In the prior post, I described CIFIA’s Limited Buydown provision in some detail.  That post covered why I think the provision will be important to carbon pipeline developers, and how it might be at risk if the program runs short of discretionary appropriations due to rising rates after accepting a large volume of loan applications.  This post is shorter and more cheerful.  It considers how CIFIA could reset the buydown if rates fall, based on recent precedents from the WIFIA loan program.

WIFIA Loan Re-Executions to Reset the Rate Lock

I wrote about WIFIA’s rate lock resets for Water Finance & Management in 2020, Resetting the Mission for WIFIA.  The first part of the article describes interest rate resets for two large and highly rated public water agencies.  They’d executed WIFIA loan commitments in 2018 when the 20Y UST was around 3.10%.  With the 2020 re-executions, their rate was lowered to the then-current 20Y UST, about 1.00%.

Why did WIFIA agree to this?  They didn’t have to.  There’s nothing in WIFIA’s statute or federal loan program guidance that would require the program to re-execute prior loan commitments just because they’ve become relatively less attractive to the borrower.

But if WIFIA hadn’t done the resets, the agencies would simply have cancelled their WIFIA loans and issued muni bonds at about 1.00% instead.  The second part of the article discusses some of the policy implications of the fact that WIFIA’s highly rated public-sector borrowers have excellent tax-exempt alternatives.  It’s likely that the resets didn’t affect the projects at all but were useful to the borrowers’ fiscal situation, which requires some explanation for an infrastructure loan program.  Of course, WIFIA was also trying to preserve its scorecard of completed loan volume.  The optics of loan cancellations are not good, regardless of reason.

WIFIA’s specific policy ambiguities aren’t relevant to CIFIA, as discussed in the prior post.  Instead, the key point for CIFIA here is that WIFIA chose to offer the resets.  More importantly, the program found a way to grant them without requiring the water agencies to unilaterally cancel their executed commitments and go through the application process all over again to obtain lower rates, a risky and time-consuming option.

Here’s a non-confidential internal technical presentation I did for WIFIA in 2019 while I was a consultant for the EPA: Rate Resets for Out-of-the-Money Loan Commitments.  I noted some of the legal and budgetary issues that might arise with the resets, but at the time I thought they’d be easily overcome, as proved to be the case.  Most of the presentation focused on options to manage resets to reduce interest rate re-estimate risk.  But as I expected, WIFIA just went with unconditional resets, which reflected the program’s willingness to give the borrowers exactly what they wanted.

In the presentation, I predicted that seven WIFIA loan commitments, totaling about $2.5 billion, would be subject to reset requests.  They were all granted, judging from statements in WIFIA’s 2020 Annual Report.  Perhaps there were more in subsequent years.  I doubt it, due to rates rising off the 2020 trough and some other factors about smaller WIFIA borrowers.  But it doesn’t matter – the precedent for unconditional resetting of undrawn construction rate locks at WIFIA is now firmly established.  Borrowers executing WIFIA loan commitments in the current rate environment can be confident that if the 20Y UST falls significantly prior to loan funding, they’ll have another bite at the apple.

Implications for CIFIA

WIFIA precedents in this area should be applicable to the CIFIA program because the two programs share many of the relevant provisions.  This is most clearly the case for post-execution rate locks since the reset of an executed but undrawn loan commitment to a carbon pipeline project would be exactly analogous to a WIFIA reset.  But I’m not sure that this post-execution option will always be as important for CIFIA’s projects as it is for WIFIA’s larger projects.

A typical WIFIA project will have a long construction period, at least five years.  WIFIA’s large and highly rated public-sector borrowers have multiple established sources of very cost-effective short-term financing which can be accessed for construction draws.  This means that they can keep the WIFIA loan commitment as a completely undrawn option until permanent financing.  That works well when the Treasury yield curve is normally sloped, which it was for 2018-2020 period preceding the first resets.  Drawn loans can’t be reset at WIFIA, but undrawn commitments can, and many early WIFIA borrowers had large undrawn commitments that were very out-of-the-money by 2020.  This created the demand for the first WIFIA resets.

In contrast, carbon pipeline construction is apparently quick once rights-of-way are finalized, perhaps two or three years.  More importantly, CIFIA loans to this sector are closer to classic non-recourse project financings, which don’t rely on large investment-grade balance sheets or other sources of debt.  A CIFIA loan commitment will probably start being drawn for construction draws as soon as it’s executed.  Rates would need to fall significantly and rapidly during a two-year construction period for a reset to be compelling on whatever remains undrawn in a CIFIA loan commitment.  That scenario might arise for some projects in certain circumstances, but the post-execution rate lock reset probably won’t be as generally important as it is at WIFIA.  Still, whenever it becomes valuable for a specific CIFIA project, I think WIFIA’s post-execution reset precedent can apply directly, and more general demand from other borrowers shouldn’t be necessary.

Resetting the Limited Buydown

A more interesting, and potentially important, question is whether the reset principle can be applied to the limited buydown.  This would useful if rates fall after the original application date but are expected to be at risk of rising again before loan execution, a recent pattern of volatility that will likely continue in these interesting times.  A CIFIA applicant will want to lock in lower rates with the limited buydown by resetting the application date.

