Peter Lynch's most portable idea, from One Up on Wall Street (1989), is that there is no such thing as a good stock in general; there are six kinds of company and each one is a different question with a different holding period and a different reason to sell. The schools already gave you the six and the map. What has been missing is the drill, and the drill is the sell trigger, because a category you cannot sell out of is a label rather than a decision.
The rule that makes this work: write the trigger on the day you buy, in numbers, and put it where you will find it. A trigger written after a stock has moved is a rationalisation.
| Category | The question that decides it | The trigger, written as a number |
|---|---|---|
| Slow grower | Is the dividend safe, and why do I own this rather than a bond | Payout above the coverage you set, or a dividend-adjusted ratio that falls below the alternative's yield |
| Stalwart | Is the price reasonable against the growth, and is the growth intact | A stated ceiling on the price-earnings-to-growth ratio, or a stated percentage gain, whichever arrives first |
| Fast grower | Is the growth profitable, self-funded and repeatable, with runway left | Same-store or same-cohort growth below a stated floor for two consecutive periods, or new-unit returns below the hurdle |
| Cyclical | Where are we in the cycle, and is the low multiple a peak-earnings trap | Capacity utilisation above a stated level combined with earnings above your own mid-cycle estimate |
| Turnaround | Can it survive, and is the fix real and underway | The justified price-to-book at the completed fix, computed in advance and compared with today's |
| Asset play | What is the hidden asset worth, and what closes the discount | A stated discount level, plus a date by which the catalyst must have happened |
Three of those triggers are worth computing rather than describing, because the arithmetic is where the discipline lives.
Advertisement
How do you calculate a sell trigger?
For a stalwart trading at 50 times earnings while growing 11%, the price-earnings-to-growth ratio is 4.55. A trigger set at 1.8 implies a multiple of 19.8, which is 60.40% below the current one. Writing that down does not tell you the stock will fall. It tells you that the trigger you thought was conservative is nowhere near today's price, which is a fact you would rather know before you buy than after.
For a cyclical at $29.40 with peak earnings of $4.20 a share, the trailing multiple is 7.0 and looks like value. On a mid-cycle estimate of $1.60 the multiple is 18.4, and a normal mid-cycle multiple of 12 implies a price of $19.20, which is 34.69% below today's. The lowest trailing multiple in a cyclical's history occurs at the top of its cycle. That single sentence is the most expensive thing an amateur has to learn, and the arithmetic above is what learning it looks like.
How do you value a turnaround stock?
For a turnaround at 1.4 times book where the fix, if it works, takes the return on equity to the cost of equity, the justified price-to-book at completion is (0.14 − 0.08) ÷ (0.14 − 0.08) = 1.00. The market is already paying a 40% premium to the value of a completed turnaround, which means the thesis needs the return on equity to overshoot to 17%, at which point the justified ratio is 1.50, before today's price makes sense. Turnarounds are bought for the gap between a broken business and a fixed one, and here the gap has already been paid for.