Every valuation model has a number that carries more weight than any other. Not the revenue forecast. Not the margin assumption. The discount rate — and specifically, the Weighted Average Cost of Capital that sits behind it. Get the WACC wrong and the entire DCF moves. A 1% error in WACC on a $50M EBITDA business can shift enterprise value by $30M or more. That is not a rounding issue. That is a credibility problem when someone challenges it in a data room.
This guide covers how a defensible WACC is actually constructed — component by component — and where the assumptions that most models rely on without question are the ones most likely to get challenged.
What makes a WACC defensible — and what doesn’t
A defensible WACC is not one that produces the valuation you want. It is one where every component can be explained, sourced, and challenged — and still holds up. That means the risk-free rate is not a number someone typed in from memory. The equity risk premium is not the first figure that appeared in a Google search. The beta is not pulled from a single data source without adjustment. And the capital structure is not the current one if it is expected to change.
Most models get parts of this right. Very few get all of it right simultaneously. The result is a WACC that looks precise — two decimal places, a tidy formula — but cannot survive fifteen minutes of scrutiny from a lender’s credit team or a counterparty’s advisor.
What we see consistently when reviewing models ahead of a financing or M&A process: the WACC was built once, early in the process, using whatever inputs were available at the time — and never revisited. By the time the model reaches a data room, the risk-free rate is six months stale, the beta is unlevered but applied as if it were levered, and the capital structure reflects the current balance sheet rather than the target structure the business will operate at post-transaction. None of these errors is obvious from the output. All of them are obvious the moment someone with transaction experience starts asking questions.
Component 1: The risk-free rate — which maturity, and why it matters
The risk-free rate is typically proxied by the yield on a government bond — in USD transactions, the US Treasury. The question is not which instrument to use. The question is which maturity.
The risk-free rate should match the duration of the cash flows being discounted. For a DCF with a five-year explicit forecast period and a terminal value, the relevant instrument is the 10-year Treasury — not the 3-month T-bill, not the 30-year bond. The 10-year is the market standard in practice for this reason: it approximates the weighted average duration of a typical discounted cash flow stream.
Where models go wrong: using a short-term rate because it is lower (which reduces WACC and inflates value), or using whatever rate was in the model template without checking whether it reflects current market conditions. A risk-free rate from 18 months ago in a model being reviewed today is not defensible.
The rate should be pulled from a current, published source — the US Treasury website, Bloomberg, or a reputable financial data provider — and the date of the pull documented in the assumptions tab.
Component 2: The equity risk premium — not a fixed number
The Equity Risk Premium (ERP) is the additional return investors require for holding equities over risk-free assets. It is also one of the most debated inputs in finance — and one of the most frequently mishandled in practice.
There are two primary approaches. The historical ERP — typically sourced from Damodaran’s annual dataset — represents the average excess return of equities over Treasuries over long historical periods. The implied ERP is derived from current market prices and consensus earnings expectations, effectively asking: what ERP does the market appear to be pricing in right now?
Neither is unambiguously correct. Historical ERP is stable but backward-looking; in periods of unusual market conditions it may not reflect current investor expectations. Implied ERP is current but volatile and model-dependent.
In practice, most transaction-context valuations use Damodaran’s implied ERP as the primary input — updated annually, widely cited, and defensible because it is a named, publicly available source. What is not defensible is an ERP of “5%” with no source, no date, and no explanation of how it was selected.
The ERP should be sourced, dated, and documented. If a counterparty’s advisor asks where the 5.2% came from, “that’s what we used” is not an answer.
Component 3: Beta — the most structurally misused input in WACC
Beta measures the sensitivity of a stock’s returns to market movements. In a DCF, it translates the company’s systematic risk into the cost of equity via the Capital Asset Pricing Model. It is also the input most frequently applied incorrectly — sometimes in ways that are not visible without knowing what to look for.
The core issue: beta as observed in the market is a levered beta. It reflects both the underlying business risk and the financial risk introduced by the company’s current capital structure. When building a WACC for a transaction — where the capital structure will change — the observed beta must be unlevered to remove the financial risk component, and then re-levered at the target capital structure.
