This is the third post in the FCRA Non-Federal Series

The Congressional directive does not mention the 2017 Report, How CBO Determines Whether to Classify an Activity as Governmental When Estimating Its Budgetary Effects, but according to the GAO Report (the subject of the next post in this series), OMB considered their FCRA loan criteria to be consistent with its views.  The CBO Report also appears to be one of two primary sources OMB used, the other being the Budget Commission’s 1967 Report.

The CBO Report is much shorter than the 1967 Report and has a specific focus on the practical issues that CBO encounters when scoring legislation that involves non-federal entities.  It describes how CBO decides whether to include ‘activities’ of non-federal entities in the budget score.  Federal program loans are not considered per se, but arguably they’re covered by CBO’s inclusion of the “cash flows related to those activities”.

Introductory Section

The introductory section of the report reiterates the ‘when in doubt, include it’ principle from the 1967 Report:

To make such determinations, CBO follows guidelines from the 1967 Report of the President’s Commission on Budget Concepts, which includes the following recommendation: “The federal budget should, as a general rule, be comprehensive of the full range of federal activities. Borderline entities and transactions should be included in the budget unless there are exceptionally persuasive reasons for exclusion.”

The report echoes another principle that the Budget Commission established as a primary reason for the separate treatment of federal credit within the budget, correctly measuring economic impact:

Although the federal budget is primarily a tool for tracking the government’s cash flows, it also serves as a measure of the scope of federal activities and their effects on the economy.  Treating the activities of some nonfederal entities as part of the federal budget, even if those transactions would not flow through the Treasury, helps to accomplish that objective.

Interestingly, CBO notes that its scoring and OMB budgetary treatment may differ, something also described in one of the report’s examples:

The Office of Management and Budget in the executive branch is responsible for recording cash flows related to enacted legislation in the federal budget.  Its budgetary treatment of activities may differ from the treatment CBO uses in its cost estimates.

General Observations About the Introductory Section:

• As discussed in the prior post in this series, ‘borderline’ ambiguity does not arise with federal program loans — they are indisputably included in the federal budget. The specific FCRA budget question is whether a program loan to a federally involved project should be given FCRA treatment or included in the general cash-based budget. For CBO scoring, it is easy to see that federal involvement in an infrastructure project can raise a different question – whether, regardless of the nominal form of ownership and the involvement of non-federal participants, cash flows related to the project’s activities beyond those pertaining directly to the federal participant should be considered federal.  The two questions both require an examination of the federal participant’s role in the project and its financing, but they are not necessarily related beyond that.

• To better reflect the economic impact of a federal action, CBO says its scoring will sometimes add the cash flows of non-federal entities.  But in some situations, perhaps subtraction of cash flows in and out of the Treasury will provide better information?  In fact, such subtraction is precisely what the federal budget’s separate FCRA section for federal credit programs was expected by the 1967 Budget Commission to do.  Since a program loan comes with a non-federal obligation to repay, including the large initial funding outflow and extended repayment inflows in a cash-based budget will distort the true picture of the loan’s economic impact.  FCRA basically nets the reversing cash flows on a present value basis and records the much smaller residual amount as the loan’s effective federal outflow or subsidy payment.  Both CBO scoring and FCRA budgeting are aiming to provide better information about federal actions, but it should be kept in mind that they work in an almost opposite way.  In effect, CBO scoring asks, ‘what non-federal cash flows should be included?’ while FCRA budgeting asks, ‘what federal outflows and non-federal inflows should be excluded?’.

• It’s not surprising that CBO scoring and OMB budgeting might differ in some situations.  In private-sector GAAP accounting, a company’s balance sheet will consolidate the non-recourse debt of a majority-owned project.  That arguably gives a truer picture of the scale of the company’s operations, but the company’s recourse liabilities (as disclosed in the financial notes) will be a much more important factor for credit rating agencies’ models.  I think CBO and OMB differences might be analogous – CBO is looking at a bigger picture and tends to consolidate activities, while OMB is focused on the budgetary cost of specific actions in a less consolidated context.

CBO’s Criteria for Identifying Governmental Activities

After the introductory section, the CBO Report states two fundamental criteria for consolidating the activities and related cash flows of a non-federal entity into a score:

1) The activity would require the exercise of the sovereign power of the federal government by or on behalf of a nonfederal entity; or

2) The activity would serve a specific governmental purpose; the entity would be directed, controlled, or owned by the government; or both of those conditions would be met.

The next two sub-sections briefly describe more about these criteria.  The first, the exercise of sovereign power, seems to be summed up here (emphasis added):

If legislation would authorize a non-federal entity to use the sovereign powers of the federal government, CBO considers the cash flows of activities related to that exercise to be federal.

The next, governmental purpose and control, is apparently applicable to a very broad scope of economic activity (emphasis added):

If a nonfederal entity would serve a [federal] governmental purpose or act on behalf of the government to satisfy a federal policy objective or achieve a regulatory outcome, CBO might consider the costs of those activities to be federal costs.

