Cost of Equity and CAPM Explained for Interviews

Introduction

Almost every other input in a valuation can be looked up. The tax rate is in the filings, the share count is on the cover of the 10-K, and the cost of debt is visible in the company's own bond yields. The cost of equity comes from nowhere. No shareholder has ever signed a document promising to accept 10.4%, no exchange quotes it, and no auditor verifies it. It is a number the analyst constructs, and every construction rests on assumptions an interviewer can pull apart in ninety seconds.

That is exactly why it gets asked about so relentlessly. Cost of equity sits at the heart of the DCF walkthrough that opens most valuation interviews, it drives the discount rate that determines the answer, and it reveals whether a candidate understands risk or has simply memorised a formula.

This post covers what the cost of equity actually represents, why it sits above the cost of debt, what each CAPM input means and where practitioners source it as of September 2026, the premiums that get bolted on top, what to do when CAPM breaks down, and the interview traps built around it.

The first thing to be clear about is what the cost of equity is a cost of. It is not a cash outflow the company records anywhere. It is an opportunity cost: the return an investor could get on an equally risky alternative, and therefore the return the company must deliver to justify holding onto shareholders' capital rather than having them sell.

Both columns describe money that capital providers demand for bearing risk. What separates them is the strength of the promise attached, and that difference explains the whole gap.

Position in the Waterfall Sets the Price

A lender holds a contractual claim. Interest is due on a stated date, principal is due at maturity, and failure to pay either triggers a legal process that lets creditors enforce against the business. Equity holders hold a residual claim: whatever is left after suppliers, employees, tax authorities and every layer of debt have been paid. In a good year the residual is large and uncapped. In a bad year it is nothing, and there is no court a shareholder can go to about it.

That asymmetry does two things to the required return. It makes equity cash flows far more volatile than debt cash flows for the same underlying business, and it means equity absorbs the first loss when anything goes wrong. Investors price both features, so the required return on equity must exceed the required return on the debt of the same issuer, always, at every point in the capital structure. A candidate who cannot state this cleanly is telling the interviewer they have never thought about seniority.

Cost of Equity

The rate of return that equity investors require in order to hold a company's shares, given the risk of those shares relative to other available investments. It is an opportunity cost rather than a cash expense, it never appears in the income statement, and because equity carries no contractual payment schedule it cannot be read off a market price the way a bond yield can. It has to be estimated, most commonly with the Capital Asset Pricing Model.

The Tax Shield Widens the Gap Further

Interest is deductible against taxable income; dividends and buybacks are not. That means a company paying 6.5% on its debt at a 25% marginal tax rate bears an effective cost of 4.875%, because the deduction returns a quarter of the interest bill. Equity gets no such subsidy. The after-tax cost of debt is what enters WACC, so the visible gap between the two costs is wider than the gap between the raw required returns.

This is where a lot of candidates over-rotate and conclude that debt is simply better. It is cheaper, which is not the same thing. Debt brings fixed obligations, covenants, refinancing risk and a bankruptcy option that lenders can exercise, and each additional turn of leverage raises the cost of the equity that sits beneath it. The cheapness of debt is bought with risk transferred onto shareholders, and CAPM is the standard machinery for pricing exactly that transfer.

The Capital Asset Pricing Model remains the default for cost of equity in banking, corporate finance and most valuation practice, not because it is a good description of how markets work but because it is transparent, defensible and universally understood. Everyone knows what you did, and everyone can argue with the inputs rather than the framework.

Where $R_e$ is the cost of equity, $R_f$ is the risk-free rate, $\beta$ is the stock's sensitivity to the market, and the bracketed term $\left(R_m - R_f\right)$ is the equity risk premium, the extra return investors demand for holding equities rather than government bonds.

