August 10, 2016

Last week’s article gave a history of the concept of the “infrastructure bank” concept in the United States. Of the two major-party candidates for President in the fall election, Hillary Clinton has made an infrastructure bank the center of her $275 billion infrastructure proposal, while Donald Trump has not specifically mentioned an “I-Bank” but has proposed doubling the amount of money spent under the Clinton proposal and has encouraged the leveraging of federal money to bring in private equity to infrastructure projects (a key feature of an I-Bank).

What kind of decisions would face the new President and the 115^th Congress in 2017 if they attempted to create a national infrastructure bank?

The first issue that must be dealt with is the fundamental concept of the structure of the infrastructure bank. From where will the I-Bank obtain the money it uses to provide financial assistance to U.S. infrastructure projects? Will the I-Bank borrow money from capital markets directly, or will the U.S. Treasury provide the capital?

Proposals under which the I-Bank raises its capital from capital markets directly include the original federal I-Bank bill introduced by Senators Chris Dodd (D-CT) and Chuck Hagel (R-NE) in 2007 (S. 1926, 110^th Congress) and the legislation first introduced by Rep. John Delaney (D-MD) in 2013 (H.R. 2084, 113^th Congress). Under these proposals, the I-Bank is a either a government-owned corporation (Delaney) or an independent federal establishment (Dodd-Hagel).

In both cases, the I-Bank would issue and sell its own bonds on the international bond market, and then uses the proceeds of those bonds as capital as the I-Bank makes its own loans and loan guarantees to fund U.S. infrastructure projects. In some scenarios the U.S. Treasury and state and local governments also make tax expenditures to support the projects funded by the I-Bank since they are not allowed to tax the interest on I-Bank bonds.

(Under the Dodd-Hagel proposal, the federal government would make a small initial appropriation for startup costs until the first bond sales went through.)

The other kind of I-Bank proposal includes every version of an I-Bank proposed by the Obama Administration, the proposal of Secretary Clinton, and the new one from the Center for American Progress. These would require the federal government to provide capital to the I-Bank, which would not be an arms-length government corporation but instead would be a federal agency or subsidiary of a federal agency to which the Congress appropriates money directly. Congress would appropriate money to the I-Bank, which would then use the appropriations as the “subsidy cost” under the Credit Reform Act for having the U.S. Treasury make direct loans or loan guarantees to fund U.S. infrastructure projects.

But since the federal government is already running massive deficits, the money to those loans to support infrastructure projects must come from the U.S. Treasury borrowing on the international bond market. And if the legislation to provide the appropriations for the subsidy money to the I-Bank does not have funding offsets, then the Treasury would have to borrow that money from the bond market as well.

The 2009 proposal by Rep. Rosa DeLauro (H.R. 2521, 111^th Congress) straddles the line between the two types of proposals. The federal government would appropriate $25 billion to the I-Bank, but the money would not be used for the subsidy cost of Treasury loans to infrastructure project sponsors. Rather, the $25 billion would be 10 percent of the total capital held by the bank. The other 90 percent would be “callable capital” in the form of a “break glass in case of emergency” requirement for the Treasury to make good on I-Bank obligations in a crisis. This is the funding model used by the European Investment Bank (EIB) and other multilateral development banks.

Whose Debt? The Dodd-Hagel and DeLauro proposals contain explicit language that debt issued by the I-Bank is backed by the full faith and credit of the United States government. Conversely, the Delaney bill says that “Interest and principal payments paid to [bondholders] shall be paid from the [I-Bank], to the extent funds are available, and shall not be backed by the full faith and credit of the United States.”

But the capital markets have always assumed that statements like the one in the Delaney bill, when used in the past for debt of government-sponsored enterprises like Fannie Mae, Freddie Mac, and the Tennessee Valley Authority, were simply political posturing and that Uncle Sam would make good on the debts if there were ever a crisis. And, in 2008, the capital markets were proven correct when the federal government took over Fannie Mae and Freddie Mac and guaranteed their debt.

So, from a conceptual point of view, it is probably best to think of any debt issued by an I-Bank as being backed by American taxpayers, just like regular U.S. Treasury debt.

