Direct Answer
Valuing cyclical companies - commodities, industrials, and homebuilders among them - is hard because a single year's earnings can be unrepresentative: peak-cycle profits make trailing P/E look artificially cheap, while trough-cycle profits make it look artificially expensive. Commonly cited responses are normalizing earnings across a full cycle (sometimes called normalized or mid-cycle earnings) and leaning on revenue- or book-value-based multiples, which are less distorted by the earnings swing.
Key Takeaways
- Cyclical earnings swing with the economy or an industry cycle, so a single year's earnings can misrepresent a cyclical company's underlying, through-cycle profitability.
- A low trailing P/E at the top of the cycle and a high trailing P/E at the bottom of the cycle can both apply to the same company, at roughly the same stock price.
- Normalized or mid-cycle earnings - a commonly cited approach that averages or smooths earnings across a full cycle - aim to produce a more representative denominator than any single year.
- Revenue- and book-value-based multiples (price-to-sales, price-to-book) are commonly cited alternatives because sales and book value tend to move less violently than earnings across a cycle.
- Where a company sits in its cycle matters as much as the multiple itself - the same P/E can mean something very different at peak versus trough.
- These are simplifications and areas of genuine debate among analysts, not a precise formula - identifying "a full cycle" and normalizing earnings within it both require judgment.
What Is Valuation for Cyclical Companies?
Valuation for cyclical companies refers to the set of adjustments analysts commonly make when the standard tools of stock valuation - trailing or forward P/E chief among them - break down for a business whose earnings rise and fall significantly with an economic or industry cycle. Commodity producers, industrial manufacturers, and homebuilders are frequently cited examples: their revenue and margins tend to track commodity prices, capital spending cycles, interest rates, or housing demand rather than growing steadily year over year.
The core problem is straightforward. A standard P/E ratio divides the current stock price by a single year's earnings per share. For a stable, non-cyclical business, that year is usually a reasonable proxy for the years around it. For a cyclical business, that single year can sit near either extreme of a multi-year swing, and the resulting P/E can say more about where the company is in its cycle than about whether the stock is cheap or expensive.
Why the Distortion Happens - and How Analysts Adjust For It
The distortion: because P/E = Price ÷ Earnings per share, and the earnings term is the part that moves the most for a cyclical company, the ratio itself becomes volatile even when the price is comparatively stable. Near a cyclical peak, elevated earnings sit in the denominator, which can push the P/E down and make the stock look artificially cheap. Near a cyclical trough, depressed (or negative) earnings push the P/E up - or make it non-meaningful - and can make the same stock look artificially expensive, even though the underlying business hasn't necessarily changed.
Normalized or mid-cycle earnings is a commonly cited response: rather than using the most recent year's earnings, an analyst estimates what the company would earn under mid-cycle, "normal" conditions - often by averaging reported (or margin-adjusted) earnings across a full economic or industry cycle, from one peak or trough through to the next. The resulting normalized EPS is then used as the denominator in place of trailing EPS:
Normalized P/E = Price ÷ Normalized (mid-cycle) earnings per share
This is a simplification, not a precise formula - there is no single agreed-upon rule for how many years constitute "a full cycle," and reasonable analysts can normalize the same company's earnings differently depending on the period chosen and how they average or weight the years within it.
Revenue- and book-value-based multiples are the other commonly cited approach. Price-to-sales (price ÷ revenue per share) and price-to-book (price ÷ book value per share) swap out the earnings denominator entirely. Revenue rarely swings as violently as earnings across a cycle - it can decline, but it's less likely to collapse toward zero or go negative the way profit can near a trough - and book value changes gradually rather than jumping from one reporting period to the next. Both multiples carry their own limitations (revenue multiples say nothing about margins; book value can be distorted by impairments, asset age, or intangibles) and are best read alongside, not instead of, an earnings-based view.
Worked Example
Hypothetical example - for education only. The company, figures, and cycle shown are illustrative and do not represent any real business.
Consider a hypothetical steel producer, "Cyclical Steel Co.," trading at $40.00 per share throughout a seven-year period during which its EPS moves with the steel-price cycle:
| Year | EPS | Stock price | Trailing P/E on that year's EPS |
|---|---|---|---|
| Year 1 (peak) | $5.00 | $40.00 | 8.0× |
| Year 2 | $4.00 | $40.00 | 10.0× |
| Year 3 | $2.50 | $40.00 | 16.0× |
| Year 4 (trough) | $1.00 | $40.00 | 40.0× |
| Year 5 | $2.00 | $40.00 | 20.0× |
| Year 6 | $3.00 | $40.00 | 13.3× |
| Year 7 | $3.50 | $40.00 | 11.4× |
At the exact same $40.00 price, trailing P/E ranges from 8.0× at the earnings peak (Year 1) to 40.0× at the earnings trough (Year 4) - an investor screening on trailing P/E alone would flag this stock as a screaming bargain in Year 1 and wildly expensive in Year 4, despite no change in price. Averaging the seven years of EPS gives $21.00 ÷ 7 = $3.00 normalized EPS, and a normalized P/E of $40.00 ÷ $3.00 = 13.3×. That normalized multiple sits closer to the middle of the observed range and is intended to be a more representative starting point than either the peak-year 8.0× or the trough-year 40.0× figure - though it remains an estimate built on the specific seven years chosen, not a precise or unique answer.
