The Capital Asset Pricing Model, beta, cost of equity, cost of debt, and how they blend into a single discount rate — the quantitative backbone behind every DCF.
Building the Discount Rate From the Ground Up
Every DCF hinges on one number that never appears on a financial statement: the discount rate. This guide builds it piece by piece — starting with what an investor requires for taking on risk, and ending with the single blended rate IU uses to discount cash flow.
Five modules — work through them in order or jump to the one you need.
"An investor doesn't hold a volatile stock out of charity. CAPM is the formula for exactly how much extra return that volatility has to promise before it's worth the risk."
You already do this intuitively every time you compare two trades. A volatile small-cap runner has to offer you a much bigger potential payoff than a sleepy utility stock before you'll touch it — otherwise, why take the extra risk? CAPM just turns that gut instinct into a number. It answers one question: given how much this specific stock swings around compared to the overall market, what annual return should an investor demand to make holding it worthwhile? That "how much it swings" measurement has a name: beta (β). It's simply how much a stock tends to move for every 1% move in the broader market.
Rf is the risk-free rate (typically the 10-year Treasury yield). (Rm − Rf) is the equity risk premium (ERP) — the extra return investors historically demand for holding stocks instead of risk-free bonds, generally 4.5–6%. Beta scales that premium up or down for the specific stock.
β = 1.0 moves in line with the market. β > 1.0 (growth, cyclicals, high-leverage names) amplifies market moves and demands a higher return. β < 1.0 (utilities, staples) dampens market moves and commands a lower return.
Risk-free rate 4.2%, equity risk premium 5.5%, and a semiconductor equipment name with β = 1.35 (above-market volatility, typical of cyclical capex-driven businesses): Ke = 4.2% + 1.35 × 5.5% = 4.2% + 7.4% = 11.6%. That's the annual return an equity investor should require to hold this specific stock.
Beta should be calculated by statistically comparing a stock's historical price moves to a broad market index over several years, not pulled from a single data provider's snapshot. It should also be re-levered for the company's actual capital structure (the mix of debt and equity it uses to fund itself) if compared across peers with different debt loads.
Next: CAPM's output is the cost of equity — Module 02 covers how it's actually applied inside a valuation.
Next Module →"Debt sends you an interest bill every quarter. Equity's cost never appears on an invoice — which is exactly why it's the cost most often ignored."
When you buy shares in a company instead of putting that money into an index fund or a savings account, you're implicitly demanding a return for giving up those safer options. Cost of equity (Ke) is just that demand made explicit: the return shareholders require to justify owning the business. CAPM (Module 01) is the standard way to estimate it. The tricky part is that, unlike an interest payment on a loan, this cost never shows up as a line item anywhere. It's a bar the company has to clear to actually be creating value for you. But nothing forces management to show you whether they're clearing it.
A company can report a growing net income while still destroying shareholder value, if the return on that equity capital falls short of what CAPM says investors require. There's no line item that flags this — it only shows up when you compare Return on Equity (ROE) or Return on Invested Capital (ROIC) against the cost of equity directly.
For smaller or thinly-traded companies, analysts often add a size premium (1–3%) on top of CAPM, since small-cap equity has historically demanded extra compensation beyond what beta alone captures. Company-specific risk premiums for concentration or key-person risk are sometimes layered on top, but should be used sparingly and disclosed explicitly.
Two companies both report 15% ROE. Company A has a CAPM-derived cost of equity of 9% — it's creating roughly 6 points of excess return for shareholders every year. Company B, a smaller and more volatile name with β = 1.8 and a size premium, has a cost of equity of 15.5% — its "identical" 15% ROE is actually value-destructive on a risk-adjusted basis. Same reported number, opposite conclusion.
Never evaluate ROE or ROIC in isolation. The only meaningful question is whether it clears the cost of the capital that produced it. That threshold is different for every company.
Next: equity is only half the capital structure. Cost of debt is the other, and it comes with a tax-driven twist.
Next Module →"Interest payments are tax-deductible. That single fact is why debt is almost always the lowest-cost capital a company can raise — up to the point where too much of it becomes dangerous."
Cost of debt is the easy half of this equation. It's basically just the interest rate a company pays its lenders — a number you can look up rather than estimate. But there's a wrinkle that matters: interest payments reduce a company's taxable income, so the government effectively subsidizes part of that interest bill. What actually belongs in a WACC calculation isn't the sticker-price interest rate. It's the after-tax version — what the debt really costs the company once that tax break is factored in.
The pre-tax cost of debt is typically estimated from the yield-to-maturity on the company's existing bonds, or from its credit rating mapped to a corporate bond spread if it doesn't have actively traded debt. The tax shield then reduces the effective cost.
The tax shield assumes the company is consistently profitable enough to use the deduction. For a company with volatile or negative earnings, the realizable tax benefit is smaller than the formula implies — that assumption should be reduced in the model, not treated as a footnote.
A BBB-rated industrial issuer has outstanding bonds yielding 6.0% and faces a 24% effective tax rate. After-tax cost of debt = 6.0% × (1 − 0.24) = 4.56%, meaningfully below its ~11% cost of equity. That's exactly why debt is the cheaper source of capital — and why capital structure mix matters so much to the blended rate.
Cheaper debt doesn't mean "more debt is better." Beyond a certain leverage point, lenders start demanding higher interest rates to compensate for the added risk, and shareholders start demanding a higher return to compensate for a rising risk of bankruptcy. So WACC eventually turns and rises again — it's U-shaped, not something that keeps falling the more debt a company takes on.