Here the analogy to WIFIA’s post-execution rate lock reset is necessarily indirect.  But the basic concepts are the same.  A limited buydown is always undrawn because, of course, it’s not a loan commitment.  In fact, given the permissive language in the provision, CIFIA’s limited buydown is not really a commitment for anything.  It simply limits what the program is allowed to do if a loan’s relevant interest rate has risen since the date on which its application was accepted.  Nothing stops an applicant from withdrawing an application and re-applying.  If accepted, the new application in effect resets the date and thereby the relevant rate.

These are basically the same conditions that made WIFIA’s reset possible.  The reset doesn’t affect drawn loans, program law is completely silent on the matter, and in any case, the borrower can unilaterally make it happen by withdrawing and going through the process again.  If anything, they’re even clearer in CIFIA’s case – limited buydowns are literally undrawable, they aren’t even a commitment, and re-application is much easier than full re-execution.

As was the case with WIFIA’s rate lock re-executions, limited buydown resets at CIFIA will likely come down to whether the program wants to offer them.  I’d expect they will (or should) be generally willing to do so, based on some of the factors described in the prior post.  Limited buydowns likely help accelerate carbon pipeline construction by facilitating aspects of the developers’ arduous rights-of-way negotiations.  As such, the provision is part of the program’s core mission.  If there’s an opportunity to improve the buydown for a specific project, doing so will presumably enhance the acceleration effect, which is better policy rationale than WIFIA’s resets seemed to have.  It will also encourage the others by demonstrating an unbureaucratic willingness to help borrowers where possible, not just where required.

Yes, withdrawal and re-application can be done unilaterally.  But that’s added transactional friction just when the developers must be fully focused the easement negotiation process.  It’s also not without risk because in a two-step process the applicant loses the original rate cap.  If something goes wrong or is significantly delayed in the re-application, project economics – or at least the certainty thereof – could be seriously impacted, resulting in a later construction start.  CIFIA can eliminate this risk by accepting an identical re-application simultaneously with the withdrawal of the original application, or some similar one-step process wherein the only change is the date.  Why not cut to the chase?

Well, there is one thing that CIFIA’s administrators need to consider – a possible gap in discretionary appropriations caused by resetting a large volume of limited buydown applications.  The risk of rising rates for limited buydowns in the program’s original applications was discussed in the prior post.  By resetting buydowns at a time when rates could, or are even expected to, rise, the program will be in the same situation.

By refusing to facilitate re-applications and discouraging unilateral action by applicants, CIFIA would presumably reduce this risk.  At some point this might be a way to ration subsidy if a shortfall looks imminent.  But at the outset of CIFIA’s operations, when there’s $2.1 billion of appropriations that Congress intended to be used to accelerate pipeline construction, an overly cautious approach will be both unnecessary and counterproductive.  I think CIFIA will come to the same conclusion and the program should be open to considering limited buydown resets for at least the next few years.

What’s At Stake?  Looking at Some Numbers

To put limited buydown resets in context, it’s worth looking at some numbers.  They’re significant.

The first chart shows what happens if a project resets a lower rate on loan execution than was expected in the original application.  In this illustration, the original application for a $1 billion CIFIA loan for a $1.25 billion project was accepted when the relevant rate was 3.75%, about the current 20Y UST.  Project economics at this rate were acceptable, with an expected 7.5% IRR on $250 million of project equity and an adequate debt service coverage ratio.

But then the 20Y UST falls to (say) 3.00%.  A loan executed at that lower rate would result in a project IRR of 10.5% if the leverage ratio was held constant.  Or the project developers could hold equity return constant and increase leverage (perhaps even from CIFIA itself) by almost $150 million.  The former would provide motivation, and the latter more negotiating headroom, to get the easement process done as quickly as possible.  But that process will still take time, and rates could rise again before loan execution.  The value of locking in the new rate with a reset application will be obvious to the project’s developers and (since it will help accelerate construction start) to CIFIA as well.

The second chart looks at a possible pattern of the 20Y UST over two years to illustrate various reset scenarios for the project described above.  The pattern is based on actual 20Y UST rates from May 2019 to May 2021, adjusted upward to about the current level, 3.75%.

• Point A:  This is the original application at 3.75%.  Project economics are good enough, but won’t tolerate a higher rate, so the limited buydown collar is an important aspect of the CIFIA application.

• Point B:  Project developers see that the 20Y UST has hit 3.00% after four months but seems headed to rise again.  They ask for a reset to lock in the numbers described above, a 10.5% IRR or about $150 million of additional debt capacity.

• Point C:  Alternatively, the developers believe rates have further to fall and wait about a year until the 20Y UST hits 2.20%, a possibly unsustainable low.  A reset at this point results in a project IRR of 12.5% or almost $300 million of additional debt capacity.

• Point D:   The developers are more cautious but can see that rates have bottomed out and will likely go steadily higher.  Still, a reset at this point results in an IRR of 11.5% or additional debt capacity of more than $200 million.

• Point E:  No reset. The execution rate of 3.40% is below 3.75%, with a resultant IRR of about 9.0% or nearly $100 million.  Still a good outcome – but not as good as it could have been with a reset at almost any point in the pattern. All three possible resets were well within the limited buydown’s 1.50% cap, which would only have started to matter for 20Y UST rates less than 1.90%.

Note that Point B would probably have a greater impact on accelerating pipeline development than Points C or D. Even though the later numbers are better, the first reset occurs early in the process. That might be something for CIFIA to consider in terms of policy objectives, but I’m sure that developers will be more persuaded by economics. In any case, the illustration shows that both project economics and CIFIA policy outcomes are improved by resets, possibly significantly.

