itself | 8.25% | 9.25% | 10.25% | 11.25% |
| 1.20 | Diversified industrial | 9.05% | 10.25% | 11.45% | 12.65% |
| 1.40 | Discretionary retail, media | 9.85% | 11.25% | 12.65% | 14.05% |
| 1.60 | Semiconductors, airlines | 10.65% | 12.25% | 13.85% | 15.45% |
| 2.00 | Highly levered cyclical | 12.25% | 14.25% | 16.25% | 18.25% |
Profiles are typical sector positions, not a substitute for your own peer set. Add any size or company-specific premium on top of these figures.
Which risk-free rate and which premium a reviewer will accept
Valuation reviewers, auditors and tax authorities care less about the exact figure than about internal consistency and disclosure. The defensible pattern is: a spot government yield matched to the cash flow horizon, sourced and dated; an equity risk premium from a named published source or a stated implied-ERP calculation; a beta whose source, window and index are disclosed; and every build-up premium tied to a specific written risk. Mixing a spot risk-free rate with a historical-average premium is standard practice, but mixing a normalised risk-free rate with an implied premium computed off the spot rate double-counts the same adjustment.
Mistakes that produce a wrong cost of equity
- Using a T-bill yield with a 20-year forecast. Short-rate volatility flows straight into your valuation. Match the bond maturity to the cash flow horizon.
- Entering the expected market return where the premium belongs. Typing 9.25 into an equity-risk-premium field turns a 10% cost of equity into roughly 15%. Switch the input mode instead.
- Trusting a raw regression beta. Standard errors are wide, short windows are noisy, and thinly traded stocks show downward-biased betas. Compare against a peer-median bottom-up beta.
- Applying a levered peer beta to a differently levered target. Unlever each comparable at its own debt-to-equity and tax rate, take the median, then relever at your target structure.
- Stacking a size premium on a company that is not small. Size premiums come from market-cap deciles; read the table rather than adding a round number by instinct.
- Double-counting company-specific risk. If you have already haircut the forecast for customer concentration, charging a premium for it as well penalises the same risk twice.
- Discounting real cash flows at a nominal rate. A nominal risk-free rate already contains the market's expected inflation, so pairing it with real cash flows understates value badly.
- Using the cost of equity to discount firm-level free cash flow. Unlevered free cash flow belongs with WACC. Only equity free cash flow and dividends belong with Re.
Where CAPM falls short, and what practitioners use instead
CAPM is a single-factor model resting on assumptions no market satisfies: everyone holds the market portfolio, borrows and lends freely at the risk-free rate, and agrees on the distribution of returns. The empirical record is mixed. The relationship between beta and realised return has been flatter than the model predicts for decades, and characteristics CAPM says should not matter — size, book-to-market, momentum, profitability — have carried persistent premia. Fama and French's three- and five-factor models exist precisely because of that evidence.
So why is CAPM still the professional default? Because it is transparent, needs three inputs, is reproducible by a reviewer, and produces answers in the same range as its more elaborate rivals for the great majority of established companies. The disagreement between a CAPM cost of equity and a Fama-French one is usually smaller than the error bar on beta itself.
Know the alternatives and when to reach for them. A multi-factor model helps when your subject loads heavily on a documented factor — a deep-value small cap, for example. The build-up method replaces beta entirely with a stack of industry, size and specific premiums, and is the norm for private companies with no usable price history. The implied cost of equity inverts a dividend discount or residual income model, solving for the rate that matches the observed share price to consensus forecasts; the Gordon growth dividend discount model does this in one line for a stable dividend payer.
Whichever you choose, run the valuation both ways. A conclusion that survives a plausible range of costs of equity is worth far more than a point estimate carried to three decimals — and if the recommendation flips between 9.5% and 10.5%, the honest output is a range, which is what the sensitivity grid in the DCF enterprise value calculator is for.
Key terms
- Systematic risk
- Risk that moves with the whole market and cannot be diversified away. The only risk CAPM pays you for.
- Idiosyncratic risk
- Company-specific volatility that vanishes in a diversified portfolio. Priced at zero by CAPM, which is why a company-specific premium needs a special justification.
- Security market line
- The straight line CAPM draws from the risk-free rate through the market portfolio: expected return rises linearly with beta.
- Levered beta
- Beta of the equity as it currently trades, reflecting both business risk and financial leverage. This is what a regression on share prices measures.
- Implied equity risk premium
- The premium solved out of current index prices and forecast cash flows rather than measured from history. Forward-looking, and it updates daily.
- Size premium
- Historical excess return of small-capitalisation stocks over what their betas predict, published by market-cap decile and added in build-up models.