« Agents of freedom | Main | Labour's cost of living problem »

August 23, 2013


Luis Enrique

hmm, a demand curve is how many people would like to buy as a function of prices. If house prices were cheap enough (say, cheaper by a factor of 1000), I'd like three houses, a nice rural pad, one by the coast and one in London would be nice. Of course the supply side would have to change rather dramatically, but a demand curve asks a hypothetical how much would you buy at this price, before thinking about supply.

A downward sloping demand curve does not ask that every individual would buy more as prices fall. Even if demand increases in aggregate as prices fall only because the marginal non home owner switches from demanding 0 to 1, that would still give a smooth aggregate curve over millions of people. Text book presentations often show a demand curve with steps being made smooth by drawing a line through it.

Luis Enrique

also, never mind *existence* what matters is *usefulness*


As you imply in the first sentence, you don't buy a house. You buy housing. I agree that they are bit lumpy and bundled as goods go but there are plenty of granularities to the product to keep demand curved in funny ways.

BTW, Europeans I know seems to think it is utterly bizarre that we measure housing stock in 'bedrooms' rather than size. All the lumpiness of our housing market isn't essential to the way the market must operate, but the bizarre way it has been constituted in this country.

Luis Enrique

this is a bug bear of mine. there's a difference between saying "aggregate demand curves (or production functions) don't exist" - which usually means you cannot analytically derive a function with aggregate arguments by aggregating many individual functions - and "aggregate demand curves (or production relations) cannot be usefully approximated by a simple function with aggregate arguments "

there are two many people out there who have read the former and leap to the latter

Luis Enrique

that's right. two.

Ralph Musgrave

“Europeans I know seems to think it is utterly bizarre that we measure housing stock in 'bedrooms' rather than size.” That’s because half of all Brits don’t understand what a square meter is. Though most of them can count up to four, which is the maximum number of bedrooms they’re likely to require.

When the Germans want to be rude about us, they refer to us as “the stupid people who live on the island”. Germans do have a point there.


@ Luis - "A downward sloping demand curve does not ask that every individual would buy more as prices fall. Even if demand increases in aggregate as prices fall only because the marginal non home owner switches from demanding 0 to 1, that would still give a smooth aggregate curve over millions of people."
This begs two questions. One is: does aggregation really give us a smooth curve? This should be established, not assumed.
Second, if non-smooth individual demand curves do give us smooth aggregate ones, don't we have to rethink the need for macroeconomics to have microfoundations?
Separately, we might question the usefulness of demand curves. I suspect that few businesses actually know the demand curves for their products (unless over a small range), but rather use trial and error.


Surely all demand curves are an aggregate rather than an individual function! If house prices fall I certainly will buy more. I sold my house at the end of 2009, and have been renting since. That is partly out of circumstance but also on the basis that I apply the basic principle that you should buy low and sell high. I have sold at the peak of the market, and look forward to being able to buy when I expect prices to be significantly lower either here, in Spain or if not in Ireland.

There are today millions of workers in Britain who cannot afford to buy because house prices are in such a gigantic bubble. They certainly would be able to demand houses if prices fell. Moreover, there are many people who would like to trade up from their existing house, but can't afford because prices have risen so much.

If you had a house you bought at £50,000 and had your eye on a £100,000 at some point in the future when you'd saved the difference, you only needed to save £50,000. If house prices have quadrupled as they have done since 1997, then although your house is now valued at £200 k, your desired house is now £400 k, leaving you with a whopping £300 k to save instead. Put another way you have become £250k worse off as a result of the rise in prices.

If prices fell far enough I might even be prepared to join the speculators in buying more than one house again on the principle of buy low and sell high, because in such conditions reversion to the mean always results in prices overshooting to the downside. Prices are estimated to be around double what they should be, but I expect on past performance prices will fall by around 80%, before correcting back up to the average over several years.