The case for limited buydown resets looks compelling, but it should not assumed to be an automatic aspect of CIFIA’s operations that can be requested at the last minute. It’s a relatively technical concept and WIFIA’s precedents are only indirectly applicable. There are always devils in the bureaucratic details, too. The possibility of resets should be brought to CIFIA administrators’ attention as soon as applications start being accepted, whether or not they look likely to be used at that point. As noted above, these are interesting times for interest rates, and much else besides. Certainty matters.

This post is the first of three on the topic. The second is here, and the third here.

The ‘Limited Buydown’ provision in TIFIA and CIFIA loan program legislation allows a loan’s construction period interest rate lock to be set at application, not loan execution, as in the WIFIA program.  That’ll be an especially useful feature for carbon pipelines, but it might use a lot of discretionary appropriations.  This long post sets out the context and risks associated with that.

WIFIA’s Off-Budget Post-Execution Rate Lock

A WIFIA loan’s construction period interest rate lock is one of its most valuable features.  In effect, it works as an interest rate option, both on construction draws and (more importantly) on the loan’s post-construction permanent financing phase.  The option is costless to the borrower, but potentially very expensive for federal taxpayers if rates rise during the lock period.  However, under FCRA budgeting methodology, this cost is not WIFIA’s problem.  The water program’s lock begins at loan execution when the loan’s interest rate is set at then-current Treasury rates.  Hence, the required discretionary appropriation (the ‘credit subsidy’) for the loan’s cost is only about its expected credit loss, a small amount for WIFIA’s highly rated borrowers.  If Treasury rates rise between loan execution and final drawdown, the discounted present value of Treasury’s funding loss shows up as a positive adjustment to the credit subsidy amount (an ‘interest rate re-estimate’) that is automatically authorized (under FCRA’s ‘permanent indefinite authority’) and receives a mandatory appropriation covering Treasury’s loss.  The accumulated balance of these mandatory appropriations is recorded in an off-budget account, and WIFIA’s discretionary appropriations are not affected.

The theory behind this special budget treatment appears to be that post-execution interest rate changes are an exogenous and uncontrollable factor for federal loan programs and will likely balance out over time, as both positive and negative interest rate re-estimates should occur.  I think the reality is much more complicated in WIFIA’s case, but that’s another topic.

TIFIA’s and CIFIA’s (Likely) On-Budget Limited Buydown

TIFIA and CIFIA loans offer the same construction period rate lock feature with the same FCRA budget treatment after execution – with one crucial difference:  For these programs, the loan’s execution interest rate can be set at what it would have been if the loan had been executed on its application date, if that’s lower than the current Treasury rate, up to a maximum 1.50% reduction.

This pre-execution rate lock (or more precisely, ‘collar’) is called a ‘Limited Buydown’ in the legislation for both programs.  The language is not very detailed in TIFIA’s case but has apparently been interpreted to include the application date trigger.  In CIFIA’s law, the application date (or a somewhat equivalent execution of a ‘master credit agreement’) is explicitly defined as the relevant start point.

The name of the provision, introduced in the TIFIA section of the 2012 MAP-21 act, is a bit odd – the ‘limited’ part is clear enough (the 1.50% limit), but the ‘buydown’ seems to derive from a home mortgage feature whereby the borrower can lower their rate with a payment.  In the loan program’s case, the lender is doing the ‘buying’.  The provision is probably more accurately described as a ‘limited loan interest rate reduction’ or more transparently as a ‘limited sub-Treasury rate adjustment’.  Perhaps a Congressional TIFIA drafter in 2012 had worked on FHA legislation and thought that ‘buydown’ was a quick way to convey the concept, since the net effect (a lower loan rate) is the same?  Or was the name intentionally a bit obscure because the provision is in essence a kind of grant in a loan program that is otherwise generally funded at Treasury cost?  In any case, for CIFIA legislation the name was almost certainly simply inherited, even if the concept clearly got more focus.

Although the pre-execution rate collar is limited to 1.50% and has relatively more optionality than the post-execution rate lock, both have substantially the same type of benefit to the borrower and cost to the federal lender.  A WIFIA loan’s rate lock over a five-year construction period when rates are rising will cost federal taxpayers about the same as a two-year CIFIA or TIFIA rate collar followed by a post-execution rate lock on a three-year construction period.  In both cases, the federal government has committed to the interest rate on a 35-year loan five years before Treasury funds it.  If rates have risen, the borrower receives the same benefit, and the taxpayer gets stuck with the same economic cost.

But there is almost certainly a huge difference in federal budgeting between the two.  As noted above, the cost of the post-execution rate lock is covered by FCRA’s indefinite budget authority and gets buried in off-budget accounts.  The cost of the pre-execution rate collar, however, is not likely to receive that free pass.  Instead, I expect it will need to be apportioned from the program’s available discretionary appropriations when the loan is executed.  FCRA language and established methodology is straightforward about estimating and budgeting for the cost of a loan (including the cost of a sub-Treasury interest rate) when it is executed.  I can’t see any reason why the pre-execution start point of the limited buydown provision would make any difference to that.

Does this matter?  Not to WIFIA and Probably Not to TIFIA

The additional on-budget cost of the pre-execution rate collar is not necessarily important per se.  The loan is going to require some credit subsidy anyway and the cost of the rate collar just adds to that amount.  However, when the aggregate additional amount of subsidy required by the full application of the limited buydown feature is high relative to the available appropriations because rates have risen, the feature could impede loan executions at the reduced rates the borrowers expected.

