Click on picture to enlarge
WIFIA’s $2.1b funding loss on about $10b of loan disbursements to date can seem abstract. In this post, I’ll make it clear that the loss involves the transfer of real resources away from taxpayers to water system ratepayers for no policy outcome whatsoever.
The above diagram shows what Congress was led to believe it was funding. The lower tier of boxes involves real resources. The upper tier is the financial allocation channel for WIFIA loans — that’s where all the complex technical stuff is, which we’ll mostly ignore for this post.
The arrows and amounts show the flows involved in the $10b of WIFIA loan disbursements through FYE25. The green arrows show the investment of cash. The blue arrows show the present value of that investment in terms of its expected cash flows discounted by the then-current market rate of return — that is, the market price.
Immediate Disbursement: What Congress Thought Taxpayers Were Getting
First, we’ll go through the steps as if the $10b of disbursements all occurred on the same days the loan commitments were made. This is necessarily hypothetical, but I think it’s what Congress had in mind in WIFIA’s legislation and funding.
- Savings: $10b in cash can be immediately converted into things in the real world. Alternatively, savers can put the cash in the US Treasury market where it will be ultimately used by someone else to buy real-world things. But the savers want a market return on their investment. The market rate reflects their valuation of real-world things they’ll get in the future in exchange for their $10b in cash today. Of course, that market rate will change, up or down, as collective perceptions of an intrinsically unknowable future change over time. But at any point in time, the market price is the best estimate of the real-world economic value of future cash flows.
- Treasury Market Rate: The 2.50% rate is an approximation of the average rate of the $10b in WIFIA loans closed 2018-2025 and then drawn 2022-2025. But it comes from the prevailing yields for all the Treasury bonds bought and sold at the time that correspond to WIFIA loan debt service payments — that is, the Treasury yield curve. As such, apart from use in setting WIFIA loan rates, it is also the fundamentally correct basis of the relevant discount rates for WIFIA loan cash flows. Note that it doesn’t matter how Treasury actually finances a specific WIFIA loan (e.g., issuing one-year bills or 30-year bonds or whatever) which will just be a tiny part of overall US deficit management. The discount rates applying to the entire repayment schedule of the WIFIA loan are what determines its economic value to taxpayers. That’s what FCRA law and budgeting is based on because that’s the best way to value the resources the federal government is allocating. The calculations are complex, but the real-world meaning is straightforward: When the federal government borrows real resources from savers, the UST discount rate provides the most correct picture of the value of the real resources involved in the transaction at that point in time.
- Federal Government (WIFIA/UST): The $10b in cash borrowed from savers is provided through inter-governmental lending by Treasury to the WIFIA program for loan disbursements. But Treasury expects to be repaid in full, with a 2.50% return, so that in turn it can repay the savers. WIFIA lends the $10b in cash to WIFIA borrowers at 2.50% on average. The borrowers are highly rated, but there’s still a small chance of credit losses (about 1%), so WIFIA values the repayment cash flows at $9.9b present value at 2.50% discount rate. The missing $100m (or $0.1b) comes from taxpayers — that’s the credit subsidy for which Congress provided discretionary appropriations and OMB apportioned as adequate for all expected loan costs.
- WIFIA Borrowers: The borrowers receive $10b in loan commitments at an average fixed rate of 2.50%, which reflects relevant Treasury yields at closings, per WIFIA statute. In this hypothetical case, they immediately draw the loans and spend the $10b on real-world things at eligible water projects. For each closing and immediate disbursement, the loan interest rate and the Treasury discount rate are effectively equal, so there is no funding loss or gain. Only the small, expected credit losses need to be covered by taxpayers.
- Water Investment Project: This is where the $10b in cash is converted into real-world assets. The expenditure will result in payments from the water ratepayers for whom the project is being built. Due to credit losses, these payments are the expected source of 99% of WIFIA loan repayment and ultimately 99% of the repayment of the savers’ $10b original investment. The missing 1%, or $100m in present value terms, comes from taxpayers.
- Taxpayers: Collectively, federal taxpayers will pay $100m in present value terms at a discount rate of 2.50%. It doesn’t matter if they pay $100m up front, or they, their children and grandchildren pay the $100m plus interest over 35 years. One way or another, they’re on the hook for a $100m transfer of their real-world resources to water projects to cover the ratepayers’ expected shortfall. For this obligation, they receive a ‘policy outcome return’ in terms of the improvement of some national public good. This is indeed abstract, but in theory it might be real in some ways and in any case, that’s what they thought they voted for. For now, we’ll assume WIFIA’s ‘narrative’ is valid and the taxpayers’ $100m did ‘unlock’ or at least ‘accelerate’ investment in water infrastructure, which will have the promised national benefit.
In this hypothetical, immediate disbursement case, everything is consistent with everyone’s expectations. Congress appropriated the correct amount of funding for the Program. OMB correctly assessed and apportioned the necessary subsidy to cover all expected costs. WIFIA borrowers spent the $10b on eligible water infrastructure and are expected to repay 99% of the loans on a present value basis. Taxpayers will cover the shortfall, but it’s small amount, for which they’ll receive a small policy return.
