Peak Earnings Risk: Normalizing EPS Before Screening P/E
Peak Earnings Risk: Normalizing EPS Before Screening P/E
After a broad drawdown, valuation screens light up. Names that traded at 25x earnings a year ago now print at 10x, and the instinct is to read the lower multiple as a margin of safety. For cyclical businesses, that instinct can point the wrong way.
Peak earnings risk is the risk that a low P/E reflects earnings sitting at a cyclical high rather than a genuinely cheap price. When margins and volumes revert, the "E" shrinks, and a multiple that looked like a bargain can expand sharply without the share price moving at all.
The problem is not new. Graham and Dodd observed in Security Analysis that earnings are strongly influenced by the business cycle, and they suggested dividing price by a five-, seven- or ten-year average of earnings (Morningstar). What has changed is how quickly modern screens surface "cheap" names, and how rarely those screens adjust for the cycle.
Why a Low P/E Misleads on Cyclical Stocks
Cyclical stocks in industrials, materials, energy, autos and semiconductors carry heavy operating leverage. Small shifts in revenue produce large swings in profit, which makes earnings the most volatile input in the P/E ratio.
Aggregate U.S. corporate profits, tracked in the BEA national accounts series on FRED, show repeated sharp contractions around recessions. At the company level, those swings are usually larger, which is what turns a routine cyclical high into peak earnings risk for any screen built on reported EPS.
Peak Earnings and the Value Trap Problem
At peak earnings, margins are stretched, pricing power is often temporary, and the market applies a low multiple because it expects reversion. A screen reads that as a discount. In practice, it is frequently a value trap: the stock is cheap only on what the business earned at the top, not on what it earns across a cycle, a classic pricing vs valuation gap.
The mirror case matters too. Damodaran notes that failing to adjust for the cycle can lead to significant undervaluation at the depth of a recession, when depressed earnings make a firm look expensive or leave it with no meaningful P/E at all (NYU Stern).
Trailing vs Forward P/E: Both Inherit the Cycle
Trailing P/E uses reported earnings, which at a peak embed the best margins of the cycle.
Forward P/E uses consensus estimates, which are anchored to recent results and can lag near a turn.
Neither one removes peak earnings risk. The trailing vs forward P/E debate is about timing within the cycle, not a correction for it.
Diagnosing Where a Company Sits in the Earnings Cycle
Before choosing a normalization method, establish position. Five signals help:
Margins versus the company's own history. An operating margin in the top quartile of its ten-year range is a flag.
Revenue versus trend. Growth well above the long-run trend often reflects pricing or volume that will not persist.
Capacity and inventory. High utilization, rising capex and industry-wide inventory builds often precede a downturn in cyclical industries.
Return on capital versus cost of capital. Returns far above the cost of capital in a commodity-like industry tend to attract new supply.
Estimate revisions. Flat or falling revisions while trailing EPS sets records suggest the market is already looking past the peak.
One flag rarely settles the question. Three or more together suggest a company is late in its earnings cycle.
Three Methods for Normalized Earnings
Damodaran frames the core question simply: what would this firm earn in a normal year (NYU Stern)? Three approaches answer it.
Multi-Year Average EPS
Average dollar earnings over a full cycle, typically five to ten years, the same principle behind the CAPE ratio. It breaks down when a company has grown or shrunk materially, because past dollar earnings no longer reflect current scale.
Mid-Cycle Margins Applied to Current Revenue
Apply the average operating or net margin from a full cycle to current revenue. This keeps the estimate at today's scale, and Damodaran points out that revenue is harder for accountants to manipulate than earnings. Mid-cycle margins also suit systematic screens, because the inputs are standardized and available for most of the listed universe.
Return-on-Capital Normalization
Apply the average return on capital to current invested capital, or average ROE to current book equity. This works best for capital-intensive businesses, where the asset base is a better anchor than sales.
All three methods share one limitation. Replacing current earnings with normalized earnings implicitly assumes the reversion happens immediately. If it takes several years, the valuation should reflect that lag.
Building a Normalized P/E Ratio Into Stock Screening Criteria
The normalized P/E ratio divides price by normalized EPS. A repeatable screen takes five steps:
Define the universe and tag cyclical sectors separately from defensive ones.