Obviously, WIFIA won’t have this budgeting issue because the pre-execution rate collar is not included in the program’s statute.  But it’s interesting to consider what might have happened if it did.  The sophisticated financial staffers at WIFIA’s highly rated water agency borrowers would have immediately seen the collar’s value and timed their applications accordingly, aiming for low points in UST rates.  WIFIA hasn’t had any problem attracting applicants without the rate collar.  With this feature, the volume of applications certainly would have increased or at least been accelerated.  The program’s loan volume is very high relative to its appropriations (more than a 50:1 ratio), a sustainable path because the high quality of the program’s borrowers requires only a small amount subsidy for expected credit losses. But there’s not much headroom.   Even a slight increase in the percentage cost of the loans would impact WIFIA’s ability to execute loan volume as planned.  The cost of pre-execution rate collars during the 2020-2022 timeframe, if they had been offered, would probably have been many times WIFIA’s available appropriations.

A loan program’s likely response in this hypothetical situation would presumably be to restrict rate adjustments under the limited buydown provision to whatever budget authority was available (if any) after an adequate amount was allocated for planned loan executions at unadjusted rates.  Such rationing is possible because the buydown language in both TIFIA and CIFIA laws is explicitly permissive, not prescriptive, for program administrators.  How such subsidy rationing would be implemented practically in this situation, especially over a multi-year period, isn’t at all clear, however.  There’d certainly be disappointed borrowers.    

That still may not have mattered in WIFIA’s case.  An un-disappointed borrower is not a policy objective for federal infrastructure loan programs – a completed infrastructure project is.  The important question about possible cutbacks in limited buydowns should be simply whether the borrower’s project can go forward or not.  The vast majority (or even all) of WIFIA’s water infrastructure projects to date would have gone forward with or without a WIFIA loan in the first place (a fundamental policy topic that should get more attention).  In that low-impact outcome context, the additional benefit of a pre-execution rate collar, or the failure to deliver that expected benefit, would have made no significant difference to US water infrastructure renewal, regardless of WIFIA borrowers’ feelings.   

The same net result is the likely outcome at TIFIA, though for different reasons.  I believe TIFIA’s limited buydown provision has been utilized at least a few times.  But I doubt that it created any serious budgeting issues because TIFIA’s loan execution volume is quite slow relative to its appropriations and much of the program’s activity took place when rates were generally falling.  Even if the program’s budget headroom for limited buydowns becomes more restricted in the future, a cutback in the rate adjustments isn’t likely to stall the borrowers’ projects.  Most of the cost benefit of a TIFIA loan comes from the avoided spread between the program’s UST rate and a BBB-ish private-sector project finance loan, which will range from 0.50% to as much as 2.00%.  That’s valuable and might be an important factor enabling a transportation project, which is the right policy outcome.  A limited buydown, in contrast, is closer to a ‘nice to have’ as opposed to ‘need to have’ with respect to the viability and timing of these types of projects.  This is especially true due to the relatively small proportion of a TIFIA loan in a project’s capital stack – the statutory limit is 49% of project cost, but program policy has further limited that to about 33%.

If TIFIA’s limited buydown becomes restricted or even zeroed-out in a budget year, that might make a difference to project economics and perhaps even to some minor aspects of final project design.  But I doubt that the effect will be enough to stop the project altogether, borrower disappointment notwithstanding.  Hence, for TIFIA, the limited buydown is likely not a critical feature to the program’s policy outcomes – it may be utilized if the budget headroom is there, but probably not a showstopper if there isn’t.

It Might Matter a Lot for CIFIA

In contrast to WIFIA and TIFIA, a possible rationing of CIFIA’s limited buydown provision will probably matter, because (1) the provision is likely important for CIFIA projects, and (2) although the new program has a high level of credit subsidy appropriations, that might get used up surprisingly quickly by even a slight rise in rates.

Apparently, though not explicitly, CIFIA is really all about midwestern carbon pipelines.  There are three big ones in development that get a lot of press, with a rough total cost of about $10 billion.  The point of these pipelines is to monetize 45Q carbon tax credits, something which is now easier and much more lucrative with the passage of the Inflation Reduction Act.  I’m guessing that other pipelines are or will soon be planned.  A lot of financing for a relatively new sector will need to be found, and project sponsors will want to move quickly.

Although the basic form of CIFIA follows its similarly named predecessors, the carbon program appears to have been designed within the framework of federal credit rules to facilitate financing for exactly this pipeline development situation.  Unlike WIFIA and TIFIA, CIFIA can offer financing for 80% of project cost, regardless of size.  And, although the projects must (of course) be creditworthy, an investment grade rating is not required.  Assuming the project developer is putting in the 20% balance as straight-up equity, CIFIA is in effect a ‘one-stop, one-step’ financing program – put in a good application, check the well-established federal crosscutter boxes, and you get approval for the only loan you need.

That’s just as well, because getting the pipeline built appears to be anything but a one-stop process.  Actual construction is likely straightforward and won’t take too long once started – about two years, even for the big ones.  But securing all the easements and rights-of-way across thousands of miles of privately owned, agriculturally productive, and savvily managed property looks like a nightmare.  Moreover, it’ll be intrinsically time-consuming.  Two or three years is publicly mentioned, but I’m sure outcomes will vary widely.  Progress mainly depends on multiple price negotiations, then on bureaucratic processes for the last steps if eminent domain is required, all with a steep trade-off between time and money.  In that process, project economics could swing wildly from expectation.

