Why Comparing Companies Across Industries Using P/E Ratios Misleads Investors

The price-to-earnings ratio gets quoted more than any other valuation metric, and it gets misused more than any other too. People treat it like a price tag. Twelve times earnings is cheap, forty times is expensive, decision made. I've sat through plenty of investment discussions where someone dismissed a stock in one sentence because its multiple looked high next to a company in a completely different industry, and that comparison almost never means anything.

The core problem is that industries have fundamentally different economics. A software company with 80% gross margins, minimal capital requirements, and a long growth runway will trade at a very different multiple than a regulated utility that grows a few percent a year and pours most of its cash back into infrastructure. Comparing their P/E ratios is like comparing a motorcycle's speed to a cargo ship's. Both numbers describe movement, but the contexts are so different that the comparison tells you nothing useful.

Why Multiples Differ by Industry

Pull up any sector-level valuation table and you'll see the same pattern year after year. Software and other high-growth sectors sit near the top of the P/E range, while banks, utilities, and energy sit near the bottom. If you look at that table and conclude that tech is overvalued and energy is a bargain, you've made the classic mistake. The gap persists for structural reasons, and understanding them is the first step to using multiples properly.

Growth rates. A company compounding earnings at 25% a year doubles them roughly every three years. At 5% growth, doubling takes about fourteen years. The market pays a higher multiple for the first company because a much larger share of its value sits in future earnings rather than current ones.

Capital intensity. Utilities, airlines, and telecom carriers have to keep shoveling earnings back into infrastructure just to stay in place. A software company can add revenue without a proportional increase in capital spending, so more of each dollar it earns is actually available to shareholders. Investors pay more per dollar of earnings when they get to keep more of those dollars.

Margin profiles. A business running 40% operating margins can absorb a serious revenue decline and stay profitable. A business at 5% margins has almost no room for error. That kind of durability earns a premium multiple, and the market is usually right to grant it.

Cyclicality. Autos, construction, commodities, semiconductors. Earnings in these industries swing with the economy, so investors apply a discount for the uncertainty. There's a nasty wrinkle here that trips up even experienced investors, and I'll come back to it below, because cyclical stocks tend to look cheapest on P/E at exactly the wrong moment.

Regulation. Regulated utilities and banks have earnings that are partly set by someone else. Rate caps and capital requirements limit the upside but also put a floor under profitability. The market prices that bounded range with a moderate, stable multiple.

The Cross-Industry Comparison Trap

Here's how this goes wrong in practice. Say a software company trades at 40x earnings and a bank trades at 10x. The instinct is to call the bank four times cheaper.

Now add the context. Suppose the software company is growing revenue 30% a year, runs 85% gross margins, has customers who expand their spending over time, and needs almost no incremental capital to scale. The bank grows at 3%, holds regulatory capital that caps its returns, competes on price in what is essentially a commodity product, and absorbs losses every credit cycle. Discount the future cash flows of both businesses and the software company at 40x can genuinely be the better value, while the bank at 10x might be fairly priced or even expensive against its real earnings power.

The E in P/E causes its own trouble. Reported earnings carry the accounting baggage of the industry they come from: depreciation schedules, stock-based compensation, loan loss provisions, one-time charges. Two companies can report the same earnings with wildly different cash economics underneath. This is why it's worth opening the actual 10-K on EDGAR instead of trusting a screener. The cash flow statement and the footnotes on unusual or non-recurring items will tell you how much of the E is repeatable cash and how much is accounting.

How to Build a Peer Group That Works

P/E ratios become genuinely useful once you compare companies in the same industry, with similar business models, at similar stages of growth. Analysts call this comparable company analysis, or comps. A defensible peer group shares:

• The same industry and sub-industry
• Similar size, by revenue or market cap
• Comparable growth rates
• Similar margin profiles
• Comparable capital structures

A practical shortcut for finding peers: open the company's 10-K on EDGAR and read the competition section, where management names its own rivals. Then pull the proxy statement and see which companies the compensation committee benchmarks executive pay against. Between those two lists you usually have a solid peer set without guessing, and you've used the company's own disclosures to build it.