Next: with both capital costs in hand, the final step is weighting them by how the company is actually financed.
Next Module →"Two companies can have identical costs of equity and debt and still land on very different WACCs — because they finance themselves differently. The blend is the whole exercise."
Think of a company as being funded by two different groups of people it has to pay back over time: shareholders (Module 02's cost of equity) and lenders (Module 03's after-tax cost of debt). WACC is just the blended, single number that represents what it costs the company to keep both groups happy. It's averaged by how much of the company's funding actually comes from each side. That mix is measured using what the company is worth in the market today — share price × shares outstanding, plus the market value of its debt — not the older numbers sitting on its balance sheet.
E is the market value of equity (share price × shares outstanding), D is the market value of debt, and V = E + D is total capital. The two weights (E/V and D/V) always sum to 100% — this is a blend, not an addition.
| Component | Value | Weight | Weighted Cost |
|---|---|---|---|
| Equity (Ke = 11.6%) | $18.0B market cap | 82% | 9.5% |
| Debt (Kd after-tax = 4.6%) | $4.0B market debt | 18% | 0.8% |
| WACC | 10.3% | ||
IU deep dives compute WACC using beta calculated from several years of historical price data (a "trailing" beta), re-levered to the company's current capital structure, plus a Treasury-based risk-free rate as of the report date and a market-value debt weight. The DCF output then gets stress-tested against a ±1.5-percentage-point WACC range, since (as Module 01 of the Valuation & DCF guide shows) small WACC changes move fair value materially more than most other inputs.
WACC is the single number where every prior module converges. Beta, the equity risk premium, credit spreads, tax rate, and capital structure all feed into it — treat it as a range to stress-test, never as a precise input to trust blindly.
Next: CAPM and WACC are the industry standard, but neither is beyond criticism — Module 05 covers the adjustments and limitations worth knowing before you trust the number blindly.
Next Module →"Every input in this guide has been a clean, single number. In practice, beta is noisy, capital structures change, and academics have spent forty years finding what CAPM leaves out. None of that makes it useless — it makes it a starting point that needs adjustment."
Here's a wrinkle most beginners never hear about: the beta number you look up for a stock isn't a pure measure of how risky its business is. It's tangled up with how much debt that specific company happens to carry right now. More debt mechanically makes a stock swing harder, inflating beta regardless of how risky the underlying business actually is. So if you want to compare two companies' riskiness fairly — or borrow a similar company's beta to value one that isn't public — you first strip the debt effect out to see the "pure" business risk underneath. That step is called unlevering. Then you add back in whatever debt level you actually want to model, a step called relevering.
Unlevering strips out a company's specific financing decisions to isolate its pure business risk. Relevering at a target Debt-to-Equity ratio (D/E) then lets you build a beta for a private company, a new segment, or your own assumed capital structure — using a peer set as the raw material.
A peer trades with a levered β of 1.50 at a D/E of 0.40 and a 24% tax rate: βunlevered = 1.50 / [1 + 0.76 × 0.40] = 1.15. Relevered to your target company's D/E of 0.20 (carrying full precision through the calculation): βrelevered = 1.15 × [1 + 0.76 × 0.20] = 1.33 — a materially different, and more appropriate, beta than either the raw peer number or a guess.
CAPM was built and tested using data from large, easy-to-trade stocks — think mega-caps, not the small-cap names that make up a lot of retail trading activity. Decades of market history show that small and thinly-traded stocks have actually earned higher returns than a plain-vanilla CAPM number would predict. The likely reason: investors demand extra compensation for the risk of being stuck in a name that's hard to exit quickly. The fix is a build-up method: start with the CAPM number, then stack additional premiums on top for what beta alone misses.
| Component | Typical Range |
|---|---|
| CAPM (Rf + β × ERP) | Base rate |
| Size premium (small-cap) | +1% to +3% |
| Illiquidity / marketability discount (private or thin float) | +2% to +5% |
| Company-specific risk (key-person, concentration) | +0% to +2%, used sparingly |
Even after those two adjustments, it's worth knowing that beta alone has never told the full story. Researchers Fama and French noticed decades ago that small companies and "cheap" stocks (ones trading low relative to what they're worth on paper) kept beating what a single beta number said they should return. It was as if the market were pricing in risks beta just doesn't capture. Their response was to add two more factors on top of beta: one for company size, one for cheapness. IU doesn't run that full academic model on every report, but the lesson underneath it matters for any retail trader: a single beta is a simplification. It tends to understate the required return on small, cheap, unloved stocks — exactly the names where getting the cost of equity wrong does the most damage to a valuation.
None of this means discard CAPM. It means treat its output as a starting estimate that needs adjustment, not a precise, final number. A small-cap industrial with a thin float deserves a build-up analysis, not a bare CAPM figure pulled from a terminal. The discipline is knowing which adjustments apply, not skipping them because the base formula is easier.
For small-cap and micro-cap names — a recurring segment of IU coverage — cost of equity is built up from CAPM plus an explicit, disclosed size premium rather than reported as a bare CAPM number. Peer betas are always unlevered and relevered to the target company's actual capital structure before being used, never applied raw.
With a fully adjusted discount rate in hand, the next step is putting it to work — the Valuation & DCF guide covers exactly that.
Continue to Valuation & DCF →