It’s easy to see why a pre-execution rate collar would be critical in this phase of carbon pipeline development.  Project economics will also be highly sensitive to the interest rate on the permanent financing.  Capping that volatile factor at the start of the process of securing rights-of-way would stabilize the numbers for pipeline cost and define clearer limits for the process’ time/money trade-offs.  Waiting until loan execution to set the permanent financing rate, when all the contracts are basically done and trade-offs set in stone, could result in some nasty surprises.  It’s no surprise that when the CIFIA legislation was being drafted, someone took particular care to define the limited buydown provision in detail – it’s intended to be used.

Mind the Discretionary Appropriation Gap If Rates Rise

Since the current interest rate environment is uncertain and expected to get worse, I’d guess that the three big pipelines (which are now in the process of rights-of-way negotiations) will put in CIFIA applications as soon as the program’s doors open. There’ll likely be some other, smaller projects applying, too.  It’s therefore possible that CIFIA will have roughly $10 billion of financing subject to pre-execution rate collars by year end, from the projects currently underway.  That’s a big number on which to set rate collars for 35-year loans fixed-rate loans.  But CIFIA has $2.1 billion in appropriations over the next five years, so the FCRA impact on the program’s budget should be easily absorbed if rates rise in the next few years, right?  Right?

I’m not so sure.  Obviously, there isn’t much specific data to work with, but if a few assumptions are basically correct, avoiding subsidy rationing at CIFIA might be a near-run thing if rates rise – even a little — over the next few years.

• Let’s assume that $10 billion of applications accepted this year result in $10 billion of executed loans towards the end of 2024, when the project developers to get their rights-of-way and contract negotiations finalized.  At that point, the program should have $1.5 billion of appropriations available, the accumulated balance of scheduled authorization (the total, $2.1 billion, becomes available over a five-year schedule, accumulating until expended).  That’s the green line in the chart below.

• Also assume that the $10 billion of loans will require 8% or $800 million to cover the subsidy for projected credit losses.  This is higher than TIFIA (about 6% for minimally investment-grade loans) and much higher than WIFIA (less than 1% for that program’s average Aa3/AA- loans).  But 8% seems about right for unrated and likely sub-investment grade project finance loans to a relatively new sector, or in any case reflective of OMB’s perception of the risk thereof.  This part of the subsidy is not related to interest rates – hence ‘Non-Interest Subsidy’ in the chart.

• The final assumption is that the applications were all set with a relevant US Treasury rate (the 20Y UST, approximately reflecting weighted average loan life) of 2.75%.  I know that’s lower than the current yield, but it’s chosen to illustrate a point about change, not levels.  And because – who knows?  A few months in the current volatile economic and political environment is a long time.  It could be the actual 20Y UST rate in the near future.  Or rates might rise dramatically instead, and the 20Y UST goes from current 3.75% to 5.25%, with roughly the same effect on the illustrative model.  One thing looks for sure – CIFIA will not be setting limited buydown rates in a calm and predictable world. 

• If the 2.75% 20Y UST rate at which the applications were assumed to be set is higher two years later at loan execution, the lowered discounted present value of the loan will require additional subsidy in an equally higher amount, per FCRA methodology.  That’s the ‘Subsidy Due to Rate Change’.

The chart summarizes what happens if the 20Y UST rises between the application and execution dates.  If no change from 2.75%, then the only subsidy required is the $800 million for projected credit losses, well within CIFIA’s available appropriations.  A rise to 3%?  A little tighter, but still some headroom.  To 3.25%, an unremarkable rise of 50 bps. over two years?  Things get interesting.

If the relevant rate rises above 3.25% in this illustration case, CIFIA will probably need to start the kind of messy subsidy rationing mentioned above.  The simplest step would be to delay loan execution a few months until the FY 2025 appropriation of $300 million becomes available.  But what if rates keep rising during and after the delay?  At 3.5%, the $1.8 billion is used up, and at 3.75%, the total subsidy required exceeds the $2.1 billion initially authorized.  Note that this is all for the original 2022 applications.  Presumably, there’ll be others in the meantime, soon requiring at least the 8% subsidy at execution for the non-interest part, for which they’ll presumably have a priority claim.

In this hypothetical scenario, the limited buydown would cause CIFIA to run out of appropriations for loans to just a handful of pipelines if rates rise by about 1% over two or three years – again, well within recent experience and only two-thirds of the 1.50% limit.

The math is no mystery.  Anyone who has ever dealt with the valuation of long-term, fixed-rate bonds and loans knows that their discounted present value, or price, is very sensitive to interest rate changes.  Was the 1.50% limit in MAP-21 determined by Congressional staffers after modelling various scenarios?  Not likely. The provision’s original proponents might have known the potential benefit of a wide collar, and even suggested the number, but the Congressional staffers probably settled on 1.50% because it sounds benign – properly prudential but not overly restrictive – until you do the math.  In any case, CIFIA got the buydown provision and its limit by inheritance, and (with the inheritors properly appreciating its value for pipeline development, and making sure details were included this time) that’s now the law.

