Jake Naughton/The New York Times

Paying for Infrastructure

by aaron klein

aaron klein, former deputy assistant secretary of the Treasury, is policy director of the Brookings Institution’s Center on Regulation and Markets. This article is adapted from Klein’s Up Front blog post on December 12, 2018.

Published December 20, 2018


Optimism that America will finally address its massive infrastructure problems is on the rise — and understandably so. President Trump never abandoned his trillion-dollar campaign pledge, and Democrats who have long supported rebuilding the country’s crumbling roads, bridges and water and sewer systems will soon regain control the House. But don’t pop any corks just yet. An infrastructure overhaul did not occur in the past two years for the same reason we’ve been in gridlock on infrastructure spending for decades: there’s no agreement on how to pay for it.

Further complicating the prospects for a successful infrastructure initiative, the optics work against building a geographically diverse political constituency to support it. For example, improving the speed at which refrigerated cargo can enter and exit the port of Seattle would provide more benefits in the long run to cattle ranchers in Montana who export their products to Asia, than to the construction workers in Seattle who would get six months of hefty paychecks. Yet the ribbon cutting for the project would occur thousands of miles away from the cattle ranches. By the same token, providing New York City’s transit authority with funds to buy new rail cars would actually create more jobs in Lincoln, Nebraska — where the Kawasaki plant that builds rail cars is located— than in New York. But Uncle Sam’s check would go to New York, while Nebraska would appear to have gotten nothing.

Overcoming such seemingly superficial political problems might seem easy. But ask yourself when you last saw a news report about an infrastructure project in another part of the country and thought: this will benefit me.

I get ahead of myself, though. Stripped to essentials, there are three ways to pay for new infrastructure: raise revenue, borrow or pretend (i.e., come up with gimmicks that mask backdoor borrowing).

Tax and Spend

Start with revenue. Transportation infrastructure has by and large been funded by “user fees” — in particular, the federal and state excise taxes on fuel that are dedicated for roads. This user fee approach historically enjoyed broad support. People understand that freeways aren’t free and that taxes collected from users for infrastructure are different from generic taxes. Republicans embraced the small “c” conservative idea that people who used public assets ought to pay for them. Democrats, for their part, overcame objections regarding both the inherently regressive nature of flat consumption taxes and the basic idea of segregating specific revenue streams from general revenue.

But this bipartisan consensus, which lasted for decades, has broken down on both sides. Republicans morphed from the party of Ronald Reagan and George H.W. Bush (who both raised the gas tax) to the party that embraced anti-government advocate Grover Norquist’s “no new taxes” pledge, which now includes the gas tax. Democrats broke their end of the bargain in 1993 when they raised the gas tax, not to pay for more roads but as a way to reduce the budget deficit after their proposal for a broad-based energy tax failed. While those gas-tax dollars were eventually restored to the federal Highway Trust Fund, the glory days were over.

Simply increasing the gas tax to account for a quarter century’s inflation and indexing the tax to the cost of living would fund a large-scale infrastructure initiative. But that would require bipartisan political will that has been lacking for 25 years.

As a result, the federal gas tax has been stuck at 18.4 cents a gallon since 1993. Simply increasing the gas tax to account for a quarter century’s inflation (which would mean 32 cents a gallon today) and indexing the tax to the cost of living would fund a large-scale infrastructure initiative. But that would require bipartisan political will that has been lacking for 25 years, not to mention repudiation of the anti-tax pledge that most Senate Republicans have inked.

Borrow and Spend

Plan B is to build on credit. But this train may have already left the station. Largely driven by unpaid-for income tax cuts and a substantial increase in military outlays, the federal budget deficit approaches $1 trillion even in a period of high employment in which prudent budgeting would create no deficit at all (or better yet, reduce the nation’s debt). Meanwhile, the prospect of adding more to the deficit for infrastructure seems to have reawakened the inner fiscal hawk in a lot of Republicans who have been anything but for the past few years. Hence President Trump’s original proposal to spend $200 billion of borrowed federal funds in the hopes of galvanizing $1 trillion of total infrastructure investment has generated little Congressional buy-in.

State and local governments do routinely borrow for infrastructure spending, maintaining separate accounts for long-term investments. But issuing long-term bonds to finance more infrastructure has pros and cons. On the plus side, it better aligns the responsibility for paying down the costs of infrastructure assets with the people who use the assets. On the down side, however, this approach does not create debt-free physical capital to bequeath to the next generation, leaving the problem of payment for the next round of investment to be solved down the road.

Distract and Spend

The third option is to resort to smoke and mirrors. This happened with the last five-year surface transportation bill when Congress raided the surpluses held by the Federal Reserve system, yet claimed the legislation wouldn’t increase the federal debt. It did add to the red ink, only the additional debt was sold to the public by the Federal Reserve, not by the Treasury, and thus didn’t officially count. (Don’t get it? That was the intention.)

An alternative (and more honest) approach to get infrastructure off the government’s books is to substitute private investment for public. That’s possible in theory, but difficult in practice. State and local governments already enjoy a large indirect federal subsidy when they borrow to finance infrastructure, since the interest received by bondholders is exempt from federal income tax. While the Trump tax cut reduced this subsidy, thereby making muni-bond-financed construction more expensive, it is hard to imagine Congress creating an even larger subsidy that made privately financed infrastructure investment competitive with the state-financed variety.

Consider, too, that private investors demand a return on investment, limiting the infrastructure that can be built with private funds to projects (like toll roads, bridges and tunnels) with projectable revenue streams. This approach has potential, but on a far narrower basis than often promised.


In spite of two years of unilateral Republican control of Washington, the Trump infrastructure initiative got nowhere. Now, with the Democrats back in the power mix, putting infrastructure in the driver’s seat would require a deal that brings on board both Congressional parties and the Trump White House. And that, in turn, would require a compromise to either raise revenue or add to the nation’s credit card — or a willingness to accept a far smaller program than either the Trump administration or Congressional Democrats want. Failure to squarely face those problematic trade-offs would likely keep us stuck in the same traffic jam we’ve been stuck in for a generation.

main topic: Infrastructure