If you run a private company or advise one, knowing how to calculate cost of capital is the difference between setting a realistic hurdle rate and silently destroying value. The shortest answer: build a weighted average cost of capital (WACC) using market-value weights for equity, debt, and any preferred stakes, then discount each component’s required return (cost of equity via CAPM or build-up, after-tax cost of debt, stated yield on preferred). For a private firm, the work isn’t the algebra—it’s estimating inputs like beta and market risk premium without a traded stock. Below I walk through a live worked example, the pitfalls I’ve hit, and a template you can steal.
What the Cost of Capital Actually Measures (and Why Private Owners Trip Up)
The cost of capital is the blended return your investors—equity holders, lenders, preferred shareholders—require to stay invested. It’s the minimum rate you should earn on new projects before destroying shareholder value. Economically, it is an opportunity cost: the return those funds would generate elsewhere at similar risk.
When I first valued a $12M revenue family-owned packaging business in 2018, I pulled equity from the balance sheet and got a WACC of 7%. That number was dangerously low because book equity ignored retained earnings market value and the real risk premium. The loan officer laughed at our model, and we lost the refinancing terms we wanted.
Most people don’t realize that for private firms, the weights matter more than the formula. Using book values for a company that has appreciated real estate on its books can underweight equity by 30–50%, artificially lowering WACC. The thing nobody tells you: your cost of capital is not a single static number—it shifts with leverage, dividend policy, and even owner psychology around risk.
In practice, finding the cost of capital means answering three sub-questions: What are the market values of each capital piece? What return does each provider require? What is the tax shield on deductible payments? If you can answer those, the calculation is straightforward.
The Two Questions Users Actually Ask
“How do you find the cost of capital?” and “How to calculate the capital cost?” are the same query phrased differently. You find it by identifying each capital component, estimating its required return, and weighting by market value. The calculation is mechanical once inputs are credible. Skip straight to the Cost of Capital Calculator if you already have those inputs; the rest of this article shows how to derive them for non-public entities.
The Core Formula: WACC and the CAPM Cost of Equity
The standard WACC expression is: WACC = (E/V × Re) + (D/V × Rd × (1−Tc)) + (P/V × Rp), where E, D, P are market values of equity, debt, preferred; V = E+D+P; Re, Rd, Rp their costs; Tc marginal tax rate. This is the answer to how to calculate the capital cost for a blended stack.
For the equity piece, the formula for CAPM cost of capital is: Re = Rf + β × (Rm − Rf). Here Rf is the risk-free rate (typically a long-term government bond yield), β measures systematic risk, and (Rm − Rf) is the market risk premium (MRP). If you are wondering “what is the formula for CAPM cost of capital?”—that single line is it, but applying it privately requires beta estimation tricks covered later.
An alternative equity model is the dividend discount model (Gordon growth), but for private firms with irregular dividends it is unreliable. CAPM remains the audit-standard despite its flaws. The constant-growth model only works post-stability.
When CAPM Fails and Build-Up Wins
For micro-cap private firms, CAPM’s beta is unobservable. Practitioners then use the build-up method: Re = Rf + MRP + size premium + industry premium + company-specific risk. I’ve used both; CAPM with a levered beta from comparables is more defensible to auditors, while build-up is faster for seed-stage startups. Neither is a silver bullet—the size premium alone can add 2–4% that public-company models ignore.
The trade-off: CAPM assumes diversified investors and a single market factor. A family owner holding 80% of shares is not diversified, so their required return may exceed CAPM. In those cases I add a 1–2% key-person or concentration premium and disclose it.
Step-by-Step WACC Calculation for a Private Company (Worked Example)
Let’s use a real-shape example from a 2022 advisory engagement: a light-industrial SME with no public ticker. Market-value estimates: equity $2.4M (owner’s stake plus retained earnings reinvestment value derived from a reverse DCF), debt $1.0M (bank term loan face), preferred $300k (issued to an angel). Total V = $3.7M.
Step 1: Weighting. E/V = 64.9%, D/V = 27.0%, P/V = 8.1%. Using book values (equity $1.1M, debt $1.0M, pref $300k) would have given 45.8% equity weight—a textbook mistake that understates cost because equity is riskier.
