In January 2020, I participated in a workshop hosted by the National Academy of Public Administration and sponsored by the Bureau of Transportation Statistics. The topic that I and two dozen other participants were there to discuss: “Improving Transportation Financial Statistics for the Future.” (It was more fun than it sounds.)

The workshop was part of an ongoing project that led to this document, a “roadmap” for BTS to use to improve its reporting of useful transportation financial statistics. Following that roadmap, BTS has since developed its Transportation Public Finance Statistics webportal, as well as monthly electronic reporting of what had been its annual Transportation Economic Trends report. Monthly reporting on Highway Trust Fund receipts and outlays is included as part of these trends.

It’s all good stuff, but something happened last week that calls into question whether or not relying on Highway Trust Fund outlays as a metric for economic activity is completely valid.

Fiscal year 2024 ended on September 30, and after the initial period of year-end accounting, the data for the year started to be reported on Treasury and program agency websites. Treasury published the year-end Monthly Treasury Statement, showing modal and selected budget account outlays (rounded to the million dollars), and Treasury also reported trust fund revenues to the penny. USDOT, through the Federal Highway Administration, published the monthly update to Table FE-1, showing actual Trust Fund Account-level outlays (to the dollar) and balances.

The first thing that jumped out of the data was that the Highway Account reported a spectacular $8.4 billion in outlays for the month of September 2024 in FE-1. Highway spending is cyclical (because: weather) and September is always the busiest month, followed by August and then October and July in some order. But $8.4 billion compares to $6.6 billion the prior September and $6.7 billion the September before.

BTS then dutifully posted those numbers, along with the Treasury tax itemization, on their Transportation Economic Trends website and other places. For anyone who isn’t overspecialized in the Highway Trust Fund, this looks like the government saying that the economic activity supported by the Highway Account of the Highway Trust Fund was at all-time record-setting levels in September 2024.

Normally, people are safe in making that assumption because almost all program outlays drawn from the Highway Trust Fund represent real-world economic activity that has already taken place. A state or local government has put a contract out for bid, a bidder has accepted, and that bidder bought materials, the state bought the land, the engineers finished their design work, and workmen were hired and built the project. Once the project (or a deliverable contract segment) was done, the state or local agency paid the contractor for the completed work, and only then did the US Treasury make the outlay. Real dollars out the door, on a cash accounting basis, to pay for real work already done in the real world.

But, as it turns out, $1.3 billion of that $8.4 billion of Highway Account outlays recorded for September 2024 does not represent real world economic activity under cash accounting. Instead, there was a $1.328 billion accrual accounting adjustment of the net “subsidy cost” to the Treasury of all outstanding loans made under the TIFIA program. TIFIA allows USDOT to make loans and loan guarantees for up to 75 years repayment period at ultra-low rates. Under the accrual accounting standards adopted in the 1990s, only the “subsidy cost” of the loan (the estimated default risk plus the difference between Treasury borrowing cost and the loan rate) shows up as part of the federal budget, and is booked as spending in the year the loan is made. At the start of this year, about $14 billion in TIFIA loans were outstanding.

But every year, that subsidy cost of those active loans gets re-estimated and shows up as either positive or negative outlays. For most budget accounts, this is an after-the-fact/Somebody Else’s Problem kind of spending or credit and gets lost in the wash of General Fund detritus at the end of a fiscal year.

But since TIFIA is funded from the Highway Trust Fund, Congress wrote section 608 of title 23, which says “If the subsidy cost of a Federal credit instrument is reestimated, the cost increase or decrease of the reestimate shall be borne by, or benefit, the general fund of the Treasury…” So, while TIFIA readjustments are shown as Highway Account outlays (made by the Build America Bureau in the Office of the Secretary), there has been a separate General Fund budget account (069-0149) since 2020 that reimburses the Trust Fund once a year for those TIFIA readjustment outlays. Both the outlays and reimbursement happen in September. The reimbursement is the only way we know how big the corresponding TIFIA readjustment outlays were.

The September 2024 TIFIA readjustment was much bigger than normal:

If TIFIA readjustments are ever going to be anywhere near this large again, then the TIFIA activity needs to be distinguished in some way from regular program outlays in public reporting, particularly if that reporting purports to use outlays as a measure of economic activity.

Why?

1. Cash accounting and accrual accounting are fundamentally different, and dollars accounted for by one method should never be fungible with dollars accounted for by another method to any significant degree. (Cash accounting is the dollar number on Grandma’s most recent monthly Social Security check; accrual accounting is that, times the actuarial estimate of how long Grandma will live, plus the economic forecast of what the annual COLA adjustments will be for the rest of her life, all totaled up and then annualized and divided by 12.) Normally, credit activity at DOT is just a rounding error on the rest of the FHWA/Department budget, so the mixing of cash and accrual numbers is relatively harmless, but not anymore. FYI, the Securities and Exchange Commission requires that publicly traded companies that use cash accounting present their non-cash accounting activities on a separate line in their cash flow summary tables.
2. As mentioned above, real program outlays represent tangible economic activity that has already taken place in the real world. TIFIA rescoring is both more abstract (what will interest rates are going to be like over the next 40 years) and more specific (what is the precise risk that the Maryland Purple Line will have to re-negotiate its loan terms), not how much pavement was installed in the U.S. last month.
3. Since the TIFIA adjustment only happens once per year, it has been recorded in the month of September, which makes for lumpy reporting. The FY 2024 TIFIA readjustment was only 2.3 percent of total FY 2024 Highway Account outlays, but it was nearly 16 percent of all Highway Account outlays in the month of September. Artificially lumpy data is bad data.

This doesn’t seem to be anyone’s fault, per se, but it needs to be addressed if TIFIA rescores are ever going to be this large again.

At any rate, this was a very long-winded way of saying that if you are looking at Highway Account outlays as a measure of tangible economic activity, the “real” September 2024 number was $7.1 billion, not $8.4 billion.