Interpreting the Result
A normalized P/E, or a price-to-sales or price-to-book multiple, is a starting point for comparison, not a verdict on its own. A useful next step is comparing the resulting multiple against the same company's own history and against similarly cyclical peers, rather than against non-cyclical businesses whose "normal" P/E range reflects a very different earnings profile.
Where a company sits in its current cycle matters alongside the multiple itself. A cyclical stock trading on a low trailing P/E immediately after several strong years can be signaling that earnings - and the multiple - are both about to move against the investor, a pattern sometimes described as a value trap. Conversely, a high trailing P/E near a cyclical trough doesn't necessarily mean the stock is expensive if earnings are likely to recover. This is a contested and judgment-heavy area of analysis: reasonable analysts can disagree about where a company sits in its cycle and how much of a recovery or decline to assume, so any conclusion should be treated as an estimate, not a guarantee, and weighed alongside balance-sheet strength, competitive position, and the broader industry outlook.
Limitations and Common Mistakes
| Mistake | Why it causes problems | Better practice |
|---|---|---|
| Screening on trailing P/E alone | A cyclical stock can screen as cheap purely because it's near an earnings peak, or expensive purely because it's near a trough. | Cross-check trailing P/E against a normalized-earnings estimate or a revenue/book-value multiple before concluding a cyclical stock is cheap or expensive. |
| Assuming one "correct" cycle length | There is no universally agreed number of years that defines a full cycle, so a normalized-earnings figure built on the wrong window can be just as misleading as a single year. | Disclose the period used, consider more than one window, and treat the normalized figure as an estimate rather than a precise number. |
| Comparing cyclical and non-cyclical multiples directly | A "normal" P/E range for a stable, non-cyclical business reflects a very different earnings profile and doesn't translate to a cyclical one. | Compare cyclical companies against cyclical peers and against their own historical range across full cycles. |
| Treating revenue or book-value multiples as a full substitute for earnings | Price-to-sales ignores margins entirely, and price-to-book can be distorted by impairments, asset age, or heavy intangibles. | Use revenue- and book-value-based multiples alongside an earnings-based view, not as a standalone replacement for it. |
| Ignoring where the company sits in its current cycle | The same multiple can mean something very different depending on whether the company is near a peak, a trough, or in between. | Weigh the multiple together with an assessment of cycle stage, industry conditions, and balance-sheet strength. |
These approaches remain simplifications of a genuinely difficult problem. Normalizing earnings requires judgment about cycle length and averaging method, and no multiple - earnings-, revenue-, or book-value-based - removes the underlying uncertainty about where a cyclical industry is headed next. Treat any single valuation output as an estimate to be stress-tested, not a precise verdict.
Frequently Asked Questions
Why is trailing P/E unreliable for cyclical stocks?
Trailing P/E divides price by the most recent year's earnings, and for a cyclical company that year can sit near the top or bottom of a swing that has little to do with the stock's underlying value. Peak-cycle earnings can make the P/E look artificially low, while trough-cycle earnings can make the same stock look artificially expensive.
What are normalized or mid-cycle earnings?
Normalized or mid-cycle earnings are a commonly cited approach that averages or otherwise smooths a company's earnings across a full economic or industry cycle, rather than relying on a single year. The goal is a more representative earnings figure than either a peak or a trough year alone, though the method for identifying and averaging a full cycle is itself a simplification and open to judgment.
Why use revenue or book-value multiples for cyclical companies?
Revenue and book value tend to swing less violently than earnings across a cycle, since revenue rarely goes negative and book value changes more gradually than reported profit. That makes price-to-sales and price-to-book commonly cited alternatives to earnings-based multiples for cyclical sectors, though they carry their own limitations and are not a universal substitute for earnings analysis.
Does a low trailing P/E always mean a cyclical stock is cheap?
Not necessarily. A low trailing P/E on a cyclical stock can simply reflect peak-cycle earnings that are unlikely to persist. This pattern is sometimes described as a value trap, and it is one reason analysts commonly look at normalized earnings or a multi-year average before concluding a cyclical stock is undervalued.
Which industries are considered cyclical?
Commodities, industrials, and homebuilders are commonly cited examples of cyclical industries, since their revenue and earnings tend to move with broader economic or industry-specific cycles such as commodity prices, capital spending, and housing demand. Other sectors can show cyclical characteristics as well, and the degree of cyclicality varies by company.
How many years should a normalized-earnings calculation cover?
There is no single agreed-upon number of years - the goal is to span a full cycle from peak to trough and back, which varies by industry and by cycle. This makes normalized-earnings estimates a contested and judgment-dependent exercise rather than a precise formula, and analysts commonly disclose the period used so the estimate can be evaluated.
How is a mid-cycle earnings figure constructed?
Common approaches average earnings across a full cycle, apply a mid-cycle margin to current revenue, or apply a normalised return on capital to the current capital base. Each requires identifying where the cycle boundaries fall, which is a judgment. The methods produce different figures, and the spread between them is a useful indication of how uncertain the normalisation is.
Why do asset-based multiples work better for cyclicals?
Book value and revenue are far more stable across a cycle than earnings, so multiples built on them do not swing with the cycle position. This makes them more comparable across time even though they say less about profitability. For a company whose earnings range from large losses to large profits, a stable denominator is the practical requirement.
How does capacity utilisation inform a cyclical valuation?
Industry capacity utilisation indicates where in the cycle the industry sits, since high utilisation supports pricing and low utilisation invites discounting. Where such data is published, it is often more informative about the cycle position than the company's own results, which lag. It also indicates whether new capacity is likely to arrive and extend a downturn.