What we see regularly in models that arrive for review: the beta has been pulled from a financial data source and applied directly as the cost of equity input — without unlevering and re-levering. In a business being acquired with significant leverage, this systematically understates the cost of equity. In a business being deleveraged post-acquisition, it overstates it. The direction of the error depends on the transaction, but the error is always present.
The correct sequence: collect levered betas from a set of comparable public companies, unlever each using the Hamada equation with each company’s own debt/equity ratio and tax rate, take the median of the unlevered peer betas, and re-lever at the target capital structure of the business being valued. This is not optional in a bank-grade model. It is the methodology.
A single-company beta pulled from Bloomberg and applied without adjustment is a shortcut. It may produce a number that looks reasonable. It will not survive scrutiny from anyone who knows what they are looking at.
Component 4: Capital structure — target, not current
WACC weights the cost of equity and cost of debt by their proportions in the capital structure. The question is: which capital structure?
For an ongoing business with a stable leverage profile, the current market-value capital structure is a reasonable starting point. For a business in a transaction — being acquired, refinanced, or restructured — the current capital structure is the wrong input. What matters is the capital structure the business will operate at on a normalized basis going forward.
But here is what most build guides don’t address: even for stable businesses, book-value weights are incorrect. Debt should be weighted at market value — which for most investment-grade debt is close to par — and equity should be weighted at market capitalization, not book equity. Using book value of equity as the weight in a WACC calculation systematically distorts the result, particularly for asset-light businesses where book equity is a poor proxy for market value.
Document which capital structure is being used and why. In a transaction context, the rationale should be explicit: “Target capital structure post-close, reflecting the pro forma leverage and equity contribution in the sources and uses.”
Component 5: Cost of debt — one step most models skip
The cost of debt is typically estimated as the yield on the company’s existing debt, or benchmarked against comparable credit instruments. What is often missed: the relevant input is the after-tax cost of debt, not the pre-tax rate.
Interest expense is tax-deductible — which means the effective cost of debt to the business is reduced by the tax shield. The formula is straightforward: after-tax cost of debt equals the pre-tax rate multiplied by one minus the marginal tax rate.
The tax rate used here should be the marginal corporate tax rate, not the effective rate from the income statement. Effective tax rates reflect past tax positions, one-time items, and deferred tax movements. The marginal rate reflects what the business will actually pay on an incremental dollar of taxable income going forward — which is the correct basis for discounting future cash flows.
The sensitivity that every WACC table needs
A single WACC output is a point estimate. Point estimates are not how investment decisions get made — they are the center of a range, and the range matters as much as the center.
Every WACC should be accompanied by a sensitivity table that shows how enterprise value moves across a realistic range of WACC inputs — typically plus or minus 100 to 150 basis points around the base case. This table does two things: it shows the decision-maker where the valuation is most sensitive, and it demonstrates that the analyst understands the model is a framework for thinking, not a machine for producing a single answer.
If you are working on a transaction or planning initiative that requires a WACC built to withstand external scrutiny — let’s talk through what that looks like.
FAQ: Building a Defensible WACC
What is the most common WACC mistake in financial models? Applying a levered beta directly without unlevering and re-levering at the target capital structure. This error systematically misstates the cost of equity whenever the company’s leverage differs from its peers or changes in a transaction.
Should I use historical or implied equity risk premium in my WACC? Implied ERP — updated annually from Damodaran’s dataset — is the standard in transaction-context valuations. It reflects current market conditions and is a named, citable source. Historical ERP is acceptable but should be documented with a clear rationale.
What risk-free rate should I use for a DCF valuation? The 10-year US Treasury yield, pulled from a current published source on the date of the analysis. Short-term rates understate the duration of the cash flows being discounted and produce an indefensible result under review.
How do I make my WACC defensible in a data room? Every component — risk-free rate, ERP, beta, capital structure, cost of debt — must be sourced, dated, and documented in the assumptions tab. A WACC where any input cannot be traced to a named source within thirty seconds will not survive a counterparty review.