Frances Coppola

I'm broadly in agreement with Luis: it would only need a fairly small proportion of potential buyers to increase demand for the demand curve to slope downwards in aggregate. However, as the housing market is mainly a secondary marketplace, if increased demand came only from non-home owners prices would rise, not fall.

Chris, I do think there are problems with microfoundations for macro. Representative agents are seldom really representative, and the risk of fallacies of composition is extremely high.


There are plenty of people sharing homes who would rather own their own. E.g. Children with parents, friends sharing a rented house, divorcees unable to afford a home each.

Luis Enrique


it's probably the case that most interesting questions about housing would require modelling separate markets (regions, house types) but even these would be aggregations on some level.

I don't understand the point about smoothness. It's an analytical convenience, not an assertion about reality. If real demand curves have large steps that would matter, if they are just a bit jiggly, approximating with a smooth curve is fine.

I find this line of attack on economic methodology puzzling. People will happily use verbal arguments in which simple relationships between aggregates are asserted, empiricists will refer to what are usually estimates of smooth parametrized functions (log Y = alpha log K + beta log L) yet as soon as somebody writes down Y=K^alpha*L^beta everybody wigs out

as for microfoundations, well it's one to thing to desire models rooted in people's behaviour, to some extent, and another thing to insist that every aggregate model can be derived perfectly from models of individuals with no skipped steps, helpful but strictly unrealistic assumptions or approximations. I'm with you on emergent behaviour etc. but models of that are microfounded too.


Lots of people need to read up on their Steve Keen (and others)who has been saying this for years. All those graphs in undergraduate textbooks are the merest fiction. The idea that macroeconomics has to be derived from microfoundations is another example of this specious bullshit.

Greg vP

What would be the textbook response to this?

An individual's demand curve for 'lumpy' goods like housing could be thought of as like a "probability of buying" at a given price. In your own case, Chris, that stays below 1 even when the price falls to 0.

Next, houses are highly differentiated products. There is not a market for housing, there are many tens of thousands of markets, for housing in particular localities and having particular characteristics. In effect there is monopoly pricing power for each house, and monopolies face downward-sloping demand curves.

With Anconan fish, various factors might cause individual buyers to buy more of more expensive fish.

Perhaps some people buy more of more expensive fish because they reason that they must be better (information asymmetries). Perhaps there are religious festivals involving fish that drive up demand temporarily (exogenous factors). Or perhaps some buyers pay more because the fish really do taste better at some times of the year (product quality variation).

With the possible exception of information asymmetry (and the herd-following heuristic used to compensate for it), there are no networks in the standard story. Only individuals.

Hmm... this all sounds rather like the epicycles that were invoked to protect the terrocentric view of the universe before Copernicus: Lakatos's "protective belt of auxiliary assumptions" that makes it impossible to refute a core proposition using evidence. (In this case, the core is methodological individualism.)


I've been playing with a model of house prices where the supply curve is vertical (at least in the short term) and the demand curve is upward sloping. (Higher prices make it more urgent to get on the housing ladder.) A disturbance from equilibrium would leaad to ever rising (or falling) prices until stopped by some exogenous factor.

Seems to fit the facts.

Jonathan Monroe

Two easy points:
1) In the market for owning homes (rather than the market for living in them) the marginal buyers are BTL speculators. We know that goods which are mainly bought speculatively don't have neat downward-sloping demand curves, hence bubbles. But if BTL investors behaved rationally, then a fall in house prices (with no corresponding fall in rents) would make BTL more profitable, causing an increase in BTL demand far larger than any increase in owner-occupier demand.

2) If you make the relevant unit bedrooms (or the more logical square metres) rather than houses, then individual demand curves do slope downward. If houses were cheaper (this is the market for houses to live in, so this applies to rents and prices), then my wife and I would have an extra bedroom for the home office. As it is, we use the attic.

The comments to this entry are closed.

blogs I like

Blog powered by Typepad