The Toilet Paper Theory

Picture two women in the paper goods aisle of a supermarket. One of them lives in a household that spends about $150,342 a year. The other lives in a household that spends about $35,046. Four times the money, near enough. They are standing in front of the same shelf, reaching for the same product, and here is the odd part. They are going to buy about the same amount of it.

The wealthier one will buy a nicer brand. The quilted kind, the one with the pattern stamped into it. But she will not put four times as many rolls in her cart. Nobody does. There is no such thing as needing four times as much toilet paper.

That sounds like a small thing to notice. It is not. It is the whole argument, and once you see it you will start seeing it everywhere.

Here is why. Some things you buy have a ceiling on them. Not a money ceiling. A body ceiling. There is an amount of toilet paper a person uses, and an amount of food a person eats, and past that point more money does not buy more of it. It buys a fancier version of the same amount. Extra income has to go somewhere, so it goes into a bigger house, a newer car, restaurant meals, tuition, and stocks. It does not go into more rolls.

Now turn it around, because this is the half that matters. At the other end there is no ceiling at all. There is a floor, and people fall through it. A household running out of money does not buy a cheaper brand forever. At some point it just buys less. And then it buys none.

The man who read other people’s grocery bills

In 1857 a Prussian statistician named Ernst Engel sat down with the account books of working families and started adding up what they spent. He was looking for a pattern, and he found one, and it has held up for nearly two centuries. The pattern is this: the poorer the family, the larger the share of its money that goes to food. Not the larger amount. The larger share.

The American numbers today are stark. A household in the bottom fifth spends about $5,498 a year on food, which eats 33.0 percent of what it makes before taxes. A household in the middle spends $9,097, which is 12.2 percent. So the middle household spends more in dollars, obviously. But look at the size of the jump. Total spending across that range more than quadruples. Food spending does not even double. And most of the increase that does happen is people eating out instead of cooking, which is not more food. It is the same food with someone else washing up.

Housing tells the same story from the other end. The bottom fifth puts 41.6 percent of everything it spends into keeping a roof over its head. The top fifth puts in 29.3 percent. Not because poor families have grander houses. Because rent is the one bill you cannot make smaller by buying a cheaper brand of it.

Economists have a clunky phrase for the thing underneath this. They ask how much of one extra dollar a household actually spends, as opposed to saves. Give a dollar to a household in the bottom fifth and about 55 cents of it gets spent. Give the same dollar to the top fifth and about 12 cents does. Sort by wealth instead of income and the gap gets wider still, roughly ten to one.

The effect even shows up inside the same person. In the stimulus experiments, people spent about 41 cents of every dollar when the check was $600. When the check was $1,200 they spent about 31 cents. When it was $1,800, about 25 cents. Same household, same week, smaller and smaller fraction. Why? Because the needs were getting filled. And needs have an end.

This is not an argument about fairness

People have known that the rich save more since Keynes, and probably since Engel. On its own that is just a remark about savings rates. What makes this different is the reason the gap cannot be closed.

Saving more could just be a choice. Someone decides to be careful with money this year and generous next year. Choices change.

But not buying a fifth roll of toilet paper when you already have four is not a choice. It is a fact about being a person. No amount of wanting produces a need you do not have. Which leads somewhere strange: the total demand for basic goods in a country is not really driven by how much money there is. It is driven by how many people can afford the basics, multiplied by roughly the same amount each.

That last line is the claim. It says something the treatise had only said about goods: that there is an order to all this, and the order runs across people too. The things that are safest when money gets tight are the last things anyone with money gives up, and the first things anyone without money gives up. Same shelf. Opposite direction.

Follow the dollar

Take one dollar away from a household at the bottom and give it to a household at the top. At the bottom, 55 cents of that dollar was going to be spent, and mostly on ordinary things made in ordinary factories: food, soap, clothes, gas. At the top, about 12 cents gets spent, and what does get spent leans toward services and nicer versions rather than more items.

So where did the rest go? It did not disappear. It went into assets.

Which means moving that dollar upward does two things at the same time, and they are the two things this framework keeps bumping into. It shrinks the part of the economy that makes things, because fewer people are buying them. And it inflates the part that trades claims, because the money has to go somewhere and there is a fixed pile of assets to bid for. Corporate profits at a postwar record of 13.25 percent of GDP while small firms struggle is what that looks like from the company side. Subprime borrowers hitting a thirty-two year delinquency record while prime borrowers are perfectly fine is what it looks like from the household side.

This is the part that deserves the word counterproductive. An economy that keeps moving money upward is not just unequal. It is quietly strangling demand for its own products, because it keeps taking purchasing power from where it would have been spent and putting it where it will not be. The companies that make ordinary things watch their sales flatten in the middle of a boom, and cannot work out why the boom is not reaching them.

When pretending becomes real

Here is where it stops being merely unfortunate. A money shortage is, at first, a kind of pretending. The factories are still there. The skills are still there. Only the claim on them was withdrawn. But leave the pretending in place long enough and it stops being pretend. It starts destroying the very thing it was only impersonating a shortage of.

At the level of a household, that has a name and a number. In 2024, 13.7 percent of American households were food insecure. That is 47.9 million people, 14.1 million of them children. By November 2025 the rate had climbed to 14.2 percent, and spiked to 16 percent in that month alone. Among households already receiving food assistance it hit 46 percent.

Nothing had run out. American farms produced a surplus in every one of those years. The food existed. The trucks existed. The people who needed it were standing in the same country as the food.

And a hungry child is not postponing a purchase. Whatever a child misses during the years their brain and body are being built, they do not get to go back and buy later. It is one of the least reversible things in all of economics. This is the pretend shortage turning into a real one, on a human timescale, and doing more damage than any scrapped machine.

Our instruments are pointed the wrong way

Something has been nagging at this framework, and it has now happened three times in two months, which is too often to be luck.

The gauge this site built to detect money shortages reads comfortably loose, at the same moment the bottom of the credit market posts its worst delinquency record in thirty-two years. The chart of corporate profits against small business failures shows the two pulling apart, and no national average can explain why. And in September 2025 the Department of Agriculture announced it would stop publishing the Household Food Security report after the 2024 edition. The one federal measurement of whether Americans are going without is being switched off.

The honest response is to treat that as a fact about our tools rather than about the world. Averages are dominated by whoever holds most of the dollars. That is just what averaging dollars does. So an economics built on averages will keep reporting that things are fine for exactly as long as things are fine for the people at the top, which can be a very long time after they stopped being fine for everyone else.

Which makes this theory two claims, not one. A claim about the economy. And a claim about why our instruments will be the last to notice.

What would prove this wrong

The name is undignified on purpose. Arguments about inequality usually get conducted in the language of fairness, where anyone can wave them away as a matter of opinion, or in the language of growth, where the effects are vague enough to argue about forever. Toilet paper does not let you do either. There is an amount a person needs. A rich person will never need more. A poor person will sometimes have none. And an economy arranged so that more of its money lands where it cannot be spent on ordinary things has not become more efficient. It has just gotten quieter about a shortage that never went away.

The treatise ends by saying that the factories are standing and the shelves are full, and whether people go without depends not on the goods but on where money sits in the order of things. This adds the missing half of that sentence. It depends on where money sits, and on where in the crowd the money is standing when the order changes.