Here’s a new article in Water Finance & Management:

Additional Context

This post adds some context to the article, specifically with respect to federal fiscal constraints.

I’ve discussed SRF leverage in several prior posts. In this one, The Limited Buydown and SRFs, I suggest that a WIFIA limited buydown might help smaller SRFs take the first step towards portfolio leverage. It’s worth noting that a limited buydown (which TIFIA and CIFIA already have) would in effect allow the Program to offer sub-UST rates in certain circumstances. More scope to utilize sub-UST rates for specific goals is therefore a matter of degree, not a violation of a founding principle, among federal infrastructure loan programs.

In another post, In Retrospect, An Ironic Criticism of the SRF-WIN Act, I point out that if WIFIA’s interest rate re-estimates were included in the budget numbers, the cost of sub-UST rates wouldn’t look that bad. This is an example of what I’m saying in the WFM article about the need for specific evaluation context for Program tools and their intended purposes. And that evaluation needs to go beyond surface appearances.

Both of those posts dealt with somewhat technical aspects of WIFIA’s capabilities in SRF leverage. Now I think a more fundamental question is emerging: In the face of federal fiscal constraints, to what extent can WIFIA leverage replace direct grants?

The question arises generally because the US economic and fiscal outlook is, well, not looking so good. Specifically, it relates to two recent developments that may effectively constrain the generous funding provided to SRFs by the IIJA in 2021. The first is the earmarking of a significant amount of annual funding in 2022. Earmarking is not an overall reduction, but it is a constraint. The second, from a couple of weeks ago, is more ominous — NACWA warns against proposed ‘radical spending cuts’ to SRFs in FY24. The 2024 budget negotiations have a long way to go, so I don’t know how real this is. But even if the proposed cuts are primarily a political bargaining chip, it shows that annual SRF funding shouldn’t be seen as a politically untouchable entitlement. In hard times, it’ll be subject to the same constraints faced by other discretionary expenditures.

The policy objective of federal funding for SRFs is to increase their loan-making capacity. That can be achieved simply & directly with annual grants. But it can also be achieved, albeit less simply & directly, with federal leverage. The existence of SWIFIA effectively recognizes that federal leverage to increase SRF loan-making capacity might require special features in WIFIA loans like the limited buydown or sub-UST rates. When federal fiscal constraints are a priority, the cost of the features can be compared to the cost of direct grant funding to achieve a given amount of loan-making capacity. The features need to actually produce the same result but also cost federal taxpayers less. If the cost is the more or even the same, what’s the point?

How Might the Numbers Look?

The chart below shows a quick illustration of how I think WIFIA leverage could balance cuts in federal grant funding for SRFs. This is a bit speculative at this point — no doubt there’s a lot of devils in the details and maybe in the main assumptions as well. But the efficiency of leverage for financial portfolios is well-established, so I’m confident that the basic ideas are valid.

The chart looks at four scenarios:

Pre-Cut Funding for Unleveraged SRFs: Let’s say a group of unleveraged SRFs expected to receive $200 million from federal grants. The marginal effect of this (ignoring details like state contribution etc. for simplicity) should be to increase loan-making capacity by $200 million. That’s the policy objective and, among SRF stakeholders, the political expectation.

Post-Cut Funding for Unleveraged SRFs: But now assume that federal funding is proposed to be cut by 50% to $100 million. The effect will be a marginal decrease in loan-making capacity compared to the pre-cut scenario of $100 million. The policy objective won’t be achieved but more importantly, there’ll be serious political pushback.

Post-Cut with WIFIA UST Leverage: With $100 million of grant funding, the SRFs could (with a lot of simplification here) borrow $96 million from SWIFIA for a total of $196 million of new loan-making capacity, a very conservative 2:1 leverage ratio. Under current WIFIA, the credit subsidy cost of the loan with a UST rate would be about $1 million. For a total federal outlay of $101 million, the cut could be basically balanced by leverage and policy objectives maintained. SRF stakeholders of course would prefer grants, but as a Plan B, it might be acceptable.