However, the duration of Treasury debt is capped at 30 years. The I-Bank proposals in which the I-Bank issues its own bonds calls for a duration of each bond that is either “greater than 30 years” (Dodd-Hagel, DeLauro) or is set at 50 years (Delaney).

Leverage. The answer to the question of “does the I-Bank borrow money directly from capital markets or not” has a direct bearing on the financial bottom line. The aggregate amount of financing that an I-Bank can provide is a function of two factors: the amount of capital the I-Bank has, and the leverage ratio – defined as “how many dollars in loans can we have outstanding for every dollar of capital we hold?”

In the proposals where the I-Bank tries to act like a real-world bank and raises its own capital, the I-Bank would presumably get to decide for itself what leverage ratio to use. At a hearing on the 2007 Dodd-Hagel proposal, investment banker Felix Rohatyn suggested a conservative 3 to 1 leverage ratio for the $60 billion raised by I-Bank bond sales, or $180 billion in total loans made by the I-Bank. By contrast, Delaney says his I-Bank would use a ratio of 15 to 1, so $50 billion in cash raised from bond sales would support up to $750 billion in federal credit assistance. (Allowing a federally-owned I-Bank to set its own leverage ratios would presumably require some kind of waiver or clarification of the Federal Credit Reform Act – see below.)

The higher the leverage ratio, the riskier the enterprise (though even a 15 to 1 leverage ratio is significantly less than those used by Fannie and Freddie towards the end). The DeLauro bill would set a maximum leverage ratio of 2.5 to 1, which she says is the same conservative leverage ratio used by the EIB.

Under the rest of the proposals, where the federal government supports the I-Bank through appropriations, the leverage ratio is effectively set by the Office of Management and Budget. Under the Federal Credit Reform Act of 1990, loans and loan guarantees made by the federal government have their “subsidy cost” re-estimated each year by OMB based on the riskiness of the loan cohort, the interest rate environment, the fees charged, and other factors. These estimates are published each year in a little-read volume of the President’s Budget called the Federal Credit Supplement.

Comparable Programs. In the Federal Highway Administration’s TIFIA credit program, the projects selected for loans over the years have performed well. The FY 2017 OMB estimate of the TIFIA subsidy cost assumed the net default rate on new direct loans would be 6.67% and the interest differential between what TIFIA charges and what the Treasury pays to borrow is a de minimis 0.06%, which gives a subsidy cost of 6.73%.

This means that $1 million of TIFIA budget authority for the subsidy cost of a direct loan (out of $275 million provided for the program in 2017 by the FAST Act) can support up to $14.86 million in direct loan principal (1 divided by the subsidy cost of 0.0673), for a leverage ratio of 14.86 to 1. (If FHWA spent the entire $275 million in FY 2017 on direct loans, the program could make $4.1 billion in loans.)

Other federal credit programs that make riskier loans have higher subsidy costs. The FY17 estimate for the Energy Department’s section 1703 loan program for clean energy projects (which has had such high-profile defaults as Solyndra and Abound Solar) assumes a net default rate of 14.65%, offset by a 1.10% difference between interest rates, for a subsidy cost of 13.55%, or a leverage ratio of 7.38 to 1.

(It bears noting that under credit reform budgeting, the money appropriated by Congress to cover the subsidy cost of a loan made by the Treasury sometimes comes from the same place where Treasury gets the money to make the loan – borrowing from international capital markets, backed by the full faith and credit of U.S. taxpayers.)

In the past, OMB estimates for the subsidy rates on the various iterations of President Obama’s I-Bank proposals have varied widely. In President Obama’s first budget (FY 2010), OMB judged that their draft I-Bank proposal would require a subsidy rate of over 50 percent on its direct loans – a leverage ratio of under 2 to 1. (This was not based on a high assumed default rate – the table only says that the total in the “all other” category was 49.61%.) Subsequent iterations of the President’s proposal have had subsidy rates of between 11.57% and 20.00%, for leverage ratios of between 8.6 to 1 and 5.0 to 1.

