A structured parking space takes about 30 square metres once you include the aisle, the ramp, and the circulation. Built above ground in a concrete deck it costs somewhere between a quarter and a third of what the same area costs as habitable space. Built below ground, with excavation, waterproofing, drainage and ventilation, it costs considerably more than the flat above it.
Nobody pays for it directly, which is the point. It is bundled into rent, into unit price, into the cost of a shop’s inventory. It is the most expensive thing in a city that appears on no invoice.
The minimum that became a maximum
Most parking minimums were written between 1950 and 1975, and the numbers came from surveys of car ownership at suburban sites in a period of rapidly rising ownership. They were then applied uniformly, including to sites with a tram stop outside.
Two things followed. First, the minimum became the norm: no developer builds less than required, and few build more, so a number invented from a small sample became the actual supply. Second, the required area started to shape the building. Below a certain site size, you cannot fit the ramp and meet the minimum without either going underground or losing the ground floor to cars.
That is why so many mid-century and late-century blocks have a dead ground floor. It is not an aesthetic decision. It is a ramp gradient.
What the space could have been
The arithmetic is worth doing on a specific site rather than in the abstract. Take an urban block of 4,000 square metres with a requirement of one space per dwelling:
- 120 dwellings requires 120 spaces, or roughly 3,600 square metres of parking — most of a full floor plate, or a full basement level.
- Built as a basement, that level costs approximately what 25 to 30 of the dwellings cost to build.
- Delivered as habitable space instead, the same area is 40 to 50 additional homes.
Reasonable people will disagree about whether those homes should be built. Nobody should be able to claim the parking was free.
Reform has a predictable shape
Cities that have removed minimums — and there are now enough of them for a pattern — see roughly the same sequence.
Developers build less parking, but not none: typically 0.3 to 0.6 spaces per dwelling in a well-connected area, which is close to what residents of comparable buildings actually own. Ground floors become usable, because the ramp disappears or shrinks. And the arguments about kerbside pressure move to where they belong, which is kerbside management — permits, pricing, and enforcement.
That last part is the bit reform advocates undersell. Removing a minimum does not remove demand for street space. It converts a construction problem into a management problem, and management requires an authority willing to charge for a public asset. Cities that skip that step get the backlash they were warned about.
The buildings left behind
There is a second-order question that architecture is only starting to take seriously: what happens to the decks already built.
Converting a parking structure to habitable use is hard for reasons that are unusually specific. Floor-to-floor heights are typically 2.6 to 2.9 metres, which is workable. Floors slope for drainage, at gradients of one to two per cent, which is not. Column grids are set out for cars — commonly 7.5 to 8.1 metres — which is generous for housing but awkwardly aligned with any sensible unit width. And the whole structure was designed with minimal fire compartmentation because it was open-sided.
The conversions that succeed tend to be to uses that tolerate a sloping floor and a hard grid: self-storage, light industrial, gyms, and occasionally studios. Housing conversions almost always end up pouring a levelling screed, and the weight of that screed is the reason many of them stop at feasibility.
Which is a reminder of the general rule: a building designed for one use with no tolerance for another is a building with a fixed end date. We built a great many of them for a vehicle fleet that is now changing faster than the structures can.