What the Formula Claims

CAPM makes one substantive economic claim: investors hold diversified portfolios, so they only get paid for risk that cannot be diversified away. Idiosyncratic risk (a plant fire, a failed drug trial, a bad quarter at one company) can be eliminated by holding enough names, so the market does not compensate you for bearing it. Systematic risk, the part of a stock's movement that tracks the whole market, cannot be diversified away, so that is what beta measures and that is what earns a premium.

Everything that follows from that claim is worth understanding, because it is the source of both the model's power and its most-criticized results. It is why a single-product biotech and a diversified consumer staples business can end up with similar costs of equity despite wildly different failure probabilities. It is why company-specific disaster risk does not, in strict CAPM terms, belong in the discount rate at all: it belongs in the cash flow forecast, as a probability-weighted expectation.

The Risk-Free Rate and Which Tenor to Use

In dollar valuations the risk-free rate comes from US Treasury yields, and the only real question is which point on the curve. The answer matters more than candidates expect, because the curve is not flat. As of early September 2026, the Federal Reserve's H.15 release put the 10-year Treasury constant maturity yield at 4.77% and both the 20-year and 30-year at 5.25% on 3 September 2026. Choosing the long bond instead of the 10-year adds roughly half a percentage point to the cost of equity before you have made a single other assumption.

The theoretically clean answer is to match duration: a DCF with a terminal value is a perpetuity, so the discount rate should reference the longest reliable government bond. The practical answer in banking is that the 10-year Treasury is the market convention, it is the most liquid and most quoted point on the curve, and using anything else invites a question about why. Both answers are acceptable. What is not acceptable is being unable to say which one you used or why.

There is a second convention worth knowing because valuation practitioners use it heavily. Kroll, whose cost of capital data is a standard reference in valuation and litigation work, recommends pairing its equity risk premium with the higher of a normalized US risk-free rate of 3.5% or the spot 20-year Treasury yield at the valuation date. The normalization exists to stop discount rates collapsing when policy rates are artificially low. In September 2026 the spot 20-year yield of 5.25% comfortably exceeds the 3.5% floor, so the spot rate governs and the normalization does no work at all.

Non-US work adds one more wrinkle. Where the local government itself carries default risk, its bond yield is not risk-free, and the standard fix is to strip the sovereign default spread out before using it. This has stopped being a purely emerging-market issue: Damodaran now treats even the US Treasury yield as carrying a small default component, and strips a US default spread of roughly 0.2 percentage points out of it before using it as the dollar risk-free rate in his country risk premium work.

The Equity Risk Premium and Where Practitioners Get It

The equity risk premium is the single largest source of disagreement in any cost of capital debate, and there are three families of estimate, none of which agrees with the others.

The numbers as of the second half of 2026 illustrate the spread. Damodaran's current data page at NYU Stern put the implied US equity risk premium at 4.14% on 1 September 2026 on his headline trailing-twelve-month measure, against a 4.75% dollar risk-free rate. His own alternative specifications on the same date ranged from 3.56% on normalized earnings to 6.05% using the average cash flow yield of the last ten years, which is a useful reminder that the model, not just the market, drives the output. Kroll's recommended US equity risk premium has been held at 5.0%. The 2025 edition of Pablo Fernandez's annual survey of practitioners and academics put the average market risk premium used for the United States at 5.5%, and the 2026 edition, covering 97 countries, was published in April 2026.

Equity Risk Premium

The additional annual return investors require for holding a diversified basket of equities instead of default-free government bonds. It is the term multiplied by beta in CAPM, so it scales the entire risk adjustment: a company with a beta of 1.2 inherits 1.2 times whatever premium you choose. Estimates for the US market generally sit between about 3.5% and 6%, with the exact figure depending on whether it is measured from historical returns, backed out of current market prices, or taken from a published practitioner recommendation.

Interviewers rarely care which number you pick from that range. They care that you can name where it came from and that you have not paired it with an incompatible risk-free rate.

Beta and the Judgment Buried in It

Beta is the coefficient from regressing a stock's returns on a market index. A beta of 1.0 means the stock has historically moved with the market; 1.4 means it has amplified market moves by roughly 40%; 0.6 means it has damped them. Statistically it is the covariance of the stock with the market divided by the variance of the market.

