As outlined in Chapter 5 of the 1967 Budget Report, the foundational principle of FCRA is that a federal loan has a substantive non-federal source of repayment. The loan is an outflow, like other federal payments. But repayment from non-federal sources is an inflow that reverses, to a precisely calculated extent, the initial outflow. Recording such reversing cash flows as expenditures and revenues would seriously distort the federal budget. Hence there is FCRA, which records federal loans on an accrual basis in accordance with well-established financial accounting procedures. It is not mysterious, nor much open to gaming, which is unsurprising since the sole purpose of enacting such special treatment for federal credit was to prevent budget games.

A substantive non-federal source of repayment is something that should be straightforward to establish objectively. It’s worth noting, however, that another federal inflow also comes from non-federal sources — federal taxes. The difference is that in a federal loan the repayment obligation is undertaken voluntarily, and loan proceeds are used (within the limits set by the loan program) in the borrower’s own self-interest. In contrast, the federal government uses its sovereign power to force tax inflows from non-federal sources and allocates the proceeds to general public goods. The former process is recorded under FCRA, the latter exclusively in the federal cash budget. The distinction is important, but I think in the overwhelming majority of federal loans it will be so glaringly obvious as to not require much, if any, confirmation.

Non-Federal Cost Shares in an Honest Land

As discussed in a recent post, FCRA Non-Federal Issue: A New Approach, federal loans for non-federal cost shares in federally involved infrastructure projects are a bit more complicated because the borrower’s self-interest is mixed with federal interests. But in an honest land, the distinction between a non-federal source’s repayment of a federal loan and payment of federal taxes should be abundantly clear even in a cost share financing.

In a normal deal, the non–federal interest may be complex, but the character of a FCRA-classified repayment (versus federal taxes) can be determined from an examination of the non-federal borrower side alone. Did they voluntarily agree to undertake the loan’s repayment obligations in order to pay for the cost share? Did they get specific, commensurate value for themselves? If ‘yes’ to both, the cash inflows to the federal government from the borrower are voluntary and self-interested. They can’t be recorded as taxes and instead belong under FCRA.

A World Devoid of Crooks?

The lapidary amendment language in HR 8127 and HR 2671 appears to be assuming that with respect to FCRA classification, the world is necessarily an honest and straightforward place. That’s somewhat understandable. Aren’t the reversing cash flows from a non-federal repayment source in themselves sufficient to force the loan into FCRA, whether or not anyone wants that, per the section’s founding principle and anti-gaming purpose? All the relevant facts confirming a non-federal characterization of the loan’s repayment source will certainly be surfaced by the credit rating agency’s extensive review. Plenty of other mischief might be occurring in the deal, but how could any of that alter such visible and objective facts?

Sadly, under current WIFIA law, I can see a way that crookedness could result in FCRA misclassification, not as a primary goal but rather as a minor side effect. This is because even manifestly non-federal sources apparently repaying a federal loan might in effect be paying federal taxes.

There are two varieties of a con that have this effect, both involving federal sovereign power, the unique factor in non-federal cost shares in federally involved projects. The first was described in the New Approach post. If an unethical federal participant leans on a non-federal source to pay for a share in a project they don’t want or find beneficial, the source is being forced through the misuse of sovereign power to pay for a federal asset, which might even look like a national public good. The payment is an unauthorized yet de facto form of federal taxation, and if there’s a federal loan in the middle, all loan repayment will in effect be federal tax payments, too. The transaction doubtless belongs in criminal (or in less extreme and more realistic cases, civil) court, but for as long as it stands, it should be recorded in the federal cash budget, and the loan can’t be classified under FCRA.

The second variety works the other way around. A non-federal crook bribes the federal participant with some unrelated quid-pro-quo to deliver to its cost share an amount of value far in excess of what the non-federal source pays. The federal participant redirects resources from tax sources to create the excess value. The cost-share payment, such as it is, can be seen as a minor, on-the-books part of the purchase price for a redirection of federal taxing power. Hence, the payment is primarily related to federal tax, not the project. I have no idea how this would be recorded in the federal cash budget prior to various well-deserved indictments, but that’s where it belongs. Definitely not under FCRA.

I hope this kind of thing is vanishingly rare. I assume it is for large scale infrastructure projects. Nevertheless, the theoretical possibility of a misuse of federal sovereign power shows that a non-federal source of repayment is a necessary but not quite sufficient condition for the FCRA classification of a federal loan to finance a cost share in a federally involved project.

A World Full of Imbeciles?

The current WIFIA Criteria don’t appear to be too concerned about crooks. Instead, they apparently assume that the world is full of imbeciles.

The so-called Criteria determine, through a series of questions and an undisclosed process (that may or may not include actual criteria), whether the project or asset being financed by the federal loan is ‘federal’ or not. The asset’s value and utility with respect to loan amount or the borrower’s motivation don’t seem to be relevant. Instead, the objective appears to be an accounting classification along the lines of subsidiary consolidation per private-sector GAAP. In a cost-share situation for a large infrastructure project, there will be many undivided and intangible assets where federal and non-federal interests interact in complex ways, but the Criteria assume the whole thing to be ‘federal’ unless proven otherwise, apparently beyond the slightest shadow of doubt. I think we know how that will turn out. The Criteria publication helpfully notes that cost shares in USACE or Bureau of Reclamation projects needn’t bother to go through the process at all — they’re simply presumed to be ‘federal’.

