Explain why beta measures only systematic risk and not total risk. Why is that the economically correct thing for CAPM to price?

What the interviewer is actually testing

Cost of equity and CAPM is a foundational DCF-building-block topic, and interviewers escalate through the same three layers as always: definitions, the "why," and dynamic application.

1. Can you state CAPM correctly and explain what each term represents in plain language? A weak candidate recites "risk-free rate plus beta times equity risk premium" without being able to explain what beta actually measures or why the risk-free rate uses a long-duration Treasury specifically. A strong candidate explains each term's economic meaning: Rf as the pure time-value-of-money floor, ERP as the market's price for bearing equity risk generally, beta as this specific stock's sensitivity to that market-wide risk.

2. Do you understand WHY beta measures only systematic risk, and why that's the correct thing to price? This is the conceptual heart of the topic and the most common follow-up: "why doesn't CAPM use a stock's total volatility instead of beta?" The strong answer invokes diversification directly: a diversified investor has already eliminated idiosyncratic risk for free, so the market doesn't compensate for bearing it, only for the systematic risk that diversification can't remove.

3. Can you handle the levered/unlevered beta mechanics and explain why you'd bother? Given a set of comps with different capital structures, can the candidate explain why you can't just average their raw betas, and walk through unlevering and re-levering? This tests whether the candidate understands leverage's effect on equity risk, not just the CAPM formula in isolation.

Signals of mastery: explaining the risk-free rate's duration-matching logic (long-term Treasury for long-duration cash flows) unprompted; giving the diversification argument for why beta (not total volatility) is priced; correctly stating the unlevered beta formula and explaining the intuition (more debt inflates levered beta for the same business risk); noting that ERP estimates vary by methodology and source. Red flags: being unable to explain what beta measures beyond "how volatile the stock is"; not knowing why systematic risk specifically is compensated; forgetting to unlever/re-lever beta when building a comp-based cost of equity.

Common traps

Trap 1: Using a short-term risk-free rate (e.g., a 3-month T-bill) instead of a long-duration Treasury yield. Equity cash flows being valued in a DCF stretch out many years (often into a terminal/perpetuity assumption), so the risk-free rate should be duration-matched to that long horizon. A short-term rate reflects short-term monetary policy conditions, not the long-run risk-free return relevant to a multi-year cash flow stream.

Say it out loud: "I'd use a long-duration Treasury yield (typically the 10-year) as the risk-free rate, because it should be duration-matched to the long-term nature of the cash flows I'm discounting, not a short-term rate that reflects near-term monetary policy."

Trap 2: Explaining beta as simply 'how volatile the stock is' rather than its sensitivity to market-wide moves specifically. Beta is not total volatility (standard deviation of returns): it's specifically the covariance with the market, scaled by the market's own variance. A highly volatile stock whose moves are largely idiosyncratic (uncorrelated with the market) can have a low beta despite high total volatility.

Say it out loud: "Beta isn't total volatility: it's specifically how sensitive the stock's returns are to market-wide moves. A stock can have high total volatility but a low beta if most of that volatility is idiosyncratic and uncorrelated with the broader market."

Trap 3: Not knowing why CAPM prices only systematic risk, and reasoning 'the company is risky, so beta should be high' without the diversification logic. This trap surfaces when candidates treat beta as a generic "riskiness score" rather than understanding the specific diversification argument that justifies pricing only systematic risk.

Say it out loud: "CAPM only prices systematic risk because a diversified investor has already eliminated company-specific risk through diversification: since that idiosyncratic risk can be removed for free, the market doesn't compensate investors for bearing it, only for the market-wide risk that can't be diversified away."

Trap 4: Directly averaging levered betas from a set of comps with different capital structures. Levered beta reflects both business risk and financial risk from leverage. Averaging levered betas from differently-levered peers blends together different amounts of financial risk, producing a beta that doesn't represent any specific, comparable level of risk.

Say it out loud: "I wouldn't average levered betas directly, since more debt mechanically inflates a company's levered beta for the same underlying business risk: I'd unlever each comp's beta to strip out capital structure, average the unlevered betas, and then re-lever using the subject company's own target capital structure."

Trap 5: Treating the equity risk premium as a fixed, universally agreed-upon number. ERP estimates vary meaningfully depending on methodology (historical vs. implied/forward-looking), time period used, and data source: candidates who quote a single ERP as an unambiguous fact rather than a range with methodology caveats look less sophisticated.

Say it out loud: "The equity risk premium isn't a single agreed-upon number: it depends on methodology, historical time period, and source, and typically falls somewhere in a 4.5%-6.5% range, though I'd want to know which convention a specific bank or textbook is using."

Trap 6: Using the subject company's own current (as-is) capital structure to relever beta, rather than its target/normalized capital structure. If the subject company's current leverage is unusually high or low relative to what it expects to maintain going forward (say, post a recent large debt paydown or a planned recapitalization), using its current D/E to relever beta produces a cost of equity that doesn't reflect its steady-state risk profile.

Say it out loud: "I'd relever the average unlevered beta using the subject company's target or normalized capital structure, not necessarily its current one, if the current leverage is temporarily unusual relative to where the company expects to operate going forward.