Unlevering Beta When Debt Is Risky

Why Debt Beta Matters in Unlevering

Unlevering beta removes the effect of financing from an observed equity beta to estimate the systematic risk of the underlying assets. That calculation is straightforward only when the debt claim is treated as having zero systematic risk. If debt holders bear market-related risk, some of the business risk is carried by debt as well as by equity. Ignoring debt beta then understates the beta attributed to the asset claim in the relevant claims-weighted model.

The central calculation uses market values, not book values, and requires betas measured against the same market return and over a consistent period. It is not a claim that debt is riskless whenever it is called debt. A debt beta can be positive even when the debt has lower systematic risk than the equity.

Definition: The equity beta $\beta_E$ measures the covariance of equity returns with the market return, divided by the variance of the market return. The debt beta $\beta_D$ is defined the same way for debt returns. The asset beta $\beta_A$ is the beta of the asset claim or asset portfolio specified by the model. In the basic claims-weighted model, total market value $V$ is $D+E$, where $D$ and $E$ are market values of debt and equity.

The General Claims-Weighted Relationship

Suppose the market value of the firm's claims is the sum of its debt and equity market values, $V=D+E$. The return on the combined claims is the market-value-weighted average of the component returns. Because beta is linear in returns, the beta of that combined claim is the same weighted average of the component betas.

\beta_A=\frac{E}{D+E}\beta_E+\frac{D}{D+E}\beta_D =\frac{E\beta_E+D\beta_D}{D+E}.

This relationship uses the current market values of the claims. It assumes that the asset beta being estimated is the beta of the combined claims represented by those market values, and that the component returns and betas refer to a consistent horizon and market index. The identity does not require debt and equity returns to be independent.

Write the total market value.
Set $V=D+E$, using market values for both claims.

Weight each beta by its claim value.
Multiply $\beta_E$ by $E/V$ and $\beta_D$ by $D/V$.

Add the weighted contributions.
The result is $\beta_A$ for the combined claim under the stated model.

Compare with the zero-debt-beta case if useful.
Set $\beta_D=0$ only as an explicit approximation, not as an automatic property of debt.

Why the Formula Holds

Let $r_E$, $r_D$, and $r_A$ denote returns on equity, debt, and the combined claim over the same period. Since the combined claim is worth $E+D$, its return is the value-weighted return on its components:

r_A=\frac{E}{E+D}r_E+\frac{D}{E+D}r_D.

For any return $r$, beta is $\operatorname{Cov}(r,r_M)/\operatorname{Var}(r_M)$, where $r_M$ is the market return and $\operatorname{Var}(r_M)>0$. Taking covariance of both sides with $r_M$, dividing by the same market-return variance, and using the fact that covariance is linear gives:

\begin{aligned} \beta_A &=\frac{\operatorname{Cov}(r_A,r_M)}{\operatorname{Var}(r_M)}\\ &=\frac{E}{E+D}\frac{\operatorname{Cov}(r_E,r_M)}{\operatorname{Var}(r_M)} +\frac{D}{E+D}\frac{\operatorname{Cov}(r_D,r_M)}{\operatorname{Var}(r_M)}\\ &=\frac{E}{E+D}\beta_E+\frac{D}{E+D}\beta_D. \end{aligned}

Thus the weights are claim values, and the inputs are systematic-risk measures. Debt's total volatility or default probability cannot be substituted for $\beta_D$: beta captures only the component of return variation that covaries with the market.

Worked Example: Unlevering With Positive Debt Beta

A hypothetical firm has market-value debt of $240 million and market-value equity of $760 million. Its equity beta is 1.25 and its debt beta is 0.20. Estimate the beta of the combined asset claim, then compare it with the estimate that would result from setting debt beta to zero. Use the basic claims-weighted model, with no separate tax-shield adjustment.

Method and inputs. The total market value is $D+E=240+760=1{,}000$ million. The equity weight is $760/1{,}000=0.76$, and the debt weight is $240/1{,}000=0.24$.

Full estimate. The equity contribution is $0.76(1.25)=0.950$. The debt contribution is $0.24(0.20)=0.048$. Therefore:

\beta_A=0.950+0.048=0.998.