I’ve discussed this rabbit hole in several prior posts, most directly here. For this one, I’ll focus on the bizarre implications of the Criteria classifying (however erroneously) a cost share as a ‘federal’ asset with respect to the characterization of the loan’s proposed non-federal source of repayment. In one sense, this is an academic exercise, since the inevitable ‘federal’ classification of the cost share will result in immediate rejection of the loan application, likely the intent all along. Still, the implications are worth considering, not so much to add to the list of the Criteria’s shortcomings (they’re damning enough already), but as context for some aspects of the Criteria’s possible replacement. It’s also oddly amusing, to be honest.

Most glaringly, the non-federal participant will certainly be very surprised to learn that the cost share it planned to finance with a WIFIA loan won’t actually be theirs, but a ‘federal’ asset instead. All those months and years of planning and analysis about the critical value the cost share will provide to the community and investors? All those hard-fought negotiations with a federal participant vigorously defending federal interests? All those presentations, public meetings, council approvals and voter referenda about investing a large amount of money into an important infrastructure project? Stacks and stacks of contract documents, engineering studies and benefit-cost analyses detailing tangible and intangible value, precisely defined undivided interests and countless rights and obligations?

Delusional! Pointless! The Criteria have determined that there is no non-federal asset in this project! What you thought was yours is in fact a national public good! You get nothing!

Since the Citeria have pronounced that WIFIA loan proceeds applied to the putative non-federal cost share will in fact be paying for a national public good, loan repayment will be a form of federal taxation, as in the case of federal extortion described above. Interestingly, extorted federal taxes should get the same treatment in the federal budget regardless of the method of payment — illicit or not, they’re tax revenue, pre-indictment anyway. But unwitting federal taxes apparently do not. If the non-federal participant paid for its cost share in cash upfront or financed it in the federally subsidized tax-exempt bond market, presumably the payment to the project would not be recorded as revenue in the federal budget, perhaps being only indirectly reflected as a reduction in the federal participant’s budget for project cost.

However, if the non-federal participant financed the same cost share with a WIFIA loan (theoretically speaking, of course), the Criteria would spring into action and in due course discover a new ‘federal’ asset that (per cash accounting) can only have been paid for by, well, federal tax revenue. Specifically, the WIFIA loan’s principal amount would (via a 100% credit subsidy) be a tax expenditure for the new-found asset, to be balanced by a new federal tax revenue stream that covers the full cost, even including UST interest on the initial deficit financing. Substantively, WIFIA isn’t really making a loan (all the initial flows are now intra-federal) but acting as a tax collection agency specializing in long-term contracts that deliver new revenue in the nominal form of ‘loan repayment’.

In a twisted way, this actually doesn’t sound so bad for the federal government. If the Criteria’s agency operators really believe that the proposed non-federal cost share is a federal asset that an imbecile is willing to pay additional taxes for, and that this reality is uniquely surfaced for inclusion in the federal budget by said imbecile’s use of WIFIA financing, why aren’t they encouraging it?

As I said, bizarre. Isn’t this the kind of budget game-inducing irrationality that FCRA was enacted to prevent?

Back in the Real World, Common Ground

Of course, the proponents of the amendment in HR 8127 and HR 2671 know that no technical financial classification or definition can be completely crook-proof, especially when there’s big money involved. My guess is that they simply couldn’t conceive of any concrete way that mischief could be involved in this case, and that generalized or vague restrictions would allow the worst elements of the Criteria to enter through the backdoor. The proponents’ basic sense that FCRA classification is straightforward and would be harmed by unnecessary qualifications is correct. But if shown a precise scenario for FCRA misclassification, I think they’d agree to include specific and limited criteria to address it.

Likewise, I am certain that the Criteria’s authors did not intend to imply that non-federal cost-share participants are imbeciles volunteering to pay more federal tax. Frankly, I’m not sure how much they thought through the FCRA issue or the Criteria’s consequences, other than what was required to obviate the issue by effectively denying all applications. My impression is that there was lot of confusion, not much real-world financial knowledge and some elements of a hidden agenda going on. But I believe that they were also motivated by a genuine feeling that the FCRA classification of WIFIA loan for a cost share in a federally involved project requires some extra steps, though they didn’t understand exactly why or what the steps should be. That feeling is justified. Again, if shown what’s actually required and how that would be consistent with FCRA law and principles, they (or some of them, anyway) might quietly agree to allow the current Criteria to be replaced if it can be done in a face-saving way.

The important point here is that there is non-controversial common ground, solidly based on FCRA law and principles, between the pro-amendment and the pro-Criteria sides. Both sides are partly right and partly wrong, and both can claim a win if the common ground is tacitly agreed to be the basis of a solution. If FCRA was less of an abstract and specialized area, the common ground would likely be more obvious. But it’s clearly not. The first challenge is to get there, the rest might be easier.

1. A non-federal asset is created when a non-federal participant pays for a share of a federally involved project with upfront cash from non-federal sources, right?

2. If the non-federal participant finances the same share by issuing a federally subsidized tax-exempt bond to be repaid from non-federal sources, that’s OK too, right? The cost share is still a non-federal asset, yes?

3. But if the non-federal participant finances 49% of the same share with a federally subsidized WIFIA or CWIFP loan to be repaid from non-federal sources, that’s — not OK? The cost share is now not a non-federal asset, but a federal asset, because — why, exactly?

4. Because the current WIFIA Criteria say so?

FCRA-issue-amendments-08182023-v2.0-InRecap-1

^{PDF of the document here}

WIFIA’s current FCRA Criteria for financing non-federal cost shares in federally involved projects are seriously and fundamentally flawed. A radical fix is needed. But what should it look like?