However, the problem is that in general WIFIA loans with UST rates don’t seem to encourage SRF leverage. Would that change if the cuts became real? I don’t know — perhaps in some cases. But I am sure that since WIFIA UST loans are currently available, simply pointing that out to angry SRF stakeholders probably won’t help. If some sort of political compromise or at least amelioration is being sought, SWIFIA features will need to be improved before suggesting them as an alternative.

Post-Cut with WIFIA Sub-UST Leverage: Assume that improvement takes the form of SWIFIA being able to offer loans at 80% of the UST rate to SRFs that will especially be affected by the cuts, are currently unleveraged, are of smaller size, etc. Would that feature result in SRF leverage and the targeted increase in loan-making capacity? Again, I don’t know but I’d guess that there’s a fair chance it would be effective. For one thing, as pointed out in the WFM article, a sub-UST rate goes to the central issue of portfolio leverage, matching SRF debt service inflows and outflows. A sub-UST rate would substantially improve those numbers. For another, 80% UST loans for select SRFs were specifically sought by SRF-WIN proponents in 2018. They might have had some different objectives back then, but the expectation that this sub-UST rate would effectively increase leverage was presumably grounded in their own, well-informed expectations. And presumably, offering the same rate SRF-WIN proponents wanted but failed to get in 2018 would garner some positive political traction.

In terms of federal outlays, 80% UST loans require about 10% in credit subsidy cost. Total outlay would therefore be about $110 million on the federal side — still a big net cut from $200 million. But the substantive compromise of offering sub-UST rates to SRFs surely has a much lower cost in terms of political capital and a far better chance of maintaining policy targets.

The ‘win-win’ approach described here should be seen primarily as an alternative to attritional, zero-sum politics — there’ll be plenty more of that soon in any case. Expanding WIFIA’s capabilities is one way water sector stakeholders can avoid some of the worst effects of the coming storm.

This post will add another concept to a prior one, CWIFP Loans to Small Dam Funds. There I outlined why I think an eligible financial entity (like a trust) can bundle a portfolio of small dam project loans to meet CWIFP’s $20 million project cost threshold without requiring full cross-default. The key element is that the entity has capitalization that will be subordinated to the CWIFP loan, thereby making the entity a substantive ‘borrower’ even though the loan’s repayment will in effect rely on cash flow from the small dam loan portfolio. I noted that such capitalization would probably be necessary to meet CWIFP’s investment-grade requirement anyway, and suggested there might be a variety of third-party sources for it.

The added concept here is that maybe the entity’s capitalization doesn’t need to come exclusively or even at all from third parties but might be provided by the small dam projects themselves. This isn’t a new idea by any means — co-operative banks and credit unions are of course a well-established feature in many sectors, especially agriculture. Small operators in a sector might find it difficult to obtain efficient or cost-effective financing individually, but a larger institution collectively owned by them can achieve economies of scale and pass on the savings. Similarly, small dam projects, especially within a state or region, seem to have a lot in common, including an apparent lack of efficient or cost-effective financing. A CWIFP loan could help address this issue, but the $20 million threshold is in effect the required ‘scale’ to access this particular ‘economy’.

How such a small dam financing co-op would work is best approached by a hypothetical illustration. Imagine a collection of ten small dam projects within a state or region that have an average cost of $2 million and the ability to start construction at roughly the same time once financing is obtained. As a bundle, this meets the CWIFP threshold and permits an application to be made for a $10 million CWIFP loan. The project borrowers are willing to cooperate with each other, but they’ll never agree to a full cross-default of their individual loans or being on the hook in any way to other borrowers. They might consider, however, contributing something to capitalize an entity for the purpose of obtaining CWIFP financing.

How Much?

The real-world question is ‘how much?’. As discussed in the prior post, the borrowing entity can’t be a shell SPV. Though it’s not explicitly stated, I think some substance is required to comply with WIFIA’s bundling rule in Section 3905.10. But the minimally adequate amount of substance should be judged in the context of what the entity is expected to do.

