Valuation multiples are the financial world's shorthand. You glance at a stock's P/E ratio and you have — or think you have — an instant read on whether it's cheap or expensive. The problem is that reading a multiple in isolation is roughly like reading a benchmark result without knowing the hardware, the workload, or what "good" even means for this use case. The number exists, it's technically correct, and it tells you almost nothing on its own.
This guide is a literacy primer, not a stock-picking system. The goal is to make you fluent enough in these metrics to know what questions to ask next.
P/E: The Multiple Everyone Knows
Price-to-earnings is calculated simply: take the share price and divide it by earnings per share. If a stock trades at $100 and earned $5 per share in the last twelve months, its trailing P/E is 20.
The intuitive interpretation is that you're paying 20 years' worth of current earnings to own a share of the business — assuming earnings stay flat forever, which they never do. That assumption is the first crack in the foundation. Markets price assets on expected future earnings, not the past, which is why forward P/E (using analyst estimates for the next twelve months) often matters more in practice.
Trailing P/E uses reported earnings. It's audited, backward-looking, and not subject to analyst optimism bias. Forward P/E uses consensus estimates. It's more relevant for valuation but inherently uncertain — analyst forecasts have a well-documented tendency to be too rosy, especially in cyclical sectors or when conditions are changing fast.
A few things P/E can't handle cleanly:
- Companies with no earnings. A pre-profit SaaS company or a biotech burning cash has no denominator. The ratio is undefined or negative, which gives you no comparison point.
- One-off items. GAAP earnings include restructuring charges, asset write-downs, and legal settlements. A single large charge can make a healthy business look far cheaper than it is, or a mediocre one look artificially cheap in a prior year.
- Accounting choices. Revenue recognition timing, depreciation methods, and capitalization decisions all affect the "E" in P/E. Two companies with identical economics can report materially different earnings.
EV/EBITDA: Comparing Across Capital Structures
Enterprise value divided by EBITDA (earnings before interest, taxes, depreciation, and amortization) is the analyst's preferred tool when comparing companies with different amounts of debt or cash.
Enterprise value is market cap plus net debt (total debt minus cash and equivalents). It represents what you'd theoretically pay to acquire the entire business — equity holders and debt holders together — after taking the cash on the balance sheet. If Company A has a $1B market cap and no debt, and Company B has a $1B market cap but $500M in debt, they look equivalent on P/E but Company B costs $500M more to actually own outright. EV captures that difference; P/E doesn't.
EBITDA backs out interest expense (so capital structure doesn't affect it), taxes (which vary by jurisdiction and structure), depreciation, and amortization. The resulting number is a rough approximation of operating cash generation before capital allocation decisions.
The critical caveat: EBITDA is not free cash flow. Depreciation is backed out, but the physical assets that depreciate still need to be replaced. A capital-intensive manufacturer with heavy equipment — call it illustratively 20% of revenue in annual capex — has very different economics from a software company with 2% capex, even if EBITDA margins look similar. Charlie Munger called EBITDA "bullshit earnings" for exactly this reason: it ignores the real cost of maintaining the business.
EBITDA also excludes working capital changes. A business that's rapidly scaling receivables (i.e., recognizing revenue before collecting cash) looks healthier on EBITDA than its cash flow statement suggests.
When EV/EBITDA is most useful: comparing telecom, cable, or industrial companies where capital structures vary widely and depreciation schedules differ significantly. When to be skeptical: software companies that capitalize engineering labor and show inflated EBITDA, or businesses using adjusted EBITDA to exclude items that are actually recurring.
A related metric worth knowing: EBIT/EV (the inverse of EV/EBIT) is used as an earnings yield in rank-based screening approaches. A higher yield means you're getting more operating profit per dollar of enterprise value. This kind of yield metric is the basis for certain quantitative screening frameworks — useful for building a universe of candidates, not as a standalone buy signal.
P/S and P/B: When Earnings Aren't Available
Price-to-sales is the go-to metric when earnings are absent or highly distorted. A pre-profit company with $100M in revenue trading at a $500M market cap has a P/S of 5x. P/S is harder to manipulate than earnings, but it says nothing about profitability or the path to it. A company with 90% gross margins deserves a higher P/S than one with 30% margins, all else equal.
Price-to-book compares market cap to the accounting net asset value (assets minus liabilities on the balance sheet). Historically useful for banks and asset-heavy industrials where tangible assets are the primary value driver. For software, consumer brands, or anything where competitive advantage lives in intangibles (code, brand equity, network effects), book value is nearly meaningless — those assets are largely absent from the balance sheet.
A Multiple Is a Question, Not an Answer
The core mistake is treating a multiple as a conclusion. "The stock trades at 12x earnings, so it's cheap" is a complete sentence that contains very little information.
Context that actually matters:
- The company's own history. Is 12x cheap relative to where this company has traded over five or ten years? Or is it compressing from 30x because fundamentals are deteriorating?
- Peers. Industry-wide valuation levels shift with interest rates, competitive dynamics, and sentiment. A "low" multiple in a sector that has repriced downward may still be fully valued on absolute terms.
- Growth rate. A rough rule: divide P/E by the expected long-term earnings growth rate to get the PEG ratio. A 30x P/E company growing earnings at 30% per year and a 10x company growing at 5% are very different propositions. Neither is obviously better without more work.
- Quality and durability. Two companies with identical EV/EBITDA can have wildly different outcomes if one has pricing power and recurring revenue and the other has neither.
A low multiple is not evidence of value — it may be evidence of a problem the market has already priced in. Cyclical companies often show their lowest P/E ratios near earnings peaks (the denominator is temporarily inflated), just before a downturn. Buying a low multiple in a deteriorating business is called a value trap. The multiple tells you the price; it doesn't tell you what the earnings are worth or how long they'll last.
What Multiples Cannot Tell You
Even when read correctly in context, multiples have structural blind spots.
Off-balance-sheet obligations. Operating lease commitments, pension liabilities, and contingent liabilities may not appear in the headline debt figure used to calculate enterprise value. Two companies can look similar on EV/EBITDA while carrying very different actual obligations.
Cyclicality. Commodity producers, semiconductor equipment makers, and shipping companies swing violently through earnings cycles. A low P/E at a cyclical peak is not a signal; it's a warning.
Accounting-driven distortions. Revenue recognition differences between companies in the same sector can make P/S comparisons misleading. Stock-based compensation — a real economic cost — is excluded from EBITDA and sometimes from "adjusted" earnings that show up in screens.
Business durability. No multiple tells you whether the competitive position is defensible, whether management is allocating capital well, or whether the industry is being disrupted. A 7x EV/EBITDA business with a durable moat may be far more valuable than a 4x business facing structural decline. The multiple captures the market's current assessment; it doesn't make that assessment for you.
Multiples are a starting point for questions. They narrow a universe of thousands of securities to a set worth examining more closely. They're not a replacement for understanding the actual business.
This article is educational. Nothing here is a recommendation to buy or sell any security. All illustrative figures are hypothetical examples only.
Originally published at pickuma.com. Subscribe to the RSS or follow @pickuma.bsky.social for new reviews.
Top comments (0)