Valuation: What a Business Is Worth

Every valuation method is the same idea wearing different clothes: a business is worth the cash it will produce, discounted for time and risk. Ratios are shorthand for that calculation. This guide shows the math, the shortcuts, and where each one breaks.

By Karanvir Singh, founder of Quantix Invest

What valuation actually is

The value of any financial asset is the present value of the cash it will hand you over its life. That is not a school of thought; it is the definition. A bond is easy because the cash flows are contractual. A stock is hard because the cash flows are uncertain, perpetual, and depend on how well a business competes for decades.

Every valuation tool falls into one of two families:

• Intrinsic (absolute): forecast the cash the business will generate and discount it back to today. Discounted cash flow analysis is the pure form.
• Relative (multiples): compare what you pay per unit of something (earnings, cash flow, book value, sales) against comparable companies or the company's own history. P/E, EV/EBITDA, and the rest live here.

Multiples are not a different theory from DCF. A P/E ratio is a compressed DCF with the assumptions hidden. That is their strength (fast, comparable) and their weakness (the assumptions are still there, you just cannot see them).

P/E: the default ratio and its limits

The price-to-earnings ratio is share price divided by earnings per share, or equivalently, market capitalization divided by net income. A stock at $60 with $3.00 of EPS trades at 20x earnings. Inverted, that is a 5% earnings yield, which is the more useful framing because it is directly comparable to a bond yield.

Three versions circulate and they are not interchangeable:

• Trailing P/E: price divided by the last twelve months of actual reported earnings. Real but backward-looking.
• Forward P/E: price divided by the next twelve months of analyst estimates. Forward-looking but built on forecasts, and sell-side estimates have historically skewed optimistic, especially eighteen months out.
• Normalized P/E: price divided by an estimate of mid-cycle earnings. The most useful for cyclical businesses, and the most judgment-dependent.

What drives a justified P/E? Three things, and only three:

1. Growth. Faster and longer-lasting growth in earnings justifies a higher multiple.
2. Returns on capital. A company that grows while reinvesting at 30% returns creates far more value per dollar of growth than one reinvesting at 8%. Growth funded by capital that earns less than its cost actually destroys value.
3. Risk. Cyclicality, leverage, customer concentration, regulatory exposure, and earnings volatility all lower the multiple a rational buyer will pay.

Where P/E fails outright:

• Negative or near-zero earnings. P/E is meaningless for a loss-making company, and enormous for one with a temporary earnings dip.
• Cyclicals. Steel, autos, semiconductors, and energy show their lowest P/E at the peak of the cycle, when trailing earnings are unsustainably high, and their highest P/E at the trough. Buying cyclicals on low P/E is a classic and expensive error.
• Different leverage. A heavily indebted company can post a low P/E simply because equity is a thin, risky slice of the capital structure. That is not cheapness, it is leverage.
• Accounting differences. Depreciation policy, capitalization choices, and one-time charges all move the E.

PEG and the growth adjustment

The PEG ratio divides P/E by the expected annual earnings growth rate in percentage points. A stock at 30x earnings growing 30% a year has a PEG of 1.0. The rough convention, popularized by Peter Lynch, is that a PEG below 1 suggests growth is not fully priced in.

PEG has real intuition behind it: growth is the main thing that justifies a high multiple, so scaling by growth is a sensible normalization. But treat it as a rough sort, not a valuation.

• It ignores how long growth lasts. Five years at 25% is worth vastly more than two years at 25%, and PEG cannot tell them apart.
• It ignores the cost of that growth. A company growing 20% by reinvesting every dollar at low returns is worse than one growing 12% while returning cash.
• It ignores risk. High-growth forecasts are also the least reliable forecasts.
• The growth input is usually an analyst consensus, and consensus long-term growth estimates have a long history of being too high.
• It behaves badly at low growth. A company growing 2% would need a P/E of 2 to hit PEG 1.0, which no functioning business ever trades at.

Price-to-book and where it still works

Price-to-book compares market capitalization to shareholders' equity: what the market pays versus the accounting value of net assets. It was the original academic definition of "value" investing and still is in most factor research.

It works well for banks, insurers, and asset-heavy financials, where the balance sheet is mostly marked to market and book value is a meaningful economic number. A bank trading at 0.7x book with a solid loan book and a 12% return on equity is saying something real.

