Kanban Is a Price System
Kanban Is a Price System
Mises in the Factory — No. 1
Taiichi Ohno used to say the idea came to him while watching American supermarkets. The customer takes from the shelf what he needs, when he needs it, in the quantity he needs. The shelf is restocked only afterwards, and only with what was taken. Nobody in the back of the store needs to know why the Tuesday shopper bought three bottles of milk instead of two. It is enough to know that they were bought.
Out of that observation came kanban, the card you now find in almost every factory that calls itself “lean”. In most of those factories, though, kanban is treated as a logistics tool: a tidy way to manage inventory. That is a mistake of perspective. Kanban is a Japanese engineer’s answer to a problem that two Austrian economists had posed for the entire economy a few decades earlier.
The problem no plan can solve
In 1920 Ludwig von Mises argued that an economy without prices for the means of production cannot calculate. The central planner may have the best intentions and the best engineers, but he has no way of knowing whether to use steel or aluminum, whether to build a railway or a road, because he lacks the one signal that compares alternative uses of scarce resources: a price born of exchange.
In 1945 Friedrich Hayek moved the problem from calculation to knowledge. The knowledge that matters for decisions does not exist in concentrated form. It is scattered across millions of people, as local, temporary details, often impossible even to articulate: knowledge of “the particular circumstances of time and place”. His example was tin. If somewhere in the world a new use for tin appears, or a mine closes, the people who use tin do not need to know which of the two happened. The price rises, and they economize. The price compresses an enormous amount of dispersed knowledge into a single number you can act on.
Now walk into a factory that runs on a traditional MRP. On Monday morning the system calculates, from a demand forecast, what each department must produce during the week. It is a central plan in miniature. On Tuesday a press slows down, a batch of material arrives defective, an experienced operator calls in sick, a customer pulls an order forward. The plan knows nothing about any of this. The people on the shop floor do.
You know the rest of the week: chasing, priority lists, meetings to “realign”, inventory piling up where it isn’t needed and missing where it is. The factory doesn’t suffer from bad planning. It suffers from having handed the decision to the one place where the knowledge isn’t.
What a card actually does
The mechanism of kanban is simple. When the downstream department consumes a container of parts, the card attached to that container goes back to the upstream department. The card says: make this item, in this quantity. Nothing else.
The upstream department doesn’t know why those parts were consumed. It doesn’t know whether a big order came in, whether downstream scrapped a batch, whether the customer changed the mix. It doesn’t need to. It receives a scarcity signal generated by real consumption, not by a forecast, and it acts. It is the same principle as Hayek’s tin: dispersed knowledge compressed into a signal that anyone can act on without understanding the whole system.
There is a second, less obvious aspect. The number of cards in circulation is limited. No card, no production. Kanban therefore also works as a budget constraint: you cannot produce what has not been consumed. Ohno considered overproduction the worst of the seven wastes, because it generates and conceals all the others. Producing to the plan, without a consumption signal, means producing against an expectation that nobody has validated.
Anyone familiar with the Austrian theory of the business cycle will recognize the pattern. When the central bank artificially lowers the interest rate, entrepreneurs receive a false signal about the availability of savings and launch investments that consumers never asked for. The result is malinvestment. On the shop floor, overproduction is malinvestment on a small scale: capital tied up in parts nobody asked for, generated by a signal that did not come from consumption.
Where the analogy breaks
It would be convenient to stop here. But an analogy is only worth as much as your knowledge of its limits, and the limits of kanban are instructive.
The card transmits a quantity, not a value. It says a part is needed, not how much that part is worth relative to another, nor what you give up producing in order to make it. A market price carries opportunity cost with it; kanban does not. It is a truncated price: it carries scarcity, not the choice between alternatives.
Moreover, someone decides the number of cards, usually production engineering. Inside the firm, as Ronald Coase observed, coordination happens by hierarchy rather than by exchange. Kanban is an island of spontaneous order inside a structure that remains planned.
Finally, kanban works well when demand is levelled and the mix is stable. That is why Toyota pairs it with production levelling. Where variety and variability are high, the card loses information, and many companies go back to the plan. Or, the best ones, look for a richer signal.
What it teaches those who lead
The first lesson concerns the question to ask. Whoever runs a struggling factory almost always asks: how do we plan better? The useful question is a different one: who receives which signal, and with what delay? Almost every flow problem I have seen in twenty years of plants was information arriving in the wrong place, or too late.
The second concerns behavior. Every time the office overrides the card with an urgent priority, it is doing what a government does with a price control: it breaks the signal. And as with price controls, the effect is not only immediate. The line learns that the card doesn’t count and stops trusting it. After a few months the kanban is still on the containers, but the factory has gone back to running on expediting.
The third concerns the direction of travel. Some companies have tried to bring something closer to real prices inside themselves. Haier, in China, broke the organization into thousands of micro-enterprises that trade services with each other under internal contracts. Morning Star, in California, one of the world’s largest tomato processors, operates without bosses: each person negotiates a written agreement every year with colleagues on what they will provide and what they will receive. These are experiments, not recipes. But they point in one direction: more signals, fewer orders.
The irony
The production system that became a synonym for efficiency owes its core to an insight that planners have never accepted: knowledge cannot be centralized; it has to circulate. As far as we know, Ohno never read Hayek. He arrived at the same point by looking at a shelf.
Anyone who runs a company chooses every day, often without realizing it, between the plan and the signal. The choice isn’t only about inventory. It is about where you allow knowledge to reside, and who has the right to act on it.
I write about operations, supply chain and Austrian economics. This is the first installment of “Mises in the Factory”. If you work in a plant and have seen a card overridden, or a plan contradicted by the shop floor, tell me below: the best stories will make it into future installments.
Versione italiana:
Write a comment