Failing in the Core Mission

If CIFIA’s available credit subsidy needs to be rationed with respect to the limited buydown, there will of course be disappointed borrowers.  As discussed above, that shouldn’t matter much in itself.  But unlike the typical projects financed by WIFIA and TIFIA, for CIFIA’s carbon pipeline projects I think the impact may go well beyond disappointment and affect the program’s policy outcomes.  Paying an unexpectedly higher interest rate on the project’s permanent financing, after all the contract negotiations have been finalized with an expectation of a lower rate, could wreak havoc on a highly leveraged project’s economics.  Whether that’s enough to stop the project altogether obviously depends on a lot of specific factors and isn’t broadly predictable.  My guess is that full showstopper situations will be rare.  Perhaps there’ll be some capital restructuring and even contract re-negotiations – painful, but not fatal.

However, if rationing ever occurs or even comes close, future CIFIA applicants may not put a lot of faith in the limited buydown provision. That may have a marginal impact on the program’s volume of applications and closed loans, the usual (and somewhat specious) soundbite measure of a loan program’s success.  But I think it will also negatively affect a much more fundamental aspect of CIFIA’s potential success – the effectiveness of the program to accelerate the volume of carbon sequestration, compared to what it would have been if CIFIA never offered limited buydowns.

As described above, a pre-execution rate collar helps pipeline developers make better – and presumably faster – decisions about the numerous time/money trade-offs that’ll be required to secure the project’s rights-of-way.  Without it, the project’s economics are less certain, and money perceived to be scarcer.  The trade-off then tilts towards extending the time involved before construction starts – lengthier, hard-ball negotiations and more chance of eminent domain proceedings.  Every delay has a quantifiable impact in terms of tons of carbon that could have been sequestered if the pipeline had been there to carry them.

The loss of 45Q revenue from those un-carried tons is the type of disappointment that goes with the territory of almost all project development when things get delayed.  But from the CIFIA program’s perspective, there is perhaps a unique policy aspect to minimizing delays.   WIFIA’s water projects and TIFIA’s transportation projects will marginally enhance a long-established stock of basic public infrastructure, the need for which is unquestionable.  A few years of delay on specific projects financed by those programs won’t make any difference to the programs’ long-term policy rationale.  In contrast, CIFIA’s infrastructure projects are all about climate change mitigation – a policy area where time is very much of the essence.

In theory, every ton of carbon not emitted into the atmosphere will mitigate climate change, and the sooner the better, starting today.  CIFIA’s policy objectives are doubtless more pragmatic, but time is still a central focus, for at least two reasons.

First, there’s the political context.  The obvious need to renew US water, transportation and other basic public infrastructure will cut through almost any level of federal political polarization or dysfunction – if anything, it might become a rare area of growing bipartisan agreement.  That’s not at all true for any federal initiatives involving climate change, to put it mildly.  I don’t think anyone can predict what future policy in this area will look like, especially for specialized climate change mitigation infrastructure, the need for which is intrinsically less visible than it is for climate resilient investment in basic infrastructure.  Stakeholders in this sector need to make hay while the sun shines, and they’ll expect CIFIA to help produce results with the level of appropriations the program currently has.  Future top-ups to CIFIA’s funding bucket are by no means assured, regardless of carbon pipeline development demand.

Second, time is really at the heart of CIFIA’s entire reason for being.  The basic economics of carbon pipelines look fundamentally creditworthy.  Carbon producers, especially ethanol plants, have an immediate and continuing need for large-scale sequestration.  The sequestration locations are many miles away, but carbon pipeline construction and operations are straightforward. Once the carbon gets there, it’s monetizable at a firm price, $85/ton, through 45Q credits.  That kind of story is eminently financeable in the private-sector debt markets, specifically with project finance bank syndicates, though the process might take some time.  Yes, a CIFIA loan will be cheaper, but the current 45Q price should make the whole operation lucrative enough – what’s the point of an additional federal program to indirectly deliver more of the same incentives?

The answer, I think, is that faster and more certain, not solely cheaper, financing is the primary purpose of CIFIA.  That’s reflected in two of the program’s notable differences to its WIFIA and TIFIA predecessors – the ability to finance 80% of project cost and the lack of an investment-grade requirement.  It’s also reflected in the evident attention that was paid to the language in the limited buydown provision.  Carbon pipeline developers will want to move fast but with as much certainty as possible.  CIFIA’s role is to deliver financing in a way that uniquely meets those two goals.  The program’s lower interest rate is simply the usual side-benefit you get with federal credit, not the main event.

I don’t know whether CIFIA was consciously designed from the start with this purpose in mind.  The program’s industry proponents probably had a clear idea about what they wanted, but began the process (again, with an eye towards speed) with pre-existing federal infrastructure loan program models and added the necessary technical-seeming modifications.  The precise federal policy rationale and objectives for CIFIA was not their problem.  I’d guess that federal policymakers didn’t put much thought into those either, and simply assumed that CIFIA would automatically incorporate the policy aspects of WIFIA and TIFIA, like a new flavor in an established consumer brand.

Regardless of CIFIA’s history or nominal policy objectives, however, the program will be evaluated by its stakeholders with respect to what it appears to be promising carbon pipeline developers in terms of faster and more certain financing.  If interest rates rise over the next few years, the limited buydown provision could be an important aspect of delivering on that promise.  In this context, any inability to execute loans at the interest rate expected by applicants due to a lack of available subsidy appropriations would be a major failure for the program.  It will be the kind of failure that has specific and quantifiable consequences for the affected projects, possibly including a delayed start to construction due to contract re-negotiation, and I’m sure that the borrowers won’t be shy about describing them.

