Getting Returns from Infrastructure

While many interest groups are promoting increased federal spending on infrastructure on the grounds that it will spur economic growth, the Wall Street Journal reports that the "benefits of infrastructure spending [are] not so clear-cut." Yet there is a simple way to determine whether a particular infrastructure project will generate economic benefits.

Spending on transportation infrastructure, for example, generates benefits when that new infrastructure increases total mobility of people or freight. New infrastructure will increase mobility if it provides transportation that is faster, cheaper, more convenient, and/or safer than before. 

In 1956, Congress created the Interstate Highway System and dedicated federal gas taxes and other highway taxes to that system. The result was the largest public works project in history and one of the most successful. Today, more than 20 percent of all passenger travel and around 15 percent of all freight in the United States is on the interstates.

Moreover, this is all new travel; the interstates didn't substitute for some other form of travel, as other highway and airline travel) have also significantly increased in those years. (Rail passenger travel decreased, but that decrease was a lot smaller than increases in other travel.) The interstates were successful because they provided transportation that is faster, cheaper (because it saves fuel), more convenient, and safer than before. 

For the past two decades or so, however, much of our transportation spending has focused on infrastructure that is slower, more expensive, less convenient, and often more dangerous than before. Too many cities have given up on trying to relieve congestion. Instead, they have allowed it to grow while they spend transportation dollars (nearly all paid by auto users) on other forms of travel such as rail transit. Such transportation is:

• Slower: Where highway speeds even in congested cities average 35 miles per hour or more, the rail transit lines built with federal dollars mostly average 15 to 20 mph.
• More expensive: In 2013, Americans auto users spent less than 45 cents per vehicle mile (which means, at average occupanies of 1.67 people per car, about 26 cents per passenger mile), and subsidies to roads average under a penny per passenger mile. By comparison, transit fares are also about 26 cents per passenger mile, but subsidies are 75 cents per passenger mile.
• Less convenient: Autos can go door to door, while transit requires people to walk or use other forms of travel, often at both ends of the transit trip.
• Less safe: For every billion passenger miles carried, urban auto accidents kill about 5 people, while light rail kills about 12 people and commuter trains kill 9. Only subways and elevateds are marginally safer than auto travel, at 4.5, but we haven't built many of those lately.

Not surprisingly, most transit projects lead to almost no new travel. Yet their backers claim this is a virtue. They have demonized the new travel generated by the interstates by calling it "induced demand." They have celebrated transportation projects that generate no new travel but merely get people to shift from one mode to another, usually more expensive, mode as "sustainable."

Even when cities spend money on roads, they often spent it making travel slower, less convenient, and more dangerous. Many cities are doing various forms of what planners euphemistically called "traffic calming," meaning narrowing streets, putting barriers in roads, and turning one-way streets into two-way streets. The overt goal is to slow down traffic, and it often has the side effect of making it more dangerous for both auto users and pedestrians.

A very simple test can determine whether any particular transportation project will be faster, cheaper, more convenient, and/or safer than before: Will the users themselves pay for it? Users will pay for real improvements in transportation; they won't pay for slower, more expensive, less convenient, and more dangerous transportation.

The Interstate Highway System was paid for exclusively out of user fees. Gas taxes aren't a very good user fee, as there was no guarantee that the users who paid the taxes were driving on the interstates their taxes were building. As it happened, the interstates mostly were paid for by users, but there is no guarantee of that for future road projects. That's one reason why it makes sense to shift from gas taxes to mileage-based user fees or tolls.

Yet there is a major push to increase gas taxes. A recent article in the Atlantic's CityLab argues that gas taxes should be increased by 70 cents a gallon because current taxes aren't paying the "true cost of driving." Yet the costs that the article says taxes aren't paying--things like congestion and auto accidents--are nearly all paid by road users in other ways, so there is no reason why gas taxes should be increased to cover those costs or any reason to think that higher taxes will reduce those costs.

Our gas taxes are "underpriced," CityLab argues, because taxes are higher in Europe. Yet the higher taxes in Europe aren't spent on roads; they mostly go for non-road activities, and not for reducing congestion and other road-related costs.

Infrastructure is important. But throwing federal dollars at it won't take care of the problems. Instead, infrastructure spending only makes sense if users are willing to pay for it. For the most part, that means infrastructure can and should be funded privately or by state and local governments out of user fees, rather than by the federal government.