The same company, two PE ratios: one says cheap, one says expensive, each waving a historical percentile. The question is not which number is right; it is that a single PE cannot cover a company. Total market cap divided by total profit — one divisor flattens every difference between segments. The first cut of valuation should not land on a calculator. It should land on scissors: split the company open first, then talk about what it is worth.
One Divisor Flattens Everything
PE is total market cap over total profit. One divisor erases segment differences — assets priced like cigar butts can sit beside businesses priced like bubbles, all folded in half and folded again inside one multiple. Someone will say valuation is not that mystical: look up the PE’s historical percentile, and cheap or expensive is plain at a glance. That argument assumes the whole company wears one face. In reality a company is often a bundle of several businesses — one earns hard money, one lives off a license, one tells a story. Reading a percentile over the bundle means pricing three different things at the same number.
So one system writes the rule in stone: no single PE or PEG, no sell-side linear optimism. Split the company into independent business segments. Find a global or domestic benchmark leader for each. Convert each segment into a market value by ratio of revenue and market share. Add all segments together to get intrinsic value. If the current market cap stands far below that sum, there is a margin of safety. Five steps; not one is optional.
Split Hengrui and the Numbers Sit on the Table
A pharmaceutical company makes the test. Hengrui splits into three businesses — oncology, anesthesia, contrast agents — benchmarked against Roche, Humanwell, and Beilu respectively. Added up, they give a reasonable center of value; set against the market cap of the time, the conclusion was a large premium. The judgment has a handrail: the constraint on benchmarking lives in the conversion ratios of share and revenue, not in imagination. The gap between the summed value and the market price is exactly where the margin of safety comes from. The finer the split, the harder to cherry-pick only the pretty segment.
The hard part is not the splitting; it is the willingness to throw the story away. Sell-side forecasts stretching to 2030 — none of it enters the valuation. Only assets, revenue, and cash flow that already exist on the ground today count. The objection goes: benchmarking measures the present, but growth stocks are bought for the future. A distant story can serve as a reason to buy; it cannot serve as a basis of value. Folding the story into the market cap means advancing the share price against profits that do not yet exist. Such advances get repaid, and repayment arrives as a decline. Cash flow belongs to today, the story to next year; two ledgers that must never share one divisor.
Benchmarks Picked by Mouth?
One more question: if the benchmark is just whoever the analyst picks, the same arithmetic can manufacture any conclusion. It sounds unanswerable, yet a segment-by-segment comparison is precisely the hardest thing to rig. Each conversion must survive scrutiny against shares of revenue; the gap after the parts are summed sits in the open. The analysis of CIMC Group is the extreme case: assume the entire marine equipment business loses everything and is written to zero, value only the land, and if the market cap still sits below the land’s equity, that is a real margin of safety. Hidden assets valued on their own, loss-making businesses zeroed before the sum — pessimism’s ruler used up first. Anyone assembling a conclusion would not dare split this way.
The worst-hit are ordinary investors who go all-in on one percentile: not lazy, just never handed a pair of scissors.
A Discount Left After the Splitting Deserves the Name
The limits of this procedure are equally clear: it measures assets already on the ground, predicts nothing, and names no price target. What it gives is a benchmark — how far the market cap sits from that base, how deep the discount. A market cap far below the sum after the splitting is the reason to act; a premium after the sum means even the loveliest story is still a premium. Scissors in hand: cut first, then add, and only then reach for the calculator.
