--- title: "Cyclical Stocks: Five Checks for Reading a Peak in the Numbers" description: "Cyclical stocks look cheapest when their profits are least repeatable. Five checks, scored out of ten, for spotting late-cycle features in a company's reported figures." author: "Marcus Rivera" category: "Investment Metrics" date: 2026-10-25T13:35:04.986Z canonical: "https://investormuse.com/blog/cyclical-stocks-peak-earnings-pe-trap" --- # Cyclical Stocks: Five Checks for Reading a Peak in the Numbers ![stock, trading, monitor](https://cdn.pixabay.com/photo/2016/11/27/21/42/stock-1863880_1280.jpg?w=1200&q=75) > Cyclical stocks look cheapest when their profits are least repeatable. Five checks, scored out of ten, for spotting late-cycle features in a company's reported figures. Flash memory did not become a different product between the summer of 2025 and the summer of 2026. It became a more expensive one. Sandisk's own results release for the quarter ended in early July 2026 shows gross margin at 84.6%, against 26.2% in the same quarter a year before, a move of 58.4 percentage points ([Sandisk, 5 August 2026](https://www.sandisk.com/company/newsroom/press-releases/2026/2026-08-05-sandisk-reports-fiscal-fourth-quarter-2026-financial-results)). Nothing in a chip factory explains a change that size. Pricing does, and that is what makes a company cyclical. Five checks below score how far into a cycle a set of reported numbers already looks. ## Ten Caution Points, and How to Hand Them Out ![](https://hsppuvezyxmkpzkgfkho.supabase.co/storage/v1/object/public/media/enrichment/ab61daa9-7ea0-4beb-a6f6-f91f26625675/5640b4ed-6315-4bac-a7c9-430d6f7c6800/632a7b1f-8479-4650-a565-6c09411e786c.jpg) This runs backwards from a quality checklist. Each of the five checks pays out zero, one or two caution points, so a company can collect ten. A zero reads as an ordinary business whose results do not hang on a commodity price. A two reads as figures that look like the top of a cycle. The total is a teaching device, not a rating, and it carries no opinion about any share. ## How Far the Margin Has Drifted From Its Own Normal A cyclical sells something priced by supply and demand, and most of its costs do not move with that price. Fixed costs act as a lever, so profit swings much harder than the price does. Put numbers on it: a unit that costs $45 to make and sells for $60 earns $15. Lift the price to $72 and the earnings per unit nearly double to $27. Drop it to $51 and only $6 is left. The selling price moved by small steps in each direction. What the owner takes home moved by multiples. Award 0 if the margin sits near its own ten-year average, 1 if it has clearly left that average in either direction, and 2 if it stands at the extreme of its own record. A decade of margin fits on one chart at Stockanalysis.com or Macrotrends, which is all this check requires. ## Price or Volume: Which One Did the Growing? Revenue growth that came from charging more is the fragile kind, because it unwinds at the speed it appeared. Growth that came from shipping more units is slower to win and harder to lose. Companies usually publish the split in the quarterly release or state it during the call, which makes this check easier than it sounds. Award 0 when volume did most of the work, 1 when the split is roughly even, and 2 when price carried the quarter. ## Who Is Building a Factory Right Now? High prices are what pay for new capacity. Boards approve a plant while margins look permanent, the plant opens eighteen months or two years later, supply arrives all at once, and the price breaks. The turn in a commodity cycle usually comes from the supply side, not from customers vanishing. The same pattern in chips specifically is laid out in our piece on [how semiconductor sales cycles actually run](/semiconductor-industry-cycles-sales-downturns). Award 0 when supply is tight and hard to expand, 1 when you genuinely cannot tell, and 2 when several rivals break ground at once. An unknown is not a zero. Score the 1 and write down that you did not find out. ## Is the "E" in That Cheap Ratio a Once-in-a-Decade Number? A price-to-earnings ratio of 8 reads as a bargain right up until you notice the denominator is the strongest year the company has ever had. Peter Lynch is widely credited with the idea that a cyclical trading on a low multiple is waving a caution flag rather than offering a discount, for exactly that reason. The wording is second-hand, summarised by [a Motley Fool write-up of his ideas](https://www.fool.com/investing/international/2012/07/24/an-investing-legends-million-dollar-insights-2.aspx) rather than quoted from him, so treat the attribution as loose and the mechanism as sound. Award 0 when the latest earnings sit close to the multi-year average, 1 when they are above it, and 2 when the newest year makes every prior year look small. A steadier habit than the single-year ratio is to divide the price by average earnings across a whole cycle, say seven to ten years, and the reasons that single-year ratios mislead are set out in [why a P/E ratio misleads more often than it informs](/why-pe-lies-more-often-than-it-tells-truth). ## Could the Balance Sheet Sit Through an Average Year? Debt taken on at a peak has to be serviced in the trough, and the trough is when cash collection gets slow and inventory stops moving. Profitable companies still run short of money for structural reasons, which we cover in [how a profitable business runs out of cash](/working-capital-why-profitable-companies-run-out-cash). Award 0 when cash exceeds borrowings, 1 when the two are close or you skipped the check, and 2 when the debt looks heavy against what the company earns in a merely average year. Stockanalysis.com and Yahoo Finance both put borrowings and cash side by side on a single screen. ## Running the Five Checks on Sandisk's Reported Quarter This is a worked example on published figures, chosen because the swings are unusually legible. It carries no view on the shares. For its fiscal fourth quarter of 2026, Sandisk reported revenue of $8,965 million against $1,901 million a year earlier, up 372%, with gross margin at 84.6% on both a reported and an adjusted basis, where the year-earlier figure was 26.2%. Full-year revenue reached $20.25 billion, up 175% ([Sandisk, 5 August 2026](https://www.sandisk.com/company/newsroom/press-releases/2026/2026-08-05-sandisk-reports-fiscal-fourth-quarter-2026-financial-results)). The first check takes a 2: a 58.4 percentage point jump inside twelve