Last month, Walter Russell Mead published an op-ed in the Wall Street Journal criticizing our fixation on infrastructure. Mead acknowledges that building canals, railroads, bridges, and the like, is a traditional function of federal and state government. But he argues that continuing to do so in the future is bad policy.

The main reason, according to Mead, is that the Internet makes the physical movement of people and goods much less central to prosperity. In the 21st century:

The challenge isn’t to move more meat [for example]; it is to move more information more effectively, and to re-engineer business practices and social organization to take full advantage of the extraordinary efficiencies that the Internet affords. The rush-hour rituals of the 20th century aren’t destined to continue to the end of time. Telecommuting, flextime and mini-commutes to satellite offices will change the way we work.

This argument is superficially appealing. It’s also dangerously misguided. Mead is correct that brick-and-mortar projects are not a cure for unemployment. But he’s wrong to conclude that they’re unnecessary because we can all just telecommute or do our shopping online. Retail, hospitality, healthcare, and other industries that require on face-to-face interaction are currently and are projected to remain among the largest sectors of employment. People need to show up for those jobs.

There’s no better illustration of this point that the ongoing effect of Hurricane Sandy. As Nicole Gelinas points out in City Journal, the greatest threat to New York’s economy is the destruction of physical infrastructure like power stations and the immobilization of its transit system.

You might imagine that Wall Street bankers, say, could just log on from home. But they can’t work when the lights are off. And they’ll starve unless an army of cooks, salespeople, dishwashers, and so on can make the long journey from New Jersey and the outer boroughs to the more fashionable neighborhoods where they’re employed. As Gelinas puts it, “the info-economy is utterly dependent on the unfashionable infra-economy. If the river annexes your subway tunnels and electrical substations, no government agency heroically intervenes, and grocery stores stay shuttered, you aren’t going to be designing social-media apps in your bedroom.”

Gelinas’s reference to government is crucial. Infrastructure systems like transport networks and the power grid  are simply too big and complicated be built and sustained by private industry alone. They require government support to function even under ordinary conditions, let alone the challenges posed by a natural disaster.

Many conservatives dismiss such support as wasteful. They’re wrong: subsidies for infrastructure are economically productive when they’re directed toward real needs, such as moving people around New York City. It is politically difficult to limit support to useful purposes: Amtrak loses money, for example, because it’s forced to provide little-used long-distance service rather than focusing on the Northeast Corridor, where it is competitive with travel by air or car. And there is such a thing as overinvestment in infrastructure: China has had problems with building too much, too fast. But fights about the best use of limited resources are worth having, rather than throwing the baby out with the bathwater.

American statesmen used to understand the role of infrastructure as a condition of prosperity rather than an end in itself. For some of the background, see James Pinkerton’s piece on the “American system”. We are, admittedly a long way from the Erie Canal. As our governments face tough decisions about where and when to spend, however, we should not let promises of a wired future blind us to our economic and political history.