The P/E Ratio Is Not Enough: Six Numbers to Read With It

The P/E ratio misleads on its own because it compresses an entire business, its growth, its balance sheet, its cash flows and its position in the cycle, into a single division problem, and almost everything useful is lost in the compression. Two companies can trade at exactly the same multiple and be completely different propositions, which means the number by itself tells you almost nothing about which one you are actually looking at.

None of this makes the P/E useless. It is a fine opening question. The mistake is treating it as a verdict. The fix is not a better single number, because no single number exists. The fix is a small set of companions that restore what the compression destroyed. There are six of them, and each one answers a question the P/E cannot.

Growth: the number that gives P/E its meaning

A P/E of 30 is not expensive and a P/E of 10 is not cheap. Neither statement means anything until you know how fast earnings are growing, because a multiple is a price paid for future earnings, and the future is what growth describes. This is the old PEG intuition: divide the P/E by the growth rate and suddenly the multiple has a denominator that speaks. A company at 30 times earnings compounding profits at 25 percent a year is being valued very differently from a company at 30 times earnings growing at 5 percent, even though the headline number is identical.

The concrete version. Take two consumer companies, both at 40 times trailing earnings. One has grown earnings per share at 18 percent a year over 2019-2024. The other has grown at 4 percent and the multiple is high because the stock is popular, not because the business is compounding. On P/E alone they look like twins. Add growth and one of them is paying forward for a machine that keeps building earnings under the price, while the other is paying the same toll for a business that is standing still.

Return on capital: the same P/E is not the same price

Two companies at the same P/E with different returns on capital employed are not trading at the same price in any meaningful sense. ROCE tells you how much profit the business generates per rupee it has tied up. A high-ROCE business can grow without constantly asking shareholders or lenders for more money, which means more of each year’s earnings is genuinely free. A low-ROCE business has to reinvest heavily just to stand still, so the E in its P/E is partly spoken for before you ever see it.

Consider two manufacturers, both at 20 times earnings. One earns a 35 percent ROCE, funds its growth from internal cash and still has money left over. The other earns 9 percent, below what a lender would charge it, and every rupee of expansion has to be financed. The first company’s earnings are worth more per unit because they arrive with optionality attached. The second company’s earnings arrive with a bill. The P/E screen puts them in the same bucket. The ROCE column takes them straight back out.

Cash conversion: is the E even real?

Accounting earnings are an opinion. Cash is a fact. Cash conversion, roughly operating cash flow divided by reported profit, asks whether the E in the P/E ever actually arrives in the bank. A company can book revenue it has not collected, capitalise costs it should have expensed, and report a profit that exists mostly as receivables and inventory. Its P/E will look reasonable right up to the point where the accruals unwind.

Picture two engineering firms at the same multiple. Over five years, one converts 95 percent of its reported profit into operating cash. The other converts 40 percent, with receivables growing faster than sales every single year. The second firm’s earnings are not lies, necessarily, but they are promises, and promises trade at a discount to cash. A P/E built on 40 percent conversion is not comparable to one built on 95 percent, no matter how similar the two numbers look on a screen.

Leverage: the multiple that forgot the balance sheet

The P/E prices only the equity. It is silent about debt, which means a company can look cheap on earnings while carrying a balance sheet that owns most of the business. This is exactly the problem EV-based multiples solve: enterprise value adds net debt to the market capitalisation, so you are pricing the whole firm, not just the slice shareholders hold. A leveraged company at 8 times earnings and a debt-free company at 12 times earnings can easily swap places once you rank them on EV/EBITDA instead.

Here is the illustrative arithmetic. Two infrastructure operators each earn 100 crore and each has a market cap of 1,000 crore, so both trade at a P/E of 10. One has no debt. The other carries 2,000 crore of borrowings. On enterprise value, the first firm costs 1,000 crore for its earnings stream and the second costs 3,000 crore for a stream of the same size, with interest rate risk thrown in for free. Same P/E. Three times the price. The multiple never mentioned it.

Cyclicality and share count: the two quiet distortions

Cyclical earnings make the P/E lie in the most seductive way possible. At the top of a commodity cycle, profits are inflated, so the P/E looks low precisely when the risk is highest. At the bottom, profits collapse, so the P/E looks absurdly high precisely when conditions are most likely to improve. A metals producer at 5 times peak earnings can be a far more aggressive bet than the same producer at 40 times trough earnings. The discipline is to normalise: ask what a through-cycle year of earnings looks like, averaged across good years and bad, and put the multiple on that instead of on whatever the last twelve months happened to deliver.

Share count is the other silent editor of per-share numbers. Earnings per share, the E itself, is profit divided by shares, and the share count is not a constant. Steady buybacks shrink the denominator, so EPS can grow even when total profit does not. Repeated dilution, through large ESOP pools, QIPs or acquisition-funded issuance, does the opposite: the company can grow total profit respectably while each shareholder’s slice barely moves. Two companies with identical profit growth of 10 percent a year diverge sharply if one is retiring 2 percent of its shares annually and the other is issuing 4 percent. Over 2019-2024 that gap compounds into materially different per-share outcomes, and the P/E, computed on today’s EPS, quietly absorbs the distortion without flagging it.

The six companions at a glance

Companion	Question it answers	What P/E alone gets wrong
Earnings growth	What am I paying for the future?	Treats a compounder and a stagnant business at the same multiple as equals
Return on capital	How efficiently does the business turn capital into profit?	Ignores that low-ROCE earnings are partly pre-committed to reinvestment
Cash conversion	Does the reported profit actually arrive as cash?	Accepts accrual-heavy earnings at face value
Leverage	What does the whole firm cost, debt included?	Prices only the equity and ignores the balance sheet
Earnings cyclicality	Is the E a peak, a trough, or a normal year?	Looks cheapest at cycle tops and dearest at cycle bottoms
Share count trajectory	Is the denominator of EPS shrinking or ballooning?	Lets buybacks and dilution silently rewrite per-share history

Never quote a P/E naked

The habit that ties all of this together is simple and slightly unfashionable: never state a P/E without stating what the E is. Trailing or forward? As reported or normalised? Consolidated or standalone? Before or after exceptional items? On today’s share count or last year’s? Every one of those choices can move the multiple by a third, and most quoted P/Es never disclose which choices were made. A multiple with an undefined denominator is not a valuation. It is a vibe with a decimal point.

This is the standard any serious research process holds itself to, and it is the standard a tool like Altys is built around: a number is only as good as the definitions and the data sitting underneath it.

Before you let a P/E influence a decision, run the checklist:

• Define the E. Trailing, forward, or normalised, and say which.
• Check growth. A multiple without a growth rate is half a sentence.
• Check ROCE. Same P/E, different returns, different price.
• Check cash conversion. Profit that never becomes cash is a promise, not an earning.
• Check leverage. If debt is material, look at EV-based multiples too.
• Check the cycle. Ask whether this year’s E is a peak, a trough, or a normal year.
• Check the share count. Look at five years of shares outstanding, not just five years of EPS.

Seven lines, two minutes, and the single most abused number in investing becomes something you can actually use.