The Metro Curitiba Never Built
Curitiba has been about to build a metro for most of half a century and has never built one. What it built instead, beginning with two bus lanes in 1974, is the system that something over 180 cities have since copied — usually without the zoning law that was holding it up.
The tube station is a glass cylinder on the central reservation of a six-lane avenue, up three steps, with a turnstile at one end and an attendant behind it. You pay there. When the bus arrives it is twenty-five metres long, hinged twice, and its floor is exactly level with the floor you are standing on. Five doors open at once, the tube empties into the bus and the bus empties into the tube, and the whole exchange takes about as long as it does on an underground platform. Then it goes, in a lane no car is allowed into, and the next one is a couple of minutes behind.
There is no metro under this street. There has never been a metro under any street in Curitiba, a city of something under 1.8 million people that had about 180,000 in 1950 and grew through the fastest urbanisation in Brazilian history without laying a metre of urban rail. That is not an accident or a shortfall. It is a decision taken in the middle 1960s and defended, by a specialist agency created for the purpose, through every administration since.
The plan that chose a shape
Curitiba's first serious master plan, drawn by the French urbanist Alfred Agache in 1943, was radial and concentric: boulevards running out from a centre, rings around it, growth in every direction at once. Very little of it was built, which turned out to be lucky, because by 1965 the city had a different plan and the opposite idea. The Plano Preliminar de Urbanismo argued that a city growing this fast should not grow outwards in a circle but along a small number of structural axes, and that everything else — where the tall buildings go, where the density is allowed, where the buses run — should be hung on those axes deliberately.
The second half of that year the city created IPPUC, the municipal institute of urban research and planning, whose job was to hold the plan steady while mayors came and went. This is the part usually left out of the story, and it is the part that did the work. In 1971 Jaime Lerner, an architect who had worked on the 1965 plan, became mayor — appointed rather than elected, because Brazil was a military dictatorship and state capitals were treated as matters of national security, which is an uncomfortable but unavoidable fact about how quickly the next few years moved. In 1972 he closed Rua XV de Novembro to cars over a single weekend, kerbs and planters in place before the shopkeepers could get to a judge. In 1974 the first express bus lanes opened on the north–south axis.
The line on the zoning map
What went in was not a bus lane. It was a cross-section, called the sistema trinário, repeated the length of each axis: a central road carrying two exclusive bus lanes with slow local traffic beside them, and one block away on either side a one-way street running in the opposite direction, taking the through traffic that would otherwise be in the middle. Then the zoning code was rewritten to match. The tallest buildings in Curitiba are permitted only on the structural axes, directly over the busways, and the permitted height steps down block by block as you move away from them. A developer who wants density has to build it where the bus already is. That is the invention. The articulated vehicles and the dedicated lanes were engineering; the coupling of a transport corridor to a land-use law was politics, and it is the reason the corridors filled with passengers instead of waiting for them.
A bus lane is a line painted on a road. What Curitiba drew was a line on the zoning map, and then spent fifty years refusing to rub it out.
Ninety seconds, and who pays for them
The network became the Rede Integrada de Transporte in 1980: one flat fare, terminals where you change lines without paying again, and therefore a deliberate cross-subsidy from short trips in the centre to long ones from the edge. In 1991 came the estação tubo, the glass boarding cylinder, which moved fare collection and the step up to the vehicle off the bus and onto the platform. In 1992 Curitiba put the first bi-articulated bus in the world into service — twenty-five metres, five doors, around 270 passengers. All three changes attacked the same number, which is the seconds a vehicle spends stationary with its doors open; get it low enough and a bus lane carries what a light rail line carries, at a fraction of the capital cost and without a decade of tunnelling first. The other half of the machine is contractual and nobody photographs it. The buses are run by private operators, but they do not collect their own fares and they are not paid by the passenger: URBS, the municipal company, takes the whole fare box into a common pool and pays each operator for the kilometres it has actually run, to the timetable it has been given. In most Brazilian cities of that era drivers were effectively paid per head, which produced the guerra do centavo — the penny war, buses racing each other to the stop, bunching, overtaking, hanging back to fill up. Curitiba removed the incentive instead of policing the behaviour, and got a network that runs to a headway because running to the headway is what it is paid for.
What travelled and what stayed
In December 2000 Bogotá opened TransMilenio, designed after a close study of Curitiba, and it became the proof that the pattern worked at eight million people rather than one and a half. After that it went everywhere. The Institute for Transportation and Development Policy published the BRT Standard in 2012, turning the Curitiba cross-section into a scored checklist a city could be graded against, and the Global BRT Data project now counts bus rapid transit running in something over 180 cities on six continents. What travelled was the corridor. The zoning code mostly did not — it is far easier to buy articulated buses and paint a lane than to rewrite a city's land-use law and then hold the line against every landowner who wants density somewhere else. The result in a good many cities is a fast bus running down the middle of a road nobody is allowed to build next to: the corridor works, the ridership never arrives, and the conclusion drawn is that bus rapid transit is a second-best solution, when what was actually tested was half of it.
The bill at home
Curitiba's own network is not in good health. Ridership peaked in the 1990s and has fallen more or less steadily since, with a step down in 2020 that has not been recovered, and the city has one of the highest rates of car ownership of any Brazilian state capital — the irony the eco-city brochures leave out. The fleet is still overwhelmingly diesel, with electric buses arriving in small numbers and late, and the metro has been announced, tendered, re-announced and shelved repeatedly since the 2000s without a shaft being sunk. The deeper problem is the boundary. The plan governs the municipality of Curitiba; the growth went to the twenty-odd municipalities of the metropolitan region around it, where land was cheap, the zoning code does not reach and fare integration is partial, so the people with the longest and most expensive journeys here are the ones the famous system was never designed for. Jaime Lerner died in 2021 and the argument he started is now about whether a fifty-year-old municipal plan can be made to work across a region that ignored it. None of which cancels what was proved on that avenue in 1974. It means the proof has an expiry date on it, like every other piece of infrastructure, and somebody has to decide to renew it.