This post is to clarify two technical points about topics frequently discussed here. I’m outlining these for the ‘avoidance of doubt’, as the lawyers say, in case some readers have questions.
1. The Difference in Value Between 35Y and 55Y WIFIA Loan Terms
In theory, a 55Y WIFIA loan should always be more valuable than a 35Y loan simply because more of the ‘flat forward’ part of the US Treasury is being incorporated into the loan commitment rate (or ‘Single Effective Rate’ as OMB zero-coupon methodology is called).
Source: InRecap analysis 010826
But in reality, it is easy to understand why big Aa3/AA- water agencies wouldn’t care that much about it, regardless of the theoretical numbers. This is because they can — and do — see their financing as an actively managed component of the capital structure. Parts will be refinanced as soon as there’s the slightest benefit in doing so (e.g., advance refundings of tax-exempt bonds when that was still available) and in effect extended in the process of doing so. The overall liability side of the complex balance sheet constantly moves into the future with about the same maturity, matching the similarly evolving asset side. In this context, the primary value of a WIFIA loan is in the interest rate management features, which are in fact handy tools for active liability management over the next five or ten years, well within the standard planning horizon.
For them, a 55Y term will still cut some basis points off the initial commitment rate, but if the WIFIA loan is expected to be reset or refinanced at some point after the construction period, that might not be considered particularly relevant. Perhaps a 55Y loan term might spark some concern among less financially knowledgeable city councilors or folks like that, regardless of penalty-free cancellation. Like ‘why are we going to pay more interest?’ or ‘why are we entering into such an unusually long-term commitment — do we need a special vote?’ and so on. Not worth that hassle. And finally, big agencies wouldn’t want their bond underwriter to raise an eyebrow over their utilization of a WIFIA feature that not only highlights (another) limitation of that market but would also have broader applicability — “yes, WIFIA loans are OK as an occasionally useful option, but don’t go too far and give people ideas about federal infrastructure finance” or something similar. The agency’s relationship with the bond market is important and likely predates WIFIA by decades.
In contrast, a large regional project for permanent water management is likely to be done as a project financing. Project financings aren’t intended to be actively managed on the liability side. Project revenues are scheduled to amortize the asset’s initial (usually Baa/BBB-ish) long-term debt, it’s assumed that refinancing such an idiosyncratic deal might be difficult, and the planning horizon is closer to the asset’s useful life. The mindset is ‘one and done’ — “let’s get this flood control system built, finance it in the cheapest and most stable way possible, pray that nothing unexpected happens, and get back to our day lives”. In this context, the 55Y term is obviously valuable — cheaper and longer than any alternative. It’s no wonder that ‘western water’ congresspeople have push for it.
Other, less ‘actively managed’ infrastructure financing situations might also benefit. Some large SRFs do leverage in a big way, and they’ve happily done a few 35Y SWIFIA loans. But the majority of SRFs don’t leverage optimally, if at all. I’m told that this primarily reflects a staffing constraint. If that’s the case, a ‘one and done’ 55Y WIFIA loan commitment with slow amortization might help if the leverage was proposed as a sort of infrequently sought-for permanent addition to the capital structure, not as another ongoing managerial obligation. If there’s interest in this, a 55Y amendment could be easily modified to include SRFs SWIFIA loans, regardless of portfolio assets (the states themselves are long-lived enough, no?).
Similarly, lower rated rural water systems or, more realistically, a portfolio of loans to such systems, might simply require ‘project financing’ because the cash flows between assets and liabilities will need to be carefully matched. No illusion of refinancings, not much margin for error, and every penny will count. My impression is that these folks are trying to replace basic — and intrinsically — long-lived stuff (pipes, sewers, levees, etc.), so the 55Y term may often be a better match for useful life and intergenerational equity. The numbers and amort schedule benefits will speak for themselves.
Of course, there are more than a few hints of a ‘class divide’ going on here between those who don’t care about the 55Y and those that do. Much more on that substantive topic in future posts.
2. Mandatory Spending Only Reflects Positive Re-Estimates
I’ve been using ‘mandatory spending’ as shorthand for WIFIA’s re-estimate issue. The handle correctly conveys the relevant consequences of the operations of FCRA machinery, but precisely speaking, it’s only one side of the equation. I’ll sketch out the whole picture here for completeness, to the best of my understanding.