A New Approach

Attempting to revise the current Criteria would amount to rearranging teacups in a rabbit hole. But the simple amendment language approach in HR 8127 and HR 2671, though clear and indisputably correct with regard to FCRA law, has its own theoretical and practical problems. I think a new approach is necessary.

Any approach to a solution at this point will certainly require amendments to WIFIA’s statute, but this is not necessarily bad. In the amendment process, Congress should also effectively be confirming its intent that WIFIA loans are meant to support water infrastructure sectors outside of state and municipal water agency projects. Such a general confirmation is important, I think, not only for CWIFP but other potential classes of WIFIA borrowers, as well. However, the amendments themselves should be narrowly focused on addressing the specific FCRA issue faced by non-federal cost-shares in federally involved projects, an area that realistically concerns only CWIFP and perhaps a few other federal infrastructure loan programs. A minimalist solution is all that’s needed in this case and has a better chance of getting through the process.

A Cost-Share is a Non-Federal Asset, not a Donation

The specific problem for non-federal cost shares in federally involved projects is that FCRA law does not define a ‘non-federal borrower’. The law’s drafters probably didn’t think it was necessary. The absence of a precise definition is not a problem for the vast majority of federal credit programs where the borrower is clearly using the loan proceeds solely for its own, manifestly non-federal purposes, and for which it agrees to be on the hook for repayment from its own, non-federal resources.

Things are inevitably more complicated in an infrastructure project cost-share situation where federal and non-federal interests are mixed. However, I think it still can be presumed that the non-federal partner is acting in its own interest and will only agree to undertake project obligations (e.g., the repayment of a WIFIA loan) to the extent they’re confident that they’ll receive, in commensurate amount, their own, non-federal value. That value, no matter how intangible or undivided within the project, will be an identifiable, measurable and contractually enforceable non-federal asset. The non-federal partner is not financing a patriotic donation, after all.

But No Funny Business

Nevertheless, the presumption of a borrower’s true and unadulterated non-federal self-interest should be validated in the application process. Large-scale infrastructure finance involves big money, complex arrangements and a lot of politics. Shocking, I know, but in the real world, budgeting games are sometimes played, undue pressure may be applied, participants are not always above-board, costs and assets may be incorrectly valued, etc. Such shenanigans can conceivably taint the character of the borrower with respect to FCRA’s implicit assumptions about non-federal self-interest being the pure motivation behind non-federal repayment. For example, an unethical federal participant might decide to ‘persuade’ a reluctant local community to pay for a share in its project with some sort of veiled threat. The intimidated locals cave to the shakedown and finance the extorted amount. Assuming the federal participant is not crooked in other ways, the resulting cost-share asset may well have value equal to the loan, but not for the local community, who can’t use it and didn’t want it. The share is, in effect, an ill-gotten federal asset. The loan’s source of repayment is nominally non-federal, but substantively federal, as a type of taxation made possible by the misuse of federal sovereign power. Obviously, this is not an arms-length deal, unless twisted arms count. And it’s not at all what FCRA law and principles were assuming.

In our imperfect world, it is therefore simply common sense to require that a WIFIA applicant for a loan to finance a cost share in a federally involved project be asked, in addition to all the other due diligence questions, to demonstrate that (1) the value of the non-federal asset it expects to receive is commensurate with the loan, and (2) that its decisions are being made independently, on an arms-length basis in accordance with its normal procedures. Honestly dealing and freely deciding applicants shouldn’t have a problem with that.

Two Low-Impact Amendments

A focused and commonsense validation process, as opposed to an amendment of eligibility, for non-federal cost-shares is what WIFIA’s FCRA Criteria should have been designed to do, and probably what Congress intended in the 2020 Directive. The current Criteria are far too deep in the rabbit hole to be salvaged, but in any case, amending WIFIA’s statutory eligibility is a more transparent path to definitively resolving what has effectively become an issue of fundamental WIFIA eligibility.

The necessary modifications will take the form of two short, low-impact amendments. The first goes to the heart of the issue and adds a specific definition for “non-Federal borrower” to WIFIA’s Section 3902 list of defined terms. The definition will start by essentially restating FCRA law and principles (i.e., a non-federal borrower using non-federal resources for loan repayment) but goes on to add the special validation provisions required for non-federal cost shares.

To avoid any ambiguity, the last part of the definition will deem a non-Federal borrower successfully complying with the rest of the definition to also be a ‘non-federal borrower’ under FCRA law, a characterization that will dovetail into WIFIA’s existing statutory budgeting procedures without further action. Note that the foregoing compliance steps (A) through (D) are substantive and certainly not deemed. Only this final step has an automatic consequence, as a way to confirm a direct connection to FCRA law, the sole goal of the definition.

Importantly, this definition is not meant to be used throughout WIFIA (where the existing term ‘obligor’ is sufficient) but only in Section 3908 (b)(8) where specific eligibility for non-federal cost shares is located. Eligibility there is already limited to an obligor with a non-federal source of repayment, which might be good enough for original policy intentions but is insufficient for FCRA purposes. The second amendment therefore replaces the qualification with the more expansive requirement that the borrower be a non-Federal borrower per the defined term, as determined by the Program.