Zero-debt-beta approximation. If debt beta were set to zero, the estimate would be $0.76(1.25)=0.950$. The approximation is lower by $0.998-0.950=0.048$, which is also the debt weight times debt beta: $0.24(0.20)=0.048$.

Check and mistake avoided. The weights sum to $0.76+0.24=1$, and the weighted components reproduce the estimate. The common mistake is to omit debt's contribution merely because debt is generally less risky than equity. A positive debt beta contributes $0.048$ here, so the zero-beta approximation is not exact.

What the Zero-Debt-Beta Approximation Omits

If debt beta is set to zero while the same market-value weights are retained, the approximation is:

\beta_A^{(0)}=\frac{E}{D+E}\beta_E.

The difference between the full estimate and this approximation is:

\beta_A-\beta_A^{(0)}=\frac{D}{D+E}\beta_D.

For positive debt beta, the full claims-weighted estimate exceeds the zero-debt-beta estimate. If debt beta is negative, the direction reverses; if debt beta is exactly zero, the two results agree. The size of the difference depends on both debt's share of total market value and its beta. A small debt beta need not imply a negligible effect when debt represents a large fraction of the claims.

Worked Example: Using a Debt-to-Equity Ratio

A hypothetical business has a market debt-to-equity ratio $D/E=0.50$, equity beta 1.40, and debt beta 0.30. Estimate the combined asset beta and compare it with the zero-debt-beta approximation. Use the basic claims-weighted model.

Method. When $D/E=0.50$, set $E=1$ unit and $D=0.50$ unit. Then $D+E=1.50$, so the equity weight is $1/1.50=2/3$ and the debt weight is $0.50/1.50=1/3$.

Full estimate. The equity contribution is $(2/3)(1.40)=2.80/3=0.933333\ldots$. The debt contribution is $(1/3)(0.30)=0.10$. Therefore:

\beta_A=0.933333\ldots+0.10=1.033333\ldots

Zero-debt-beta comparison. The approximation is $0.933333\ldots$, so the omitted contribution is $1.033333\ldots-0.933333\ldots=0.10$. This equals the debt weight times debt beta, $(1/3)(0.30)=0.10$.

Check and mistake avoided. The ratio $D/E=0.50$ is not the debt weight. The debt weight is $D/(D+E)=0.50/1.50=1/3$. Using 0.50 directly as the weight would overstate debt's contribution. The calculated weights $2/3$ and $1/3$ sum to 1.

When a Tax-Adjusted Formula Is Appropriate

The basic claims-weighted identity describes the beta of the combined debt-and-equity claim. In a tax-deductible-debt framework, a common unlevering formula instead targets the beta of the operating assets before the financing tax shield. Under the conventional approximation that the tax shield's systematic risk tracks debt's risk, with a constant marginal tax rate $T$, the relationship is:

\beta_E=\beta_A+(1-T)\frac{D}{E}(\beta_A-\beta_D).

This formula is not obtained by simply replacing $D$ with $(1-T)D$ in every claims-weighted identity. It is a separate tax-adjusted relationship with a different asset-beta interpretation and additional assumptions. Rearranging it gives:

\begin{aligned} E\beta_E &=E\beta_A+(1-T)D\beta_A-(1-T)D\beta_D,\\ \beta_A &=\frac{E\beta_E+(1-T)D\beta_D}{E+(1-T)D}. \end{aligned}

The rearrangement is useful because it shows the debt-beta term explicitly. It also shows why the tax-adjusted zero-debt-beta approximation is $\frac{E\beta_E}{E+(1-T)D}$, not necessarily the basic-model approximation $\frac{E}{D+E}\beta_E$. Use a formula only after identifying which asset beta is intended and whether its assumptions fit the situation. If tax shields have materially different systematic risk from debt, the stated tax-adjusted relationship may not apply.

