When One Great Year Makes a Stock Look Cheap: The Cal-Maine Lesson

← All Articles

A trailing price-to-earnings ratio is only as honest as the earnings underneath it. Feed a screen a single blockbuster year and it will hand you a low multiple that looks like a bargain, even when the underlying business is priced richly against what it can normally earn. The trailing P/E is not lying, but it isn't telling the whole truth. It is just answering a narrower question than most investors think they are asking. It tells you what you are paying for last year's profits, not what you are paying for the profits the company can sustain.

Value investors inherited this warning directly from Benjamin Graham, who argued that earnings should be averaged over a full cycle, ideally seven to ten years, before you draw any conclusion about value. He understood that a business caught at the top of its cycle can look statistically cheap at exactly the moment it is most expensive. Nowhere has that been more vivid recently than in the egg aisle.

Cal-Maine: a 5x multiple that was never really 5x

Cal-Maine Foods (NASDAQ: CALM), the largest shell egg producer in the United States, is the textbook case. In fiscal 2025 (the year ended May 31, 2025), the company reported diluted earnings of $24.95 per share on net income of roughly $1.22 billion. That was not a good year in the ordinary sense. The prior year, fiscal 2024, produced diluted EPS of $5.69. So, earnings did not grow by a respectable double-digit percentage. They rose a whopping 338% in twelve months.

The driver was not brilliant management or a durable competitive shift. It was highly contagious bird flu outbreak. Repeated bird flu outbreaks forced the depopulation of tens of millions of commercial layer hens across the industry, egg supply collapsed into a period of strong holiday demand, and prices spiked. Cal-Maine's net average selling price per dozen jumped to $3.13 for the fiscal year from $1.93 the year before. When your product nearly doubles in price and your own flocks stay largely intact, profits explode.

Here is where the trap springs. When that $24.95 sits in the trailing window, screens flatter the stock badly. On a trailing basis the shares have screened near 5x earnings, against a ten-year median closer to 11x. To a mechanical value filter, that reads as one of the cheapest large caps in the market. Anyone who bought on that signal was, in effect, paying for bird flu to recur every year forever.

In case anyone was unaware, it did not. As supply recovered, egg prices fell and earnings reverted fast. By the third quarter of fiscal 2026, Cal-Maine's diluted EPS had dropped to $1.06, down roughly 90% from the comparable quarter a year earlier. The stock, which had traded well above $100 during the boom, has since drifted back toward the high $80s. The 5x multiple was never a real 5x. It was a normal multiple wearing an abnormal year's clothing.

Run the check the other way and the picture is clearer. This is a company that earned $5.69 in fiscal 2024, that has bought capacity through the ISE America and Echo Lake Foods acquisitions, and that is building out a prepared-foods segment to smooth its results. A reasonable normalized earnings figure lands somewhere around $7 to $8 per share, not $25. Put the recent price against that normalized number and the multiple is roughly 11x to 13x, right in line with the company's own long-run median. Not a screaming bargain. Not obviously overpriced either. Just a fairly valued cyclical, which is a very different investment case than "5x earnings."

This is a category problem, not a Cal-Maine problem

Cal-Maine is memorable because eggs are concrete, but the distortion shows up anywhere earnings swing with a cycle the company does not control.

Commodity producers are the obvious group. Oil and gas explorers, fertilizer and chemical makers, copper and gold miners, lumber and pulp companies: their profits track a spot price, and a single year near the top of that price can compress a trailing multiple into low single digits. The stock looks cheap precisely because the commodity is expensive, which is usually the worst time to underwrite the earnings as permanent.

Cyclical industrials rhyme with this. Machinery, trucking, steel, auto suppliers, and semiconductor-equipment names all earn peak margins when demand and pricing align, then give much of it back in the downturn. Homebuilders are perhaps the cleanest parallel to Cal-Maine. A strong housing year can leave a builder trading at six or seven times trailing earnings while forward earnings are set to fall as rates, affordability, or the order book turn. In each case the trailing P/E is measuring the crest of a wave and quietly assuming the tide never goes out.

The common thread is simple. When the swing factor sits outside management's control, a great year tells you almost nothing about earning power, and the trailing multiple built on it tells you even less.

How GrahamGrade screens for it

Our screener is built to refuse this trap rather than fall into it. When a ticker survives the quantitative Graham filters, we do not stop at the reported trailing figure. We look at the earnings history across the cycle and ask whether the most recent year is representative or exceptional. A year that towers over the surrounding ones, especially one traceable to a supply shock, a commodity spike, or a one-time event, gets normalized rather than taken at face value.

In practice that means recomputing the valuation against a mid-cycle earnings estimate, flagging when a low trailing P/E depends on a single outlier year, and surfacing the gap between the headline multiple and the normalized one. A stock at 5x trailing that becomes 13x normalized is not disqualified, but it is reframed honestly, so the number you act on is the one that reflects the durable earning power. That is the difference between a screen that finds cheapness and a screen that finds the appearance of cheapness.

Where normalization can go wrong

Normalization is a judgment call, and judgment cuts both ways. The central risk is that you smooth away a real, permanent improvement and dismiss a genuinely cheap stock as a mirage. If Cal-Maine's prepared-foods and specialty-egg push actually does raise its earnings floor, then a normalized figure anchored to the old commodity business will be too low, and the stock will look more expensive than it is. Structural change is exactly what a naive cycle-average misses.

The error runs the other direction too. Picking the wrong baseline, over-weighting a weak year, or assuming mean reversion in a business that has genuinely re-rated will make you pass on winners. And normalization tells you nothing about timing: a cyclical can stay near its peak far longer than a cycle-average model expects, and being early can feel identical to being wrong for years.

The honest position is that normalized earnings are an estimate, not a fact. The goal is not false precision. It is to avoid underwriting one exceptional year as if it were the run rate, while staying open to the possibility that the business really has changed.

The takeaway

A low trailing P/E is a question, not an answer. Before treating it as cheap, ask what the company earned across the whole cycle and whether last year belongs to that pattern or stands apart from it. For cyclicals and commodity producers especially, the single most useful habit a value investor can build is to normalize earnings before believing the multiple.