The balance of the report is taken up with descriptions of scoring examples and case studies in the context of these criteria.  None specifically involve major public infrastructure projects with federal participation, though one about power transmission systems and two about federal facilities are potentially indirectly relevant.

General Observations About CBO’s Criteria

• The two criteria seem to be very consistent with the 1967 Report’s fundamental external discipline principle, the foundation of both general federal budget inclusion and FCRA’s exception for a program loan’s reversing cash flows.   If the federal government is the primary cause or controller of an activity at a non-federal entity, then the cash flows related to that activity are not effectively subject to non-federal external discipline and should be included in federal budget metrics like CBO’s scoring.  This would also be the case when the activity in question a federal program loan – if the federal government is the primary source of the loan’s repayment or is compelling non-federal entities to repay it, then external discipline is lacking, FCRA budgetary treatment will not apply, and all the loan’s cash flows should be reflected in the federal budget.  In terms of principle applied to a hypothetical clear-cut case, CBO scoring and FCRA budgetary treatment would come to the same conclusion.

• But the scope of the descriptive language for the criteria is very broad.  A literal interpretation would produce some odd results.  Regarding federal sovereign power, CWSRFs and DWSRFs are obviously non-federal, state entities.  But they owe their existence to federal CWA legislation and their continued capitalization to legislated federal funding that comes with any number of strings attached.  Are the cash flows of SRF loans therefore ‘related to the exercise’ of federal power?  When an SRF issues a tax-exempt municipal bond to leverage its loan portfolio, should those bonds be considered federal debt [1]?  When SRF leverage comes from the SWIFIA program, should the program’s loan be ineligible for FCRA budgetary treatment [2]?

• A literal interpretation of the government purpose and control criteria has the same problem.  If a local water authority is constructing a CSO system in compliance with a federal consent decree, does that make it a ‘federal’ project?  If the authority issues municipal bonds to finance the project, are the bonds ‘federal’ because the authority was arguably ‘compelled’ to raise local taxes to repay them [3]?  If WIFIA makes a loan to the authority for the CSO project in the same circumstances, should the loan receive FCRA treatment [4]?

• To be fair, the CBO Report’s scoring examples (the bulk of the document) make it clear that CBO applies their two criteria in limited and realistic ways.  They don’t make odd decisions in reality. CBO’s pragmatic approach is the correct context in which to consider how the report’s criteria should be applied to the FCRA non-federal issue. The 1967 Report’s budget principles and specific criteria like those in the CBO Report need to be anchored in the real-world to be useful and efficient for policy implementation.

CBO’s Criteria in the Context of a Public Infrastructure Project Financing

The best way, I think, to provide a realistic anchor for the FCRA non-federal issue is to describe the basic elements of major public infrastructure project financings that are likely both (1) to apply to a federal loan program and (2) to have a federal participant, and then to see where and how CBO’s criteria might be applied.

The main distinguishing characteristic of a classic project financing is that the primary source of repayment for the project’s debt is not the project owners, but other creditworthy entities that purchase the project’s output under long-term contracts or (more typically for public infrastructure) the communities that benefit from the project’s existence and operations.  The latter agree to pay taxes or user fees that will cover the project’s operational costs and amortize the financing of the project’s construction costs.  Since the project’s debt is not fundamentally recourse to the project’s owners (who are also often its developers and managers), it’s usually called ‘non-recourse debt’ — but in fact the debt has a lot of recourse to project’s contractual counterparties, cash flows, and other security.

The relationship between the project’s beneficiaries and the non-recourse lenders, though usually mediated by the project’s owners or managers, is critical to the project’s success.  There are any numbers of factors related to the type of power and control that the CBO criteria describe.  The non-recourse lenders will require that the beneficiaries execute lengthy, long-term contractual obligations to pay for the project.  In addition to debt service, payments must cover detailed specifications for project maintenance and operations.  In turn, the beneficiaries require that the lenders provide cost-effective financing on acceptable terms that are consistent with their ability and willingness to pay for the project.  Most importantly, it is a very long-term relationship – each side will rely on the other to perform for decades.

The role of the project’s owners and manager is of course also critical to its success.  But this is most intense during the initial planning, development, and construction phases of the project’s life.  For large-scale public infrastructure like roads and waterways, most of the value is realized by successful construction completion and subsequent O&M activities are relatively routine.

In this kind of classic project financing for major public infrastructure projects, the beneficiaries are local or regional communities that agree to pay taxes or user fees for the value they expect to receive.  Federal involvement is primarily at the initial development stages, with a possible continuing role in project ownership and management.  Non-recourse lenders can include federal infrastructure loan programs if the project meets threshold and transaction quality eligibility requirements [5].  The nature and interaction of these three parties will determine whether a program loan will be subject to external discipline for OMB’s FCRA budgetary treatment and consistent with CBO’s scoring criteria for non-federal activities.

