What it is
The PEG ratio adjusts a stock's P/E for its earnings growth rate:
PEG = P/E ÷ Earnings Growth Rate (%)
Worked example: a stock at a 24× P/E growing earnings 20% per year has a PEG of 24 / 20 = 1.2. The same 24× P/E on a no-growth business has a PEG of infinity — the growth-adjusted valuation is meaningless because there's no growth in the denominator.
The metric was popularized by Peter Lynch in One Up on Wall Street. His shorthand: a PEG of 1.0 is fair; below 1.0 is potentially cheap; well above 1.0 is overpriced relative to the growth you're underwriting.
What "good" looks like
The Lynch bands hold up reasonably well:
- Under 0.5: Either deep value or a value trap. The earnings growth in the denominator might be reverting from a high base, or about to roll over. Worth investigating, not assuming.
- 0.5–1.0: GARP (growth at a reasonable price) sweet spot. The cleanest growth-at-a-fair-price bargains the model rewards tend to live here.
- 1.0–1.5: Fair. The market has priced the growth roughly correctly.
- 1.5–2.0: Expensive. You're paying up for the growth — it had better arrive.
- Above 2.0: Growth premium territory. Justified for category leaders with durable moats; speculation otherwise.
Why it matters
P/E in isolation is misleading because it doesn't know whether you're buying a stagnant utility or a 30%-grower. PEG bridges the gap by encoding the trajectory:
- It lets you compare across growth rates. A 35× P/E on a 30% grower (PEG 1.2) is statistically the same value-quality call as a 14× P/E on a 12% grower (PEG 1.2). Without the growth adjustment you'd default to thinking the lower P/E is cheaper.
- It catches the "story stock" trap. A meme that's running on narrative will have a sky-high P/E and underwhelming revenue/earnings growth — its PEG explodes. The ratio is the simplest discipline against paying $50 today for $1 of growth you've been promised over the next decade.
- It rewards compounders. A business growing earnings 15% per year for 10 years sees its earnings compound 4×. The market often discounts this back to a reasonable P/E ahead of time — and the PEG reads as "fair" the whole way up.
Common gotchas
PEG is the most-abused ratio in retail finance, in large part because no two sources compute it the same way:
- Growth rate ambiguity. Forward EPS growth (5-year analyst consensus)? Trailing earnings growth (year over year)? Revenue growth as a proxy when earnings aren't there? Different sources pick different denominators. We use the data provider's PEG when it supplies one, and otherwise derive it as
P/E ÷ analysts' expected EPS growth for next year. The derived figure is taggedest.on the grade card so you know which version you're reading. Either way, there is no PEG when analysts expect next year's earnings to be no higher than last year's. - Cyclical companies. A cyclical at the bottom of its cycle has zero or negative earnings growth. PEG goes to infinity or negative. The ratio simply doesn't apply.
- Acquisition-driven growth. A serial acquirer can show 30% YoY earnings growth that's entirely from M&A, not organic. PEG looks great until the acquisitions stop or unwind.
- Coverage gaps. Forward EPS estimates require sell-side analyst coverage. Names below ~$2B market cap often have one or two analysts covering them — or none. We surface a "n/a PEG" rather than fabricate one. Our model treats this as neutral (not penalizing the name), unlike screeners that mechanically rank such names as worst. A PEG left blank because analysts expect no earnings growth is different: there is no growth to pay for, so the score counts it as the most expensive in its sector.
- Banks, insurers, REITs. For sector-structural reasons their PEG often doesn't exist or doesn't mean what it means for industrials/tech. So PEG isn't on their grade card at all; they're scored on a separate financial-strength card instead.
How we use it
PEG is one of the five primary grades on the card, and a growth-adjusted P/E is a core input to the score's Value pillar. On the grade card, the letter follows fixed bands close to the Lynch shorthand: 0.5 or below = A, up to 0.8 = A-, up to 1.2 = B+, up to 1.5 = B, up to 2.0 = C+, up to 3.0 = C, above 3.0 = D. The score doesn't use those bands, and it doesn't use the card's PEG either; see "The score's version" below. A cheap growth-adjusted price on a business that is actually growing is exactly the "growth at a reasonable price" setup the quality-growth model is built to find.
