Why an observed beta has to be unlevered before you can use it
Beta measures how much a stock's return moves with the market, and it picks up two sources of variation at once. The first is business risk: how cyclical the products are, how much of the cost base is fixed, how sensitive demand is to the economy. The second is financial risk: interest is a fixed claim that gets paid before shareholders, so borrowing amplifies whatever the assets do. Two companies running identical operations will show different equity betas purely because one of them borrowed more.
That makes a raw published beta almost useless as a comparable. If you are valuing a private company, a division, or a business you intend to recapitalise, the beta you can observe belongs to somebody else's balance sheet. Unlevering strips the borrowing out and leaves the asset beta — the risk of the business itself, which is genuinely transferable between companies in the same industry. Relevering then adds back the leverage you intend to carry.
This is the standard route to a bottom-up beta: take five to ten comparable companies, unlever each one at its own debt-to-equity ratio and tax rate, average the asset betas, then relever the average at your target structure. Averaging asset betas rather than equity betas is the point of the exercise, because averaging equity betas averages other people's financing decisions along with the business risk. The result feeds straight into the CAPM cost of equity and from there into WACC.
The Hamada equation, term by term
Robert Hamada derived the link in 1972 by combining the capital asset pricing model with the Modigliani–Miller propositions on capital structure. The result says the equity beta is the asset beta scaled up by a leverage factor:
βL = βU × (1 + (1 − t)·D/E)
Read the factor from the inside out. D/E is the market value of debt over the market value of equity — market value, not book, because beta is a market-based measure and book equity can be negative or meaningless after years of buybacks. (1 − t) shrinks the leverage effect because the tax shield on interest is a partial offset: the government absorbs t of every interest dollar, so debt transfers slightly less risk to shareholders than its face amount suggests. Add 1 because an all-equity firm has D/E = 0 and its equity beta simply is its asset beta.
Invert it to unlever: βU = βL ÷ (1 + (1 − t)·D/E). Because the factor is always at least 1 whenever D/E and t are non-negative, the asset beta is never further from zero than the equity beta. That is the whole intuition — leverage magnifies, and unlevering undoes the magnification.
Two assumptions worth knowing you are making
Hamada's derivation assumes the debt beta is zero: lenders bear no market risk. That holds well for investment-grade debt and badly for distressed debt, where lenders effectively own the business. It also assumes a fixed dollar amount of debt whose tax shield is discounted at the cost of debt.
If a company instead rebalances to a constant ratio of debt to value — which is how most large firms behave and how most DCF models are built — the tax shield is as risky as the assets, the tax term drops out, and you get the Harris–Pringle form: βU = βL ÷ (1 + D/E). That is the second option in the calculator. At a D/E of 0.6 and a 25% tax rate the Hamada divisor is 1.45 against the Harris–Pringle 1.60, so an equity beta of 1.20 unlevers to 0.8276 one way and 0.7500 the other — a gap of 0.078, small next to the standard error on the underlying regression.
Worked example: unlevering a comparable at 1.20 and relevering to 0.35
You are valuing a private specialty manufacturer that intends to run at a debt-to-equity ratio of 0.35. Your closest listed comparable has a published beta of 1.20, market debt of $600 million against market equity of $1,000 million (so D/E = 0.60), and a marginal tax rate of 25%. Your own marginal rate is 21%, the long government bond yields 4.3%, and you are using an equity risk premium of 5.0%.
- Build the comparable's leverage factor. 1 + (1 − 0.25) × 0.60 = 1 + 0.75 × 0.60 = 1.45.
- Unlever. βU = 1.20 ÷ 1.45 = 0.8276. The comparable's business risk, with its borrowing removed, is below the market average even though its equity beta is above it.
- Build your target leverage factor. 1 + (1 − 0.21) × 0.35 = 1 + 0.79 × 0.35 = 1.2765.
- Relever. βL* = 0.8276 × 1.2765 = 1.0564.
- Cost of equity. 4.3% + 1.0564 × 5.0% = 4.3% + 5.282% = 9.582%.
Check the direction against intuition. Your relevered beta of 1.0564 is below the comparable's 1.20 because you plan to carry less debt — 0.35 against 0.60 — and less debt means shareholders absorb less amplification. Had you targeted 1.00 instead, the factor would be 1 + 0.79 = 1.79 and the relevered beta 0.8276 × 1.79 = 1.4814, lifting the cost of equity to 4.3% + 7.407% = 11.707%. That 2.1-point swing on the cost of equity comes entirely from the financing decision, with the business untouched.
Run the same comparable through the Harris–Pringle form for contrast: βU = 1.20 ÷ 1.60 = 0.7500, and relevering at 1.35 gives 1.0125. Slightly lower than the Hamada answer, and the gap widens as leverage rises.
How to read the asset beta you get
An asset beta is a statement about the business, so test it against the business. Regulated utilities and consumer staples cluster well below 1.0 because demand barely moves with the cycle. Capital goods, semiconductors, airlines and construction sit above 1.0 because both demand and operating leverage swing hard. If you unlever an industrial comparable and get 0.35, or unlever a water utility and get 1.5, the input beta is more likely wrong than the industry.
Judge the relevered beta by the change from the comparable, and let the direction come from the leverage comparison rather than from habit. If your target debt-to-equity exceeds the comparable's, the relevered beta comes out above the comparable's equity beta; if it is lower, the relevered beta comes out below. Equal factors give you the comparable's beta back — which is a useful sanity check that you entered the ratios in the same units, since a D/E of 60% and a D/E of 0.60 are the same number and 60 is not.