More importantly, such a failure may impact perceptions among existing and future applicants about what overall level of certainty the program provides for the terms of an executed loan, not only with respect to the limited buydown, but other aspects of loan economics that aren’t precisely prescribed and required in the program’s legislation.  If CIFIA fails to fully deliver on the limited buydown rate, what else might the program fail to deliver?  Timely loan execution?  OMB and other approvals?  Acceptance of environmental reports?  Common-sense loan covenants and documentation?  The list goes on and on, all based on a long history of borrowers’ experience of things that go wrong at federal loan programs.

If carbon pipeline developers lose the faith that CIFIA will deliver speed and certainty, the program will have failed in its core mission.  I see the limited buydown provision as a central part of what the program is expected to do.  It’s not the only thing, but the provision is relevant to the first and perhaps most difficult part of the show, securing easements and other rights-of-way for pipeline construction.  CIFIA’s stakeholders, especially including program administrators, need to keep an eye on how rising rates might affect available discretionary appropriations.  However technical and abstract these FCRA mechanics might seem, a budgetary shortfall at CIFIA is a gap that’s well worth minding.  

This new op/ed connects two topics that I frequently comment on here:

• The first part summarizes the big picture context — why federal infrastructure loan programs should expand their loan product capabilities, not just capacity, to prepare for more difficult times ahead.

• The second part describes one very specific example of expanded loan product capability — the extension of WIFIA’s maximum post-completion loan term from 35 to 55 years.

In a recent post about AWWA’s 2022 SOTWI survey, I note that a loan with a subsidized rate should be extended by the borrower for as long as possible. That certainly makes intuitive sense — other things being equal, it obviously makes sense to receive the benefit of a subsidy for as many years as they’ll give it to you. Of course, things are never exactly equal when it comes to subsidized debt — there are always strings attached. But if you’ve already dealt with the upfront strings for your project to qualify, and ongoing compliance is not too burdensome, then the principle of ‘longer is better’ will apply, especially if prepayment at par is generally costless. Those conditions seem to characterize debt from the three sources of subsidized infrastructure financing highlighted in the AWWA survey, municipal bonds, SRFs and WIFIA.

Still, it’s worth looking at some hypothetical numbers to elucidate the nuances. That starts with assumptions about the basic subsidized rates offered by each of three, relative to a market rate. For simplicity, the market rate here is always a 0.75% spread over a long-term US Treasury rate that’s the same for loan terms from 15 to 60 years. That’s also the discount rate for the present value (PV) benefit analysis.

The project is assumed to have a five-year construction period and a very long useful life. After completion, project debt will amortize on a level-payment schedule for the balance of the total term. Loan term is the main variable in the analysis.

From there, equally simplified assumptions for the three financing sources:

• A tax rate of 22.5% for the municipal bonds, for an offered interest rate that gets the investor back to an after-tax market yield. If the market rate is 3.75%, for example, the tax-exempt bond rate will be 2.91%. Total term is limited to the market-based 30 years.

• SRFs vary widely, but for demonstration, we’ll assume here that this source will offer a rate that’s 75% of the market rate. Again, if the market rate is 3.75%, the SRF rate will be 2.81%. Total term is limited by prudential state policy to 25 years.

• WIFIA is straightforward. By statute, the loan rate is ‘not less’ than the US Treasury yield for a loan’s weighted average life. We’ll ignore the one basis point added by the program. A market rate of 3.75% with a 0.75% spread means a 3.00% US Treasury yield, and hence a 3.00% WIFIA rate. Total term is limited by statute to 35 years post-completion, for 40 years in this case. An extended WIFIA case, with 55 years post-completion, is also considered.

When debt from these three sources is discounted at the market rate, there’s always a PV benefit, and that PV benefit increases the longer the subsidized debt can be amortized after construction completion.

With full debt term from 15 to 60 years as the variable, and a long-term US Treasury rate of 3.00% applied throughout, here’s what we get:

As you could guess from their lower rates, SRF or muni bond financing has a better benefit than WIFIA in years 15 through 25 and 30, respectively, when they hit their limits. WIFIA loans are more beneficial starting from a total term of about 35 years, and materially better at the 40-year maximum. Extended WIFIA loans would deliver a much higher PV benefit, almost 14% of project cost with a 60-year term.

These results reflect the fundamental point I was making in the AWWA survey post — compared to SRFs and muni bonds, the WIFIA program can offer much longer loan terms, something that might intrigue water system CFOs for various pragmatic reasons (e.g., lower annual debt service for the first ten years), but is justified by a strict PV analysis, too.

No surprises, so far. But changing the UST interest rate in the analysis uncovers an interesting nuance. If the long-term UST is assumed to be 1.00%, for a market rate of 1.75%, we get a very different picture:

In this case, WIFIA loans are always better, and a by a lot. That’s because the market spread is a relatively large portion of the total rate when UST rates are ultra-low. Both the SRF discount and the bond tax-exemption operates on the market rate, while WIFIA cuts to the chase and eliminates the spread altogether.

But at higher UST rates, this effect completely reverses. With a long-term UST assumption of 5.00%, a WIFIA loan always delivers much less PV benefit, regardless of term, current or extended:

Again, subsidy mechanics explain the result — the SRF 25% discount and muni tax-exemption apply to the whole interest rate, while WIFIA is stuck at the UST rate, which in this case is a much bigger proportion of the market rate.

