P/E — price to earnings
What it measures¶
P/E is the price investors pay for each unit of a company's yearly profit. It is the share price divided by the earnings per share (net income divided by the number of shares) — or, the same thing for the whole company, the market capitalisation (share price × number of shares) divided by net income. A P/E of 20 means investors pay 20 times one year's profit; read the other way up, the profit is 5% of the price.
Why it is on the page¶
It is the most quoted valuation multiple. On the Fundamentals page it opens the valuation rows of the ratio tree; on the Workspace it is one of the four multiples of the Valuation strip, described against the company's own past and its sector. A high P/E says investors expect profit to grow; a low one says they expect it to stall or fall — the multiple measures that expectation, not the business itself.
Thresholds, their source and the population they were fitted on¶
The summary never scores valuation, so P/E has no pass or fail. It is placed by position only:
| Reference | What it shows | Source |
|---|---|---|
| own five-year band | above its 90th percentile, within, or below its 10th percentile — the band of the P/E at the company's last five fiscal year-ends | the company's own stored history; at least three year-ends are needed |
| sector | above, within or below the middle half of similar companies — within 10% of their median where only a median is published | Aswath Damodaran's January industry tables |
| 15 times | the ceiling Benjamin Graham set for a cautious investor, on three years' average profit | Graham, The Intelligent Investor |
| 22.5 | Graham's ceiling for P/E multiplied by P/B (price to book value) | the same |
- Fitted on: Graham's lines come from his experience of US industrial companies in the mid-twentieth century and were never tested statistically.
- Evidence grade: folklore (Graham's lines); practitioner convention (sector medians).
A worked example¶
Apple, on the data stored on 25 October 2026: a P/E of 45.3, against an own band whose middle was 30.5 and whose 10th to 90th percentiles ran from 23.7 to 36.4 over four fiscal year-ends — the strip reads above its own band. No sector median was stored, so the sector position was not shown.
Honeywell International, on the data stored on 26 October 2026: no P/E at all. Its share count has fallen 48% since FY2025, far more than its June 2026 split explains, so today's market value and its FY2025 statements describe different numbers of shares — the strip prints that reason in place of a figure. (On 25 October the same data gave 14.8, just under Graham's 15 line: a figure priced across two share bases, which the strip no longer forms.) Nor is there an own band: its FY2021–FY2025 year-ends ended before the split, and their share counts are not priced against the split-adjusted closes.
What “No verdict” means here¶
P/E never carries a verdict word: valuation is described, never scored, and never enters a percent. The row reads not applicable when the profit is zero or a loss — a P/E of a loss has no meaning — and insufficient data when the share price or the profit is not stored. The Valuation strip says band not available when fewer than three year-ends can form the band, with the reason; a P/E without a figure says why in its place.
What the verdict is and is not¶
Above its band says investors pay more for each unit of profit than they did at most of the company's recent year-ends. It is not a judgement that the shares are priced too high or too low, not a forecast of the price, and not an instruction to trade the shares. Profit can be depressed or inflated in a single year, which moves the P/E without any change in what investors pay.