That last item is important. Status as a non-Federal borrower shouldn’t be automatic, but subject to programmatic discretion, no matter how objectively clear that status might be in particular cases. This is not unusual. The same is true with respect to credit approval for very high-quality applications that obtain a AAA rating in their agency opinion letters. Broad lender discretion just goes with the territory in project finance. Of course, it also means that OMB will have to sign off on the program’s decision. That’s as it should be, and they doubtless will ask very hard questions. However, OMB’s FCRA grilling now will necessarily be guided and bounded by the requirements of a precisely relevant statutory definition. This won’t be a rabbit hole. There are rules.

What the Amendments Will Do

Operating together, the amendments outline a process to establish eligibility for a specific sub-set of applicants. Most EPA WIFIA and perhaps many CWIFP applicants can ignore it. But if you’re planning to finance a cost share, you’ll need to go through Section 3908 (b)(8), as before. Now, however, just showing that your source of repayment is non-federal isn’t sufficient. You must be a qualified non-Federal borrower to be an eligible obligor of a WIFIA or CWIFP loan. The definition shows what you’ll need to demonstrate for the program to consider approval.

I’m assuming that the typical non-federal cost share here involves a relatively big loan to pay for a substantial (though not majority) share of a major project, though I think the process established by the amendments should be workable for a range of possibilities. In a typical situation, the applicant should have the material necessary for the non-Federal borrower demonstration mostly in hand. The credit agency review for the opinion letters will naturally focus in-depth on the loan’s repayment sources. A thorough benefit-cost analysis of the cost share will almost certainly have been completed, either for private-sector investors or public-sector authorities representing local taxpayers. Those distinctly non-federal folks expect to see a convincing case for the value they’ll receive from the cost share relative to their money expended, which is the same case required under the definition.

It should also be straightforward for the applicant to show that its decisions were made independently on an arms-length basis, especially for a public-sector applicant that has had to obtain council approvals or successful voter referenda to proceed with the deal. If this isn’t straightforward for any reason — well, the applicant will need to expect more questions and possible rejection. An ‘above suspicion’ standard is not an unreasonable position for the program to take under the circumstances — there are likely plenty of honest and straightforward infrastructure cost-share situations out there that need financing. The program should prioritize these and let more marginal or complex situations find financing elsewhere.

What the Amendments Won’t Do

Perhaps as important is what the amendments won’t do. They won’t require an unnecessary and completely FCRA-unrelated determination of whether a federally involved project is a ‘federal project’ or something else — the amendments’ minimalist focus is solely on the non-federal share for which the non-federal applicant is seeking a loan, and specifically with respect only to the FCRA classification of the loan itself. They won’t alter WIFIA’s FCRA procedures or deny OMB the final word in this aspect of the program’s budgeting matters. They won’t establish a FCRA precedent for anything other than non-federal cost share situations in federally involved infrastructure projects, a very specialized area covered by only a few loan programs. And hopefully, enacting them won’t require a zero-sum fight in which someone must be publicly proved wrong, and where only one side can win, but both may lose. The amendments will occupy a small, commonsense-oriented place in the middle ground, which is where this issue belongs. Why make any solution a bigger deal than it has to be?

A Graceful Exit for the Current Criteria?

I may be missing something about the way laws work here, but I think that even if these amendments (or the ones in HR 8127 and HR 2671) are enacted, the current Criteria will still be out there and in force. Recent ‘until expended’ appropriations for WIFIA and CWIFP contain specific conditions about certified compliance with the Criteria — they’ll still be on the books, right? Couldn’t the pro-Criteria side simply say that the new definition and modified cost-share requirements are all very nice, but the applicants will have to jump through the Criteria’s hoops as well? And mention, oh by the way, the non-statutory modification of 3908 (b)(8) in footnote 4 is still applicable?

I’d be fairly certain that an additional step is required to defuse the current Criteria and close the rabbit hole for good.

The simplest way, and perhaps what the proponents of HR 8127 and HR 2671 were thinking of, is for Congress to add more law that essentially deems approval by the program and OMB under the new amendments to be compliance under the Criteria wherever that’s required in existing funding authorizations, and then exclude funding conditions (or anything else) related to the Criteria from new legislation going forward. Simple and definitive — but very public and a bit harsh to the pro-Criteria side, no?

It may make sense to consider face-saving alternatives for this step, if only to preserve political energy for substantive battles about WIFIA eligibility. One approach is simply to note that if FCRA-related parts of WIFIA law are amended, the Criteria should naturally be modified to conform to that new law. A short, judgement-free Congressional directive could instruct the relevant parties work together off-stage to produce criteria that are consistent with the processes outlined in the amendments. Some useful refinements and clarifications to the approval process may even come out of that, though the real point is to give the pro-Criteria side an opportunity to gracefully back down in private. Since the new criteria will be consistent with the new law, there needn’t be any further fuss about the issue. Everybody wins.

But if the pro-Criteria side wants to continue fighting to preserve their rabbit hole version of FCRA reality — well, then it’s back to the public arena, where they will likely lose. This is because while WIFIA and CWIFP eligibility for real public infrastructure matters to real people, the current Criteria aren’t worth defending for any reason since they can easily be replaced by something that’s demonstrably better. I believe the pro-Criteria side knows this, and if they’re offered a diplomatic off-ramp, they’ll take it.

In late July, S.2443 Energy and Water Development and Related Agencies Appropriations Act, 2024, was passed out of committee and put on the Senate calendar. I assume it’ll become part of a must-pass consolidated appropriations bill later this year.