Reading that as a single objective number is the mistake. Every beta is the output of at least four choices:

Two providers running different conventions on the same stock routinely publish betas that differ by 0.2 or more, which translates into roughly a full percentage point of cost of equity at a 5% premium.

The most common adjustment is the Blume adjustment, which pulls a raw beta toward 1.0 on the empirical observation that betas mean-revert over time. The standard weighting is two-thirds raw and one-third market: a raw beta of 1.45 becomes 0.67 multiplied by 1.45, which is 0.97, plus 0.33, giving an adjusted beta of about 1.30. Bloomberg's default beta screen is adjusted, which is why a beta pulled off a terminal without checking will usually sit closer to 1.0 than the underlying regression does.

In practice bankers rarely use a single company's own regression beta. They take a comparable company set, unlever each comp's beta to strip out the effect of its capital structure, take the median, and relever to the target's structure. That process gets a full treatment in our guide to levered versus unlevered beta, and the practical question of what a data provider is actually handing you is covered in how to tell if a beta is levered or unlevered. The important point here is why the detour exists: an observed beta belongs to a specific company with a specific debt load, and using it unadjusted silently imports that company's financing decisions into your valuation.

Cost of equity questions surface in nearly every valuation interview: Work through CAPM, WACC and DCF technicals with worked answers, start practicing interview questions for free and find the gaps before an interviewer does.

Take a mid-cap US specialty chemicals manufacturer. The comparable set is four listed peers whose median unlevered beta works out to 0.95. The company's target capital structure is a debt-to-equity ratio of 0.45 on market values, its marginal tax rate is 25%, and it borrows at 6.5% pre-tax.

Step 1: Relever the Comp-Set Beta

The median unlevered beta describes the business risk of the peer group with financing stripped out. Relevering puts our company's own leverage back in:

Work the bracket first: 0.75 multiplied by 0.45 is 0.3375, so the bracket is 1.3375. Then 0.95 multiplied by 1.3375 is 1.2719, which rounds to a levered beta of 1.27. The company is modestly more sensitive to the market than the average stock, and about a quarter of that sensitivity comes from its debt rather than from what it makes.

Step 2: Apply CAPM

Now feed the three inputs in. We will use the 10-year Treasury constant maturity yield of 4.77% from 3 September 2026 as the risk-free rate, and Kroll's 5.0% recommended US equity risk premium:

A cost of equity of 11.1%. Note how much of that is the risk adjustment: 6.35 points of the 11.12 come from beta multiplied by the premium, so the risk term is larger than the risk-free term.

Now watch what two other defensible conventions do to the same company. Substitute Damodaran's implied premium of 4.14% and the cost of equity becomes 4.77% plus 1.27 multiplied by 4.14%, which is 4.77% plus 5.26%, or 10.0%. Follow Kroll's own pairing rule instead and use the spot 20-year yield of 5.25% with the 5.0% premium, and you get 5.25% plus 6.35%, or 11.6%. Same company, same beta, same day, and a range of 1.6 percentage points across three internally consistent methods. That range is the honest answer to "what is the cost of equity", and saying so out loud is far stronger than defending a single decimal.

Step 3: Roll Into WACC and Watch the Valuation Move

A debt-to-equity ratio of 0.45 means debt is 0.45 divided by 1.45, or 31% of total capital, with equity at 69%. The after-tax cost of debt is 6.5% multiplied by 0.75, which is 4.875%. Putting the pieces together using our 11.12% cost of equity:

That 9.2% WACC is the discount rate for unlevered free cash flow, and the full build is set out in our step-by-step guide to calculating WACC. Notice that cost of equity contributes 7.67 of the 9.18 points. For a typical corporate capital structure, WACC is mostly cost of equity wearing a hat, which is why interviewers who want to test your discount rate go straight to CAPM rather than to the weights.