It works badly, and increasingly so, for modern businesses. Under current accounting, research and development, software development, brand building, and training are largely expensed as incurred rather than capitalized. So the assets that actually produce the profits of a pharmaceutical, software, or consumer-brand company do not appear on the balance sheet at all. Meanwhile, goodwill from acquisitions does appear, inflating book value for serial acquirers. The result is that price-to-book has drifted from a measure of asset value toward a measure of "how intangible is this business," which is not the same thing.

This is one plausible reason the classic price-to-book value factor performed poorly through much of the 2010s. Researchers have proposed fixes, notably capitalizing intangible spending to build an adjusted book value, and those adjusted measures have looked better. The lesson generalizes: a ratio is only as good as whether its denominator still measures what it used to.

EV/EBITDA and enterprise value

Market capitalization prices the equity. Enterprise value prices the whole business, independent of how it is financed:

Enterprise value = market capitalization + total debt + preferred stock + minority interest minus cash

The intuition: if you bought the entire company, you would pay for the shares, assume the debt, and get to keep the cash on the balance sheet. Enterprise value is what the business itself costs.

Pair enterprise value with a pre-interest profit measure. The standard is EBITDA (earnings before interest, taxes, depreciation and amortization), giving EV/EBITDA. Because both numerator and denominator sit above the debt line, the ratio lets you compare a debt-free company to a leveraged one on equal footing, which P/E cannot do.

Worked comparison. Two identical businesses, each with 100 of EBITDA and 40 of operating income after 60 of depreciation. Company A has no debt. Company B carries 400 of debt at 6% (24 of interest).

On P/E, Company B looks 16% cheaper. On enterprise value, the two businesses cost exactly the same. Company B's lower P/E is entirely a leverage artifact, and the equity holder in B is taking materially more risk for it. This is why private equity buyers, credit analysts, and anyone valuing an entire company work in enterprise value.

EBITDA has its own famous flaw. Charlie Munger's objection was that calling depreciation a non-cash charge is fine only if the equipment never needs replacing, which is false for anything physical. For capital-intensive businesses, prefer EV/EBIT (which keeps depreciation as a cost) or EV/free cash flow. EBITDA is most defensible for asset-light businesses whose depreciation genuinely overstates required maintenance spending.

Free cash flow yield

The most honest quick number: free cash flow divided by market capitalization (or, in enterprise-value form, free cash flow to the firm divided by enterprise value). A company with a $10 billion market cap producing $600 million of free cash flow has a 6% free cash flow yield.

Why it beats earnings yield in most situations: free cash flow already accounts for capital expenditures and working capital, and it is much harder to manipulate than reported profit. It is the cash genuinely available to pay dividends, repurchase shares, retire debt, or accumulate.

Two cautions. First, free cash flow is lumpy; a year with a new factory can look terrible for reasons that are good. Average over three to five years. Second, high-growth companies deliberately depress free cash flow by investing, so a low yield is not automatically bad. The question is always whether the investment earns a return above the cost of capital.

Free cash flow yield is also directly comparable to a bond yield, which makes it useful for the only comparison that really matters: what else could this money be doing?

DCF intuition: a worked model

Discounted cash flow is the honest version of valuation because it forces the assumptions into the open. It has three parts: forecast cash flows, discount them, and add a terminal value for everything after the forecast.

The discount rate reflects both time and risk. Practitioners typically use the weighted average cost of capital, blending the after-tax cost of debt with an estimated cost of equity. Precision here is false comfort; a sensible range (say 8% to 11% for a stable business) beats a spuriously exact 9.37%.

Worked example: Harbor Systems. A stable software and services company. Current free cash flow is $200 million. Assume 8% annual growth for five years, then a 2.5% perpetual growth rate afterward, and a 9% discount rate. There are 100 million shares outstanding and $300 million of net cash on the balance sheet.