One section of the bill adds $2.2 million in funding to WIFIA’s program account for dam safety projects, in effect boosting CWIFP’s dam funding to $77.2 million. More importantly, the section also expands the scope of where the funding can be used to include a closely related type of water management infrastructure, levees:

WATER INFRASTRUCTURE FINANCE AND INNOVATION PROGRAM ACCOUNT
For the cost of direct loans and for the cost of guaranteed loans, as authorized by the Water Infrastructure Finance and Innovation Act of 2014, $2,200,000, available until expended, for safety projects to maintain, upgrade, and repair dams identified in the National Inventory of Dams…Provided, That amounts made available under this heading in this Act shall also be available for projects to construct, maintain, upgrade, and repair levees and ancillary features with a primary owner type of state, municipal, county, private, or other non-Federal entity [1].

Interestingly, the funding’s scope for levees is somewhat broader that it is for dams, which is limited to ‘safety projects’ and seems to exclude new construction. The language for levees doesn’t have the safety qualifier and includes new construction. I would guess that this expansion probably reflects some more-or-less thought-out practical considerations for the levee sub-sector.

It’s certainly possible that the extra funding and/or the levee language could be modified or deleted as the bill progresses through the full process. But I find that hard to imagine in the context of everything else that’s going on in Congress during these interesting times. They’ll be quite busy this fall. So, I think it fair to assume that this section will be eventually enacted as is, and its potential impact on CWIFP financing volume is worth considering for inclusion into CWIFP follow-up dam finance recommendations.

Estimating Potential CWIFP Levee Borrowers

As you’d expect, I don’t have any more real-world knowledge or experience with levees than I do with dams. Which is to say, basically none. But with that caveat, I think I can develop some very approximate estimates of the numbers of potential borrowers in the levee sub-sector using the same process I followed for CWIFP dam borrowers. And fortunately, the same data sources I used for the dam estimates are available (albeit in less detail) for levees. The first is the USACE’s National Levee Database, which is very closely modelled on FEMA’s National Inventory of Dams. The filters in the NLD are not as extensive as those of the NID, but adequate enough to make some estimates here. The second is from the Association of State Dam Safety Officers, which unsurprisingly covers levees in addition to dams and has a separate organizational section dedicated to the subsector. Their 2018 report, A Summary of Risks and Benefits Associated With
the USACE Levee Portfolio, is the main source I use here.

There are two things to note. First, in contrast to dams, which often have primarily local economic purposes (e.g., hydroelectric, reservoirs, recreation, etc.), levees seem to be mainly about regional flood control and inland navigation, both of which are (I think) primarily federal responsibilities. The USACE is the central federal agency for this task (as levee database sponsorship and the focus of the ASDSO report indicate) — and of course, it’s the agency where CWIFP operates. This circumstance undoubtedly will result in more and likely stricter application of WIFIA’s notorious FCRA Criteria, as will be discussed further below.

Second, this analysis will be abbreviated relative to prior one for dams. Partly, this is due to the immediately available data — to dig out more would require the kind of sectoral expertise I don’t have. I also think even a short analysis for the levee sub-sector is sufficient to add to recommendations for the primary focus, CWIFP dam finance.

Levee Length and Federal Involvement

According to the NLD, there are 6,837 levees in the US, with 25,411 of total mileage and an average age of about 60 years.

Like height is for dams, length seems to be the primary characteristic we need to consider for levees, as it will probably determine project cost. Also, for the FCRA Criteria reason noted above, federal involvement (there appear to various shades) should be baked into the analysis at the outset. The NLD has filters for length in miles and for various categories of ‘Authorization’, including Local, USCE and other federal agency involvement. Using these filters, we can categorize both aspects:

As with dams, most levees appear to be quite small, with a dramatic drop-off in numbers for those over 15 miles. Unsurprisingly, federal involvement is strongly correlated with length. Small levees are mostly fully local, while the majority of larger ones have some federal (primarily USACE) involvement.

Levee Project Cost Per Mile

I don’t have clear data about the percentage of levees with safety and other issues, or what type of project might be necessary to fix the issues. I’ll simply assume that at some point over the next few years all levees will need some type of repair, rehabilitation, restoration, expansion, etc., as you’d expect for essential infrastructure with an average age of 60 years. Obviously, a more in-depth analysis would provide a better picture, but the assumption will do for now.

Data about the cost of various project types is also not easily available (I’m beginning to see that the granular and useful detail summarized in ASDSO dam report for the prior analysis is very much the exception, not the rule). But there is one data point about total project cost of the USACE project portfolio in the ASDSO levee report, an average of $21 billion in aggregate. Using the total mileage associated with that portfolio, 14,400, I get an estimate of about $1.5 million per mile for a (possibly?) typical levee project. That number is obviously rough-and-ready, and doubtless will vary greatly depending on the project, but the portfolio size is a significant proportion of the total (though it likely skews to larger projects, as noted above) so I’m pretty sure that the estimate reflects some reality.

Combining the $1.5 million/mile cost estimate with the length & federal involvement data above, and putting it into the three categories of project size I used the in the dam analysis, yields the following picture:

The total aggregate cost of all 6,837 possible levee projects in the next few years would be about $37 billion, of which $21 billion is in the USACE portfolio, as noted above. Importantly for our purpose here, almost all of the small projects (under $2m) are local, and the great majority of large projects (over $20m) have federal involvement. Medium size projects ($2-20m) are a mix of local and federally involved.

CWIFP Potential Levee Borrowers

As in the prior dam analysis, with an approximate taxonomy of possible levee projects, we can estimate the volume of CWIFP financing that might be relevant to this sub-sector’s needs.