Worked Example: Applying the Tax-Adjusted Relationship

A hypothetical firm has market-value equity of $800 million and debt of $200 million. Its equity beta is 1.10, debt beta is 0.30, and marginal tax rate is 25%. Assume the tax-adjusted model just described applies. Estimate its operating-asset beta, compare it with the tax-adjusted zero-debt-beta estimate, and report the difference to three decimal places.

Method and assumptions. Use $\beta_A=[E\beta_E+(1-T)D\beta_D]/[E+(1-T)D]$. The calculation assumes a constant marginal tax rate and the specified debt-like risk treatment of the tax shield.

Full estimate. First, $(1-T)D=(1-0.25)(200)=0.75(200)=150$ million. The equity term is $E\beta_E=800(1.10)=880$. The debt term is $(1-T)D\beta_D=150(0.30)=45$. The denominator is $800+150=950$. Thus:

\beta_A=\frac{880+45}{950}=\frac{925}{950}=0.973684\ldots

Zero-debt-beta approximation. Setting $\beta_D=0$ leaves the same denominator and removes the debt term:

\beta_A^{(0)}=\frac{880}{950}=0.926315\ldots

Difference and check. Compute the difference from the unrounded values: $\frac{925}{950}-\frac{880}{950}=\frac{45}{950}=0.047368\ldots$, which rounds to 0.047 to three decimal places. The full estimate is higher by the debt contribution divided by the model's denominator. Subtracting the separately rounded estimates, $0.974-0.926=0.048$, gives the wrong rounded difference; round only after subtracting the unrounded estimates.

Interpretation and Exam Traps

Debt beta measures systematic risk borne by debtholders; it is not a synonym for the debt's yield, credit spread, default probability, or total volatility. A high promised interest rate does not by itself determine debt beta. In a beta calculation, use an estimated debt beta only when it is supplied or can be estimated under the stated model; do not infer it directly from the debt's interest rate.

The two formulas in this tutorial answer related but distinct questions. The basic claims-weighted equation applies to the beta of the combined market-value claims. The tax-adjusted equation is a conventional approximation for operating-asset beta under tax-shield assumptions. Mixing their numerators, denominators, or interpretations can yield a number that has no clear model behind it.

Using book values. The weights in these relationships are market-value weights unless a problem explicitly specifies a different model.
Treating all debt beta as zero. This is an assumption, not a definition. State it and quantify its effect when debt is risky.
Using $D/E$ as a debt weight. Convert the ratio to $D/(D+E)$ in the basic claims-weighted equation.
Comparing rounded estimates. Subtract the unrounded estimates first, then round the difference at the requested precision.
Ignoring the target definition. Decide whether the requested beta is for combined claims or operating assets under a tax-adjusted model before selecting a formula.

Key takeaway: In the general claims-weighted relationship, debt contributes $\frac{D}{D+E}\beta_D$ to the combined-claims beta. A zero-debt-beta approximation omits exactly this term. In tax-adjusted unlevering, use the corresponding formula only under its stated assumptions and for its intended operating-asset beta.

Check Your Understanding

For each question, identify the model before calculating, retain unrounded values until the final answer, and check the market-value weights.

A hypothetical company has market debt of $150 million, equity of $850 million, equity beta 1.20, and debt beta 0.10. Under the basic claims-weighted model, calculate asset beta and the zero-debt-beta approximation. (Answer: $\beta_A=1.035$; zero-beta approximation $=1.020$.)
A firm has $D/E=0.40$, equity beta 1.30, and debt beta 0.25. Under the basic model, calculate the claims-weighted asset beta. (Answer: $\beta_A=1.000$.)
A hypothetical business has market debt of $400 million and equity of $600 million, with equity beta 1.50 and debt beta 0.40. Under the basic model, calculate asset beta and the amount by which it exceeds the zero-debt-beta estimate. (Answer: $\beta_A=1.060$; difference $=0.160$.)
Under the tax-adjusted relationship, a firm has $E=900$, $D=100$, $T=0.20$, $\beta_E=1.00$, and $\beta_D=0.30$, with dollar values in the same units. Calculate operating-asset beta and the tax-adjusted zero-debt-beta estimate. (Answer: $\beta_A=0.9429$; zero-beta estimate $=0.9184$.)

What You'll Learn