Terminal value. Using the perpetuity growth formula, terminal value at the end of year 5 equals year 6 cash flow divided by (discount rate minus perpetual growth):

Year 6 free cash flow = 293.9 x 1.025 = 301.2
Terminal value = 301.2 / (0.09 minus 0.025) = 301.2 / 0.065 = $4,634m
Present value of terminal value = 4,634 x 0.650 = $3,012m

Result.
Enterprise value = 973 + 3,012 = $3,985m
Equity value = 3,985 + 300 net cash = $4,285m
Per share = 4,285 / 100 = $42.85

Now look at what that model actually says. The terminal value is 76% of the total. Three quarters of the answer comes from a single assumption about cash flows in year six and beyond, discounted at a rate you also guessed. That is not a flaw in this particular model; it is true of nearly every DCF of a going concern.

Which is why sensitivity analysis is the model, not an appendix to it. Hold everything else constant and vary two inputs:

The same business is worth $31 or $71 depending on two assumptions that no one can know, both of which sit inside any reasonable range. A one-point change in the discount rate moves value roughly 20%. That spread is the honest output of a DCF.

Reverse DCF: what is the price assuming?

The most practical use of DCF machinery is to run it backward. Instead of forecasting cash flows to produce a value, take the current market price as given and solve for the growth rate the market must be assuming. Then ask one question: is that assumption reasonable?

Worked example. Harbor Systems actually trades at $70, not $42.85. Market cap is $7,000m, so enterprise value is 7,000 minus 300 net cash = $6,700m. Keep the 9% discount rate and 2.5% perpetual growth. What five-year growth rate justifies $6,700m?

Working through it, roughly 22% annual free cash flow growth for five years gets there: year 5 free cash flow of about $540m, a terminal value near $8,500m, and a total present value in the neighborhood of $6,700m. So the market is pricing Harbor to nearly triple its cash generation in five years.

Now the question is concrete and answerable with business judgment rather than spreadsheet judgment. Does Harbor have the market size, the competitive position, and the pricing power to grow cash flow 22% a year for five straight years? If it is a dominant platform in an expanding category with 40% incremental margins, maybe. If it is a mature company facing well-funded competitors, almost certainly not.

Reverse DCF converts valuation from "what do I think it is worth" (where your own biases enter freely) into "what does this price require" (where you are checking a specific claim). That is a much better use of an afternoon.

Why cheap can be a trap

A low multiple is information, not a conclusion. Markets are not efficient in a strong sense, but they are not stupid either, and a stock at 6x earnings usually got there because a large number of informed people concluded the earnings will not persist.

The classic value traps:

• Structural decline. Print newspapers, physical video rental, department stores, and directory publishers all traded at very low multiples for years while earnings ratcheted down. The multiple was low against an E that kept shrinking. Cheap on last year's numbers, expensive on next year's.
• Peak cyclical earnings. A homebuilder at 5x earnings in a housing boom is not cheap. Normalize earnings across a full cycle before applying any multiple to a cyclical business.
• Leverage disguised as value. A company at 8x earnings with net debt at 5x EBITDA has an equity stub whose value can go to zero on a modest downturn. Always look at enterprise value, not just equity value.
• Value destruction at the top. Cash-rich, cheap-looking companies run by managers who serially overpay for acquisitions can stay cheap forever, because the market correctly discounts the cash for how it will be spent.
• Accounting that flatters. Earnings inflated by aggressive accruals produce a low multiple against a number that will not repeat.
• Governance and jurisdiction. Structural discounts for weak minority-shareholder protections can persist for decades and are not obviously mispricings at all.

The mirror error is treating quality as a free pass. Excellent businesses have been bought at prices that took a decade to grow into. The Nifty Fifty of the early 1970s were mostly genuinely great American companies; investors who paid 50x to 90x earnings for them still suffered years of poor returns while the businesses did fine. Price matters even when the company is superb.

Market-level valuation and CAPE

Valuation applies to whole markets too, and the best-known tool is the cyclically adjusted price-to-earnings ratio (CAPE, or the Shiller P/E), popularized by Robert Shiller. It divides the real (inflation-adjusted) price of the index by the average of real earnings over the previous ten years.

The ten-year averaging is the whole trick. Single-year index earnings crash in recessions, so an ordinary market P/E spikes to absurd levels exactly when stocks are cheapest. In 2009, trailing S&P 500 P/E briefly exceeded 100 because earnings had collapsed, while CAPE was around 13 and correctly signaled a good entry point.