I make the same assumptions as used in the dam analysis with respect to CWIFP and project size, and for the same reasons. Projects under $2m in size are probably not relevant to CWIFP demand, while those over $20m can benefit from direct CWIFP loans, generally in the manner provided by WIFIA loans. Projects in the $2-20m range could benefit from indirect CWIFP loans that involve project combination or lending by intermediary project funds.

For the levee sub-sector, however, I add another type of categorization that I fear may be an important element in whether CWIFP will be able to offer financing for otherwise completely eligible large and medium size projects — federal involvement that results in disqualification under the FCRA Criteria.

I assume that local projects (per the NLD Authorization filter), won’t have an issue with FCRA. But those with some sort of federal agency involvement per the NLD filter will run right into it. Two aspects of the category would seem to intensify the potential for serious FCRA issues. The first is that the overwhelming majority of federal involvement is related to USACE activities. The second is that the filter itself references some type of ‘authorization’. Federal involvement itself would likely trigger the application of the FCRA Criteria, but the USACE and authorization aspects are further and specifically constrained by the Criteria’s infamous Footnote No. 4 [2]. Obviously, the severity of the issue will depend on the facts of a particular project, and I don’t know or completely understand everything that’s going on in the NLD filters, but in the overall context of CWIFP’s FCRA travails, I wouldn’t be too optimistic. At the least, federal involvement per the NLD filter is a flashing red warning light, and hence I’ve included a FCRA category for large and medium size projects.

Conclusions

The same conclusions for the dam analysis generally apply to non-FCRA levees. For these, the Senate language simply adds a closely related infrastructure sub-sector to CWIFP’s funding base and scope of demand. This appears to be relatively important for CWIFP direct loans — $2.7bn of volume across about 90 levee projects would be a significant addition to the $10.7bn of volume and 420 projects estimated for CWIFP direct loans in the dam analysis. For medium size non-FCRA levee projects, however, the potential impact looks much more marginal, since medium size dam projects dominate volume and number of projects in that sector.

But the biggest area of possible demand — CWIFP direct loans to levee projects with a potential FCRA issue, about $14.6bn of volume across 240 projects — is also the most problematic. For many of these projects, a FCRA disqualification might be objectively justified. The ASDSO levee report notes that the USACE portfolio includes 190 ‘Federal’ levees that are operated by the Corps without local involvement. Aggregate possible project cost for this group might total about $6bn, most likely primarily for projects over $20m.

As they are, I assume these possible ‘Federal’ projects will rely on federal sources for debt repayment, which legitimately precludes them from federal financing, and not just because of FCRA treatment either. But what if a local community proposes a cost-sharing arrangement to improve the safety of their section of a Federal levee? Is CWIFP financing for the local share (including fully local repayment sources) of the project eligible? That would pass WIFIA’s Section 3908 (b)(8) cost-share test and be consistent with FCRA law and fundamental principles, as well. But apparently neither of these matter, per FCRA Criteria’s Footnote 4, which overrides them both.

Other groups of possible levee projects with a lesser degree of federal involvement might get through the FCRA Criteria gantlet, depending on the specific facts. But realistically, I think any project with USACE involvement that includes any kind of Congressional authorization (that is, most?) is presumed guilty and can be written off. That’s probably the bulk of the $14.6bn and $5.4bn possible project volume for CWIFP direct and indirect loans for levee projects, respectively, or about two-thirds of overall potential CWIFP financing volume for the sub-sector.

In one sense, this is not surprising. As noted above, by their nature levees are often national-level, federal infrastructure assets and as such, extensive USACE involvement is inevitable. Given the FCRA Criteria’s focus (express and implied) on the USACE and CWIFP loans, two factors effectively overlap to disqualify a lot of otherwise completely eligible projects in the levee sub-sector. Likely this was an unintended consequence — it’s hard to imagine OMB having a specific animus against US levees, per se.

But in another sense, the unintentional nature of the FCRA Criteria’s exclusion of so many levee projects from federal financing demonstrates a surprising amount of carelessness about the real-world impact that OMB’s flawed and arbitrary FCRA rules might have. That’s the real problem. How many fewer local cost-share projects in this sub-sector will get done? How much more cost will federal taxpayers bear for essential levee projects due to this? How many efficiencies and economies of scale between USACE operations and CWIFP financing will be lost? Did OMB even think about these numbers? Of course not. Perhaps they should be made to.

_____________________________________________________________________________________________

Notes

[1] I am assuming that ‘amounts made available under this heading in this Act’ was intended to mean that CWIFP funding for levee projects is limited to $2.2 million, though it can also get used for dams. But perhaps the full $77.2 million was meant also to be available for levees? Does the ‘in this Act‘ refer to the specific $2.2 million amount or to the program account ‘heading‘ where the full $77.2 million will be? It’s a theoretical question until the program gets going and actual demand can be assessed, and perhaps it’ll be clarified in that context. The important thing is that the levee sub-sector got added with any amount of funding, opening the door to applications.

[2] Here is the footnote, with emphasis added. For a future analysis, it would make sense to develop a precise comparison between how OMB intends the word ‘authorized’ to be understood and the exact scope and meaning of ‘Authorization’ in the NLD filter. Unfortunately, I suspect there’s a lot of overlap.

 WIFIA authorizes loans to support local cost-sharing requirements. See 33 U.S.C. 3908(b)(8) (“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.”). 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. Project applicants are encouraged to review all applicable statutory requirements before seeking WIFIA financing.

