Lesson 7 of 7
Why cheap stayed cheap
The cheapest tenth of US companies by P/E has a median Altman Z of 1.66 against 3.52, thinner interest cover, and unusually fat current margins.
By Javier Sanz
Founder of ValueScreener. Built Ninety Nine, a securities services provider, and sold it to a tier-one global bank; ran operations and internationalization at Alpaca, the brokerage API.
Who writes thisEditorial processLinkedIn (opens in a new tab)X (opens in a new tab)
In one minute
- The cheapest tenth of the market is not a random sample of companies with low prices. It differs measurably from the rest.
- Its median distress score is half the rest of the market's, its interest cover is thinner, and its current net margin is nearly twice as fat.
- Three of the four recorded reasons a multiple stays low are visible in that one table, before any company is opened.
"Cheap for a reason" is the oldest sentence in value investing and it is usually said without a measurement behind it. Here is the measurement.
The cheapest decile against the rest of the market
- Median Altman Z, cheapest tenth1.66
- Median Altman Z, the other 1,8723.52
- Median interest cover, cheapest tenth3.69
- Median interest cover, the others5.64
What the cheap end is made of
| Median of the group | P/E under 9.03 | The rest |
|---|---|---|
| Altman Z-Score | 1.66 | 3.52 |
| Interest coverage | 3.69x | 5.64x |
| Net margin | 19.76% | 10.54% |
| Piotroski F-Score | 5 | 5 |
Three of those four rows say something and the fourth says something too.
The balance sheets are weaker. A median Altman Z of 1.66 against 3.52 is the difference between a score in the range the model associates with distress and one comfortably outside it. The cheap end carries more of the companies where survival is a live question.
The interest bill is closer. Operating profit covers it 3.69 times at the median against 5.64 elsewhere. That is not a crisis, and it is a much smaller cushion against a bad year.
The margins are unusually high. This is the least obvious row and the most useful. The cheapest tenth earns nearly twice the net margin of the rest, which is what a peak-earnings denominator looks like in aggregate: the multiple is low because the current year is unusually good, not only because the price is low.
The accounting quality is the same. The Piotroski score's median is 5 in both groups, so whatever separates the cheap end, it is not a broad difference in the nine accounting tests behind that score.
The four recorded reasons, and where each one shows
- The earnings were a peak. Visible above in the margin row, and company by company in the gap between a trailing multiple and a ten-year one.
- The balance sheet is doing the discounting. Visible in the distress score and the interest cover. A company whose equity might be diluted or wiped out is priced for that, and a multiple cannot see it.
- The earnings are not repeatable as cash. A multiple divides by reported profit, which is not what the owner receives. A company converting little of its profit into cash will keep a low multiple for as long as that lasts.
- The discount is structural. A control structure, a single customer, a pending legal matter or a business in decline. None of these is a temporary mispricing and none of them is in the ratio.
What the evidence does and does not support
The published record on buying the cheap end as a strategy is a separate subject, with its own drawdowns and its own decay after publication, and it has its own lesson. What the table above supports is narrower and more useful: the cheap decile is a group with measurably weaker balance sheets and measurably better current margins than the rest of the market, so the prior for any individual member of it should be that at least one of those two is true.
Checking which one takes one filing. That is the work a multiple was never going to do.
Check yourself
What separates the cheapest tenth of the market from the rest, on the measurements above?
- Worse accounting quality on the Piotroski score
- Weaker balance sheets, thinner interest cover, and unusually high current margins
- Nothing measurable; the difference is only in price
Show the answer
Answer: B.
The distress score and interest cover are materially weaker and the net margin is nearly twice as high. The accounting score's median is the same in both groups.Why is a high net margin in the cheap decile a warning rather than reassurance?
- High margins attract regulation
- In aggregate it is what a peak-earnings denominator looks like: the multiple is low because the year was unusually good
- Margins are easy to manipulate
Show the answer
Answer: B.
A low multiple can come from a low price or a high denominator. A group with margins twice the market's is showing the second.Which of the four reasons a multiple stays low is invisible in the ratio itself?
- Peak earnings
- All four, which is why the ratio is a starting point rather than a finding
- Only the structural discount
Show the answer
Answer: B.
The multiple is one price over one year of profit. Every reason it is low sits outside that fraction, in the balance sheet, the cash flow statement or the disclosures.Sources
- The product's own store, read 2026-09-16: 2,078 US-listed common equities with a computed price to earnings ratio on prices last stored 2026-09-15, split at the published tenth percentile of 9.03.
- Altman, Journal of Finance, 1968: the five-factor distress score whose medians are compared above.
- Piotroski, Journal of Accounting Research, 2000: the nine accounting tests whose median is unchanged between the two groups.
What this leaves out
- This is a cross-section on one date, not a study of what happened next. It says what the cheap end is made of, not what it returned.
- The split uses the published tenth percentile of a cohort deduplicated to one row per company, applied to listings, so the two counts are not the same statistic.
- Every median hides a wide spread. Individual companies in the cheap decile carry strong balance sheets, and the table does not identify them.
- Nothing here is advice about any security.