The regression behind the input beta deserves as much scrutiny as the algebra. A beta estimated over two years of weekly returns against a narrow local index is a different animal from five years of monthly returns against a broad one. Published betas are often adjusted toward 1.0 using a Blume-style shrinkage, and if yours already is, you are unlevering an adjusted figure. Standard errors on single-stock betas are large, which is precisely the argument for averaging asset betas across several comparables instead of trusting one.
Do not stop at the beta. The cost of equity it produces has to be plausible next to the company's borrowing cost: equity is a residual claim, so the cost of equity must exceed the pre-tax cost of debt from the after-tax cost of debt calculator. If it does not, one of the two rates is wrong.
How the leverage factor moves with debt-to-equity and tax rate
| Debt-to-equity | t = 0% | t = 21% | t = 25% | t = 35% |
|---|---|---|---|---|
| 0.00 | 1.0000 | 1.0000 | 1.0000 | 1.0000 |
| 0.25 | 1.2500 | 1.1975 | 1.1875 | 1.1625 |
| 0.35 | 1.3500 | 1.2765 | 1.2625 | 1.2275 |
| 0.50 | 1.5000 | 1.3950 | 1.3750 | 1.3250 |
| 0.60 | 1.6000 | 1.4740 | 1.4500 | 1.3900 |
| 1.00 | 2.0000 | 1.7900 | 1.7500 | 1.6500 |
| 1.50 | 2.5000 | 2.1850 | 2.1250 | 1.9750 |
| 2.00 | 3.0000 | 2.5800 | 2.5000 | 2.3000 |
| 3.00 | 4.0000 | 3.3700 | 3.2500 | 2.9500 |
The 0.60 row at t = 25% gives 1.4500 and the 0.35 row at t = 21% gives 1.2765 — the two factors used in the worked example above.
Debt-to-equity, not debt-to-capital
The Hamada factor takes D/E. A target often arrives stated as a debt-to-capital weight instead, because that is the form WACC uses. Convert before you enter it: a 30% debt weight means D/V = 0.30 and E/V = 0.70, so D/E = 0.30 ÷ 0.70 = 0.4286, not 0.30. Entering 0.30 there understates leverage and hands you a beta that is too low. The debt-to-equity ratio calculator does the conversion if you have the balances rather than the weights.
Pitfalls that produce a wrong beta
- Using book values for D/E. Beta is a market measure. Use the market capitalisation for equity and market value — or amortised cost as a proxy — for debt.
- Mixing a ratio with a percentage. A D/E of 0.60 and a D/E of 60% are the same thing; entering 60 where a ratio belongs turns a 1.45 leverage factor into 46 at a 25% tax rate. Use the unit switch on the field.
- Confusing debt-to-equity with debt-to-capital. A 30% debt weight is a D/E of 0.4286. This is the single most common arithmetic error in a bottom-up beta.
- Averaging equity betas instead of asset betas. Unlever each comparable at its own leverage and tax rate first, then average. Averaging equity betas averages other firms' financing choices.
- Applying the zero-debt-beta assumption to distressed credit. Once lenders bear business risk, unlevering pushes too much risk into the equity beta and the asset beta comes out too high.
- Relevering with the comparable's tax rate. Unlever with the comparable's marginal rate, relever with yours. Two different rates, and the calculator keeps them separate for that reason.
- Ignoring excess cash. A cash-heavy comparable has a diluted asset beta, because cash carries a beta near zero. Some practitioners use net debt or make a separate cash adjustment; say which you did.
Where this sits among the ways to get a beta
There are three routes to a beta and they are not equally reliable. A regression beta from the company's own stock returns is direct but noisy, unavailable for private firms, and reflects the historical capital structure rather than the intended one. A bottom-up beta — the method on this page — trades a single noisy estimate for the average of several, which cuts the standard error and lets you specify the leverage. An accounting or fundamental beta built from earnings variability is a last resort when no listed comparable exists.
Two refinements are worth knowing. The Fernandez and Miles–Ezzell formulations add a debt beta or a one-period rebalancing assumption, and they matter when leverage is high enough that lenders share business risk. Adjusting for operating leverage matters when your fixed-cost structure differs materially from your comparables', since a business with a higher fixed-cost share has a higher asset beta at identical revenue volatility — the mechanism the degree of operating leverage calculator quantifies.
Whatever you produce, remember it is one input among three in a cost of equity. The risk-free rate and the equity risk premium carry their own uncertainty, and in most valuations the premium assumption moves the answer more than the beta does. Test the whole build-up for sensitivity before you defend a discount rate to three decimal places, and if you are valuing outside your home market, add sovereign risk explicitly through the country risk premium calculator rather than burying it in beta.
Key terms
- Levered (equity) beta
- The beta you observe from a stock's returns. It reflects business risk and the amplification added by that company's borrowing.
- Unlevered (asset) beta
- Beta with the effect of financial leverage removed — the systematic risk of the operating assets alone. Comparable across firms in the same industry.
- Hamada equation
- The 1972 result linking levered and unlevered beta through the factor 1 + (1 − t)·D/E, derived from the CAPM combined with Modigliani–Miller.
- Bottom-up beta
- A beta built by unlevering several comparable companies, averaging their asset betas, and relevering at the target capital structure.
- Debt beta
- The systematic risk borne by lenders. Assumed zero in the Hamada form, which is reasonable for investment-grade debt and poor for distressed debt.
- Harris–Pringle relevering
- The no-tax form βL = βU(1 + D/E), appropriate when a firm rebalances continuously to a constant debt ratio so the tax shield carries asset risk.