Compute normalized EPS, using mid-cycle margins on current revenue, or a multi-year average for companies whose scale has been stable.
Compare trailing P/E with normalized P/E. When the trailing multiple sits well below the normalized one, treat it as a flag for peak earnings risk.
Apply the cycle diagnostics above as a secondary filter.
Re-run on a fixed schedule, so the screen updates as revenue and margins change, and pair it with systematic sell signals for current holdings.
The point of these stock screening criteria is not to forecast the turn. It is to make sure a "cheap" signal survives a cycle-adjusted denominator.
Conclusion
A low multiple is only informative if the earnings beneath it are representative. For cyclical names, the most dangerous P/E is often the lowest one on the screen.
Normalized earnings will not time the cycle. They do remove the single largest distortion in a valuation screen, and they make the conversation easier when an investment committee or a client asks why a "cheap" name is off the list. The practical shift is to treat peak earnings risk as a standing filter, not a judgment call made one name at a time.
Turn This Framework Into a Rule With Surmount Wealth
Knowing you should normalize earnings is easy. Doing it consistently across hundreds of names, every quarter, without a spreadsheet breaking somewhere, is not.
Surmount Wealth lets advisors and portfolio managers turn a thesis like this one into an automated, rules-based strategy. You can start from a prebuilt strategy library or build your own rules from scratch. It runs on top of existing brokerage accounts, with no asset transfers and no coding.
A Hypothetical Illustration: The "Cycle-Adjusted Value Screen"
This is a hypothetical concept for illustration only. It is not an existing Surmount strategy, it is not a recommendation, and no performance is implied or should be inferred.
Universe: U.S. large- and mid-cap equities, with cyclical sectors tagged.
Normalize: Estimate EPS as the ten-year average operating margin applied to trailing revenue, adjusted for tax and share count.
Flag: Exclude names whose trailing P/E sits more than a set threshold below their normalized P/E, since that gap suggests earnings near a cyclical high.
Confirm: Keep only names where at least two of the earnings-cycle diagnostics read neutral.
Refresh: Re-screen and rebalance quarterly, after each earnings season.
Assumptions and limitations
It depends on the quality and consistency of reported financial data.
Ten years of history may not span a full cycle for every company.
The thresholds are arbitrary and would need testing and review.
Normalization assumes margins revert, which may not happen if an industry changes structurally.
Any backtest of a rule like this is hypothetical and does not indicate future results.
Why Advisors Use Surmount Wealth
Prebuilt or custom: Start from a strategy library or encode your own thesis.
No code required: Build rules like the one above without writing software.
Test before going live: Check how a rule set behaves on historical data first.
Keep existing custodial relationships: Connect to current brokerage accounts with no transfers.
Automated, on schedule: Screens, rebalances and rule changes run without manual rework.
Consistent and documented: The same rules apply across accounts and are written down for compliance review.
Want to see what your own screening rules would look like on autopilot?
Book a demo with Surmount Wealth →
Disclosure: Surmount Wealth provides technology tools for building, testing and automating rules-based strategies. This content is for informational and educational purposes only. It does not constitute investment advice or a recommendation to buy or sell any security. The strategy described above is hypothetical, has not been traded, and is presented only to illustrate platform functionality. Hypothetical and backtested results have inherent limitations and do not reflect actual trading. Past performance does not guarantee future results.
FAQ: Peak Earnings Risk
What is peak earnings risk?
It is the chance that a low P/E reflects cyclically high earnings rather than a cheap price. When margins revert, the multiple can expand without the share price moving.
Why are cyclical stocks cheap at peaks?
Peak earnings inflate the "E," so the P/E shrinks while the market prices in reversion. That is why a low multiple on a cyclical name can signal a value trap.
How do you calculate normalized earnings?
Common methods average EPS over five to ten years, apply mid-cycle margins to current revenue, or apply average return on capital to today's capital base.
When is a normalized P/E ratio useful?
It is most informative for cyclical sectors such as materials, energy and industrials, where profits swing widely across the earnings cycle. Stable businesses usually need less adjustment.
Is trailing or forward P/E better?
Neither corrects for the cycle. Trailing vs forward P/E differ only in timing: one embeds peak margins, the other relies on estimates anchored to recent results.
