“With real interest rates at zero or below, shouldn’t the federal government just borrow the money to pay for something worthwhile, like a big infrastructure initiative?”

Of course, in a time when the federal government is running systemic and voluminous deficits, the federal government is, in reality, already issuing bonds to pay for infrastructure alongside everything else. (In the just-ended and COVID-inflated fiscal year 2020, 48 cents of every dollar of federal outlays was supported by borrowing, not by federal receipts.)

But there are two separate sets of issues involved – one set being federal borrowing for the specific purpose of infrastructure funding, and the other set being the deficit financing of infrastructure as part of the general deficit financing of the entire federal government.

Infrastructure-specific federal borrowing.

There is a difference between the federal government issuing bonds in support of the entire government, of which infrastructure is a part, and the federal government issuing bonds specifically to support federal infrastructure spending.

Issuing federal bonds for a specific purpose (infrastructure or otherwise) is not compatible with the way the federal government finances itself generally or finances large infrastructure programs specifically.

Generally. Here are a few examples of proposals to finance large federal infrastructure programs via the issue of special-purpose bonds, all of which were unsuccessful:

• 1955 – The “Clay Committee” (President’s Advisory Committee on a National Highway Program) appointed by President Eisenhower recommended funding the proposed Interstate Highway System by creating an independent Federal Highway Corporation and allowing that Corporation to issue $25 billion in bonds, to be repaid by annual Congressional appropriations approximately equal to current law gasoline and other highway user taxes.
• 1959 – The Bureau of Public Roads proposed to pay for Interstate Highway construction cost inflation by allowing the Highway Trust Fund to issue special bonds to be repaid by dedicated tax receipts after Interstate construction was completed.
• 2003 – Senate Finance Committee leaders proposed to reduce the portion of federal motor fuel taxes dedicated to the Mass Transit Account of the Highway Trust Fund from 2.86¢/gal. to 0.5¢/gal., eliminate general fund appropriations for mass transit, and make up the difference having by Treasury issue a special 20-year bond (that paid federal income tax credits in lieu of interest), with the proceeds of bond sales to be deposited in the Mass Transit Account.
• 2017 – Rep. Peter DeFazio (D-OR) introduced the Investing in America: A Penny for Progress Act, requiring the Treasury to issue special 30-year “Invest in America” bonds, with the proceeds from bond sales to be deposited in the Highway Trust Fund, and the bonds to be repaid with annual motor fuel tax indexation increases.

All of these proposals for the issuance of special-purpose federal bonds to pay for infrastructure failed to become law, and they all had one other thing in common – strong institutional resistance from the Treasury Department. Treasury believes that the federal government should only be issuing one kind of security – the all-purpose Treasury bill/note/bond (the name depends on the duration) used as the means of financing to fill the federal deficit, because that is the security that gets the absolute lowest market interest rate. Treasury believes that every arm of the federal government should then go to Treasury for money, not to the bond markets.

This testimony that Treasury gave before the President’s Commission to Study Capital Budgeting in 1998 is as true today as it was then:

Prior to 1973, the Federal Government actually had multiple issuing authorities.

In this environment, many federal agencies issued their own securities to fund their programs. At the time, we had agencies competing in the market with each other and with the Treasury. This led to an inefficient use of Federal credit. Federal agency debt and federally guaranteed debt carried higher interest rates than that issued by the Treasury, sometimes carrying interest as much as two percentage points higher than on comparable Treasury securities.

To address these concerns, the Treasury requested that Congress create the Federal Financing Bank. All agencies are now required to borrow from the Federal Financing Bank and no longer issue securities directly to the market. This allows the various programs to be financed at the Treasury’s lower cost of funds. This important innovation in federal finance was actually initiated by a young Treasury official named Paul Volcker. He later went on to hold a slightly more important post for the nation.

In conclusion, while there may be aspects of capital budgeting that are applicable to the Federal Government, we do not believe that this would include any form of segregated financing.

Specifically. The federal government has chosen to fund the bulk of its spending on infrastructure programs through federal trust fund accounts. In fiscal year 2019, federal outlays on construction and rehabilitation of major public physical investment (non-defense) totaled $97.6 billion, and about $60 billion of that came from the four transportation infrastructure trust funds:

• Highway Trust Fund
• Airport and Airway Trust Fund
• Harbor Maintenance Trust Fund
• Inland Waterways Trust Fund

Federal trust fund accounts are visibility exercises – a way to track the receipts of specific federal excise taxes over time and visibly correlate those receipts with the spending from specific federal budget accounts that are supposed to provide direct benefits to those who paid those excise taxes.

The creation of those four trust funds was also, in each instance, part of the political deal necessary to convince those interests that would be paying increased taxes to accept those targeted tax increases – by ensuring that their tax payments could not only be used to pay for benefit that set of taxpayers and prohibiting “diversion” of those taxes.

In the instances of all four federal trust funds for infrastructure, another of the selling points of the legislation creating the trust funds was that such a trust fund would be “deficit-proof.” The Senate Finance Committee only approved the bill creating the Highway Trust Fund after chairman Harry Byrd (D-VA) successfully amended the bill to add what he called a “pay-as-you-build” provision that would automatically reduce Interstate funding apportionments to states to match short-term projections of dedicated tax receipts.

Of course, the Highway Trust Fund ran through “deficit-proof” 12 years ago and never looked back. Since first running out of money in September 2008 (see the whole sordid story here), the HTF has required almost $154 billion in bailout transfers from general revenues, which, when spent, is the same as deficit financing. (Some of those transfers came in bills that purported to pay for the transfers over a ten-year period, but the budget scorekeeping rules being used encourage the use of fanciful “pay-fors,” because if the pay-fors never materialize down the line, it’s too late, the money they were paying for has already been spent.)