Step 2: Cost of debt. The loan carries 6.2% coupon. After the 21% federal rate per the IRS, Rd×(1−Tc) = 6.2% × 0.79 = 4.90%. Use marginal, not effective, rate. If the firm operates in a state with 5% tax, marginal becomes 25.2%, shield slightly higher.
Step 3: Cost of preferred. The angel’s term sheet specifies 8% cumulative dividend, non-participating. Rp = 8.0%—no tax shield, so it enters gross. If preferred is convertible, see later section.
Step 4: Cost of equity via CAPM. Rf from 10-year U.S. Treasury yield = 4.1%. We selected five public comparables in industrial machinery with average levered beta 1.15. Unlevering: βu = 1.15 / (1 + (1−0.21)×0.4) ≈ 0.93. Re-levering to our D/E of 0.42: βl = 0.93 × (1 + 0.79×0.42) ≈ 1.22. MRP assumed 5.5% (historical long-run per NYU Stern data). Re = 4.1% + 1.22×5.5% = 10.81%.
Step 5: Blend. WACC = (0.649×10.81%) + (0.270×4.90%) + (0.081×8.0%) = 7.01% + 1.32% + 0.65% = 8.98%. That’s the hurdle rate. If you want to skip the math, the Cost of Capital Calculator reproduces this in seconds.
Downloadable Template and Input Sheet
I’ve packaged the above into a one-tab spreadsheet: it forces market-value cells, auto-applies the tax shield, and flags if book weights exceed market by >10%. The template also includes a comparator tab for beta derivation. Use it before your next board meeting to avoid the book-value trap I fell into years ago.
Estimating Inputs When You Have No Public Market Data
The content gap most articles ignore is private-firm input estimation. You can’t pull beta from Yahoo Finance. Here’s the field approach I use:
- Risk-free rate: Always use the 10-year government bond of your operating currency. For USD, the U.S. Treasury site publishes daily yields. Don’t use 3-month T-bills; maturity mismatch biases WACC downward by 1–2%.
- Beta: Identify 3–5 public comparables with similar leverage and margins. Unlever each (βu = βl / [1 + (1−Tc)×D/E]). Average βu, then re-lever to your capital structure. In our example, comparables’ βl ranged 0.9–1.4, showing dispersion.
- Market risk premium: Use 4.5%–6.0% depending on source; I default to 5.5% for U.S. small caps but add a size premium of 2.0%–3.5% from Duff & Phelps data (acknowledge uncertainty—these are estimates, not facts).
- Company-specific risk: For family firms with key-person dependency, add 1%–2%. Document the rationale in a memo.
The thing nobody tells you: a 0.2 beta error changes WACC by ~1.1% for a 65% equity-weighted firm. That swings project NPV by millions over a 10-year horizon. Treat input estimation as the real work, not an afterthought.
In one engagement, a client insisted their beta was 0.5 because “we are stable.” Comparables said 1.1. We compromised at 0.9 with a written note; later a downturn proved the higher beta right. Be conservative.
Startups and SMEs Without Positive Earnings
If your client has no earnings history, CAPM beta is meaningless. I switch to a venture capital method: estimate required return via target IRR of comparable exits (often 25%–40%). That’s not WACC but a cost of equity proxy. Be honest about the limitation—conventional WACC is inappropriate pre-scale, and using it can approve ruinous projects.
Handling Preferred, Convertible, and Retained Earnings
Preferred equity is a hybrid. In WACC it sits between debt and equity. Use its stated dividend yield as Rp; if it’s participating, adjust for expected conversion value. Convertible notes complicate: split into debt and option components via binomial trees if material. For a $300k preferred with 8% dividend, we used straight Rp = 8%.
Retained earnings are not a separate cost component. They are part of equity (E) and carry the same Re as contributed capital. The misconception that retained earnings are “free” because no dividend is paid destroys mid-market models—opportunity cost applies.
Most people don’t realize that cumulative preferred dividends in arrears increase effective cost if the firm plans to cure them; discount that stream at the preferred rate but flag liquidity risk. In a 2020 distress case, we added 150 bps to Rp because arrears were two years deep.
The Convertible Note Edge Case
When preferred or debt is convertible, you cannot just use the coupon. I allocate value: $1M note with 5% coupon, 20% conversion premium. Using a Black-Scholes proxy, $120k of value sits in the equity option. The remaining $880k is debt at 5%×(1−Tc). This raises WACC slightly versus treating it all as debt. Ignore this and you understate cost by 30–50 bps.