Bottom line
PEG is a sharper P/E. It turns "this multiple looks expensive" into "this multiple looks expensive given the growth you're underwriting." Used properly, it's the single best filter for catching paying-up-for-growth mistakes early — and for catching reasonably-priced compounders that look superficially expensive on P/E alone. Just know where the denominator came from, and don't trust any PEG figure you can't audit.
A growth-adjusted P/E leads our value pillar — and it's scored relatively
The Bull Rankings value pillar has no absolute cutoffs. There is no rule saying "PEG under 1 is cheap." Instead, a stock's growth-adjusted P/E is ranked as a percentile against its own broad sector, and the cheapest name in that sector scores 1.
It carries a weight of 1.0 — the joint-heaviest read in the value pillar, alongside sector P/E, which is only used when there is no growth-adjusted P/E. Everything else is lighter: FCF yield versus sector at 0.4, P/E versus the stock's own history at 0.7, and an EV/EBIT sector read as a light balance-sheet-aware tilt.
That ordering is the whole GARP thesis expressed as weights. A growth-adjusted price leads, because paying 30× for a business growing 30% is a different proposition from paying 30× for one growing 5%, and a raw multiple can't tell those apart. Non-growth-adjusted reads are deliberately kept as tilts so they can't override the thing the screen exists to find: a quality grower that looks expensive on a flat multiple and reasonable on PEG.
Why relative and not absolute
The conventional "PEG below 1" rule fails in a way that's easy to miss: it embeds an assumption about what growth is worth, and that assumption differs enormously by sector.
Software businesses systematically carry higher PEGs than utilities, and not because software investors are irrational — the cash flows are more durable, the margins higher, the reinvestment runway longer. An absolute threshold quietly turns a screen into a sector bet, loading it with whatever industry happens to be cheapest at the moment.
Ranking within sector removes that. A name scoring well on PEG here is cheap relative to the companies it actually competes with for capital, which is the comparison that means something.
The score's version
The score divides the trailing P/E by revenue growth, for every company: the P/E on reported earnings over the last twelve months, divided by revenue growth over the same period.
That is the exact figure our twelve-year backtest was run on. No history of analysts' estimates exists to rebuild an earnings-growth PEG for past years, so the backtest has always used revenue growth. Until October 2026 the live score used the data provider's PEG for most companies, a number whose formula isn't published, while the backtest tested a different one. Using the backtest's version everywhere means the published results describe the model the site actually runs.
Two rules sit on top of it:
- A company reporting a net loss reads as the most expensive. It has no earnings to pay for, and leaving the read out would hand the value pillar to whatever was left: in a September 2026 test, free cash flow yield alone lifted a loss-maker from 555th into the Top 30.
- So does a company whose earnings analysts expect to be no higher in two years. A growth-adjusted price with no growth behind it runs to infinity.
The trade-off is real, and it's why the card doesn't show this number. Revenue growth flatters a fast-growing business with thin margins: Oscar Health's revenue-based figure has read under 1 when its earnings-based PEG was above 9. The card shows the familiar earnings-based PEG; the score uses the one with twelve years of evidence behind it.
The trap in the denominator
PEG's weakness is that the growth rate is an input someone has to choose, and small changes to it move the ratio a lot.
On the card, we use the data provider's PEG where it exists, and where it doesn't, we derive one from analysts' forward EPS growth estimates. Where the figure is derived rather than supplied, the page labels it — the metric strip marks it est. and the note says the PEG is derived.
That labeling exists because of a bug worth admitting. An earlier version silently fell back to a revenue-growth proxy whenever the data provider didn't supply a PEG, which produced values that looked authoritative and reproduced the P/E-over-revenue-growth arithmetic exactly — OSCR displayed 0.83 and DLO 0.48 on that arithmetic alone. Removing the proxy changed the published figure for a large number of companies, and the label now exists so a reader can see which kind of number they're looking at.
A second bug taught a second rule: a rebound is not growth. When analysts expect earnings to drop this year and recover next year, next year's growth rate is mostly the recovery. In October 2026 Uranium Royalty showed a PEG of 0.02: its earnings were expected to fall 98% this year and then rise 460% from that low, which still left them 90% below last year's. So when earnings are expected to dip this year, we divide by the yearly average over this year and next, measured from last year's earnings. When even that shows no growth, there is no PEG, including when the data provider supplies one.
If you take one thing from this page: a PEG is only as good as its growth estimate, and a site that won't tell you where the growth number came from is asking you to trust the least reliable half of the calculation.