• The most important point here for both FCRA classification and CBO scoring is that there are two distinct types of project cash flows related to different activities – those related to project development, construction, and management of the project, and those related to (and often legally segregated for) non-recourse debt repayment.  This distinction means that CBO scoring criteria should not be applied in an undivided way, as if all the project’s cash flows were necessarily related.  Rather, the scoring decision should be more nuanced (as it appears to be in the CBO Report’s real-world examples) and include or exclude non-federal project cash flows towards the aim of providing the most accurate picture of federal involvement.  Non-recourse debt service will usually be an infrastructure project’s largest post-completion cash flow — incorrectly including these amounts (which should not be distinguished between those for private-sector debt and for a program loan, if there is one) would be misleading with respect to the scale of federal involvement in the project.

• With this distinction in mind, CBO scoring criteria can first be applied at the project’s partnership or JV entity level, where the federal participant and non-federal participants will interact directly.  At this level, there will be many factors related to the project’s planning, design, development, authorizations, and construction in which the federal participant has a dominant or even exclusive role [6].  That may be the basis for including non-federal cash flows related to those activities in the CBO score.  But note that even if all the cash flows at this level are included, that does not necessarily mean that the cash flows related to the repayment of non-recourse debt should also be included.

• To determine whether non-recourse repayment cash flows should be included in CBO scoring, the relationship between their source, the project’s beneficiaries, and the federal participant should be separately considered.  This is also the most important focal point for OMB’s FCRA treatment classification.  Are the primary beneficiaries of the project non-federal entities? It’s a bit hard to imagine that they would be otherwise for a large-scale public infrastructure project — perhaps if the project exclusively serves a major military facility?  Most usually, the beneficiaries will be communities represented by a sub-national state, local or regional authority.

• After it is established that the source of repayment is non-federal, more nuanced questions can be asked.  Most importantly, is the federal participant compelling the beneficiaries in some way, unrelated to the enforcement of widely applicable federal laws, to repay project debt?   This may be indirect — if the federal participant requires some aspect of project design, scope, siting, construction, etc. that is not consistent with standard public infrastructure benefit-cost principles and makes rejection of the project by the ‘beneficiaries’ practically impossible, that might be a form of compulsion.  This too is hard enough to imagine on the project’s physical level, much less with respect to the federal participant’s policy objectives and the beneficiaries’ legal rights.  But the question, and others like it pertaining to any power and control dynamics that might exist between the federal participant and the project debt’s source of repayment, can be asked in the context of the CBO criteria.

• OMB should be asking similar questions about the federal participant’s relationship with the project’s source of repayment for FCRA loan budgeting, if the project has applied for a federal infrastructure loan. Of course, these appear to be included in some form in OMB’s current FCRA criteria for WIFIA. But I think a more explicit and precise approach is necessary to avoid over-inclusion and ambiguity. Looking at the CBO Report gives a sense of the pitfalls to be avoided — if CBO’s brief description of their criteria is too literally interpreted, the results are odd and misleading, but their more expansive examples show limited and realistic applications of the criteria. This kind of realism needs to be surfaced for the FCRA non-federal issue, whether for improved criteria or a statutory fix. Much more on this topic in future posts in this series.

_____________________________________________________________________________________________

Notes

[1] JCT, the CBO’s tax section, estimates the tax revenue impact from legislation that might change the amount of the tax-exempt bonds issued in future. This includes estimates for WIFIA’s (supposed) effect of increasing bond issuance. The broader point is that including the bonds’ repayment cash flows in scoring isn’t being considered, even though JCT is presumably modelling them for the tax impact.

[2] Like WIFIA, the tax revenue impact of SRF leverage with tax-exempt bonds is explicitly included in CBO scoring. Again, plenty of focus on these cash flows.

[3] In theory, this is not just a CBO scoring issue — if the creditworthiness of tax-exempt debt is ‘substantively’ supported by the federal government, the tax-exempt status may be questionable. Forcing local taxpayers to repay a bond would have this effect. There will be more on this in a future post in this series.

[4] This is a real case: Decatur Priority Areas Sewer Assessment and Rehabilitation Program Consent Decree Packages. There is no publicly available information indicating that OMB ever questioned the FCRA treatment of the DeKalb CSO loan.

[5] In the four programs that generally limit loan size to 49% of project cost (TIFIA, WIFIA, SWIFIA and CWIFP), a program loan is not likely to be the sole source of non-recourse debt for a large infrastructure project. If CBO scoring includes the program loan to a federally involved project, presumably the private-sector debt will also be included.

[6] If the federal participant is the source of specific legislative authority to enable the project, that’s a major factor to consider with respect to initial project cash flows — equity investments by non-federal participants, for example. But once in place, that doesn’t mean that all of the project’s cash flows were enabled by the specific project legislation in any meaningful way. For the repayment of project’s non-recourse debt, the agreement by the non-federal beneficiaries to pay taxes, and by the non-recourse lenders to provide financing, are much more important with respect to non-federal external discipline and FCRA classification.