Valuation

What Happens to WACC as You Add More Debt?

Article5 min read
In this article
  1. Start with the formula
  2. The part people miss: equity gets riskier
  3. Without taxes, they cancel
  4. Taxes are what actually make debt cheap
  5. What it’s missing is that debt stops being safe
  6. Answering it in an interview
  7. Takeaways

As debt rises, WACC falls, and then it rises. Most candidates say “it goes down, because debt is cheaper than equity” and stop there. That answer is half of one of the three things that happen, and an interviewer who wants to find the ceiling of your knowledge will ask the follow-up.

Adding debt does three things to WACC:

  • It shifts weight toward the cheaper component. Pushes WACC down.
  • It makes the remaining equity riskier, so beta and the cost of equity rise. Pushes WACC up.
  • Past a point it makes the debt itself riskier, so the cost of debt rises too. Pushes WACC up.

The first two very nearly cancel1. What tips the balance is the tax shield, and what eventually reverses it is distress.

Start with the formula

EE is the market value of equity. DD is the market value of debt. VV is the two added together, the total capital in the business. KeK_e is the cost of equity, KdK_d the cost of debt, and tt the tax rate.

WACC=EV×Ke+DV×Kd×(1−t)\text{WACC} = \dfrac{E}{V} \times K_e + \dfrac{D}{V} \times K_d \times (1 - t)

Adding debt raises D/VD/V and lowers E/VE/V. If KeK_e and KdK_d stayed fixed, WACC would fall mechanically, because KdK_d after tax is smaller than KeK_e. That is the normal answer.

The problem is that KeK_e does not stay fixed.

The part people miss: equity gets riskier

Debt has a prior claim on cash flows. Lenders get paid first, in good years and bad. Everything left over belongs to equity, and “everything left over” is more important as an idea when there is a fixed payment ahead of it.

That shows up in beta. Levered beta is the unlevered business risk geared up for the capital structure:

βL=βU×[1+(1−t)×DE]\beta_L = \beta_U \times \left[ 1 + (1 - t) \times \dfrac{D}{E} \right]
Ke=rf+βL×Equity Risk PremiumK_e = r_f + \beta_L \times \text{Equity Risk Premium}

So as D/ED/E rises, beta rises, and the cost of equity rises with it, in a straight line.

Without taxes, they cancel

Take a business with an unlevered beta of 1.0, a 4% risk-free rate and a 5.5% equity risk premium. Assume the debt is riskless, so it costs 4% too, which is what makes the relevering formula above valid here. No taxes.

D / VD / ELevered betaCost of equityWACC
0%0.001.009.5%9.5%
25%0.331.3311.3%9.5%
50%1.002.0015.0%9.5%
75%3.004.0026.0%9.5%

Look at the last column. WACC does not move. At 75% debt the cost of equity has gone from 9.5% to 26%, and the weighted average is unchanged.

That is Modigliani-Miller (the idea that enterprise value is independent of capital structure in a frictionless environment). No one we’ve met in finance has brought up this theorem by name, but pretty much everyone knows it in practice. Capital structure does not create value on its own. Every dollar of “cheap” debt you add is paid for by the higher return equity now demands.

Taxes are what actually make debt cheap

Now run the same company at a 25% tax rate. Interest is deductible, so the government funds part of the interest bill.

D / VLevered betaCost of equityWACC
0%1.009.5%9.5%
25%1.2510.9%8.9%
50%1.7513.6%8.3%
75%3.2521.9%7.7%

Now WACC falls. Note where the benefit comes from: not from debt being “cheaper,” but from the deductibility of interest. That is the whole gap between the two tables.

Notice the beta formula already incorporates this. The (1−t)(1 - t) term in the relevering equation is why levered beta rises more slowly here than it did in the no-tax case: 3.25 instead of 4.00 at 75% debt.

With these assumptions, WACC falls forever and the optimal capital structure is 100% debt, which is obviously wrong. The model is missing something.

What it’s missing is that debt stops being safe

Both tables assumed the cost of debt stays at 4% no matter how much you borrow. Real lenders don’t behave that way. Ratings fall, spreads widen, covenants tighten, and at some point all lenders refuse to lend.

Hold everything else constant and let the cost of debt rise with leverage:

D / VCost of debtCost of equityWACC
0%4.0%9.5%9.5%
25%5.0%10.9%9.1%
50%7.0%13.6%9.4%
75%12.0%21.9%12.2%

WACC falls, bottoms out somewhere around 25% debt in this example, then climbs steeply. That U-shape is the real answer, and the bottom of the U is what people mean by an optimal capital structure.

Three things drive the right-hand side of the curve:

  • Lenders reprice. Higher leverage means a worse rating and a wider spread.
  • The tax shield stops working. A company with too much interest relative to earnings cannot use the whole deduction, and interest deductibility is capped in the US anyway2.
  • Distress has its own costs. Customers and suppliers get nervous, management spends time on the balance sheet instead of the business, and investment gets deferred.

Answering it in an interview

The fast version: as debt rises, WACC falls because you are weighting toward a cheaper, tax-deductible source.

Only add detail when prompted: after a certain point, more debt increases WACC because equity gets riskier and eventually the debt does too.

If they ask why equity gets riskier, go to beta: debt has a prior claim, so levered equity carries more risk, and you relever beta by [1+(1−t)×D/E]\left[1 + (1-t) \times D/E\right].

Takeaways

WACC falls then rises as you add debt. Three forces: a bigger weight on cheaper debt pushes it down; a rising levered beta and cost of equity push it up; a rising cost of debt pushes it up harder once leverage gets serious. The minimum of the resulting U-shaped curve is the optimal capital structure.

Footnotes

  1. Almost, but not exactly, and how close depends on which relevering formula you use. The one used in interviews and in most models, βL=βU×[1+(1−t)×D/E]\beta_L = \beta_U \times [1 + (1-t) \times D/E], assumes debt carries no market risk. Price the debt above the risk-free rate while using it and the two effects stop cancelling: WACC drifts upward with leverage even at a zero tax rate, because equity is charged for all of the business risk while lenders are separately paid a spread for absorbing part of it. The same risk gets billed twice. Damodaran gives the general version, which nets the debt’s own beta back out: βL=βU×[1+(1−t)×D/E]−βD×(1−t)×D/E\beta_L = \beta_U \times [1 + (1-t) \times D/E] - \beta_D \times (1-t) \times D/E, or equivalently βL=βU+(βU−βD)×(1−t)×D/E\beta_L = \beta_U + (\beta_U - \beta_D) \times (1-t) \times D/E. Run the table below through that, with no taxes and a debt beta of (Kd−rf)/ERP(K_d - r_f) / \text{ERP}, and WACC is flat at 9.5% for any cost of debt, not just the risk-free one. Neither formula is wrong; the simple one is only valid when debt is close to riskless, which is why the no-tax table prices it at 4%. One caution if you do estimate a debt beta: a credit spread also pays for expected default losses and illiquidity, and neither of those earns a beta, so dividing the whole spread by the equity risk premium overstates it. See Estimating Risk Parameters, footnote 6, which runs across pages 26 and 27. ↩
  2. Section 163(j) caps the deduction at 30% of adjusted taxable income. Disallowed interest carries forward indefinitely, so for most companies this is a timing cost rather than a lost shield. It only becomes a real loss if the capacity never comes back. ↩

Ready to practice?

Put the concept to work on real interview questions.

Browse Questions