This is the first post in the series, Dam Finance Recommendations for CWIFP

According to the Association of State Dam Safety Officials, there are about 83,600 non-federal dams [1] in the US. Nearly 70% of them require some sort of work, the aggregate cost of which ASDSO estimates to be about $157 billion. The extent to which funding that huge expenditure might benefit from CWIFP financing essentially defines the current universe of potential Program borrowers.

CWIFP’s $75 million of budget authority for dam projects translates to about $7.5 billion in CWIFP loans or (with the usual 49% limit) about $15 billion in total project cost. Compared to $157 billion, on the surface it would seem that CWIFP should have plenty of potential borrowers relative to its resources. After all, that has been EPA WIFIA’s consistent experience since it began operations in 2017 with roughly the same budget authority for an infrastructure sub-sector with the same scale of potential eligible borrowers and required aggregate project costs.

But with a closer look, things are not so simple. Dam and dam project characteristics vary much more widely than those of WIFIA’s water agencies and their capital expenditures. In particular, a lot of dam projects that are otherwise completely eligible will not meet CWIFP’s various specific and statutory thresholds for size and credit worthiness. To determine — even very approximately — the Program’s pool of realistic potential borrowers requires a more granular classification of US dams and dam projects, in effect a specialized ‘taxonomy’ of the sector.

Even a quite limited taxonomy designed to identify certain types of dam project financing would ideally involve a lot of real-world dam sectoral expertise and experience, neither of which I have. That caveat should be borne in mind for the following discussion. Fortunately, however, there are two excellent data sources that, when used together, appear (at least to me) to provide the basic real-world metrics which I can then connect to financing concepts. The first is an invaluable recent report from ASDSO, The Cost of Rehabilitating Dams in the U.S. This report not only summarizes the overall numbers but outlines the sophisticated methodology behind them in some detail. The second is FEMA’s National Inventory of Dams, a comprehensive online database with very effective search filters.

The Basic Numbers Using ASDSO’s Logic Diagram

I start with the logic underlying ASDSO’s methodology, which their report summarizes in the chart below. Click on any of the graphics here to enlarge.

Applying the first two steps in the logic as filters in the NID database, I get the following numbers for size and age of US non-federal dams.

The results correspond to the general impression that most US dams are old and on the small size. The drop-off in numbers about 50 feet in dam height is pretty dramatic, though.

I use ASDSO’s next two logic streps for dam condition and expected remediation as additional filters in the NID database to estimate what capital work US dams seem to require. For simplicity, I label all potential works as ‘projects’ even though some may not ordinarily be described as such (e.g., repairs).

About 58,000 dams currently need some type of project, ranging from repair to extensive rehabilitation. Here I make an assumption that is outside the scope of the ASDSO report but might be relevant re potential CWIFP borrowers: The dams that require rehabilitation include those that ideally would simply be removed. I’m assuming this because dams slated for removal are probably (though not necessarily) in a poor/unsatisfactory condition, either as the reason for removal or the result of long neglect of an unwanted fixture. I’m guessing that, if correct, it’s likely truer of smaller dams than larger.

Dam Projects by Aggregate Cost

Two more steps beyond ASDSO’s logic diagram are required to connect the data to the financing aspects relevant to CWIFP. The first is to assess the cost of the projects, as that metric is what requires funding and might benefit from financing. The ASDSO report has got us covered there, too. Here are their estimates of project cost for each dam size and project type category:

Simply multiplying the average cost estimates by the numbers in each category shows us how the $157 billion total cost is distributed. My numbers include another assumption about dam removal, that the cost of removal is essentially the same as that of rehabilitation. I think this makes sense since both involve extensive work and although removal won’t need to include the specific reconstructive aspects of rehabilitation, it will likely require removal of a lot of debris, restoration of the site, landscaping of the drained reservoir, etc.

You can see that despite the low number of 50+ foot dams, their aggregate project cost as a proportion of the total is a bit higher. But not by much.

The second step is more subtle but central to the analysis. Dam height is perhaps an indicator of several things that might be relevant to CWIFP financing, especially in connection with the potential urgency of completing projects for safety reasons. But height only roughly correlates to the primary aspects of a project’s financing. For example, CWIFP requirements for minimum $20m in project cost don’t really correlate with dam height because it depends on the project type — a medium dam’s rehabilitation project might be over $20 million whereas a large dam’s repair project won’t. Other financing aspects will have similar overlaps between dam height categories. A reclassification of aggregate project cost is required for our purpose here. I use three project cost size categories that I think are especially relevant to the Program, as further explained below.

With the numbers resorted in the three categories, the story gets clearer.

• Small Project – under $2m: Almost all dam repair and a significant percentage of retrofits will fall below $2m in project cost. There are a lot of projects in this category — almost 28,000 — with an aggregate cost of $33bn. But the weighted average project cost (WAPC) in this category is only about $1.3m.

• Medium Project – $2m to $20m: In contrast to small projects, almost all of the major dam work — rehabs and removal — will fall into the medium $2m to $20m range, along with the big majority of retrofits. There are even more projects in this category, about 30,000, with an aggregate cost of $114bn and WAPC of $4.2m. Clearly, this segment is the core of US dam funding requirements and therefore potential financing benefits.

• Large Project – over $20m: Finally, large projects of over $20m in cost are only a small percentage of the total, with an aggregate cost of about $10bn and WAPC of nearly $27m. Most importantly, there are only about 400 projects in this category.

