There is a debate in the blogosphere regarding Soos and Egans new ebook about Australian housing bubbles. Quite apart from their overall charting of housing bubbles back over nearly two centuries and the economic theory they use to understand it in one chapter they criticise the assumptions used by free market urban theorists such as Edmund Glaser, Ryan Avent and Matt Yglesias and known as the ‘Urban Containment Hypothesis’ (page 657 onwards). Note I hesitate to use the term ‘market urbanism’ as sometimes used to describe this concept, as this often covers those who believe that cities would and should be denser without planning controls, no I am using the term here specifically to cover the ‘urban containment hypothesis’ that controls on city size make housing costs/urban containment policies push house prices above levels to which they otherwise would be, and hence fuel bubbles. The solution, propagated in several market urbanist books in the last few years is simply to let cities grow, upwards and outwards, and this would prevent housing bubbles forming. Soos and Egan criticise this position noting that bubbles formed in housing markets well before contemporary planning controls. They note however that then there was
“….an inability to build homes outside the immediate city centre given limited transport options; travelling by foot, horse, or carriage….”
Soos and Egaan are criticising the sole focus on the supply side of the urban containment thesis, wanting greater focus on credit creation. Cettinly the empirical evidence is that credit gropwth is much more importaant than supply conatrsints (see this recent study by the OBR), however there are valid criticisms to the potential conclusion that this does not ultimately arise because of supply constrants, some of which are picked up by Phil Best who states
‘the paradigm shift in land markets that occurred with automobile based development, which for the first time brought sufficient land within transport system reach, in the process diminishing economic land rent.’
He quotes Robert Murray Haig’s much cited (1926) “Towards an Understanding of the Metropolis” which first theorized that the growth in automobile use would push down land rent and suppress bubbles.
Here I wish to focus on this very specific point. If physical supply constraints can create bubbles and their removal can postpone or remove them should we be removing urban containment policies?
Lets start with the hypothetical but unrealistic position of no supply side constraints on housebuilding at all. There are no inventory constraints, no regulatory constraints, no finance constraints for developers, so any increase in demand is immediately reflected in an increase in demand. If this situation persisted and everyone expected it to persist there would be no scope for any speculation, save from the effects of a natural disaster. In the real world supply does not react instantly to demand, building work takes a long time. Homer Hoyt and Henry George both suggested it was this friction in the supply of land which creates the conditions for land price speculation.
There are two kinds of frictions. One is the demand elasticity of supply, the slope of the supply curve, which restricts the flow of new housing services. Only a small part of the stock of housing in any year relates to newly produced stock, what is mostly traded is existing assets. The second is absolute physical constraints, such as where a city hits the sea or shortages of construction materials which pose an absolute restriction on supply. Beyond this point the supply curve is flat, increases in supply does not result in an increased demand. This kinked nature of the supply curve for new housing is rarely remarked on. It is important though because it means that bubbles can both by created by rigidities in supply and be burst by expected shortages in supply or there removal. Lets take a case where a speculator assumes that supply will be fixed, and hence all increases on demand go to economic rent on a block of existing land. This gets fed into future options price on that land, speculation. Then if there is an unexpected new boost to supply that option price will be too high, people may have borrowed too much, the bubble may pop. Another example; there is an expectation there will be a steady new flow of land, but there is an expected supply constraints, of the kind experienced in Florida in the 1920s where a sink ship prevented building materials coming to Miami, that also burst a bubble as those who has borrowed to buy land could not build out sell houses and service their loans. So we can see that if supply is greater or less than that to which speculators expected in taking out loans can burst a bubble.
The hypothesis that the long post war period of steady growth without housing bubbles was in part due to the automobile, which significantly shifted the point at which the housing supply curve hits the vertical is probably correct. Here we are taking about bid rent curves, where people trade the generalised cost of time commuting against lower housing costs. The problem is transport networks have limits so beyond a certain physical line people will not commute. The day has only so many hour, you cant commute for 25 hours in a day, people need to eat and sleep etc. Beyond a certain point travel costs will exceed savings. This physical limit of expansion will be much larger in a city that opts for transit and high density around transit nodes as the benefits of lower rents will not be eroded by traffic congestion lengthening commutes. It will also be increased if there is state funded low income housing increasing supply. With the collapse of state funded affordable housing and less funding for infrastructure the social safety net against housing bubbles has been removed, there appears to be no direction left for public policy to prevent housing bubbles other than to deliberately promote sprawl, as has been the case in the recent neo-liberal shift in housing and planning policy in the England.
There is a problem however as sprawl dramatically changes the balance between the urban economies and dis-economies of aggregation. I have written about this in the past as ‘the broken city model’. For cities undergoing rapid car orientated growth eventually congestion will vitiate any advantages of expansion. New supply constraints will impose themselves. So it may have been the case that the rapid urban expansion in the years after the Great Depression may have delayed the previous boom bust cycles but the length of this cycle and the length of the great moderation was also elongated by welfare urbanism, investment in public transport, new towns, urban renewal and and affordable housing, combined with policies of urban containment, that maximise the urban economies of scale whilst minisming the diseconomies caused by growth. Now we are relaxing policies of containment whilst reducing transit investment, urban renewal and investment in affordable housing. A very potent and dangerous combination as it increases the potential for severe supply constraints caused by unplanned growth to reassert themselves and so increasing the potential for boom and bust. Contrast this for example with jurisdictions like Hong Kong, and Singapore, with have successfully introduced macroprudential controls on borrowing whilst at the same time investing in public housing and public transport. Indeed as Bloomberg notes it is likely that these policies have only been successful because the publicly funded housing has provided an alternative to the foreclosed option of home ownership. Sprawl simply replaces shorter smaller bubbles with more widely spaced and bigger ones of much greater systemic danger, that danger is increased when the sprawl is unplanned and not accompanied by public investment.
Soos and Egans are right to draw attention to the crude static equilibrium assumptions of the Alonso/Muth bid rent curve urban models used in regional science. What this draws attention to though is the need to replace these with a dynamic disequilibrium model of urban growth, credit expansion, city growth, bust and potentially city decline. One which would owe more to Homer Hoyts sector theory aligned to his sadly almost forgotten model of real estate cycles which underlay it, one with a more sophisticated model of the urban economy not assuming equilibrium and accounting for both spatial economies and disceconomies, a private and public sector.