People prefer neo-traditional buildings

It seems obvious to me—and I think to most people—that housing built since the 1930s is by and large much less attractive than housing built before. But if this is true, and if we are much richer now than we were in the 1930s and before, then why would we build, buy and live in housing we don't like? We have some sort of market in housing; surely if we really all preferred traditional housing styles we'd just buy it.

A new paper (slides) provides the answer—at least if we can assume the UK and the Netherlands are similar in this respect. The authors look at a large database of new-builds and sales and compare similar neo-traditional houses to houses with some traditional features and those with none. They find that, even controlling for a wide range of features, fully neo-traditional houses sell for 15% more than fully non-traditional houses. Houses with references to tradition sell for 5% more. We might reasonably speculate that truly traditional houses sell for yet more.

In their words:

Popular reports on the housing market often refer to attractive style characteristics of houses. In the case of the Netherlands specifically housing from the 1930s is very popular. It is, however, difficult to disentangle the attractive vintage effects of the dwelling from (often inner city) locational amenities.

This paper studies exactly these attractive physical features without the confounding influence of age and location effects by studying newly built houses in newly developed neighborhoods only. A rich data set of housing transactions in the Netherlands is enriched with style characteristics of houses on 86 (Vinex) housing estates across the Netherlands. This resulted in over 60,000 transactions between 1995, the starting point of the development of these sites, and 2014.

The hedonic price model that has been estimated shows a significant price premium of 15% for pure neo-traditional styles and 5% for referring to traditional styles. Various robustness checks confirm that these results are partly, but not entirely, driven by, e.g., unobserved differences in quality between houses with different building styles.

The riddle is why—if this premium exists—do developers not build mainly or purely in the neo-traditional style, to reap profits from satisfying market demand. Why do developers only build neo-traditionally—why don't they really try and ape the creations of the past? The authors blame tight regulation of both the volume of production and the inputs; local authorities effectively prescribe modern styles and proscribe prerequisites for traditional design. They find that construction cost has only a marginal or negligible impact, by contrast.

This research is especially plausible, as it turns out London and the Netherlands probably are similar in the relevant respect: some of our housing problems are caused by similar factors. Everyone wants to live in beautiful terraces, right? But new housing usually looks nothing like the most popular existing stock.

Nicholas Boys Smith and Create Streets provide an elegant explanation (pdf): building codes in London make popular traditional housing—which is very dense and could be sold very profitably—near-impossible. They blame six key barriers including: bans on recycling dead space between buildings into gardens; universal lift requirements; illogical value calculations; staircase width rules; and excessive wheelchair requirements; as well as many others with smaller individual impacts.

Top-down planning ruins cities, wherever it is tried. If we loosened regulations on the volume of building, and the type of buildings that could be built, then we could massively increase London's density while simultaneously providing the sorts of dwellings people actually want to live in.