months is about as far from a normal range as a margin series travels ([Sandisk, 5 August 2026](https://www.sandisk.com/company/newsroom/press-releases/2026/2026-08-05-sandisk-reports-fiscal-fourth-quarter-2026-financial-results)). The second also takes a 2, because the company said roughly one third of the quarter-on-quarter revenue increase came from volume and two thirds from higher pricing ([Sandisk, 5 August 2026](https://www.sandisk.com/company/newsroom/press-releases/2026/2026-08-05-sandisk-reports-fiscal-fourth-quarter-2026-financial-results)). The third takes a 1, since the release says nothing about what rival producers are building, and an unknown is a 1 under the rule above. The fourth takes a 2 purely on the shape of the earnings, with reported quarterly operating income of $7,037 million against $18 million in the comparable quarter a year earlier, and the fifth takes a 1 because this example never looked at borrowings. The release does note a further $14 billion of buyback authorisation, lifting the remaining total to $15.5 billion, and whether repurchasing shares at a cyclical high creates anything turns on nothing except the price paid, a judgement we set out in [what a margin of safety actually protects](/margin-of-safety-explained-price-to-fair-value) ([Sandisk, 5 August 2026](https://www.sandisk.com/company/newsroom/press-releases/2026/2026-08-05-sandisk-reports-fiscal-fourth-quarter-2026-financial-results)). That totals 8 of 10. Chairman and chief executive David Goeckeler said in the same release that the company's technology and products are positioned to produce ["durable free cash flow"](https://www.sandisk.com/company/newsroom/press-releases/2026/2026-08-05-sandisk-reports-fiscal-fourth-quarter-2026-financial-results). The claim is testable rather than false, and the test is simple: a margin that climbed this far on price has not yet shown what it does when price turns. ## What Your Total Does and Does Not Say Between 0 and 3, the cycle is not where your reading time belongs with this company. Between 4 and 7, a few cyclical traits are showing, and the next move is to revisit anything you scored a 1 and go and find the missing fact. Between 8 and 10, several peak-like traits are present together, which describes the figures in front of you and says nothing whatever about next quarter. ## The Lynch Rule Is Right About the Ratio and Wrong About the Timing The familiar version of the cyclical warning gets used as a clock, and that is where it fails. A high score tells you the denominator in the ratio is probably unrepeatable. It does not tell you when the unrepeatable part stops repeating, and the historical record is blunt about how badly people estimate that. In the National Bureau of Economic Research's dating of US business cycles, expansions since 1945 have averaged 64.2 months against contractions averaging 10.3 months, and the expansion that began in June 2009 ran for 128 months before it ended ([NBER business cycle expansions and contractions](https://www.nber.org/research/data/us-business-cycle-expansions-and-contractions)). So the position here is that a peak score belongs in the valuation work and nowhere near a calendar. Use it to decide which earnings number you are willing to divide a price by, which is a question you can answer today from a ten-year margin chart. Do not use it to decide that a turn is imminent, because the length of a good run is the one variable in this exercise nobody has ever forecast well. Once an industry's capacity additions are publicly announced with delivery dates, as large chip and refining projects often are, the supply side becomes partly visible and the timing question stops being pure guesswork. Even then it is a window of quarters rather than a date, and the broader cycle still moves on its own schedule, as we trace in [how economic cycles shape a long-term approach](/how-economic-cycles-shape-investment-strategy). ## Three Things This Score Cannot See It cannot see a structural shift in demand that resets an industry for good, and sometimes one arrives. It cannot see how long a peak will hold, which the NBER record shows can run to years ([NBER business cycle expansions and contractions](https://www.nber.org/research/data/us-business-cycle-expansions-and-contractions)). And it cannot see a management team that spent the fat years repaying debt and buying nothing expensive, which is the company a high score treats most unfairly. Use the five checks to ask better questions, never as a verdict. ## Cheapest on Paper, Dearest in Practice A cyclical company looks most attractively priced at the exact moment its earnings are least repeatable, so judge it against a full cycle of results rather than against its strongest three months. *This article is for educational and informational purposes only. It is not financial advice, and Investor Muse is not a registered investment advisor. Historical figures describe the past and do not predict future results. Speak with a qualified professional before you make decisions about your money.* *Position disclosure: neither Investor Muse nor its editorial team holds a position in any company, fund or security named in this article.* ## FAQ ### What makes a stock cyclical? Its sales and profits move with a commodity price or with the wider economy rather than with steady demand. Memory chips, steel, shipping, airlines, homebuilding and heavy machinery are the usual examples, because the product is largely interchangeable and the price is set by supply. ### Why is a low P/E a warning sign on a cyclical? Because the earnings in the denominator are often the best the company has produced. If those earnings fall back toward their own average, the same share price is suddenly attached to a much higher multiple without the price having moved at all. ### How do I tell whether growth came from price or volume? Most companies state the split in their quarterly results release or say it on the call. Sandisk, for instance, reported that roughly one third of its sequential revenue growth in the quarter to early July 2026 came from volume and two thirds from higher pricing. ### Can this score tell me when a cycle will turn? No, and treating it that way is the main misuse. NBER's dating of US cycles shows expansions since 1945 averaging 64.2 months, with the 2009 expansion running 128 months, so a peak-like set of numbers can stay peak-like for a very long time. --- Source: https://investormuse.com/blog/cyclical-stocks-peak-earnings-pe-trap