Nobody buys twice as much toilet paper because they got a raise. Nobody buys half as much because they lost a shift. The package sits in a closet. You use what you use.
That is the whole theory. It also has a consequence that turns up in quarterly filings, which means it can be caught being wrong.
Fix the amount a household can use, and the companies selling it hit a wall. They cannot grow by selling more. So growth has to come from somewhere else. Either they charge more for the same amount, or they take a customer off a rival.
So here is the tell. Revenue rising while units fall. Find that across a whole category. Then find the missing units at the cheap end of the shelf. If both show up, the theory is doing real work. If volumes grow along with prices, it is wrong, and it should be said to be wrong.
The first half of 2026 is the first clean test. Here is what it says.
The category answers yes
Across American grocery, the first six months of this year. National brands sold 0.5 percent fewer units than a year earlier. Over the same six months they collected 2.2 percent more dollars.
That gap is the entire prediction in one line. Fewer things left the shelf. More money came in.
And the units did not vanish. They moved. Store brands sold 0.2 percent more units. Their dollar sales stayed flat. That is what taking share at a lower price looks like in arithmetic. Store brands are now 23.8 percent of every unit sold in an American grocery store. The highest share ever recorded.
The companies answer no
Now the complication. It belongs on the table, not left out.
Go to the filings of the companies themselves and the picture does not look like the chart above. Procter and Gamble reported organic sales up 3 percent in the January to March quarter. Two points of that came from volume. Kimberly-Clark, which actually makes the toilet paper, reported something stranger still. Volume-led growth of 4.1 percent. And price investments of negative 1.5 percent.
Read that plainly. The largest tissue maker in the country grew by selling more, and it did that by charging less.
Taken at face value, that is the opposite of the prediction. A theory that says necessity firms must grow on price has just been handed two firms growing on volume.
Two things are worth saying before rescuing anything. These are global companies, and their reported figures blend dozens of countries with an American grocery aisle. And organic sales is a measure built for investors, not for this argument. Neither of those is a refutation, but neither is an excuse either.
Which one is wrong
The two readings cannot both be describing the same thing, so look at what the phrase price investment actually means.
It means cutting the price. A company does not describe a price cut as an investment when demand is strong. It describes it that way when it has lost units and wants them back.
So ask why Kimberly-Clark would give up 1.5 points of price to buy 4.1 points of volume. If households could simply be sold more tissue, the company would raise price and sell more at the same time. That is how a firm behaves in a growing market. It did not do that. It traded margin for units. That is how a firm behaves against a fixed quantity of demand somebody else has taken.
That is the ceiling asserting itself from the other direction. In 2022 it looked like firms taking price because volumes could not grow. In 2026 it looks like firms giving price back because volumes would not return. Both are the same wall seen from opposite sides.
And the category numbers are the arbiter here, because they count the aisle rather than the corporation. In the aisle, national brand units fell.
The part that does not reverse
Here is the finding that changes this from a cyclical story into a structural one.
Ninety-four percent of shoppers say they will keep buying store brands even if grocery prices come down.
Trading down has always been understood as a recession behavior. Money gets tight, people buy the cheaper thing, money loosens, they go back. That rhythm is well documented and it has broken. Private label share climbed through 2024. It climbed through 2025. It kept climbing into 2026, through a period nobody would call a downturn.
This matters for the theory specifically. A capped-demand argument and an income argument make the same prediction during a squeeze and different predictions afterward. If households were merely short of money, the units would come back when the money did. They are not coming back.
The trade-down is also no longer a poor household’s behavior. Dollar General reports customers trading down at an accelerated rate. The fastest growth is among households earning more than $100,000 a year. A ceiling that only bound the bottom decile would not show up in six-figure households.
The tell beneath the tell
There is one move left to a company that cannot raise price and cannot lose units. The toilet paper aisle is where you can watch it happen.
Shrink the unit.
Quilted Northern took its mega roll from 295 sheets down to 255. Across the category, a mega roll that carried 264 sheets in 2022 now carries 224. And the sheet itself got narrower, from the 4.5 inches that was standard to 3.9 inches on most brands today.
| The same roll | Then | Now | Change |
|---|---|---|---|
| Sheets per mega roll | 264 | 224 | −15% |
| Sheet width | 4.5 in | 3.9 in | −13% |
| Paper in the package | Fewer sheets, each narrower, compounding | about −26% | |
Those two reductions multiply. Fifteen percent fewer sheets. Each one 13 percent narrower. The roll holds about a quarter less paper than the one that sat under the same name a few years ago. The price on the shelf may not have moved at all.
This is the theory in its purest form. And it happens in the exact product the theory is named after. Demand sits at something like a household’s actual need. Price cannot rise without losing the sale. So the package quietly gives you less. The ceiling gets paid anyway.
Where this leaves the claim
The prediction survives, in a more careful form than it was made.
What the theory got right is the category. Dollars and units have separated. The separation runs in the predicted direction. And the missing units are exactly where it said they would be, at the cheap end of the shelf.
What it got wrong is assuming this would show up in any one manufacturer in any one quarter. It does not, because firms respond. A company facing a ceiling takes price one year and gives it back the next. Its reported volume swings. The constraint underneath has not moved at all.
What it did not anticipate is shrinkflation. That is a third way the ceiling gets paid. It is also the only one invisible in both the price line and the volume line.
What would prove this wrong
Three things, and they are all in public quarterly data.
National brand units turn positive for two halves running, with dollars growing faster still. That is a real volume recovery, and the capped-demand claim is in trouble. Private label unit share falls back below 22 percent. Then the ratchet was cyclical, and the 94 percent was people describing an intention they did not keep. Sheet counts start rising while shelf prices hold. Then the ceiling has slackened, and the companies have room they did not have.
None of those is present today. All three are checkable in ninety days, which is the point of writing the prediction down.
Two cautions. The 26 percent figure is arithmetic on two separately reported changes, and their periods do not line up exactly. Treat it as the scale of the effect, not a measurement. The 94 percent is survey evidence, which records what people say they will do. The share data is the stronger claim. It records what they did.