Usually when people discuss housing they focus on the first-order costs. Too few houses means higher housing costs, which lowers people’s standard of living. Since rent or mortgage payments make up most people’s biggest single expenditure after tax, especially for poorer people, reducing housing costs seems like one of the best ways of helping to raise people’s standard of living.
These high rents and house prices are a big enough problem on their own. But prices change people’s behaviour, and expensive housing costs could have significant effects on where people choose to live. This in turn could have significant effects on productivity and GDP per capita. I will first try to explain the mechanisms behind this and then the empirical evidence around it.
Chelsea is nice but expensive, so I choose to live in grottier but cheaper Stockwell, south London, instead. But Stockwell expensive compared to Hull – so many people are choosing to live in Hull when, if housing costs were the same, they’d prefer to move to London.
That might matter economically because in London there are more and better job opportunities, in general, than in Hull. Larger cities tend to be more productive per worker than smaller ones. The costs of matching workers and firms with each other in mutually beneficial ways are smaller, so businesses have a larger pool of workers to choose from. Knowledge transfers become easier too, with workers, entrepreneurs and managers being better at learning from others when they’re close to them.
So expensive housing keeping workers out of London is harmful for perhaps two other reasons – it stops them from getting the job that they would be most productive in, and on aggregate may be preventing business innovations from taking place that would raise the productivity of workers around them, as well as their own. Sheer size is not the only thing that matters here – a good computer programmer might be better off in Silicon Valley than in New York City, because the population of relevant jobs, firms and workers is still larger in San José than in larger-overall NYC.
Empirically we have quite a lot of evidence in support of this view. Not only are big cities more productive than smaller ones, a study from Spain shows that workers who move from small to large cities gain a wage premium when they do so and accumulate better experience as time goes by – experience which persists even if they leave. In the United States, productivity per worker rises by 11% with each doubling of city size, “whether from 10 to 20 thousand or from 1 to 2 million”.
However, this average effect is subject to a lot of variation between cities, and seems to be driven by skilled workers – unskilled metro areas do not tend to become more productive as they get bigger. Work by Ed Glaeser suggests that the bulk of the evidence supports the knowledge transfer theory that cities are more productive because they allow people to learn from one another more easily. It is not that cities automatically become more productive as they get bigger, but that they create more opportunities for entrepreneurship and innovation that raises productivity.
This doesn’t mean that there would be no benefit from poorer people moving to those cities, where wages are generally higher even at the bottom of the skill distribution, just that the productivity gains may be lower overall. And there is no suggestion that low skilled workers make anyone worse off.
Back to our computer programmer. If she and thousands of others like her are prevented from moving to where they’d be most productive, we would expect not just individual effects but noticeable effects on overall economic growth. Even if she was a less skilled worker, her income could be much higher working in many jobs in a relatively high productivity city.
The total effect of this ‘spatial misallocation’ in the United States has been estimated by Chang-Tai Hsieh and Enrico Moretti to be somewhere in the order of 13.5% of GDP, more than two Great Recessions’ worth of growth. This is driven mostly by regulatory constraints on housing supply in places like New York, San Francisco and San Jose. Cutting planning regulations to the level of the median city, making it cheaper and easier for people to move to where they’ll be most productive, could boost US GDP by 9.5%.
A more recent paper by the same authors, which looked at 220 metro areas, found such a large GDP estimate that I can hardly believe it myself: that housing supply constraints may have lowered aggregate US GDP growth by almost 50% between 1964 and 2009. The authors also note that one way of mitigating the effects of tight constraints on housing is to have a good transport network, like London’s – though London’s commuter train network is now highly constrained by the green belt and in-city building regulations. Other recent evidence points to high housing costs stopping lower-skilled workers from moving to more productive parts of the US.
A rough calculation by London YIMBY, author of an excellent recent paper co-published with the ASI, suggested that the GDP hit to the UK of our own housing constraints might be in the region of 25-30%. These should be taken with a pinch of salt but suggest that ‘spatial misallocation’ is at least as important a part of the housing story as high housing costs themselves are. It may be that highly skilled Northerners are mostly not put off moving to London, but less skilled ones are and could have much higher earnings if not for the price of housing.
We have suggested various solutions to this problem for many years. But the purpose of this post is to illustrate that expensive housing is something that makes us all poorer, even rich homeowners, through lower productivity, lower GDP per capita, and less money available to pay in tax. The first order effects of tight supply constraints and expensive housing are bad enough, but the second-order effects may be even worse.