The problem with propping up an insolvent trust fund with general fund bailouts every couple of years is that the hole keeps getting deeper. The first year that the Highway Trust Fund ran out of money and needed a bailout was 2008, a year in which the HTF ran an operating deficit of $6.6 billion. As tax receipts stayed flat but spending got annual inflation adjustments, the Trust Fund ran an $11.9 billion deficit in 2018. Last year, the deficit rose to $15.2 billion, and current projections have those annual deficits rising to $20 billion per year in 2025 and $25 billion per year by 2030. Each bailout means that the next bailout has to be even bigger.

If Congress is going to perpetuate the trust fund model, it needs to get away from borrowing as the means of keeping the trust funds solvent, not get more reliant on borrowing.

Deficit financing of infrastructure

The federal government is already borrowing money to pay for infrastructure – it’s just doing so in an ad hoc, haphazard way, running multi-trillion-dollar general fund deficits while using more borrowed money to keep propping up the Highway Trust Fund. As shown above, borrowing to fund more infrastructure really can’t be done with specific issuances of dedicated federal bonds, and probably shouldn’t be done through transportation trust funds.

The other option is simply to create new spending, either through the Appropriations Committees or through a bill from the authorizing committees that creates new mandatory budget authority out of the general fund. There are a variety of rules in the budget process that make this difficult (the budget process was put into place to prevent non-Appropriations committees from creating too many new spending programs, but in return, the budget process also makes it hard for the appropriators to fund programs more than one year in advance), but if this is truly a once-in-a-generation infrastructure initiative to use all-time-low interest rates to rebuild the nation, perhaps those rules should be waived.

But deficit financing has its own problem – the inability to correlate interest rates of today with what the interest rate will be when spending actually takes place. In a time of systemic deficits (like today), if Congress appropriates $100 billion for infrastructure tomorrow, or creates $100 billion in new non-appropriated direct spending for infrastructure, that doesn’t immediately necessitate any new borrowing by the Treasury. The borrowing only takes place when that money is actually spent, and Treasury has to hold bond sales to make ends meet. Infrastructure can spend so slowly that, just because interest rates are zero today when Congress appropriates the money, doesn’t mean that interest rates will still be zero three or five years from now, when that infrastructure money is actually spent and the Treasury has to issue bonds to cover it.

If deficit financing of federal infrastructure spending can’t be reliably used to capture the current ultra-low interest rate environment, what other options exist?

Federal credit programs

There is one method currently in use that allows infrastructure borrowing that captures today’s interest rates – so long as the federal government is not the ultimate borrower. The federal TIFIA credit program (for highway and mass transit projects) and the federal RRIF loan program (for railroad projects) allow non-federal entities to build infrastructure projects using something close to the Treasury borrowing rate.

For example: a state or municipality borrows $1 billion from the Department of Transportation to build a new subway line. DOT draws that money from the Federal Financing Bank within Treasury, and Treasury then sells $1 billion of bonds on the open market to make the deal possible. The savings are passed on, and the interest rates are locked in (today’s 30-year TIFIA loan rate is 1.61 percent).

Essentially, every time the federal government makes a TIFIA or RRIF loan, Treasury just borrows on the bond market and DOT then signs the proceeds over to the ultimate non-federal borrower.

Capital spending – who owns the asset?

One of the fundamental problems with attempting to use capital budgeting principles in federal infrastructure programs is asset ownership. Of that $97.6 billion in federal outlays on construction and rehabilitation of major public physical investment (non-defense) in 2019, less than 20 percent went to build or maintain physical assets owned by the federal government.

The other 81 percent of the money, including most of the programs overseen by the Department of Transportation, took the form of grants-in-aid to state and local governments for the construction or maintenance of physical assets owned by the state or local government – highways, bridges, subways, airports, ports. (Some grants went to private entities, usually railroads, whose privately owned assets also serve a public purpose.) The non-federal party builds or improves their own asset, using federal standards, and then the federal government reimburses them for between 50 and 100 percent of the cost.

Most federal spending on infrastructure capital is better thought of as spending on somebody else’s capital infrastructure. (Even if asset depreciation were possible – and no state or local governments actually use depreciation in their budgets – the federal government could never depreciate assets it did not own.)

Borrowing that is directly tied to a specific physical asset or a specific set of physical assets is more properly done by the entity that owns the asset.

However, there are areas where significant physical infrastructure assets are owned by the Department of Transportation. One is the collection of radars, radios, and computers that run the U.S. air traffic control system (and the buildings that house them), which are owned directly by the Federal Aviation Administration. Another is the Northeast Corridor – the railroad track, bridges, catenary, switching and stations between Washington Union Station and New Rochelle, New York, and from New Haven, Connecticut to Providence, Rhode Island and Springfield, Massachusetts, all of which is owned by Amtrak (which is an off-balance-sheet entity ultimately owned by the federal government).

A new concept in capital budgeting for infrastructure was proposed several years ago by the career staff of the Office of Management and Budget and included in the Trump Administration’s infrastructure proposals. The proposal was limited to federal spending on public buildings and called for a “revolving fund” with a one-time infusion of $10 billion in mandatory budget authority from general revenues (and presumably deficit-financed).

This $10 billion could then be used to purchase or construct new federal office buildings for federal agencies (a minimum project size of $250 million was proposed). The agency using the building would then repay the revolving fund for the cost of the building over a 15-year period using its annual appropriations from Congress. Once the full cost of the building is repaid to the revolving fund, the money could be used again, to buy or build a new building.

The concept of a capital revolving fund could be used to accelerate the implementation of “NextGen” air traffic control at the FAA and could also be used as the financing mechanism to fix the $40 billion capital backlog of the Northeast Corridor.