Textbook CAPM stops after three inputs. Practice rarely does, because the model demonstrably underprices certain kinds of company. The standard response is to add explicit premiums, and each of them is contested.

Size and Company-Specific Premiums

The size premium is the extra return demanded for smaller companies, on the observation that small-capitalization stocks have historically delivered returns above what their betas predict. Kroll publishes size premia by market-capitalization decile, and they are used routinely in valuation work for smaller businesses. They are also genuinely disputed. The premium appears to have shrunk substantially since the effect was first documented, and researchers have argued both that it largely vanishes once you control for company quality and that the growth of private markets has removed much of the small-company return history from public datasets. Knowing that the debate exists is worth more in an interview than knowing a decile figure.

The company-specific risk premium is the more uncomfortable one. It is an added spread for factors like customer concentration, key-person dependence, thin management depth or a single-site manufacturing footprint. In strict CAPM terms these are diversifiable risks that do not belong in the discount rate at all. In practice they get added anyway, most often when valuing a private company where the buyer will not be diversified and the risks are real to that buyer. The honest framing is that it is a judgment adjustment, typically expressed in whole or half points, that should be documented rather than buried.

Country Risk Premium in Cross-Border Work

Valuing a business with meaningful emerging-market exposure using a US risk-free rate and a US equity premium understates the required return. The usual fix is a country risk premium added to the CAPM output, or in more careful work applied in proportion to the share of revenue or assets actually exposed to that country rather than to the whole company.

Country Risk Premium

An additional return that investors require for holding equities exposed to a specific country's political, economic and currency risk, over and above the premium demanded in a mature market such as the United States. The standard estimation route starts from that country's sovereign default spread, measured from its foreign-currency bond yields or its credit default swap spreads, then scales it upward to reflect the fact that equities are more volatile than government bonds. Damodaran publishes country risk premiums for most of the world and updates them at least annually.

The scaling step is where candidates get caught. A sovereign default spread prices the risk of a government not paying its bonds; equity in that country is riskier than that, so the raw spread is multiplied by the ratio of equity market volatility to government bond volatility before it is used. Quoting a CDS spread as if it were the equity premium skips the entire adjustment.

CAPM has been empirically challenged for decades. Beta explains far less of the cross-section of returns than the theory implies, low-beta stocks have persistently outperformed what the model predicts, and the model cannot say anything useful about a company with no traded equity. Practitioners have responded with alternatives rather than abandonment.

The Build-Up Method and Multifactor Models

The build-up method dispenses with beta entirely. Start with the risk-free rate, add the equity risk premium, add a size premium, add an industry premium if one is available, and add a company-specific premium. It is the standard approach for small private companies where no meaningful public comparable exists and where a regression beta is simply unavailable. Its weakness is obvious: without beta there is no systematic mechanism scaling the market premium to the specific business, so the answer is only as disciplined as the analyst stacking the layers.

Fama-French models go the other way, adding factors rather than removing them. The three-factor version supplements the market factor with size (small minus big) and value (high minus low book-to-market); the five-factor version adds profitability and investment. Empirically these explain more of the variation in returns than CAPM does, which is why they dominate academic work and quantitative investing. They are rare in banking because they require estimating a factor loading for each factor, they need a factor return series to multiply those loadings by, and a client is unlikely to accept a discount rate that cannot be explained in one sentence.

Backing Out an Implied Cost of Equity

The last approach inverts the problem. Instead of estimating a required return and using it to value the company, take the company's current share price as given, forecast the cash it will return to shareholders, and solve for the discount rate that reconciles them. In its simplest form, using the Gordon growth version of the dividend discount model, the cost of equity is the forward dividend yield plus the expected growth rate.

This is standard practice for regulated utilities, where dividends are stable and the growth rate is heavily constrained by the regulator, and for banks, where dividend policy is a central part of how the equity story is told. It is also the cleanest available sanity check on a CAPM output: if CAPM says 11% and the market's own pricing implies 8%, one of the two is telling you something.

