Common valuation traps and manipulations

Most bad valuations do not look bad. They look rigorous. They have a discounted-cash-flow model with ten years of projections, a page of comparable companies, a neat sensitivity table. Everything is in its place. And the whole edifice was built backwards — the author started with the answer they wanted, usually the price they hoped for, and then arranged the assumptions until the model agreed.

This is the most important thing to understand about valuation abuse: the manipulation is rarely a lie in the arithmetic. The sums add up perfectly. The dishonesty lives in the inputs — the growth rate chosen, the peers selected, the costs quietly left out — each one nudged a little, none obviously wrong, all pointing the same convenient way. A valuation is only ever as good as its assumptions, and assumptions are where a motivated mind does its work.

This module is a field guide to the handful of moves that do most of the damage. Learn to spot them and you gain two things: you stop fooling yourself with your own optimistic models, and you become very hard to fool with someone else's.

There is a reason these traps recur. A valuation has many assumptions, and each one can be pushed in a direction that raises or lowers the answer. When someone wants a high value — a promoter selling shares, a fund talking its book, or just you, having already fallen for a stock — every ambiguous choice gets resolved the same optimistic way. No single choice looks like cheating. But a dozen small nudges, all leaning uphill, compound into a valuation that is wildly detached from the business, while every individual step looks defensible.

The defence is not cynicism — assuming every model is a fraud — but a checklist of the specific places optimism hides, applied to your own work first. Because the person a valuation most often fools is the one who built it. You already decided you liked the stock; the model is where that feeling goes to get dressed up as analysis. Numbers need a story, and a story needs numbers — and the traps are exactly the points where a flattering story is allowed to override an honest number, or a made-up number to prop up a hopeful story.

The rest of this module names five traps. They are not exotic. They are the ordinary, everyday ways ordinary people talk themselves into paying too much.

One — anchoring to a target. Anchoring is letting an early, often irrelevant number silently set your estimate. You hear a price target of ₹500, or you remember the stock's old high, and every piece of analysis that follows quietly bends toward justifying it. The tell is a valuation that arrives at a suspiciously round, familiar number — the model was steered to the target, not the target discovered by the model. The fix is to build your value estimate before you look at the price or anyone's target, so the anchor cannot reach in.

Two — cherry-picked comparables. When valuing by multiples, the "comparable companies" are chosen, and choosing is where the mischief lives. Cherry-picking comps means selecting only the peers that flatter — the three most expensive firms in the sector — and quietly dropping the cheaper ones that are just as similar. The average multiple of a hand-picked set proves nothing except that the author could pick. The fix is to demand the whole relevant peer set and a stated reason for any exclusion.

Three — the hockey-stick forecast. This is the most common trap of all. A hockey-stick forecast shows modest, checkable growth in the first year or two — the years the forecaster will actually be held to — and then a sharp bend upward in the distant years that no one will ever test. Growth runs 8% next year, then magically accelerates to 25% in years three through ten. Since most of a DCF's value comes from those far years, the whole valuation rests on the part of the forecast least likely to be true. The fix is to ask one blunt question of every acceleration: what specific, named thing changes to make growth jump — and why hasn't it happened yet?

Four — double-counting. The same good thing gets paid for twice. You value the company's cash-generating business with a DCF, and then add the cash on its balance sheet — but that cash is already producing the interest income inside the DCF, so you counted it in two places. Or you credit a company for a growth investment in your forecast and apply a premium multiple for that same growth. Each optimistic assumption, fair on its own, is silently stacked on another that already contains it.

Five — ignoring dilution and stock-based pay. A company that pays its staff in freshly issued shares is quietly getting bigger in share count every year. Dilution — the shrinking of each existing share's slice as new shares are issued — means the profit you care about is profit per share you own, and that can fall even as total profit rises. Stock-based compensation — paying employees in shares rather than cash — is a real cost that many models omit precisely because no cash left the building. It did not leave as cash; it left as a piece of your ownership. Value the per-share stream after dilution, and treat share-based pay as the cost it is.

Watch a single valuation collect four traps at once. illustrative

A note argues a composite software company is worth ₹500 a share against a market price of ₹300. It looks thorough. Read it with the checklist and it comes apart.

First, the ₹500 is the exact figure a brokerage published as its target six months ago — the model was anchored to it. Second, the growth forecast runs 12% next year, then bends to 30% a year from year four onward, on the reasoning that "the addressable market is enormous" — a hockey-stick with no named mechanism, and about three-quarters of the ₹500 comes from those distant years. Third, the note then cross-checks with "comparable" companies — the four most expensive names in the sector, with a dozen cheaper, similar firms left out — a cherry-picked set that conveniently agrees with the DCF. Fourth, the company pays a fifth of its wages in new shares each year, which the model treats as free because no cash moves; the share count is quietly rising and the note values total profit, not profit per share you would actually own — ignored dilution and SBC.

Strip the traps out. Hold growth to a defensible 12–14%, use the full peer set, count the stock-based pay as a cost and the dilution against per-share value, and remove the anchor entirely by valuing from the business up. The honest range lands far below ₹500 — closer to today's price than to the target. Nothing in the note was arithmetically wrong. Every number added up. The manipulation was entirely in which numbers were chosen, and it all leaned the same way: uphill. Remember, in the end, price is what you pay and value is what you get — and a valuation engineered to reach a hoped-for price has quietly stopped measuring value at all.

A checklist of traps is a powerful defence, and it has its own failure modes.

It cannot make a clean model correct. You can remove every trap — no anchor, honest growth, full peer set, dilution counted — and still be wrong, because the future simply did something no careful model foresaw. A trap-free valuation is more honest, not more clairvoyant. Rigour is a floor under your thinking, not a guarantee of the answer.

It cannot replace judgement about the business. Some of these "traps" are, occasionally, justified. A young company's growth genuinely can accelerate when a new plant comes online; that is a hockey-stick with a real mechanism, not a lie. The skill is not to ban the steep forecast but to demand its reason — and sometimes the reason is good. Applied mechanically, the checklist becomes its own trap: dismissing every optimistic assumption as manipulation is just pessimism wearing the badge of scepticism.

It cannot tell you the manipulator's intent. A backwards-built valuation might be a deliberate sell-side pitch, or it might be an honest analyst who fell in love with the stock and never noticed the assumptions all leaned one way. You usually cannot tell which, and it does not matter for your decision — a flattering model is equally dangerous whether the flattery was cynical or sincere. Judge the assumptions, not the motive you cannot see.

1. Mistaking thoroughness for honesty. A long model with many tabs feels trustworthy. Length is not honesty; a hundred pages of assumptions all leaning uphill is more dangerous than a rough estimate that admits its ignorance.

2. Checking the arithmetic, not the inputs. People audit whether the sums add up — they always do — and never audit whether the growth rate, the peer set and the discount rate were chosen fairly. The fraud is upstream of the arithmetic.

3. Letting a round target become the anchor. Once ₹500 is in your head, every fact you meet gets read as evidence for or against ₹500, instead of building the number from scratch. Value first, look at the target later — if at all.

4. Treating stock-based pay as free. Because no cash leaves, share-based compensation slips out of the model, and the steady rise in share count is never charged against per-share value. It is a cost; it is simply paid in ownership instead of rupees.

5. Applying the checklist only to others. The easiest valuation to manipulate is your own, for a stock you have already decided to like. The traps are hardest to see when the person laying them is you.