A National Infrastructure Bank?

Proposed bank can fill a niche, but current proposal needs to be refocused

One of the hottest ideas in infrastructure politics these days is the proposal now in Congress for a National Infrastructure Bank. Presidential candidate Barack Obama, New York City Mayor Michael Bloomberg, and many other big-city mayors have all endorsed the idea, as have a number of trade associations favoring increased infrastructure investment.

The leading vehicle is S. 1926, introduced last summer by Sen. Chris Dodd (D, CT) and Chuck Hagel (R, NE). It would “target large capacity-building projects not adequately served by current financing mechanisms” including “mass transit systems, housing properties, roads, bridges, drinking water systems, and wastewater systems.” The NIB would be set up as an independent government entity, modeled after the Federal Deposit Insurance Corporation, with a five-member board of directors appointed by the President and confirmed by the Senate. Minimum federal investment per project would be $75 million. The legislation calls for the bank to issue $60 billion in long-term (up to 50-year) bonds.

Projects would be selected by the NIB’s board on the basis of “national or regional significance,” with the amount of federal investment determined on a “sliding scale” based on the type of infrastructure, location, project cost, current and projected usage, non-federal revenue, promotion of economic growth and community development, reduction in congestion, environmental benefits, and land-use policies that promote smart growth.”

My initial reaction to this proposal is “Huh?” There’s no question that this country has not been investing enough in either rebuilding and modernizing existing infrastructure or adding much-needed new capacity. But is a new federal entity of this sort a sensible response?

One clue that this is mostly smoke and mirrors is the paltry $60 billion amount. With estimates of infrastructure funding shortfalls at or above a trillion dollars, this seems like the proverbial drop in the bucket.

Second, it is hard to see how this entity would constitute anything like a bank, in the normal meaning of the term. A bank is an enterprise that lends money, on a sustainable basis. That mean the borrowers have to pay it back, with interest, so that the bank can remain a going concern. If you look at what the sponsors list as the kind of financing the NIB would provide, you get the following list:

  • Direct subsidies [grants]
  • Direct loan guarantees
  • Long-term tax credits
  • General-purpose bonds
  • Long-term tax-credit infrastructure bonds.

Conspicuously absent from this list is revenue bonds-i.e., finance that is based upon the users paying for the services provided by the new infrastructure. One of the crying needs in U.S. infrastructure investment is better targeting of investment to projects that provide significantly more benefits than costs (i.e., are not Bridges to Nowhere). Yet the NIB proposal, as written, seems to ignore sound principles of project finance, such as user-fee-based financing.

Moreover, most of the articles and speeches by proponents of NIB have tended to single out federal tax-credit bonds as the main funding vehicle they envision. This idea has been around for nearly a decade without going anywhere, but its proponents keep trying. Unlike a revenue bond, where the principal and interest are repaid out of revenues from fees paid by users (water bills, tolls, etc.), with a tax-credit bond the interest would be paid by the government’s general fund, while the principal would be repaid by setting aside a portion of the initial bond offering and investing that in Treasury securities sufficient to pay off the principal at maturity. Thus, this is a way of having the federal government’s hard-pressed general fund go even further into debt to fund infrastructure.

Back in 2002, the Government Accountability Office compared federal tax-credit bonds with conventional federally tax-exempt bonds, TIFIA direct loans, and outright federal grants. GAO estimated the total (societal) costs and the federal government costs of each of these alternatives, over a 30-year period. GAO concluded that “Federal costs would be the highest under the tax credit bond alternative.” (GAO-02-1126T, p. 9). That same year then-Treasury Secretary John Snow said that if a tax credit bond proposal were enacted by Congress, he would recommend a presidential veto.

The idea of targeting investment to “projects of national and regional significance” has merit, but it’s questionable whether an entity beholden to the President and Congress would really be able to do that. As part of SAFETEA-LU, Congress actually created a program called Projects of National and Regional Significance. It appropriated $1.7 billion for the PNRS program, intended to fund a handful of projects of great benefit but with such high costs or benefits beyond the jurisdiction of a single state or region that those entities could or would not fund them. What happened? Members of Congress proceeded to earmark the entire $1.7 billion, parceling it out to 25 highway, railroad, intermodal, and transit projects.

If Congress wants to assist the financing of infrastructure projects on a sustainable basis (i.e., as a real bank would), it could expand a number of existing programs that already do this-in transportation, we have TIFIA, Private Activity Bonds, and State Infrastructure Banks, and in water and wastewater there are State Revolving Funds. All of these are intended as self-sustaining activities, in which revenues from the funded projects pay back the loans, which enables new projects to be funded.

Rather than simply dismissing the NIB proposal as the wrong direction for expanded and smarter infrastructure investment, I consulted a number of experts in infrastructure finance and asked their assessment. All basically agreed with my critique of the existing proposal-but all of them also argued that since the legislation already exists and has some political momentum, the best approach for critics might be to propose better content for the measure.

The gist of these people’s overlapping comments was that there is a role for the federal government to do more to encourage sound investment in large infrastructure projects-at least those like highways, bridges, water and wastewater systems in which a user-fee revenue stream is feasible. There is already nearly $200 billion sitting in infrastructure investment funds, looking for good projects. But there are very few large-scale projects ready to go. What’s need is a lot of detailed (and costly) pre-development work to establish basic feasibility, get environmental clearance, and do preliminary design.

Funding those pre-development efforts for user-backed projects of national or regional significance could be done on a soft-loan basis by such an entity-soft in the sense that the loan would be paid back (on a subordinated basis) if the project ultimately gets financed and built, but could be forgiven if the project proved non-viable and did not get financed. The need to make this “bank” self-sustaining would serve as a restraint on funding pre-development work on highly speculative projects.

How much of a difference could this make in “deal flow” for the billions in private capital that is looking for viable investments? One colleague provided this optimistic numerical example. Assume the $60 billion in initial funding for this re-purposed NIB were parceled out over a decade to 500 projects at an average of $120 million each in pre-development work. If that $120 million was about 10% of the average project’s total capital costs (i.e., $1.2 billion each), it would generate $600 billion in total construction (or reconstruction). That’s pretty impressive leverage.

Another expert, Dana Levenson of RBS, gave the analogy of one of the more successful government-sponsored enterprises, PEFCO (Private Export Funding Corporation). PEFCO started out with a mix of government and private sector capital; it assists with the financing of exports, both as a direct lender and as a secondary market buyer of export loans originated by lenders, and is currently leveraging its capital at a 50:1 ratio. Levenson suggests that on an initial (public plus private) capital of $20 billion, a National Infrastructure Bank modeled after PEFCO could leverage up to $1 trillion in infrastructure. (Even with 25:1 leverage, it could jump-start $500 billion in projects.) But all of that would have to be infrastructure with a user-fee revenue stream to pay back the loans.

In short, I’m persuaded that there is a niche that an NIB could fill, to help this country take advantage of the huge pool of investment funds seeking to invest in U.S. infrastructure. Whoever ends up as our next president should seek to refocus current national infrastructure bank proposals into this sustainable form.

Robert Poole is director of transportation at Reason Foundation. An archive of Poole’s work is available here and Reason’s transportation research and commentary is here.