That’s pretty limited, at least initially. The trust will apply for a single fixed-rate, long-term loan with very flexible terms from a buy-and-hold, policy-oriented federal program. If the application is successful, the entity will on-lend the proceeds on the exact same fixed-rate, long-term basis to a specific group of ten, otherwise completely CWIFP-eligible, small dam projects. There aren’t many moving parts in this picture — the only important risk is a credit default by one of the projects. The entity’s capitalization should therefore be evaluated primarily with regard to a substantive mitigation of that risk.

Assume for the moment that each one of the project loans could obtain a minimal investment-grade rating of Baa3/BBB- [1]. The NPV of the expected default loss associated with such a rating for a public infrastructure project loan is (very) roughly 5% of the loan amount. If the borrowing entity was capitalized with $500,000 of equity to cover expected losses from the $10 million loan portfolio, that would certainly seem to be substantive. After all, WIFIA and the other federal infrastructure loan programs do exactly the same exercise per FCRA rules to determine the adequacy of the required credit subsidy for their own loan portfolios. If the approach is good enough for OMB, why shouldn’t it be good enough for this small dam financing co-op?

For the project sponsor, $50,000 per project doesn’t sound like much — but remember we’re only talking about a $1 million loan each. Since it’s extremely unlikely that all the loans will default, they’ll likely get most or all back when the trust winds up, though that will be decades in the future. Perhaps a smaller amount will work if the portfolio is somewhat diversified? Or grant funding or philanthropic assistance could be sought for the purpose? All that, and many other angels and devils in the details, will need to be considered and worked out in any real-world situation. The only point here is that adequate capitalization for the purpose of Section 3905.10 should not be a large amount, probably in the range of 5% of the CWIFP loan.

A Revealing Simplification

It’s worth noting a possible simplification to co-op capitalization that doesn’t change the substance of the approach. For the very limited purposes of the CWIFP borrower entity, upfront cash capitalization to cover expected loan losses doesn’t seem to be necessary. The capital is there only to offset final losses after default, work-out and recovery. For public infrastructure project loans, this process happens very rarely, very slowly and takes a long time to complete. There’s no point in keeping $500,000 in cash available for that purpose. Presumably, a capital call or similar undertaking by each of the co-op’s members for their share would be sufficient.

That simplification also reveals another perspective on the co-op approach with respect to Section 3905.10. As discussed towards the end of the prior post, hardliners might insist that full cross-collateralization and cross-default is required from each borrower in the project bundle. Well, okay — I can see the principle of some substantive connection between the project loans. But why does it need to be an unlimited cross-collateralization or cross-default obligation? Why wouldn’t a limited obligation that reflects some important substantive risk aspect of the project loans be sufficient? Especially if that limited obligation is assessed in a very similar way to federal FCRA budgeting for loan program credit loss reserves? No one expects a 100% credit subsidy for investment-grade infrastructure project loans to be necessary [2]. Why a hard line against small dam borrowers with projects that are demonstrably consistent with CWIFP policy objectives but a soft line for the amount of credit subsidy required from federal taxpayers?

Of course, the capital call described above for the co-op CWIFP borrower is essentially a limited cross-collateralization obligation directly to the CWIFP loan. If one works, the other should, too. There may well be other reasons to develop a small dam financing co-op, not the least of which is the option or the need for third-party capital to be involved in the mix. But at the very least, it strikes me that a co-op approach might be a useful way to start the narrative that project bundling for small dams ought to be actively encouraged at CWIFP in the expectation that hardliners will inevitably surface sooner or later.

_____________________________________________________________________________________________

Notes

[1] This assumption is probably not realistic in many cases, which means that the third-party capitalization discussed in the original post will likely be necessary and the substantive nature of the borrower entity will automatically follow. Other, more classic reasons for a co-op approach — control, cost-savings, scale economies, etc. — would still be applicable, especially with respect to getting the ball rolling.

[2] Well, perhaps in a way some hardliners do.