What CAPE can do. The evidence for a relationship between starting CAPE and subsequent ten-year real returns is reasonably solid. Historically, decades that started with CAPE in the low teens tended to deliver strong real returns, and decades that started with CAPE above roughly 30 tended to deliver weak ones. The mechanism is not mysterious: you are buying a stream of earnings, and the price you pay per unit of earnings mechanically affects your return.

What CAPE cannot do. A great deal, and this is where it is most often misused.

• It has no timing ability whatsoever. CAPE exceeded its historical average in the early 1990s and stayed above it for most of the following thirty years. An investor who sat out on that signal missed one of the great compounding periods in market history. Elevated CAPE has persisted for a decade or more at a stretch.
• The comparison base drifts. The average CAPE since 1881 includes long stretches with different accounting rules (notably post-2001 goodwill impairment treatment, which depressed reported earnings), higher dividend payout ratios, different corporate tax rates, and an index dominated by capital-intensive industrials rather than asset-light technology firms. Comparing today's CAPE to an 1881-to-present average is not a clean comparison.
• Interest rates matter and CAPE ignores them. A given earnings yield is worth more when the risk-free alternative pays 1% than when it pays 6%. Attempts to fix this (the "excess CAPE yield," which subtracts the real bond yield from the CAPE earnings yield) are more defensible than raw CAPE for cross-era comparison.
• The statistical evidence is weaker than it looks. Overlapping ten-year windows massively overstate the effective sample size. There are only about a dozen genuinely independent decades in the modern US record. And the relationship has been fitted repeatedly in-sample, which is exactly the condition under which apparent predictability tends to fade out of sample.
• US survivorship. Much of the evidence comes from the single most successful equity market of the twentieth century.

So what is the practical use? Treat CAPE as a rough dial on expected returns, not a switch. High CAPE is a reasonable argument for tempering your return expectations, saving more, and keeping your asset allocation honest. It is not a reason to move to cash. Market timing based on valuation has a poor empirical record, largely because "expensive" can get much more expensive for many years, and missing a modest number of the best days does severe damage to long-run returns.

Putting it into practice

A workable process, roughly in order:

1. Understand the business first. How does it make money, from whom, and what stops a competitor from taking it? No amount of multiple arithmetic substitutes for this.
2. Normalize the inputs. Use multi-year averages for cyclical earnings and lumpy capital spending. Decide which non-GAAP adjustments you accept.
3. Pick the right ratio for the business. EV/EBIT or EV/free cash flow for capital-intensive and leveraged firms. Price-to-book for banks and insurers. Free cash flow yield for mature cash generators. Price-to-sales, used cautiously and only alongside a view on eventual margins, for early-stage growth.
4. Compare three ways: against direct competitors, against the company's own ten-year history, and against the market. Divergence between these is where the interesting questions are.
5. Run a reverse DCF. Find out what the price already assumes.
6. Write down what would prove you wrong. A specific, falsifiable condition, decided before you buy, is the single most useful discipline in the whole exercise.
7. Size the position for your uncertainty. Valuation ranges are wide. Position sizes should reflect that.

And the honest framing that should sit alongside all of this: the same public filings you are reading are being read by thousands of full-time analysts with faster data and better models. Skill in valuation makes you a more informed investor and a much harder person to sell a story to. It does not reliably produce market-beating returns, and the long-run evidence on active stock selection after costs is discouraging. Most people are best served by broad, low-cost, diversified index funds, with individual analysis as a small and deliberately bounded part of the whole. This is education, not individualized financial advice.

Common mistakes

• Using one ratio. Every multiple has a blind spot. Triangulate with at least two, ideally one earnings-based and one cash-based.
• Anchoring on a historical average multiple. Businesses and rate environments change. "It always traded at 18x" is not an argument.
• Applying P/E to cyclicals at the peak. The most reliable way to buy a cyclical at the worst possible moment.
• Ignoring the balance sheet. Two companies with identical P/E ratios and different net debt are not similarly priced.
• Treating DCF output as precise. Report a range, name the swing assumption, and be honest about the terminal value share.
• Confusing a falling price with improving value. If the earnings power fell faster than the price, the stock got more expensive on its way down.
• Using CAPE to time the market. Useful for calibrating expectations, poor for deciding when to be invested.
• Forgetting costs and taxes. A 2% valuation edge disappears quickly under trading costs, spreads, and short-term capital gains rates.