Society got very excited about the invention of the car. From the 1950s onwards especially, governments around the world began huge road-building projects, bulldozing swathes of cities for huge multi-lane dual carriageways. This frenzy has subsided, and it’s largely clear how this ideological expansion reformed our cities for the worse.
Highways are a crucial part of a nation’s infrastructure. Turnpikes—privately-funded and toll-financed roads—were, alongside the private waterways network, a key part of the earliest stages of England’s industrialisation. Their high tolls—the rents earned by those who built them—were incentives for the creation and rapid expansion of the world’s first, and for a long time the greatest, railway network. So much for natural monopolies: price competition was so vigorous that Mancunian workers and industrialists (and eventually the city corporation) funded a canal that made the city, 40 miles inland, the world’s third biggest port.

But it’s all too easy to go from thinking something is beneficial in particular contexts to thinking we should expand it indefinitely, even outside of that context. In post-war Britain, infrastructure investment was not driven by the profit motive—it was driven by a conviction that cars were the future. Birmingham is, of course, an especially prominent UK example of this. But it happened almost everywhere, and London was only just saved. The US midwest provides even more extreme examples. Of course, the more heavily roads and parking were subsidised, the more rapidly transit declined.

An insightful short newsletter, which I saw online like a blog post, applies this insight to broadband.
Our rampant enthusiasm for involving computers in everything reminded me of the last century’s most important technological revolution, the automobile, which turned out differently than it began. Like computers today, cars were once so novel and promising that we “did the things that let us use cars” rather than the inverse.
That honeymoon period lasted long enough (and for some it hasn’t ended yet) to entirely and permanently remake the infrastructure of cities, the most familiar example being the wave of support that enabled Robert Moses to slice New York apart with expressways. By the time the car’s real problems became visible—traffic, pollution, physical danger, and social atomization—it was too late to reverse its complete dominance. Most of the rebuilding was already done.
Right now, we’re building a digital infrastructure that rivals the physical one we built for cars, just as fast and with equally uncritical enthusiasm. If freeways prioritized cars at the expense of people, this prioritizes bits similarly, no less because we’ve all begged for it. Of course, it’s still too early to say what the internet version of gridlock and global warming will turn out to be once its saturation of society is complete.
In the eyes of this author, it’s Zuckerberg and Jobs we should be worrying about, but I’m not sure. See, like private road, rail, and waterways, Facebook and Apple have automatic feedback mechanisms that help guard against both overexuberance and underexploitation: profit. Yes, private firms make crazy decisions, but when they do they get wound down, automatically. They only stay alive, assuming a decent institutional framework, when they satisfy preferences. Uber’s venture capitalist investors can only subsidise riders while they have money in the bank. Historically we can see that government-funded boondoggles have not faced this discipline.
So I’m sceptical about the clarion calls I hear absolutely everywhere that we should spend billions of pounds, or hundreds of billions around the world, building broadband because the internet is the “technology of the future”. It probably is, but unless we have incredibly solid evidence that broadband has large net positive externalities, then we should let it roll out privately; when people are willing to pay how much it costs society to build. Real technologies of the future fund themselves.