Building Up, Pricing Out

AFew propositions in urban economics look safer than the claim that adding homes to a city makes housing cheaper. The mechanism is elementary: supply rises, competition for each vacancy falls, and rents follow. If that were all there is to it, the cities permitting the most new flats since the early 1990s ought to be the cities where rents have risen least. A synthesis published in 2019, pooling permit records from 1994 onwards across 47 metropolitan areas, found the pattern only in part. Twenty-six of the metropolitan areas showed rents rising more slowly than the national average after a decade of heavy building, fifteen showed no divergence at all, and six showed rents pulling ahead. The elementary mechanism predicts one response everywhere; one response everywhere did not appear. What has been argued over since is not whether supply matters, but at what scale and for whom it matters.

BMarisol Enríquez, an urban economist at the Institute for Metropolitan Studies in Rotterdam, holds that the synthesis measured the right thing at the wrong scale. Between 2003 and 2021 her group tracked advertised rents across 214 postcode districts in four countries, matching every district that received a completed building to a control district of similar age, income and transport access. Within 250 metres of a newly finished market-rate block, rents fell by a median of 4.6 per cent over the following three years, against 0.8 per cent in the matched controls. Metropolitan averages, on her account, drown this signal: a city can add flats in three districts while demand rises in thirty, leaving the citywide series flat although every street near a crane has grown cheaper. She is careful about scope. The effect appeared only in districts whose metropolitan vacancy rate stood above three per cent; below that figure, nearby completions produced no measurable movement in rent at all.

CThe account makes a prediction, and Enríquez has tested it. If competition for vacancies is what rents answer to, lifting the caps on flats per plot should deliver both more flats and cheaper ones. In 2015 four provincial cities in her sample raised permitted densities on the same day, and her group followed 1,180 eligible plots for six years. Completions ran to 9,300 units, roughly triple the previous rate, but nine tenths of them stood on plots already held in single ownership. On land divided among four or more owners the reform yielded 41 new units in six years, because no developer could assemble a site. Building, in other words, was effectively confined to ground where the ownership question had already been settled. Enríquez therefore treats permission to build densely as a licence for supply rather than an engine of it, and accepts that on fragmented land the licence goes unused for reasons her price model does not contain.

DIngrid Salvesen, who studies housing tenure at the Nordic Centre for Urban History in Bergen, replies that a rent series records prices rather than people, and that the two can move in opposite directions. Working from lease registers covering 38 districts, she followed the households themselves instead of the flats. In districts where advertised rents fell after new building, the share of tenants drawn from the lowest income quintile fell by 11 percentage points over the following decade, and the households that departed went on to pay a median of 9 per cent more in the districts they moved to. What looks like a district growing cheaper, she argues, can be a district exchanging one population for another, the saving accruing to arrivals and the cost to those who leave. She concedes two things. Her registers cover a single metropolitan area, and they record leases newly signed rather than every household in residence.

EThe dispute stopped being academic in 2022, when the rules governing density bonuses were rewritten. A developer seeking extra storeys must now show that the scheme will not cut the number of low-income households resident in the district, and applications that cannot produce the evidence are refused; the older procedure had required the municipality to demonstrate that displacement was likely before the bonus could be withheld. Salvesen points out, uncomfortably for her own position, that a district can lose its poorest tenants only once, so the mechanism she describes has a floor beneath it and cannot go on producing displacement for another century. Enríquez answers that a floor at least arrives on a schedule, whereas the licence she describes can be withdrawn by a single quarter of falling vacancy. Both regret the same habit: that the literature keeps reporting the advertised rent of empty flats, a figure the departure of poorer tenants can move on its own, instead of what a given household pays ten years later.

FWhat is emerging is not a verdict but a division of territory. In metropolitan areas that grow quickly and can build outwards as well as upwards, among them the Dutch Randstad and the Texan triangle, the supply account explains most of what happens to rent and turnover explains the remainder; in small historic cores hemmed in by protection orders, where a decade of building adds perhaps two per cent to the stock, the composition account is left in possession of the field. The awkward cases lie in between, and nobody has yet proposed a division both camps will sign. A larger question is untouched by any of it: whether the flats now rising in dense districts are still serving as primary homes thirty years on, or have drifted into second homes and short lets that no rent series records. No sequence of measurements yet runs long enough to say. What both researchers reject is the assumption written into the earliest housing targets, namely that a count of completions can be read as a report on affordability.