Economists call for rigorous cost-benefit analysis before spending money on a project
Rather than building new lanes to ease traffic in an urban area, two economists suggest in a new paper that it might make sense to consider congestion pricing.
A new paper by a pair of economists says the gains from infrastructure spending aren’t always clear-cut and recommends that policymakers examine the costs and benefits of each project.
“If we are going to commit a significant amount of resources to new infrastructure projects or to maintain our existing infrastructure, bringing some discipline to the way we decide what we’re spending on is an important element of this,” said James Poterba, an economist at the Massachusetts Institute of Technology, who co-wrote the paper with Edward Glaeser of Harvard University.
The paper is set for publication Wednesday by the Aspen Economic Strategy Group, a division of the nonpartisan Aspen Institute. Mr. Poterba is a member of the group. Congress and the White House are now working on a bill that would include about $579 billion in new infrastructure spending.
The common way to determine the country’s infrastructure needs is to calculate how much it would cost to relieve congestion and to improve the roads, bridges, airports and transit systems that already exist. The economists’ approach instead looks at projects individually to assess whether they are needed.
In some cases, the authors write, the best solution doesn’t involve construction at all. Rather than building new lanes to ease traffic in a dense urban area, it might make sense to consider congestion pricing, which charges drivers a variable fee depending on the time of day, they write.
Mr. Poterba recommended a voucher or tax-rebate system for lower-income households to ensure they aren’t disproportionately hurt by the fees.
The cost of repairing an unsafe bridge in a remote area with very little traffic may exceed the benefits, they write. In that case, the most economically efficient solution might be to close or demolish it. It might also make more sense to link cities with rapid buses on dedicated lanes rather than build new rail lines. Satellite broadband or 5G network access might be a good alternative to laying fibre optic cables to provide high-speed internet access to rural areas, they write.
The paper suggests establishing an independent entity such as an infrastructure bank that would perform cost-benefit analyses on projects and that would fund only those where benefits exceed the costs. That would help keep political considerations out of the process, they write.
The studies would take into account a project’s financial costs as well as its impact on the environment, maintenance requirements and long-term economic impact.
But there are significant hurdles to cost-benefit analyses. It is often hard to estimate a project’s final cost because its design can change due to community opposition or environmental considerations.
Identifying the benefits of a project also is complicated, because measuring benefits depends on how much it will be used, which is difficult to predict in advance.
“You have to be careful you’re not being highballed with rosy projections about what the demand for utilization will be,” said Mr. Poterba.
Such cost-benefit analyses likely would recommend focusing more on maintaining existing infrastructure rather than new projects, they write. About 47% of highway spending now goes to maintenance, they find.
Officials sometimes prefer spending on new projects over maintenance because of a “ribbon-cutting bias,” Mr. Poterba said, “where you can point to the thing and say it wasn’t there before my time and now it’s there”.
The paper also expressed scepticism about public-private partnerships, in which a private entity finances the project in exchange for future revenue from tolls or other charges. Low-interest rates in recent years mean that governments can often borrow more cheaply than the private sector, they write.
“Some state and local governments may be attracted to these partnerships because they relieve current cash flow constraints,” the authors write. “But they may come at a price in terms of the long-term cost of infrastructure services”.