Note · Opinion, not rated
A low multiple is not a cheap stock.
A price to earnings ratio has two parts, and only one of them is a fact. The price is observed. The earnings are an output of wherever the business happens to sit in its cycle, which means the multiple is at its most flattering exactly when the business is at its most dangerous.
The denominator is not a constant
Everybody checks the numerator. The price is on the screen, it is precise, and it is genuinely what you pay. Then it gets divided by a number that moves 10 times across a normal cycle while the business underneath does not change at all.
Here is why the denominator swings so hard. Most of a manufacturer's cost base is fixed in the short run: the plant, the maintenance, the salaried staff, the depreciation. When volumes rise a little, almost all of the extra revenue falls to profit. When volumes fall a little, almost all of the shortfall comes out of profit. Revenue breathes. Earnings convulse.
One company, one price, one cycle
Worked example
100 million shares, mid-cycle revenue of $1,000M, a business earning a 8% operating margin across a full cycle, worth a fair 15 times mid-cycle earnings. The price is held at $12.00 the entire way through, which is fair value in every single row
- Revenue, peak against trough
- 1.35 times
- Earnings per share, peak against trough
- 10.15 times
- Operating leverage, the amplification
- 7.5 times
The business is worth exactly the same at every point below. The price never moves. Only the reported multiple does.
| Phase | Revenue | Margin | EPS | Reported P/E | Paying, on mid-cycle earnings |
|---|---|---|---|---|---|
| Trough | $850M | 2.0% | $0.17 | 70.6x | 15.0x |
| Recovering | $950M | 5.5% | $0.52 | 23.0x | 15.0x |
| Mid-cycle | $1,000M | 8.0% | $0.80 | 15.0x | 15.0x |
| Strong | $1,080M | 11.5% | $1.24 | 9.7x | 15.0x |
| Peak | $1,150M | 15.0% | $1.73 | 7.0x | 15.0x |
Read the last two columns against each other. The stock is fairly priced in every single row, and the reported multiple runs from 70.6x to 7.0x. A screen would flag it as dangerously expensive at the bottom and as a deep value opportunity at the top, and it would have the sign backwards both times.
The trap, priced
Now let a buyer act on the peak-year number. The stock at the top screens at 7.0x, and somebody decides that is too cheap to ignore, so they will pay up a bit and still get a bargain at 10 times peak earnings.
Ten times peak earnings is $17.25. Against mid-cycle earnings of $0.80, that is 21.6 times. Against a fair value of $12.00, they have overpaid by 43.8%, while believing they bought at ten times earnings.
They did buy at ten times earnings. That was the problem. The earnings were the peak ones.
And then the cycle turns, which is what cycles do. Earnings fall back toward mid, the multiple they actually paid is revealed, and the stock does not decline because the market became irrational. It declines because the denominator went home.
The inverse is the interesting half
Run the trap backwards and it becomes an opportunity, which is where the money is and why almost nobody takes it.
At the trough the same fairly priced stock shows 70.6 times earnings. That number is disqualifying on any screen, in any committee meeting, and in most people's stomachs. It is also the correct price. A buyer who understood the denominator would be paying a fair 15 times what the business earns in a normal year, at the moment of maximum discomfort, with the entire recovery still in front of them.
This is the practical reason cyclicals are so persistently mispriced. The information required to see through the multiple is not hidden, complicated, or expensive. It is just extremely unpleasant to act on, because the screen, the headlines, and the recent price chart all say the same wrong thing at the same time.
What to do instead
Normalize the denominator before you divide by it. Ask what margin this business earns across a full cycle, not what it earned last year. Ten years of margins tells you more than any single year's earnings, and it is a twenty minute exercise.
Ask which phase the current number represents. Not "is the multiple low" but "low against what." A multiple is a claim about the denominator, and if you have not formed a view on the denominator you have not formed a view.
Be suspicious of a cheap cyclical and curious about an expensive one. That is uncomfortable and it is the entire edge. If the reason a stock looks cheap is that earnings are unusually high, the cheapness and the earnings will disappear together.
Check what the price already assumes. A reverse discounted cash flow makes this concrete, because it forces the peak-versus-normal question into the open rather than leaving it buried inside a ratio.
This is standard practice on every report here. It is exactly the argument in the Nike initiation, where a twelve-year low turned out not to be the same thing as cheap once earnings were put back on a normal footing. Sometimes normalizing makes a stock look better than the screen says. Sometimes it overturns the view we walked in with. The model result governs, not the thesis we arrived with, which is the only way the exercise is worth doing.
A multiple is not a measurement. It is a ratio of a fact to an opinion, and the opinion is in the denominator.