WASHINGTON - Seven years ago this month, the Interstate 35 West bridge collapsed into the Mississippi River during evening rush hour in Minneapolis. Thirteen people were killed; more than 100 were injured. That same summer, a steam pipe exploded at Lexington Avenue in Midtown Manhattan. The scalding pillar of 400-degree steam shot more than 1,000 feet in the air. The crater it left was 35 by 40 feet. The New York Post described the explosion as "Midtown's Volcano."

America's infrastructure problems are not limited to these two memorable episodes. The New York Times reports that there were more than a dozen steam pipe explosions in New York City in the 20 years before the volcano. According to the Federal Highway Administration, only about half of the vehicle miles traveled on federal-aid highways have pavements with good ride quality, and one in five bridges are classified as deficient. The World Economic Forum's report on global competitiveness ranks the United States 19th out of 148 economies in the quality of our overall infrastructure. As far as infrastructure goes, our "shining city upon a hill" isn't as bright as many other advanced economies.

This is something we all know.

Yes, the claims of an "infrastructure crisis" by President Barack Obama, many liberals and the American Society of Civil Engineers - not exactly a disinterested party - are exaggerated. And yes, existing laws and regulations turn infrastructure projects into boondoggles that take an order of magnitude longer to complete than necessary and cost more than they should. So conservatives are rightly skeptical of infrastructure projects.

But conservatives should put that skepticism aside and proceed - as always, with apprehension and great prudence - with a program of infrastructure investment.

We shouldn't follow the left's approach to infrastructure stimulus, calculating the number of jobs we'd like to create, using that to estimate the amount of government spending we need. Throw in politics, and that is a proven recipe for poor outcomes - the 2009 "stimulus" package was very poorly designed and properly gave infrastructure a sour taste.

Instead, a conservative approach to infrastructure would begin with a question: What are some projects that we actually need to fund? We all know by now that "shovel ready" projects are rare. So we should take some time to actually figure out which projects offer the highest value to society. We should put in place a multi-year program of infrastructure investment, not sugar-economics Keynesian stimulus.

Who should pick the projects? Primarily states and localities, where most of the responsibility and relevant knowledge lies. They should be free to figure out how to fund their share of the costs, as well. After identifying high-social-value projects - upgrading shipping ports and airports; repairing power grids, bridges, roads, and schools; working to shorten commute times; upgrading everything with 21st-century technology - we should move (some of) that spending from the future into the present. A multi-year program will help growth and employment over the next few years, when the economy will probably still need a boost.

More important, a conservative infrastructure program should prioritize a conservative goal: Helping the working class to rise. What, specifically, could this help look like? We know that urban areas characterized by a high degree of socioeconomic segregation often have relatively low mobility rates and high unemployment rates. One way to support employment and earnings is to spend money on transportation infrastructure to connect low-income workers with jobs.

Buses are great because they're flexible, cheap and use existing roads. Additionally, we could spend more money to build more sophisticated transportation networks - more roads, maybe rail; roads that function as dedicated bus lanes - to support working-class Americans in their noble effort to earn their own success in the labor market.

By significantly decreasing commuting times from lower-income neighborhoods and exurbs we would effectively increase the number of jobs available to low-income workers. And the quality of life for low-income workers with long commutes would increase.

Improving infrastructure generally and helping low-income workers specifically are useful functions of government, as opposed to proliferating middle-class entitlement programs, one of government's favorite activities.

In fact, middle-class entitlement programs are a good place to start looking for the federal government's share of the money to pay for these infrastructure projects - projects that are valuable to society, help the working class to rise and help low-income Americans lead flourishing lives, which include the dignity of work.

Michael R. Strain is a resident scholar at the American Enterprise Institute. This column was first published by The Washington Post.