CWIFP Potential Borrowers

With a basic — and hopefully approximately accurate — taxonomy in place, we can draw some conclusions about CWIFP’s potential dam project borrowers. Two conclusions for small and large projects are relatively straightforward:

• Small projects are not likely to be a significant source of CWIFP demand: CWIFP’s project size threshold is an obvious limitation, but it’s perhaps not the most fundamental. Infrastructure capital projects in the $1-2m range are usually funded out of cash flow or other internal resources by substantial entities. They have the credit quality to borrow, but discrete external financing for small amounts is usually not worth the transaction costs. Entities that can’t fund such a small project out of internal resources are unlikely to be creditworthy enough to access cost-effective market financing, much less meet CWIFP’s investment-grade threshold. And these won’t want to (or simply can’t) consider anything too complicated. They’ll be looking for grants, soft loans from philanthropic or state funds, etc. Yes, in theory, some sort of scale could be achieved with a portfolio of small loans to this category, but any leverage involved would probably need to be very basic, which excludes a relatively sophisticated CWIFP loan and its subtle benefits. I could be wrong, but I think the Program can (at least at the beginning) basically ignore this project size category.

• Large projects are likely to be financed and to benefit from direct CWIFP loans: Capital projects above $20m are likely to be undertaken by substantial, creditworthy entities which can arrange cost-effective financing to fund the cost. Most, if not essentially all, of their large projects will pass the Program’s eligibility tests. For these potential borrowers, a direct CWIFP loan is an obvious option that can deliver a range of benefits not found in their private-sector alternatives. In many ways, they’re similar to WIFIA’s water agency borrowers, and the success of the EPA program should help encourage large dam project borrowers to take a close look at CWIFP. I’d predict that there’ll be significant demand from this group. However, in contrast to the potential scale of demand from WIFIA’s pool of qualified potential borrowers sustaining that program’s consistent success, there just aren’t many potential large dam project borrowers and they don’t need that much financing, relative to CWIFP’s $7.5bn loan capacity. Direct loans to large dam projects should obviously be an important part of CWIFP’s development and policy outcomes, especially in the Program’s early phases. A lot of WIFIA experience will neatly translate to large dam loan origination and execution at CWIFP, jump-starting operations and the first closings. But then what?

The third conclusion concerns medium-sized dam projects, by far the largest segment of potential demand for funding and financing in the sector, and it is less straightforward:

• Medium-sized projects might be an indirect source of CWIFP borrower demand through project bundling and loans to dam funds: Obviously, a single project with a cost of less than $20m won’t meet CWIFP’s minimum size threshold [2]. That precludes a direct CWIFP loan for the project, but there are two paths for medium-sized dam projects to indirectly obtain CWIFP financing. The first is ‘project bundling’ or a combination of medium-sized projects with aggregate cost exceeding $20m that submits a single application. This is not a practical path for small projects, but it may be for medium-sized ones, e.g., a combination of just two $10m projects would get there. And at the medium size, CWIFP’s creditworthiness threshold for the combination is more likely to be achieved. The second path is through CWIFP direct lending to public (e.g., SRFs) or private funds that then on-lend to smaller projects. However, as discussed in this post, CWIFP Loans to Small Dam Funds (where ‘small’ there is ‘medium-sized’ in this taxonomy), neither path is simple in practice for either borrowers or the Program. I think CWIFP will need to push the interpretation of current law (as WIFIA has done in several cases for other loan features) to make the project combination path feasible. Lending to public funds will rely on their specific eligibility rules for dam projects [3] and the situation is a bit ambiguous for private dam funds, were they even to exist in scale. And I’m sure I’m missing a lot of real-world devils in the details. Indirectly financing medium-sized dam projects will almost certainly require sustained effort, outreach and innovation by CWIFP. But that would seem to be justified in terms of the Program’s objectives. This segment of the dam sector is not only the largest by far in terms of aggregate cost of necessary work, medium-sized projects are where CWIFP financing could have the most impact and demonstrable additionality [4]. Even if a lot of work is required to get there, the policy payoff could be transformational, not only for CWIFP but federal infrastructure lending in general.

_____________________________________________________________________________________________

Notes

[1] I’m not sure how ASDSO is defining ‘non-federal’ here compared to how it will be applied in WIFIA’s FCRA Criteria. Absent any amendment or clarification of these fundamentally flawed Criteria, I’d guess that a number of ASDSO’s non-federal dams will be considered ‘federal’ by OMB (especially larger ones with some type of federal ownership or management history) thereby precluding from them from CWIFP financing.

[2] The WIFIA statute makes an exception to the $20m threshold that I assume is also available to CWIFP. Project cost can be as low as $5m for small communities when the project is eligible under CWSRF and DWSRF rules. In theory, this might be applicable for medium-size dam projects that are involved in some way with clean or drinking water management. Perhaps storm run-off management or connected to a reservoir? But I don’t know how frequently this could work in practice — I doubt it’s significant.

[3] WIFIA’s ability to lend to SRFs under SWIFIA is specific to their purpose under the CWA. Re the note above, maybe that’ll occasionally include some dam projects. But if an SRF were to expand their eligibility beyond CWA rules to include more dam projects that were also eligible under CWIFP (except for size, perhaps credit rating), I’m not sure that the Program is authorized to make a SWIFIA-type loan for that portfolio. Another detail to clarify.

[4] Demonstrable additionality — clearly enabling a project that would not have happened otherwise — is not easy for a loan program with institutional investment-grade standards. WIFIA’s track record in this is not great, to put it mildly. CWIFP will probably face the same issue in its loans to large dam projects. But medium-sized projects are a very different world, one in which a new type of policy-oriented finance might unlock a lot of potential.