One of the action items from the Dam Finance Roundtable was to develop a summary of “dam finance recommendations, both for US dam finance in general and CWIFP in particular.” I don’t have sufficient real-world expertise in the dam sector to add much to the ‘general’ part of these recommendations without falling into a Dunning-Kruger trap. But as you’d expect on this site, I do have some specific ideas for CWIFP that stem from my thinking about WIFIA and other federal infrastructure loan programs.

For clarity and to establish context, I’m going to approach this exercise in a series of three posts that start with questions:

1. What is the sub-set of US dam projects that can, or potentially could, realistically benefit from CWIFP financing?

The US has a lot of dams, many of which are in need of some type of repair, rehabilitation or removal. But only a sub-set could benefit from what a federal infrastructure loan program like CWIFP can realistically offer. Using data from various sources like the National Inventory of Dams and the ASDSO reports, I think the basic scale and characteristics of this sub-set can be roughly determined. In effect, that will help define CWIFP’s ‘customer base’ and their needs & limitations.

2. What are the current policy objectives that guide CWIFP implementation or modification with respect to dam finance?

I’m making the general assumption in this exercise that recommendations must be consistent with furthering current CWIFP policy objectives. Of course, these could change as things develop and a possible type of recommendation involves policy refinement or redirection, especially in connection with additional funding. But this requires real-world sectoral expertise that’s beyond my pay grade. Instead, I’ll try to summarize those CWIFP current policy objectives which seem to be especially relevant to meeting the needs of the program’s customer base in the immediate future.

3. For potentially relevant dam projects, what realistic & near-term actions could be taken to improve CWIFP’s ability to achieve its policy objectives?

These are the recommendations. There’ll be two types: (1) minor legislative refinements in the WIFIA statute that could be enacted relatively soon and are especially applicable to dam projects, and (2) outreach and innovative development that’ll likely be required for smaller and more challenging dam projects. The latter might also include possible future legislative changes as a realistic, but not necessarily near-term, goal.

As I did for the FCRA Non-Federal Series, I’ll highlight & link each question as the posts are completed over the next few weeks.

When a federal credit program funds a loan during a budget deficit (i.e., always now), the process will ultimately involve a marginal increase in federal Treasury debt equal to the full amount of the loan. The federal budget impact per se will be much smaller (only the FCRA credit subsidy), especially for large-scale infrastructure loans, e.g., about 1-2% of loan amount for WIFIA. But that doesn’t matter if the concern is the size of the national debt in relation to GDP — a $1 billion infrastructure loan will have the same impact as a $1 billion infrastructure grant.

There is one important difference of course. The federal loan will be repaid over time with interest (only the minor credit subsidy amount is intended as a permanent transfer) whereas the grant is out the door forever. In effect, the Treasury debt issued to fund the loan is largely self-liquidating since it is matched with an interest-earning and amortizing asset.

That’s probably not too important with respect to the size of the national debt at any point — after all, a lot of federal discretionary spending is meant to have a positive economic impact over time, so in theory debt issued for that is self-liquidating, too. Maybe even multiple times over. Yes, the ‘in theory’ qualification is quite an understatement, but the point is that there’s not necessarily a bright-line distinction between a federal credit financial asset and other federally created economic ‘assets’ in terms of the federal balance sheet.

But when the concern over US federal debt is focused on the inexorable need to pay interest on the issued Treasuries, the picture changes. Arguably, that’s the real problem with federal debt — ultimate impact and future repayment are tomorrow’s somewhat hypothetical and politically spinnable problem once the debt is issued. But paying scheduled interest requires writing checks that cut into other discretionary spending, which is very much a real-time problem.