Now let’s look at what actually happened.
Delayed Disbursement: What Taxpayers Actually Paid and What They Got
Click on picture to enlarge
$10b of WIFIA loan commitments made 2018-2022 were actually disbursed roughly over 2023-2025. Part of the delay was intrinsic — drawdowns must be spent for construction expenditures, which of course follow a years-long schedule. But some significant part was the fact that most WIFIA borrowers have cheaper short-term financing alternatives, and they simply exercised their delay option. During the delay, UST rates steadily rose, from an average of the relevant yields at about 2.50% to 3.75%.
- Savings: Savers were of course still willing to invest the $10b in Treasuries, but now they required an average rate of 3.75%. Whatever they were willing to accept a few years before doesn’t matter in the slightest — 3.75% is the market rate that reflects the (new) best estimate of the economic value of future cash flows.
- Treasury Market Rate: The 3.75% rate is no longer relevant for the interest rate on the WIFIA loans being disbursed — those rates were fixed at commitment years before. But it very relevant with respect to what Treasury now has to pay savers to borrow $10b and therefore is the relevant discount rate for WIFIA to use in valuing WIFIA loan repayment cash flows.
- Federal Government (WIFIA/UST): The federal government must honor WIFIA loan commitments on the contractual terms. WIFIA had to lend $10b to the WIFIA borrowers at an average rate of 2.50%. The present value of the loans’ repayment cash flows at the now-current 3.75% discount rate was lowered by $2.1b to $7.8b. The shortfall is covered by $2.1b value of additional real resources transferred from taxpayers. The accounting characterization of the transfer (i.e., as mandatory spending) is not relevant for this exercise. In real-world terms, it is simply a funding loss — WIFIA was forced to fund loans with an average interest rate of 2.50% while borrowing funds from Treasury and ultimately from Savings at an average rate of 3.75%.
- WIFIA Borrowers: They got the deal they signed up for. But note, if things had gone the other way — interest rates falling further instead of rising — these highly rated borrowers would not have drawn the loans. That too is valid: the cancellation option was also part of the deal.
- Water Project Investment: As before, the $10b will be spent on water projects. But now the present value of ratepayer payments to cover the expenditure is only $7.8b. In effect, the $2.1b is a federal grant indirectly subsidizing these ratepayers.
- Taxpayers: Federal taxpayers did not get the deal their representatives signed them up for. Assuming, as we did above, that the $100m apportioned for credit losses was acceptable because water infrastructure investment would be ‘unlocked’ or ‘accelerated’ with an associated national benefit, what did taxpayers get for the additional $2.1b to cover funding losses? At disbursement, the water projects were already under construction and their designs finalized — whatever national benefit was expected from WIFIA loan commitments would already have been realized. The additional $2.1b value of taxpayer resources subsidizing the loans would not change that. Instead, as noted above, the value went as zero-sum transfer payments to the projects’ ratepayers, who may or may not have needed it. In any case, such localized subsidies are not part of WIFIA’s stated policy objectives, which are focused on physical water infrastructure.
To put it simply, in terms of real resources, the $10b of WIFIA loan commitments disbursed 2023-2025 cost federal taxpayers an unexpected additional $2.1b that delivered zero policy return. From taxpayers’ perspective, it is simply a loss.
To be crystal clear, it is a realized loss — not an ‘accounting adjustment’, or a ‘book loss’, or a ‘mark-to-market revaluation’ or anything like that. The cash is out the door. WIFIA borrowers got $10b in cash but will repay in PV terms only $7.8b — a great deal, no? But there’s no ‘free lunch’ — taxpayers are in turn irrevocably obligated to use their real resources to pay in $2.1b in additional taxes. These values can be re-evaluated with different discount rates as rates change in the future, up or down, but at the time the cash was transferred, they were the actual cost, both economically and per FCRA law for budgeting purposes. Nothing abstract about that.
(But…but…but, interest rates go down as well as up, so in theory some future $10b of disbursements might provide $2.1 of funding gains for taxpayers? That would be a transfer payment from water ratepayers to taxpayers. So over time, it’ll all even out, right?
Sorry, nope — never happen. As noted above, and written above extensively here, the loan commitments contain cancellation options and WIFIA’s highly rated borrowers will exercise them.)
True, to the extent that WIFIA’s funding losses are transfer payments that make some people better off (e.g., Silicon Valley water ratepayers save a few dollars each month, which they might appreciate if they notice it), they’re not a deadweight loss in economic terms. But from a policy perspective, a transfer without any consideration of need or other policy outcome might have as well been done more efficiently and fairly through a lottery. The ‘policy return’ for federal taxpayers is in effect zero.
As is typical for this kind of loss, bad analysis and bad budgeting — in effect, bad pricing — were the cause of this misallocation of real resources. That must be addressed at WIFIA and OMB. Corrections and solutions are where all the technical financial and budgeting stuff comes in. But the real-world purpose of that process must be kept foremost in mind.