Within a well-built group, an outlier multiple is a real signal. If five similar enterprise software companies trade between 30x and 40x and a sixth sits at 20x, that gap deserves investigation. Maybe the market is pricing in a problem you haven't found yet, like customer concentration or a product transition going badly. Maybe the market is wrong and you've found an opportunity. Either way you now have a specific question to answer instead of a vague hunch.

Better Tools for Cross-Industry Comparisons

Sometimes you genuinely need to compare across industries, say when allocating across sectors or screening a broad universe. Use metrics built for the job.

EV/EBITDA. Enterprise value over EBITDA strips out differences in capital structure and tax treatment, so a leveraged industrial and a debt-free software company sit on more even footing. Sector differences don't disappear, but the distortion shrinks. Be careful with capital-heavy businesses though, since EBITDA ignores the capex they can't avoid paying.

PEG ratio. The P/E divided by the expected earnings growth rate. It's a blunt tool that leans entirely on a growth forecast, but it at least forces you to ask cheap relative to what growth. The convention treats a PEG near 1.0 as roughly fair value. Treat that as shorthand, not a law of nature, and remember the forecast is doing all the work.

Free cash flow yield. Free cash flow divided by enterprise value. Cash is much harder to dress up than accounting earnings, so this measure cuts through depreciation policies, stock compensation, and the other non-cash noise that varies across industries.

Return on invested capital. ROIC measures how efficiently a business turns capital into profit, which makes it a quality question rather than a price question. It's one of the few cross-sector comparisons I trust, because it reveals which business models are structurally better at compounding capital regardless of what the stock costs today.

The Cyclical P/E Trap

In cyclical industries, the P/E ratio moves backwards from what intuition expects. At the top of the cycle, earnings are temporarily inflated, so the ratio looks low right when the stock is most dangerous to buy. At the bottom, earnings are crushed or negative, so the ratio looks huge, or meaningless, right when the stock may be closest to a bargain.

Walk through a hypothetical. Say a homebuilder earns $10 a share at the peak of a housing boom and the stock trades at $80. That's 8x earnings, which screens as a bargain. Then the cycle turns. Earnings fall to $2, the stock drops to $40, and the same screen now shows 20x. The stock got cheaper while the multiple got bigger, and anyone who bought at the cheap 8x is sitting on a 50% loss.

For businesses like this, analysts use normalized or mid-cycle earnings, meaning an estimate of what the company earns averaged across a full cycle, instead of the trailing twelve months. If you don't want to build that estimate yourself, at least average the last several years of earnings before trusting the multiple.

A Few Mechanical Traps While You're at It

Beyond the industry problem, watch for the mechanics of the ratio itself.

Trailing versus forward. Trailing P/E uses the last twelve months of reported earnings. Forward P/E uses analyst estimates for the next twelve. Screeners and news articles mix the two freely, and for a fast-growing or cyclical company they can be very different numbers. Know which one you're looking at before you compare anything.

One-time items. A big asset sale, a legal settlement, or a write-down can inflate or crater a single year of earnings and make the ratio meaningless. The income statement footnotes in the 10-K flag these. Strip them out before you lean on the multiple.

Negative or tiny earnings. A company earning a few cents a share can show an enormous P/E while being perfectly healthy, and a company losing money shows no P/E at all. In both cases the metric simply doesn't apply, and you should reach for revenue multiples or cash flow instead.

How I Actually Use P/E

My rule is that a P/E only means something after I can answer three questions. What do this company's true peers trade at? How does its growth compare to those peers? And how much of the reported E converts into cash? If someone tells me a stock is cheap at 12x, my first response is compared to what. If the answer is the overall market, there's more digging to do. If the answer is a company in a different industry, the comparison hasn't told us anything yet.

Valuation is always relative. Twenty times earnings can be expensive for a slow-growing utility and cheap for a fast-compounding software business, and both can be true on the same day. Get the peer group right and the ratio starts carrying real information. Get it wrong and you're just comparing two numbers that happen to share a name.