Secretary Clinton’s proposal has not yet been fleshed out with legislative language, but her campaign document apparently assumes a 9 to 1 leverage ratio ($225 billion in federal credit assistance based on $25 billion in federal appropriations), which equals a subsidy rate of 11.11%. This is close to the 11.57% for direct loans in President Obama’s FY 2016 I-Bank proposal, which was the one most recently available when Secretary Clinton’s proposal was released.

The Center for American Progress, a progressive-leaning think tank full of people who have worked with Secretary Clinton (or who will probably wind up working in a new Clinton Administration) recently issued a comprehensive infrastructure proposal that included a plan for Congress to appropriate $125 billion over ten years to fund an I-Bank. However, it is impossible to estimate the total amount of money that could be leveraged by the $125 billion, because the innovative proposal from CAP calls for the I-Bank to spend that $125 billion not just on the subsidy cost of traditional direct loans and loan guarantees but also on below-market-rate-interest, zero-interest, and negative-interest loans as well as outright grants.

The word “leverage” is also used in another way. Most I-Bank proposals limit the percentage of total project cost that can come from an I-Bank loan to 50 percent or less, in hopes that private equity could also take an interest in infrastructure projects (as well as the state or municipality providing some of the money). A 2012 Congressional Budget Office study of I-Bank proposals noted that “By lowering the cost of borrowing, an infrastructure bank might induce additional private investment, but the amount of that investment would probably be limited by the fact that private-sector investors would require a rate of return comparable to what could be earned on other investments and that transportation projects for which such a return could be earned are probably limited.”

Loans vs. Grants. The CAP proposal and the CBO reminder that private equity demands a market rate of return bring up another fundamental issue with structuring an I-Bank: should it stick to making and guaranteeing loans that can be repaid in a manner that covers the I-Bank’s costs, or should the I-Bank also give money away for free?

The first three years of the Obama Administration’s I-Bank proposals called for the I-Bank to make a mix of grants and loans. This may have been, in part, a reaction to the extensive earmarking of federal programs by Congress, especially in the 2005 SAFETEA-LU surface transportation law, as well as a push-back against the distribution of transportation funding via formula (formulas which in many cases had become outdated and politicized).

However, Congress got rid of earmarking in 2011, and the 2012 MAP-21 surface transportation law got rid of the Department of Transportation’s discretionary surface transportation programs as well (except for the TIGER program, which is funded by the Appropriations Committees, not by the surface transportation bill). At that point, it was clear that if Congress wasn’t going to let the Department of Transportation make discretionary grants, it certainly wasn’t going to create a new arms-length federal entity and let it make discretionary grants, so the Administration dropped the grant-making ability from subsequent iterations of its I-Bank proposal.

Allowing an I-Bank to make grants would have some advantages. The CAP proposal notes:

since the [I-Bank] would also offer grant funding, the authority would serve as a one-stop shop for sponsors to receive a bundle of loan and grant assistance. This would substantially simplify what can sometimes be a complex and disjointed system of federal programs housed in multiple agencies or, in the case of transportation, modal administrations within the same department.

The mixture of grants and financing also allows state and local governments to apply for project support and then negotiate the most appropriate combination of support without having to specify the exact outcome on the front end.

But in the real world, banks aren’t in the business of giving away free money. Grant programs would be incompatible with the models where the I-Bank raises its own money from capital markets, and policymakers have been hesitant to allow new grant-making agencies to be established which are well-insulated from political pressure. Also, grant-making (as well as the ability to make zero- or negative-interest loans, as proposed by CAP) would add another level of financial complexity to the I-Bank structure – there might have to be separate appropriations accounts for grants, low/zero/negative interest loans, and “real” loans.

(In any case, the CAP proposal would require new amendments to the Federal Credit Reform Act, which the House and Senate Budget Committees would have to approve. CAP wants to make its I-Bank a “revolving fund,” where loan repayments could be used to fund the subsidy cost of new loans. As the CBO study noted, “Allowing loan repayments to be used for new loans— without any additional appropriation to cover the subsidy costs of the new loans—would raise the effective FCRA subsidy cost of the original loans to 100 percent (the same as for grants).” And this report from the Center on Federal Financial Institutions notes that one of the goals of credit reform was to prevent federal revolving funds from making loans.)

