Thursday, June 26, 2008

Andrew Leigh on house prices & the value of a quality public education

Andrew Leigh is an amazingly active economist and does work of real social value. The current Economic Record has a piece by Andrew and his colleague Ian Davidoff where he values public school education in the ACT by looking at the effects of better than average test scores on house prices. A preprint of the whole paper is here.

I have a question for Andrew. If you get an education benefit from locating in a suburb that provides access to better-than-average public schools that means that house prices should rise to internalise that benefit. But doesn't it also mean that house prices should subsequently grow more slowly to account for the non-residential benefit? For example suppose real estate is increasing at 7% per annum on average everywhere. If you get some education benefit from living in a living in a particular location (say it is worth 2% of the value of the house) doesn't arbitrage mean that house prices should grow at the slower rate 7%-2% = 5%.

I wonder about this because I observe house prices in suburbs like Kew and Balwyn in Melbourne. The prices of these houses are high partly because they are very near good public and private schools. My theory suggests rates of capital appreciation should be slower in these suburbs - a prediction at variance with the facts. These suburbs are galloping away in terms of rates of capital accumulation.

Am I confused? I have asked many people about this over the years and remain none the wiser. The best discussion I had on the topic was with Ted Sieper a decade ago - he was adamant that prices in suburbs offering education benefits should grow more slowly than the market as a whole.


Joshua Gans said...

Harry, house prices are only observed on trading. So that means that you have to consider who the buyer and the seller is at the time. If everyone has identical but for the ages of their children, there would be a large incentive for those who children move out of school to sell to a family with children about to move into school.

However, in that transaction, the school quality control only impacts on demand. While the general state of growth elsewhere in the city impacts on willingness to sell. For that reason, you would expect the house price growth rate to be the same in all suburbs.

That said, Andrew bought a house a little while back after writing but before publishing this paper. Hmm, any inside rents there?

Andrew Leigh said...

Harry, that's a superb line of thinking - and one I confess hadn't occurred to me before now (I blame the fact it doesn't appear in the literature).

It sounds to me as though your reasoning is perfectly sound, though I suspect it would be difficult to see in the data. As Robert Frank points out, rising top incomes (as we've seen in Australia) have the effect of pushing up demand for the best schools. So that'll work against the expected slower appreciation of the school component (which will be very hard to discern in the data).

Andrew Leigh said...

Looks like Joshua and I commented simultaneously.

As to our own purchase, I'm embarassed to say that we were so relieved to find a decent house in our price range that I didn't even bother to look at the quality of the neighbouring schools.

hc said...

Don't get it Joshua. The implication of your argument is that I can send my kids to a quality public school for nothing provided I am not wealth constrained.

Suppose I have $1m in wealth and I can live in Balwyn by paying $1m for a house there and send my kids to a very good school, say, Balwyn High. Alternatively I can live in the same quality house in Bundoora for $400,000 and invest the $600,000 in real estate but have to pay $20,000 annually to send my kids to a private school since they are ineligible for Balwyn High.

Unless the rate of return on housing is lower in Balwyn I am always better off living there. I get the market rate of return on my whole $1 million compared to getting the same return less the cost of the private school in Bundoora.

Markets don't 'work' here to eliminate by arbitrage the profitable move to Balwyn.

derrida derider said...

I'd suggest it's an easier hypothesis to test by looking at the US data, because school quality is far more variable there (a reminder of the very real effects central versus local funding can have) and is known to be a very conscious factor in location decisions there.

Kaleberg said...

Why should the rate of appreciation slow? The value of a good education may be rising more rapidly than the value of a good place to live. Good school systems are hard to scale and require a long term communal commitment to maintain and improve.

Developers are more likely to put in a golf course than a good school system as an incentive. Golf courses are cheap when compared to good school systems.

The effect of having a good school system is similar to having waterfront property. It is much easier to create houses than waterfront, so waterfront property rises in value more rapidly than non-waterfront property. There is no slowing in growth as the differential is internalized by the pricing.

The private school argument ignores the fact that good private schools are under no obligation to accept your child. In fact, the ability to select and expel students is what allows private schools to maintain their quality. Public schools, even good public schools, have to accept all children in their district. You might win by taking the gamble, but you might also go for a sure thing.

rabee said...


Doesn't the Modigliani-Miller theorem state that
two firms that are identical except for dividends have equal in market value?

So I don't see your investment analogy.

rabee said...


A quick back-of-the-envelope calculation based on
MM shows me that the "rate of growth"
should be independent of benefits accrued from being close to a good school.

[On the other hand. Is the rate of growth house prices in high rent price areas lower than the rate of growth in low rent price areas? HMMM]

The Modigliani-Miller theorem was surprising at the time.

hc said...

Rabee, Doesn't the MM theorem on the irrelevance of dividend policy simply stem from the fact that, in the absence of taxes, profits retained by the firm will be invested at the market rate of interest and give equivalent net worth to the investor through capitasl gains.

Does this have any applicability here?

Don Harding said...


You are assuming that the school location benefit is fixed. But its determined by the additional expected discounted earnings of your children which are determined ultimately by labour augmenting tech change.

The value of an additional unit of housing (the house price) is equal to marginal rate of substitution between housing services and leisure which is ultimately determined by technology.

Thus house prices grow at the same rate in suburbs with poor schools as good schools because the two components of house prices are ultimately determined by the same thing (ie technology).

The relative price of houses reflects initial conditions and this will be influenced by differences in the quality of schools etc.

Francis Xavier Holden said...

In this model (house prices and public education) what happens before children are say 10 years old and after secondary school - say youngest 18. The "good" suburbs are, usually, older and have a fixed housing supply.

Shouldn't they eventually block up. Why do people sell in these suburbs?

The transaction costs involved in shifting from a shocking downmarket Commodore suburb like Bundoora to 4WD BMW country like Balwyn seem high when combined with the increased capital costs if it is just to get a freebee private school equivelant education.

conrad said...

"Shouldn't they eventually block up. Why do people sell in these suburbs?"

To retire -- I live in one of these suburbs, and that's what the guy I bought my house off was doing.

The other thing that sent the prices through the roof it appears is speculation. Whilst economists might think through this all logically and rationally based on theory, I think in times of irrational speculation, many investors think that good schools equal a secure long term investment (somewhat irrespective of the initial price), including better rent in the long term due to schools get worse.