And it’s beginning to look like that that problem has arrived and is expected to get much worse. Here’s a recent analysis by the Committee for a Responsible Federal Budget: Interest Costs Will Grow the Fastest Over the Next 30 Years. The graphic tells the story — net interest cost lurked in the shadows when interest rates were suppressed even as federal debt rapidly grew. Now it’s back with a vengeance:

This gloomy picture provides some context for considering the expansion of federal infrastructure loan programs in anticipation of even harder times ahead. Unlike the economic returns from other federal ‘assets’, the interest paid on federal loans should provide a solid and precisely predictable offset to the interest cost of the Treasury debt issued to fund them. I think this is actually imbedded in FCRA accounting mechanics whereby the interest that the loan programs collect from borrowers gets credited to their intragovernmental liabilities with Treasury (that’s how it works for the infrastructure loan programs I’m familiar with anyway). So even if a $1 billion loan has the same balance sheet impact as a $1 billion grant, the impact on Treasury’s net interest cost should be very different. For infrastructure loan programs offering interest rates based on UST yields at closing, the offset should be almost complete [1]. So, unlike other federal spending, expanding federal infrastructure loan programs shouldn’t add much to that steep red line.

There is a huge caveat, however: New or significantly expanded loan programs need to be well-designed and carefully implemented. Solyndra-like disasters vaporize the interest offset and in effect convert loans into unintended and demonstrably bad grants. But overly risk-averse or bureaucratic programs will fail to have any real impact. Not easy to get it right. Since the need appears to be predictable and increasingly imminent, why not start now?

_____________________________________________________________________________________________

Notes

[1] Well, that was the legislative intent. For infrastructure loan programs that offer rate locks for long construction periods, the story is more complex: WIFIA’s FCRA Re-Estimate Elephant

This post is an addendum to the FCRA Non-Federal Series and my other FCRA analyses for federally involved projects. It adds some thoughts about a short paragraph for the use of WIFIA loan proceeds in current law. The provision touches on loan repayment sources in a way that is consistent with my prior FCRA analyses. However, it’s worth reviewing in a little more depth not only for completeness (I should have included it earlier) but because a careful reading might shed some light on OMB’s thinking behind a footnote in their published FCRA criteria.

Section 3908 (b)(8)

Section 3908 of the WIFIA statute describes various guidance and rules for the program’s secured loans. This is where the financial criteria for Transaction Eligibility are found, as I used those terms in the FCRA Non-Federal No.1 post.

Paragraph (b)(8) in this section is (somewhat implicitly) about federal cost-share situations, in effect federally involved projects:

(8) Non-Federal share
The proceeds of a secured loan under this section may be used to pay any non-Federal share of project costs required if the loan is repayable from non-Federal funds.

The provision simply confirms that using loan proceeds for a local cost-share in a federally involved project is permitted as long as the source of repayment is non-federal. It wasn’t developed (or perhaps even thought about much) for WIFIA specifically — the exact same language is found in exactly the same place in TIFIA law. So most likely a cut-n-paste. [1]

That’s straightforward enough on the surface and the repayment proviso is completely consistent with our conclusions here about FCRA budgeting treatment for federally involved projects. But note that the (b)(8) provision (and Section 3908 in general) is not really about FCRA or any type of budgeting by WIFIA. Most directly, the provision reflects a policy decision that WIFIA loans can be used for this purpose (presumably because it will help deliver more of the sought-for infrastructure) with a proviso that can also been seen as policy (or prudential?) decision — to ensure that the financed cost-share really is an external share, not some sort of intra-federal shell game. If the purpose of a cost-share is to benefit federal taxpayers by reducing the federal resources involved in a big project, the reduction has to be ultimately sourced from non-federal taxpayers, whether paid for upfront or over time by debt repayment. As such, including the proviso would make sense even if FCRA didn’t exist.

I can imagine that the drafters of TIFIA law thought about their (b)(8) paragraph in this policy context. They seem to have decided that valid cost-sharing in federally involved projects can promote win-win outcomes and their new loan program should be explicitly permitted to help. But they also wanted to avoid an opening for internal federal games and added the proviso. WIFIA inherited their decision, perhaps thoughtlessly or simply because such an obviously good idea didn’t require much thought. In any case, Congress agreed in both cases. In effect, (b)(8) is a kind of statutory eligibility in TIFIA and WIFIA that has nothing to do with FCRA per se.