The decision of whether or not an I-Bank tries to act like a real bank (loans only, and having the bank be self-supporting in the long run) also determines the kind of projects that can be supported. In order to repay a loan, there has to be a long-term revenue stream – and that means that a lot of projects won’t qualify. The CBO report says that “The number of projects that would be good candidates to receive a loan from a federal infrastructure bank as envisioned in recent proposals is probably limited, at least in the short term. In principle, such a bank could identify and support large-scale projects that have substantial economic benefits for which users could be charged directly, so only a little federal assistance would be needed to cover the expected costs. By encouraging such user charges, the bank would make the available federal funds go farther.”

Defining “Infrastructure.” Another key issue for policymakers when establishing an I-Bank: what kinds of infrastructure can receive funding from the I-Bank? All of the proposals for a federal I-Bank made in recent years allowed investments in all manner of transportation infrastructure. Many of them also allowed investments in energy projects (the electrical grid was a personal priority of Rep. DeLauro) and in navigation, water supply and wastewater infrastructure projects (although Congress is already funding state revolving funds for the latter two, which are the equivalent of water-only state infrastructure banks).

Some more recent proposals would also allow the I-Bank to invest in broadband internet infrastructure projects (DeLauro and Delaney) as well as a broader category of environmental infrastructure projects (DeLauro). The Dodd-Hagel plan would have allowed investments in public housing projects, and the Delaney bill would allow investments in construction of public schools and libraries.

Water and electric systems are a natural fit for an I-Bank because they tend to attract ratepayers over many decades to slowly repay debt. The same logic applies to broadband Internet infrastructure. The issue of construction of public schools and libraries is a bit trickier – the operation of those tends to be subsidized by the annual budget process of states and localities, not any kind of user revenue stream.

Whither Muni Debt? Asking Congress to provide $10 billion or $25 billion up front to endow a National Infrastructure Bank sounds like a lot of money. But that amount pales in comparison to the amount of money Congress already provides to state and municipal governments to finance their own loans, much of which already goes to infrastructure. The Joint Committee on Taxation estimates that the “tax expenditure” of revenue foregone by Congress due to the tax exemption given to general purpose state and local government bonds will average $37.5 billion per year over the fiscal 2015-2019 period, and this does not include the tax exemptions for special purpose bonds like private activity bonds and Build America Bonds.

There is a strong argument to be made that giving state and local government lower borrowing rates via the tax exemption for their interest is not the most efficient way to subsidize these projects, because too much of the subsidy ends up in the pockets of rich bondholders. Under that theory, direct federal loans or loan guarantees, such as from an I-Bank, would be more efficient. But none of the serious I-Bank proposals have taken the next logical step and proposed tampering with the 800-pound gorilla that is the muni debt interest tax exemption – all pending proposals would add federal I-Bank credit on top of the existing tax-free municipal bond market.

In addition, the recent action by Congress to allow Puerto Rico to restructure its various debts is widely seen as the “canary in the coal mine” for debt restructuring that will be needed by a growing number of U.S. states and municipalities in the future as unfunded pension liabilities continue to mount. When considering an I-Bank, Congress might also consider the wisdom of layering hundreds of billions of additional municipal debt on top of the existing debt that is already breaking some areas.

Other Issues. There are numerous other issues that must be decided when crafting an I-Bank. A system of corporate governance must be established. Will members of a Board of Directors be named by the President and confirmed by the Senate? How much authority will they have?

The statute will likely prescribe rules for project selection. Expect conservatives to push for stricter economic cost-benefit analysis criteria for evaluation while liberals insist on the consideration of more “externalities” like the environmental and quality of life benefit of a project.

And there a host of other technical issues. (In particular, see the CAP proposal’s interesting take on the relative level of I-Bank debt subordination.)

Conclusion. No matter who wins the presidential election, federal infrastructure spending will be central to next year’s political conversation, and proposals to establish some sort of infrastructure bank will be at the forefront. It is in everyone’s best interest to start thinking about the big conceptual “I-Bank” issues delineated above as soon as possible.

We have a one-page PDF tabular summary of all I-Bank proposals here, and an article with paragraph summaries of previous I-Bank proposals is here.