Now that some in Congress have shown that they take technical budget accounting matters at WIFIA extremely seriously, these diligent lawmakers should be very interested in directing the CBO to take a look at what appears to be a glaring error in Program scoring.
The Joint Committee on Taxation makes a major assumption about the projected impact of the WIFIA Program on federal revenues, which in turn makes a significant difference in CBO Cost Estimates. Here’s the relevant language from CBO’s June 28, 2018 scoring for S.2800:
JCT is assuming classic loan program ‘strict additionality’ here. Their expectation is that an infrastructure project that otherwise would not proceed will be able to do so only with a WIFIA loan. And since most water infrastructure is built by public-sector agencies and financed with tax-exempt debt, tax-exempt bonds are expected to be issued for the 51% of project capital cost that’s not covered by the WIFIA loan. Hence, a $49 million WIFIA loan will cause the additional issuance of $51 million of long-term tax-exempt bonds. Since in the absence of WIFIA, the project wouldn’t have proceeded and the bonds would not have been issued, the bonds’ tax revenue impact must be included in Program budget scoring.
JCT’s logic is completely correct. And at the Program’s outset, it was also completely reasonable for them to assume that the Program would develop along these lines, based on the general applicability of additionality objectives behind loan program policy and specifically on actual outcomes at TIFIA, the transportation loan program on which WIFIA was very closely modelled.
Now Let’s Look at Results So Far
Now that WIFIA has three years of actual results, including over 90 qualified and selected applications totaling $15 billion, the facts underlying this assumption should be re-examined. InRecap’s preliminary analyses based on publicly available information of WIFIA’s 2017-2019 results are here, here and here.
It turns out that the vast majority of WIFIA borrowers are in fact highly rated public water agencies with access to the tax-exempt bond market. In specific cases, their plans for financing the proposed projects include internal cash, SRF loans, grants, etc. But usually the 51% non-WIFIA balance is expected to be wholly (or at least partly) provided by tax-exempt bonds. And in contrast to more opportunistic sources like SRF loans or grants, the muni market is predictably the lowest cost alternative for large-scale, long-term financing that these agencies’ have. In the absence of WIFIA, tax-exempt bonds are realistically the most likely alternative. Hence, the second leg of JCT’s assumption looks fundamentally correct.
The vast majority of water infrastructure projects being financed, however, do not appear to be contingent on a WIFIA loan. In many cases they appear to be completely non-optional as necessary parts of water system operation (e.g. replacement of failing pipes) or legally mandated (e.g. in connection with storm water consent decrees). Even for those projects that appear to be somewhat optional, the driving factors appear to be related to major gains in efficiency and other engineering metrics or achieving environmental and climate-related goals – factors that are unlikely to be influenced by a slight difference in financing cost.
Most importantly, highly rated public water agencies can proceed with non-optional projects, or decide on optional projects, with the confident expectation that sufficient cost-effective financing will be available in tax-exempt market. A WIFIA loan may be a useful enhancement to the project’s financing, but it is unlikely to be a critical factor to the project itself. The volume of water infrastructure projects that will proceed and require financing is actually an exogenous variable, not influenced by the availability of WIFIA loans. Strict additionality is simply not applicable.
A New and More Nuanced Assumption is Needed
Using these updated facts, JCT’s logic now leads to the opposite conclusion. If anything, a WIFIA loan is most likely to replace tax-exempt bonds in projects selected by the Program, resulting in fewer muni bonds being issued than otherwise would be the case and an increase in federal tax revenues.
Obviously, a major adjustment of JCT’s estimates and CBO scoring for the WIFIA Program is required. Significant numbers are involved – the estimated $2.6 billion revenue reduction in S.2800 may in fact be a revenue increase of similar magnitude.
But a note of caution is warranted here. Strict additionality is clearly not a tenable assumption based on WIFIA results so far. This does not mean that its polar opposite, strict substitution of tax-exempt bonds, should be automatically adopted as a singular assumption about WIFIA’s impact on federal revenues going forward (as much as WIFIA proponents might like that). WIFIA loan applications contain a lot of data that can be analyzed to provide a more precise picture of what Program borrowers are doing. In addition, the Program is still relatively new — different trends in WIFIA borrower characteristics and financing plans may emerge.
Since the Program is operational, a more nuanced model of its interaction with the tax-exempt bond market, as well as other frequently used sources of capital (e.g. SRFs), could be developed and updated as additional data became available. For CBO scoring purposes, the model’s output doesn’t need to be complicated. Even if it’s a single metric — e.g. “75% of WIFIA loans result in a substitution of tax-exempt bonds, 10% in an addition and 15% with no impact” — that’s better than a single either-or assumption about additionality or substitution.
As noted, big numbers are involved. As with the federal-ownership budget issue, taxpayers deserve a realistic a picture of what’s happening at WIFIA.