Skip to content
Rabin Poudyal
Rabin Poudyal
← All models

Economics and incentives

Supply and demand

Price settles where what sellers will offer meets what buyers will take; hold it anywhere else and a shortage or a glut does the adjusting instead.

Alfred Marshall, Principles of Economics (1890), who drew the two curves and compared them to the blades of a pair of scissors.

As price rises, buyers want less and sellers offer more. The price where the two quantities match is the one that clears the market. Marshall's point was that neither side sets the price alone: "We might as reasonably dispute whether it is the upper or the under blade of a pair of scissors that cuts a piece of paper, as whether value is governed by utility or cost of production." A shock to either side moves the price. A cap held below the clearing price leaves buyers wanting more than sellers will supply, a shortage rationed by queues, luck or connections. A floor above it leaves a glut.

It holds because price does two jobs at once: it tells people what is scarce, and it rewards them for responding. A high price pulls in new sellers and pushes marginal buyers out, and both responses shrink the gap. Suppress the price and both responses stop while the underlying scarcity remains, so the adjustment moves to quantity. The usual pattern is slow, which is why a price cap can look successful for years: existing supply does not vanish on day one, but new supply is not built and old supply leaks away. The San Francisco rent control study below shows this time lag in numbers.

The curves are not observable, so the model predicts direction well and magnitude poorly; how much quantity moves depends on elasticities that have to be estimated. Markets with a few sellers, with buyers who know less than sellers, or with costs that land on third parties do not clear at the socially right price, and that is where regulation has a case. And demand is willingness plus ability to pay, not need, so a market can clear while people go without. None of this rescues a price cap from the quantity response; it only says that the right fix is often on the supply side, such as building more, rather than on the price.

Example

In November 1994 San Francisco voters narrowly passed a ballot measure extending rent control to small multi-family buildings of four units or fewer, but only those built in 1979 or earlier. Rebecca Diamond, Timothy McQuade and Franklin Qian compared tenants and landlords of these buildings with those in similar buildings built from 1980 on. Tenants who gained rent control were nearly 20 percent more likely to still be at the same address years later. But landlords of the controlled buildings cut the supply of rental housing by 15 percent, by selling to owner-occupiers and redeveloping, and the authors estimate this pushed citywide rents up by 5.1 percent. The paper appeared in the American Economic Review in September 2019.

Where it applies

When a rate limit, a free tier or an internal chargeback price sits below what capacity really costs, you get queues and quota fights rather than lower spend; the quantity adjusts. The same logic prices graphics processing units (GPUs) and model application programming interface (API) calls: when demand for inference outruns supply, waiting lists and rationing by priority appear until capacity or price moves.

Think about it

Where in your work is a price or a quota held fixed, and what shortage or glut is doing the adjusting instead?

Applied it to something real? Log it.

Test yourselfSan Francisco extended rent control to small buildings in 1994. What happened to the supply of rental housing and to citywide rents?

Controlled landlords cut rental supply by 15 percent, by selling to owner-occupiers and redeveloping, and citywide rents rose about 5.1 percent, while protected tenants stayed put nearly 20 percent more often.

Related models

Sources and further reading