Common Mistakes That Skew Your Cost of Capital
From auditing 40+ private models, these recur:
- Book vs market weights: Using carrying value of equity overstates debt ratio, understates WACC. Always mark equity to estimated enterprise value less debt.
- Ignoring taxes: Forgetting the (1−Tc) shield on debt. At 21% rate, that’s a 1.3% absolute drop on a 6% loan—material.
- Mixing maturities: Pairing 10-year Rf with 3-month commercial paper Rd creates mismatch.
- Double-counting risk: Adding size premium in both beta comparables and build-up.
- Using marginal vs statutory tax: If state taxes apply, marginal can be 25%–28%; using only federal understates shield.
- Old comparables: Using 2010 betas in 2024 rate environment; refresh quarterly.
Rule of thumb: if your private-company WACC comes out below 8% in today’s rate environment, re-check every input. It’s usually a weighting or tax error.
A Practical Checklist & Input Matrix for Private WACC
Below is the decision matrix I hand clients. It forces a deliberate choice on each input source.
- Equity value source: Recent transaction? Appraisal? DCF reverse? (Choose one, document)
- Debt value: Face value if floating; discount if distressed.
- Preferred: Liquidation preference or stated value?
- Rf: 10-yr govt yield date: ______.
- Beta method: Comparables unlever/re-lever / Build-up / None (VC).
- MRP + size: Base MRP ____ + size ____ = total equity risk premium ____.
- Tax rate: Federal + state marginal ____.
- Cross-check: Run Cost of Capital Calculator for sanity.
This framework is the information gain competitors miss: it turns “how to calculate cost of capital” from a textbook into an audit-ready process. Pair it with the calculator for live sensitivity.
Sensitivity Analysis: Stress-Testing Your Private WACC
A single point estimate is dangerous. I run a tornado: vary beta ±0.2, MRP ±1%, tax ±3%. In our example, WACC ranges 7.9%–10.1%. That band should frame capital budgeting, not the 8.98% midpoint alone.
One-Way and Two-Way Tornado
One-way: shift each input alone. Two-way: beta and debt weight together, because leverage changes beta. Most models skip this; the thing nobody tells you is that WACC is convex in leverage—too much debt raises beta and can increase WACC despite shield.
Marginal Tax Rate Trap and Jurisdictional Nuances
The tax term (1−Tc) looks simple but hides traps. In the U.S., federal 21% plus state 4–8% yields 24–28% marginal. For pass-through S-corps, the shield may sit at owner level, not entity—so WACC should reflect owner’s personal rate on interest. I’ve seen models apply corporate shield to an S-corp incorrectly, overstating value by 5%.
Cross-border private firms must match tax jurisdiction to deductibility. A Dutch holding with U.S. ops needs layered rates. Always cite the IRS for federal and local authority for state.
Validating Your WACC with Reverse DCF
A neat trick: infer implied WACC from a recent transaction multiple. If a similar firm sold at 8x EBITDA, back out the WACC that equates present value to enterprise value. This cross-check caught a 150 bps error in our 2022 case. It’s not perfect but anchors expectations.
When to Use a Different Method (Trade-offs)
WACC is right for stable cash flows. For a pre-revenue startup, use staged VC targeting. For a highly leveraged buyout, consider APV (adjusted present value) to separate tax shields from operating risk. I’ve seen PE firms reject WACC because debt schedule changes yearly—APV gave cleaner optics.
Also, if you operate cross-border, currency mismatches in Rf and cash flows require a parallel currency WACC. The honest limitation: no single formula captures private-firm illiquidity perfectly. You approximate, disclose, and stress-test.
Putting It All Together
Calculating cost of capital for a private business is 20% math, 80% defensible input hunting. Start with market values, estimate equity cost via CAPM with comparable betas, shield your debt, and respect preferred. Avoid book-weight laziness. The worked example here (8.98% WACC) is a template you can adapt. Download the sheet, fill the matrix, and you’ll have a number that survives scrutiny.
If you remember one thing: the question “how do you find the cost of capital?” is answered not by a calculator but by rigorous, documented input work. Do that, and your capital decisions will finally rest on solid ground.