7.4 Market-Based Valuation: Price and Enterprise Value Multiples

7.4 Market-Based Valuation: Price and Enterprise Value Multiples

Core Insight: Equity valuation combines relative market comparisons with absolute economic accounting models. Market multiples evaluate whether an asset is cheap or expensive relative to comparable peers or fundamental drivers, while Residual Income (RI) models recognize that accounting book value creates real economic value only when a firm generates returns in excess of its cost of equity capital.

---

1. Market-Based Valuation Multiples Framework

Relative valuation relies on two distinct methodologies:

1. Method of Comparables: Comparing a firm's market multiple to a benchmark group of peer firms or historical averages. Stocks trading at multiples below peers are considered undervalued if fundamentals are identical.
2. Method Based on Forecasted Fundamentals: Deriving justified price multiples directly from discounted cash flow equations (Gordon Growth Model).

Justified Price Multiples Derived from Fundamentals

By substituting Gordon Growth relationships ($P_0 = \frac{D_1}{r - g} = \frac{E_1(1 - b)}{r - g} = \frac{E_0(1 - b)(1 + g)}{r - g}$), we derive justified multiples that reflect intrinsic value:

$\text{Justified Forward P/E } \left(\frac{P_0}{E_1}\right) = \frac{1 - b}{r - g}$

$\text{Justified Trailing P/E } \left(\frac{P_0}{E_0}\right) = \frac{(1 - b)(1 + g)}{r - g}$

$\text{Justified Price-to-Book (P/B) } \left(\frac{P_0}{B_0}\right) = \frac{\text{ROE} - g}{r - g}$

$\text{Justified Price-to-Sales (P/S) } \left(\frac{P_0}{S_0}\right) = \frac{\text{Net Profit Margin}_0 \times (1 - b)(1 + g)}{r - g}$

Where:

• $(1 - b)$ = Dividend payout ratio.
• $b$ = Retention rate ($1 - \text{payout ratio}$).
• $r$ = Required rate of return on equity.
• $g = b \times \text{ROE}$ = Sustainable growth rate.

Core Principle of Justified P/B: If a company's expected $\text{ROE} = r$, its justified $P/B = 1.0$. If $\text{ROE} > r$, the firm earns economic rents and its justified $P/B > 1.0$. If $\text{ROE} < r$, the firm destroys shareholder value and its justified $P/B < 1.0$.

2. Enterprise Value (EV) Multiples

Enterprise Value represents the total market value of a company's core operating business, independent of capital structure:

$\text{Enterprise Value (EV)} = \text{Market Value of Common Equity} + \text{Market Value of Preferred Stock} + \text{Market Value of Debt} - \text{Cash and Cash Equivalents}$

Why EV/EBITDA is Preferred over P/E in Capital-Intensive Sectors

1. Capital Structure Neutrality: EV and EBITDA are both pre-interest metrics. Comparing two identical operating businesses with different leverage ratios distorts P/E, whereas EV/EBITDA remains neutral.
2. Accounting Method Neutrality: Differences in depreciation and amortization methods (straight-line vs. MACRS, differing asset useful lives) distort Net Income and P/E, but EBITDA is unaffected.
3. Handling Negative Earnings: Firms in cyclical troughs often report negative Net Income (rendering P/E meaningless), but maintain positive EBITDA.

3. Justified Valuation Multiples Summary Table

4. Alternative Price Multiples, Normalized Earnings & Predicted P/E

The two approaches to using any multiple

The two are complementary. A stock trading at a P/E of 11 against a peer median of 16 looks cheap by comparables; if its justified leading P/E from fundamentals is 10, it is not.

The alternative price multiples and when each is used

Earnings yield (E/P) inverts the P/E. It is used precisely because it remains meaningful when earnings are zero or negative — a stock with an E/P of −0.02 is comparable across a screen, whereas its P/E of −50 is not. Ranking by earnings yield is therefore the standard screening convention.

Normalizing earnings per share

A cyclical company's trailing P/E is systematically misleading: at the trough, earnings collapse and the P/E looks enormous; at the peak, earnings are inflated and the P/E looks cheap. This inverse pattern is the Molodovsky effect. Two normalization methods:

1. Method of historical average EPS. Average the company's EPS over a full cycle, typically five to ten years. Simple, but it ignores changes in the size of the business over that period.
2. Method of average return on equity. Multiply the average ROE across the cycle by the current book value per share. This adjusts for growth in the equity base and is generally preferred when the company has grown materially.

Worked example. A cyclical industrial has current book value per share of 40.00 and reported EPS over five years of 5.20, 1.10, −0.40, 3.60, and 6.10 on book values of 26, 29, 28, 33, and 40. Its current share price is 62.00.

• Historical average EPS: $(5.20 + 1.10 - 0.40 + 3.60 + 6.10)/5 = 3.12$; normalized P/E $= 62.00/3.12 = \textbf{19.9}$
• Average ROE: annual ROEs are 20.0%, 3.8%, −1.4%, 10.9%, 15.3%, averaging 9.70%; normalized EPS $= 0.0970 \times 40.00 = 3.88$; normalized P/E $= 62.00/3.88 = \textbf{16.0}$

The trailing P/E on the peak year's 6.10 is only 10.2 — which would flag the stock as cheap at exactly the wrong point in the cycle. The average-ROE method gives the higher normalized earnings because it reflects the larger equity base the company now operates.

Underlying earnings (also called persistent, continuing, or core earnings) strip non-recurring items — restructuring charges, asset sale gains, litigation settlements — from reported earnings. This is a different adjustment from cycle normalization and the two are often applied together.

Predicted P/E from a cross-sectional regression

Instead of assuming a functional form, an analyst can regress observed P/E ratios across a group of companies on fundamental drivers:

$\widehat{P/E} = b_0 + b_1(\text{Dividend payout}) + b_2(\text{Expected growth}) + b_3(\text{Beta})$

Substituting the subject company's own fundamentals produces its predicted P/E, which is then compared with its actual P/E. The gap is the estimated over- or under-valuation.

Its limitations are precisely those developed in Quantitative Methods and are frequently the point of the item set:

• the relation is captured within one sample and one period, so it is unstable out of sample;
• multicollinearity among the fundamental drivers inflates the standard errors of the coefficients;
• the estimated relation does not persist across time or across markets;
• the regression is fitted to whatever mispricing existed in the sample, so if the whole sector was overvalued the predicted P/E is overvalued too.

The PEG ratio

$PEG = \frac{P/E}{\text{Expected earnings growth rate in percentage points}}$

A stock with a P/E of 24 and expected growth of 20% has a PEG of 1.2. The convention is that lower is cheaper, and PEG scales P/E so that companies with different growth rates can be ranked against one another.

Its weaknesses are examinable:

• it assumes a linear relation between P/E and growth, whereas the justified P/E is non-linear in growth;
• it ignores risk, so a high-growth, high-beta company looks identical to a low-risk one at the same PEG;
• it ignores the duration of growth — a company growing 20% for two years and one growing 20% for ten years have the same PEG.

Using multiples for terminal value

A multiple can replace the Gordon growth formula at the end of a multistage discounted cash flow model:

$TV_n = \text{Terminal-year metric} \times \text{Exit multiple}$

The requirement is that the multiple be a mature-company multiple appropriate to the terminal year, not the subject company's current multiple. Section 7.7 develops the choice between the two terminal value methods and the cross-check between them.