Footnote 4 in OMB’s Criteria

With the above policy objectives in mind, and FCRA out of mind for the moment, we can revisit OMB’s footnote in the 2020 Criteria:

(4) WIFIA authorizes loans to support local cos-sharing requirements. See 33 U.S.C. 3908(b)(8) …. However, such a loan that would finance a project that is in whole, or in part, a project authorized by Congress for the Army Corps of Engineers or the Bureau of Reclamation to construct would not meet the Federal asset screening process…

In effect, this footnote modifies Section 3908 (b)(8) by adding another proviso that rescinds the permission to use WIFIA loan proceeds for cost-share situations which involve two specific Federal agencies, the Corps and the Bureau, regardless of repayment source. The fact that OMB adds a rationale about their decision that projects involving these agencies will never pass their criteria is not very relevant.

Obviously, Congress can modify a loan program’s statutory eligibility (they probably should do that more often) but what’s OMB’s position here? Certainly, their job is to interpret statutes in terms of operational management and add prudential guardrails to their implementation, especially when large-scale federal spending is involved. Some sort of guidance about how federal loans to federal cost-share situations should be managed (again, purely from a policy implementation perspective) would seem to be fully justified in light of the potential, in itself reflected by (b)(8)’s proviso, that games might be played. OMB’s Circular A-129 about loan program management is full of such guidance and in effect, sets forth a lot of rules.

Let’s assume for argument’s sake that such guidance about WIFIA’s (b)(8) eligibility is especially warranted for two the Federal agencies mentioned in the footnote. You could imagine a set of procedures and check lists that applied whenever these two agencies were involved, with particular emphasis on confirming certain facts. The totality of this could present a very high bar for potential WIFIA applicants to overcome, and as a practical matter, many if not all would take a pass. But this approach would still be consistent with interpreting statutory eligibility, not modifying it.

In contrast to this hypothetical, I think OMB’s footnote 4 crosses the line into statutory modification. The footnote is written with definitive language (“would not meet”) that might be applicable in very simple, black-and-white financing situations. Something self-evidently outside Congressional intent that no one would disagree with — e.g., “cost share financing for terrorist organizations would not meet our screening process”. But for the complex, sophisticated multi-party project financings regularly developed for the kind of infrastructure project that these agencies would be involved with, how can OMB claim that all possible variations of loan use and repayment sources will fall afoul of their ‘process’? The factual pre-judgement is so broad that the footnote becomes an unsupported and unjustified assertion — in effect, a modification of WIFIA’s statutory eligibility.

Footnote 4 would be bad enough if it showed up in a policy-oriented circular like A-129. But here it is in criteria that are solely meant to clarify a FCRA budgeting issue. Yes, (b)(8)’s proviso is consistent with a correct interpretation of FCRA law, but as discussed above, there are other reasons to include the proviso from a policy perspective, unrelated to FCRA budgeting. And if the published FCRA criteria are working properly for all cases, why is it necessary to single out loans involving two specific agencies for differential treatment?

Unfortunately, I cannot help thinking that OMB had already made a decision that WIFIA loans for projects involving the Corps or the Bureau needed to be prohibited for some reason which might be related to budgeting but likely not FCRA per se. The FCRA criteria presented them with an opportunity to make a stealthy statutory modification in a context that’s so intrinsically technical and complex that few would dare question it. If this background is basically true, that doesn’t mean the tactic was clearly and malevolently plotted out. Confusion about FCRA and what the criteria were meant to accomplish probably played a big role, and a longstanding wish (very possibly well-meaning) became father to the specific footnote. That’s consistent with the rest of the criteria’s Alice in Wonderland nature, as described in this recent post.

Regardless of intent, however, the overall conclusion about OMB’s criteria in that post applies especially to footnote 4 — just junk the lot and start over.

_____________________________________________________________________________________________

Notes

[1] Interestingly, the provision doesn’t appear in CIFIA law, which was drafted after the FCRA criteria were published. This might reflect some impact of the FCRA issue — though I doubt it. CIFIA is different in other ways and (most importantly) that program’s advocates would not have been contemplating federal involvement in private-equity backed, 45Q monetizing, profit-maximizing pipeline projects, in contrast to public transportation & water infrastructure projects.