Get the complete guide: Download our comprehensive 160-page PDF, access the IB Interview Guide covering valuation frameworks, DCF mechanics and the cost of capital questions interviewers return to every cycle.

Cost of equity questions almost never stop at the formula. The formula is the setup; the follow-up is the test.

What Happens to Cost of Equity When Leverage Rises

This is the most reliable trap in the entire topic, because the intuitive answer collides with a half-remembered fact. Candidates know that adding debt can lower WACC, and they generalize it into "leverage lowers the cost of capital, so cost of equity falls too." It does the opposite.

More debt means more fixed obligations ranking ahead of shareholders, so the residual claim becomes more volatile and equity holders demand more. Mechanically, levered beta rises with the debt-to-equity ratio, and a higher beta feeds straight into CAPM. Cost of equity rises monotonically with leverage. WACC can still fall over some range because the tax shield makes cheap debt cheaper still, but that is a statement about the blend, not about either component. Push leverage far enough and the cost of debt starts stepping up across rating bands too, distress costs become material, and WACC turns and rises. The trade-offs behind that turning point are laid out in our guide to capital structure decisions.

A clean answer names all three effects in order: cost of equity up, cost of debt eventually up, WACC down then up. Most candidates get one of the three.

Negative Betas, Historical Returns and Unobservable Numbers

Three more questions come up often enough to prepare properly.

"What if beta is negative?" Taken literally, CAPM says the cost of equity falls below the risk-free rate, because an asset that rises when the market falls is a hedge and investors will pay for that insurance by accepting a lower expected return. That is the correct model answer and you should give it. Then give the practical one: a negative beta on an operating company is almost always a statistical artifact rather than a real hedge, produced by a short window, an unrepresentative index, or a handful of outlier observations. Gold miners and some precious-metals names are the classic candidates, and even they usually print small positive betas over long windows. The professional response is to widen the estimation window, check the comp set, and use the comparable-company median rather than accept an output that implies equity is safer than Treasuries.

"Why not just use the stock's historical return?" Because realized returns are a terrible estimator of expected returns. Annual equity return volatility runs around 20%, so even with fifty years of data the standard error on the mean is roughly 20% divided by the square root of 50, which is about 2.8 percentage points. A confidence interval that wide is useless as a discount rate. There is a second, subtler problem: past returns include the effect of the discount rate itself changing. A stock that re-rated upward delivered strong historical returns precisely because its cost of equity fell, so plugging that high historical return back in as the required return gets the sign of the relationship backwards.

"Cost of equity cannot be observed, so how do you defend your number?" By showing your work and showing the range. Name the risk-free rate and the date you took it. Name the equity risk premium source. Explain where beta came from and whether it was relevered. Then present the sensitivity, because the honest answer to a client or an interviewer is not a point estimate but a band with the drivers labeled. The same discipline applies to the discount rate you choose in the first place, which is why knowing when a valuation calls for cost of equity rather than WACC matters: unlevered free cash flow is discounted at WACC to reach enterprise value, while levered free cash flow or dividends are discounted at the cost of equity to reach equity value directly.

Cost of equity rewards a particular kind of candidate: one who is comfortable saying that a number is an estimate, and then defending the estimate anyway. The mechanics are two lines of arithmetic that anyone can memorise in an afternoon, so nobody is testing whether you can add a risk-free rate to a product. They are testing whether you know that the risk-free rate has a tenor, that the equity risk premium has a source and a convention attached to it, that beta is a regression with four dials on it, and that the answer you produce is a range with a midpoint rather than a fact. Get the arithmetic right, then get the judgment right, and the DCF questions that hang off it become considerably easier to handle.

Before the interview, build one full cost of equity for a company you know: pick the tenor, name the premium source, relever a comp-set beta, and write down the range three conventions produce. Being able to say "between 10 and 11.6 percent, and here is why" is the answer that ends the question.