Why GARP screens use two growth rates, not one
The classic PEG ratio trusts a single growth number. Here is how a GARP screen forces revenue and earnings growth to agree before trusting either.
The PEG ratio was built to fix a real problem: a low P/E can mean "cheap" or it can mean "declining," and the ratio alone doesn't tell you which. Dividing P/E by an expected growth rate was supposed to separate the two. In practice, PEG has a second problem it never solved — it takes a single growth number on faith, and that number is usually the easiest one to distort.
One growth line isn't enough
Textbook PEG most often uses forward EPS growth as its denominator. EPS is also the line item a company has the most tools to move without the underlying business changing much: a buyback shrinks the share count and lifts EPS with no new revenue behind it, a prior-year write-down creates a low base that makes next year's rebound look like growth, and one-time tax or litigation items can swing the number either way. A stock can show 40% forward EPS growth off a depressed base while revenue is flat, and a plain PEG ratio will call it cheap.
Revenue growth is harder to manufacture. It requires selling more, at some price, to someone. That makes it the more honest of the two lines — but it isn't sufficient by itself either, since revenue can grow while margins collapse and earnings go nowhere. The fix isn't picking the better of two flawed numbers; it's requiring both to actually agree.
Making both lines earn it before either counts
Zorro's GARP screener preset builds its growth-adjusted PEG on exactly that requirement: instead of dividing forward P/E by one growth estimate, it divides by the slower of trailing revenue growth and EPS growth. If revenue is up 35% but EPS is up only 12% — margins compressing as the company grows — the ratio uses 12%, not 35%. A stock can't get a flattering GPEG by having one metric run hot while the other lags; both lines have to be moving together for the ratio to look good.
The growth figure used as the denominator is also capped at 40% and requires a floor of at least 3% before it counts at all — a company compounding faster than 40% a year in either metric is priced as a special case either way, and anything below 3% doesn't qualify as "growing" in the first place. A stock with negative or near-zero growth on either line simply returns no GPEG value and drops out of the ranking rather than distorting it.
Financials get routed around this test entirely. A bank or insurer's "revenue" line is interest and premium income, not a comparable measure of unit sales, so the screen's isFinancial check — detected primarily by whether the stock reports a current ratio at all, since a lender's balance sheet has no conventional current-asset-versus-current-liability split — lets those names qualify on EPS growth alone rather than forcing a revenue comparison that wouldn't mean the same thing.
The floors around the ratio matter as much as the ratio
A clean GPEG number by itself still isn't sufficient to call something GARP — growth-at-a-reasonable-price, not growth-at-any-price. The screen adds gates on either side of it: forward P/E has to sit between 8 and 35, operating margin has to be positive, and return on invested capital has to clear 10%. Each of those floors exists to kill a specific failure mode the ratio alone lets through. Drop the ROIC floor and a bank coming off a loss year rides its earnings rebound to a near-top rank on cheap math with nothing behind it. Drop the forward-P/E floor and the list fills with names sitting at a cyclical earnings peak — memory chips, tanker shipping — that look cheap only because this year's number won't repeat. Drop the EV/EBITDA floor and a distressed shipper trading at under three times EBITDA can outrank companies with real balance sheets, because a market pricing in default risk also prices in a low multiple.
None of these are redundant with the growth-rate fix above. A two-line growth test stops a company from gaming the numerator's input; the surrounding floors stop the whole formula from being gamed by capital structure or cycle timing instead.
Why a fresh listing can't run this screen yet
One consequence of requiring two growth lines to agree is that a stock needs a real prior-year base for both of them — you can't compute year-over-year revenue growth or EPS growth without a year-ago quarter to compare against. That's also why a company that just showed up in the recent-IPO list won't clear a growth screen like this for roughly a year after its debut: there's no comparable trailing period yet, on either line, for the ratio to be built from. The same operating-history requirement is why fresh listings get evaluated on entirely different terms in the meantime — first-day pop, lockup timing, where the stock sits against its offer price — rather than on growth or value screens that need four quarters of history to even compute. It's the same discipline that shows up in free cash flow yield done properly: a metric is only as good as the guardrails stopping it from being computed on data that can't support it, and a screen that returns "no result" for a name it can't fairly judge is doing its job, not failing at it.
The broader lesson travels past GARP specifically. Any ratio built on a single growth or earnings input inherits whatever can be done to that one input. The fix is rarely a smarter ratio — it's requiring a second, harder-to-move number to confirm the first before the ratio gets to run at all.