FCRA law just talks about re-estimates in general as a ‘change in program costs’:
504 (f) RE-ESTIMATES.–When the estimated cost for a group of direct loans or
loan guarantees for a given credit program made in a single fiscal year is
re-estimated in a subsequent year, the difference between the re-estimated
cost and the previous cost estimate shall be displayed as a distinct and
separately identified subaccount in the credit program account as a change in
program costs and a change in net interest. There is hereby provided
permanent indefinite authority for these re-estimates.
OMB Circular A-11 Part 5 On Federal Credit (pp, 547-631) splits the ‘change’ into either a ‘positive’ re-estimate (drawdown rate higher than commitment rate, higher cost) or a ‘negative re-estimate (vice-versa, lower cost) and spells out different treatment for each, as would seem to be allowed under 504 (f):
- A positive re-estimate “must be recorded against permanent indefinite budget authority available to the program account for this purpose.” So, the higher cost goes straight into the PIA account, which automatically incurs mandatory appropriations, which I think is generally turned into mandatory spending either simultaneously with funding the loan (the ‘spending’ is the top-up of an otherwise deficient Treasury account) or very shortly thereafter.
- A negative re-estimate, however, “will be recorded as offsetting receipts, which will offset the total budget authority and outlays of the agency and the budget subfunction of the program. However, at the discretion of the OMB representative with primary responsibility for the program, a special fund receipt account may instead be established.” In this case, the cost is less than expected — what do to with the ‘windfall’? From 504 (f), the lower cost must be recognized, but it is not specific as to how. I think OMB took a utilitarian approach, saying in effect, “dump the gains back into the general program budget, unless we decide some special treatment is necessary.”
All quite workable. But note there is a disconnect between the re-estimate gains and losses — they don’t cancel out. In theory, shouldn’t the program ‘pay back’ the losses before it can use the gains? That’d seem to be fairer to taxpayers and arguably provide a more useful budget metric. However, I can see how that might not be so simple. The losses are immediately realized, and the mandatory spending is sent to Treasury. If gains occur the next day (or the next minute), I assume you can’t ‘reverse’ the mandatory spending easily, if at all — the mighty PIA has been invoked, after all. Practically speaking, probably the best you can do is just keep the gains at the program and use them, as OMB prescribes. Notice that taxpayers won’t get any benefit at all if you can’t, whereas if the gains are generally usable, in theory fewer program discretionary appropriations will be required in future. Over time, things should more or less net out with respect to value, and of course detailed records are kept so a precise net position can be calculated if required.
We should keep in mind, notwithstanding the WIFIA case, that both the drafters of FCRA law and OMB rule makers likely assumed that interest rate re-estimates would always be a relatively minor item — most loans would be drawn quickly relative to interest rate changes, and the gains/losses would be quite minor relative to the big discretionary numbers involved with, say, risky student loans. Kind of a bookkeeping annoyance, certainly not worth a lot of complex policy. Or so perhaps it seemed at the time.
Here’s the important clarification: It is clear from the above that WIFIA’s mandatory spending numbers are only half the picture. It is logically possible that the program incurred $1.6b of re-estimate losses while racking up $1.6b of re-estimate gains, rendering this topic a non-issue. But of course, the program didn’t — and won’t. Ever.
First, there’s no evidence from the numbers I can see (mostly, WH Budget Technical Appendices) that some big lumps of negative re-estimates are appearing in the program’s “total budget authority”. The numbers below, FY24-FY26 include some bits and pieces, but generally the levels correspond to current discretionary funding or to carryover balances of unused discretionary funding. The carryover balances track almost exactly WIFIA’s actual loan volume 2018-2025. The big mandatory numbers that appear in FY24 and FY 25 do not carryover anything material, presumably since they’re spent the same year. Although WIFIA’s mandatory spending is only half the picture, there is no evidence that the other half is material — and so the shorthand description works.
Second, there’s a more fundamental reason that WIFIA only incurs positive re-estimates. Consider the question: In what planet would Aa3/AA- water agencies with excellent access to tax-exempt bonds ever drawdown a penalty-less loan commitment with an interest rate materially higher than their alternatives? None. Back in the real world, of course, these agencies will draw on in-the-money loan commitments all day long. The behavior is entirely predictable and (from their perspective) rational, but it was not anticipated by FCRA law nor the OMB rules. More on this topic in future posts, too.
But if you can’t quite tear yourself away from arcana yet, here’s the appendix to the 2021 WIFIA cost analysis that creates a model showing how ‘loan commitment drawdown optionality and correlation’ inevitably leads what we’re seeing at WIFIA now:
The-Economic-Cost-of-WIFIAs-Current-Loan-Portfolio-Part-1_-The-Potential-Cost-of-Interest-Rate-Re-estimates-2021-Appendix-A-1