This is of course a completely abstract analysis. But I think the basic assumptions reflect (very roughly) the reality of most situations involving these sources of subsidized infrastructure financing. The broad direction of how their respective PV benefits change with longer terms and different rates is likely accurate, even if the specific numbers aren’t particularly meaningful. A few observations:

• The analysis shows what you already knew – longer terms mean more PV benefit – but it also literally illustrates something that’s important to much basic water infrastructure renewal. Look at the big empty space in the charts where the SRF and bond lines end! Not all infrastructure projects have a useful life of 60 years by any means, but surely a lot of most essential stuff gets into the 40- or 50-year time frame. Subsidized debt should encourage a long-term perspective in infrastructure renewal. I can see the prudential logic behind state SRFs’ limited terms, especially since these funds often focus on smaller, more equipment intensive projects. But why is the primary federal financial subsidy for large-scale state & local public infrastructure structured as a tax-exemption utilizable only by retail investors, a class that requires high liquidity in 30-year markets? Yes, I know — a rhetorical question. For now.

• The CBO’s latest budget and economic outlook projects 10-year Treasury rates for the period 2023-2032 that average about 3.2%. Long term Treasury rates for the period will be higher on average, but probably around 3.5%. This means that the first chart in this post, where the PV benefit from the three subsidized source differ primarily by term, is likely the most relevant one. In that context, combining the various sources will result in the maximum PV benefit available. WIFIA loans are limited to 49% of project cost in any case, but the 51% balance can be composed of faster-amortizing SRF loans and muni bonds, a structure that WIFIA explicitly permits and generally encourages. From what I’ve seen in the real world over the past few years, this is often exactly what systems are doing. More on WIFIA combinations here and here.

• Most fundamentally, I think the analysis highlights two very different policy approaches to subsidizing US infrastructure finance. Interest rate subsidies from SRF discounts and the muni bond tax-exemption are essentially transfer payments. The discounts are ultimately paid for by federal and state grants, the tax-exemption by less transparent tax expenditures. In contrast, federal loan programs in the WIFIA, TIFIA and CIFIA series offer loans to qualifying borrowers that are expected to cover their cost of Treasury funding (in theory, anyway), and require only a small appropriation for credit losses, given the borrowers’ credit quality. In these programs, taxpayers are primarily incurring the opportunity cost of not lending at market rates. I don’t see this as merely academic distinction. Transfer payments are the blunter tool — more powerful, with a greater impact, good or bad. The opportunity cost approach effectively requires loan programs to differentiate their products, not only from market debt but from other, transfer payment-based financing. That’s particularly challenging for WIFIA, as the analysis shows. But if the differentiation is based on actual federal comparative advantages (e.g., very long-term lending), positive outcomes represent a real improvement in economic efficiency. That is a sustainable approach in several ways, not just fiscally. It also requires innovation at the interface between real-world goals for physical infrastructure and the influence of finance on them. Lot of policy scope there — and ‘innovation’ is a word that’s right in all three programs’ name.

In my last update of the economic cost of WIFIA’s portfolio, The Cost of WIFIA’s FYE 21 Portfolio as of June 30, 2022, I noted that the analysis didn’t include loan commitments added since FYE 2021 or additional drawdowns. Since then, I added the 27 loans closed between FYE 2021 and 5/26/22, for a total portfolio at 6/30/22 of $15.3 billion, roughly in line with WIFIA’s website information. I also assumed that 25% of the portfolio had been drawn, versus the previous assumption of 12.5%, and that realized re-estimate losses on the drawn loans were about $100 million, versus the $25 million reported for FYE 2021.

As expected, the weighted average interest rate on the portfolio loans and loan commitments increased to about 1.8%, versus about 1.5% at FYE 2022 since rates have been rising in the interim.

Based on all that, if all WIFIA loan commitments had been drawn on 6/30/22 at the then-current 20Y UST of 3.4%, the realized funding loss would have been about $2.4 billion, the same estimate as before, as I had thought it would be. Although the portfolio’s weighted average rate is higher than at FYE 2021, portfolio volume is also greater, resulting in about the same estimated loss.

It’s not quite all bad. Note that if interest rates start to fall back to the 2.0% range, potential funding losses on the 6/30/22 portfolio will decrease more quickly than they would have with the FYE 2021 portfolio (the dotted line in the chart) since many of the recent loans will be at about that rate. A 20Y UST of 2.0%, for example would only result in a $300 million loss or 2% of the portfolio — instead of the $600 million or 5% that would have been incurred by funding the FYE 2021 portfolio. That’s much better, especially regarding the optics — but this result assumes WIFIA doesn’t reset the higher rate loan commitments downward. Given the program’s reset precedents, I’d imagine there’ll be some tough discussions if rates fall a lot, and resets will probably happen.

Of course, all of these economic cost analyses are only as good as the assumptions I’m making in the absence of direct data. The size and the closing dates of the loans are well-disclosed by WIFIA and there’s a ton of accurate UST rate data for those closing dates. I’m also pretty sure about the typical loan WAL of 20 years and FCRA methodology. But the amount of loan commitments drawn, and their realized re-estimate losses, are frankly guesstimates. Some information about the FYE 2021 portfolio did surface in White House budget appendices last year, but only indirectly and on a projected basis. Nevertheless, that data was broadly consistent with my assumptions at the time. It’ll be interesting to see what the next budget might disclose, especially regarding my assumption of $100 million of realized re-estimate losses for FY 2022. I’d predict that’s a minimum.