Historically there has been an underlying confusion among market participants regarding the uses of project finance. From inception, project finance has been a way for financially weak players to raise capital without depending on their own balance sheets. The natural extension of this approach came to be called leveraged economics, and involved the valuation of capital projects using return on equity rather than return on capital. Over time, doubts grew about the validity of these approaches. These doubts spurred sponsors and lenders to devise other applications, such as mitigating large project risks, financing joint ventures between unequal participants, mitigating agency issues, and addressing underinvestment. Yet, the original rationales did not fall into disuse. Many sponsors continue to use project finance to shield their own credit capacity and leverage project returns. Indeed some – most famously Enron – abused these rationales with catastrophic consequences.

There exists no conceptual consensus about the validity of off-balance sheet financing or levered economics. Many sponsors that pursue these aims do so on the basis of hiding their project debt using accounting artifice. High quality firms ignore project financing options that could create value because they do not believe in the validity of its uses.

How did conditions get to this present juncture? Right from the beginning there were questions about how off-balance sheet most project loans are. Then project finance evolved, becoming more limited recourse in nature. These changes, and periodic debtor meltdowns, reinforced sceptics’ view of the product. Still, deals based on the historic rationales continued to close. Consequently, we can say that the oldest issues in project financing – whether loans should be regarded as separate from consolidated debt, and whether leveraged economics are valid – remain unresolved.

This article will revisit project finance’s historic uses in the light of more recent financial theory. It will suggest that viewing project finance as funding with embedded options sheds light on the two unresolved issues – indeed, it positions them more clearly to be used properly.

Project financing ends up having a qualified potential to expand debt capacity, and most leveraged economics end up looking less attractive than sponsors might wish. This should not be surprising. Quantification of contingent commitments means acknowledging that these liabilities have costs and that project finance’s utility will vary with the sponsor’s ability to make use of its optionality. Accepting these limits however, should enable sponsors to discuss their project loans more candidly with auditors and ratings agencies, without engaging in self-deceptions.

The origins of the debate

Project financing’s special claims originated when sponsors – often individuals with an adventurous streak – had nothing of their own to pledge. Lenders frequently took an equity stake to compensate for their loans’ high risk. This rough-and-ready version of project finance was at least pure: lenders could only look to a project’s cash flow and assets for repayment. Sponsors felt with justification that such debts were ring-fenced legally and thus separate from other debts for which they were liable. If the project went badly, sponsors would walk away from loans in a heartbeat. Their equity was all that they risked, meaning that leveraged economics was a valid framework for assessing their ventures.

By the late 1970s, two events – the massive inflow of Petrodollars into the London bank market and the development of North Sea oil – changed the market. The former created cheap Eurodollar funding, for which banks sought outlets. The latter attracted a new class of sponsors, oil majors that used project financing for new reasons. BP used a production payment financing for the Forties Field, even though it had a quality balance sheet and substantial financing capacity. It used project finance on Forties to hedge its operating and tax risks – and AAA-rated Exxon followed suit.

These changes raised new questions. How should sponsors evaluate project economics for oil fields that they finance using non-recourse loans? How should these loans appear in financial statements? Should the ratings agencies count these loans when calculating financial ratios? The answers were not immediately obvious. Would BP and Exxon surrender their oil reserves rather than fund project overruns? Would they walk away from the projects if temporary cash flow deficiencies threatened defaults? Sponsors rarely provided clarity.

Partial answers have emerged. Project finance lenders became more adept at extracting limited and/or contingent guarantees from sponsors to compensate for commodity price volatility. To keep loan costs down, some sponsors provided informal assurances that they would remedy any default. In other cases, sponsors voluntarily stepped up when faced with the risk of losing the assets in a default. Today project finance banks routinely ask sponsors whether their projects are strategic, a polite way of asking whether the sponsor will support them in the event of trouble.

These developments constitute a formidable case against the idea that project loans should be segregated from a sponsor’s consolidated debts. Academics and some practitioners argued as follows:

Almost all project loans are now limited recourse in nature. Sponsors are always on the hook, to some degree. Moreover, it is impossible for outsiders to gauge the degree of sponsor support. Accounting disclosures say little about completion or operating deficiency supports, and nothing about informal borrower assurances. Voluntary sponsor rescues occur with some frequency. The safest course then is to assume that project debts should be counted as sponsor obligations unless the sponsor can prove otherwise.

One would expect the ratings agencies to be most interested in assessing the degree to which project debt should be segregated from a sponsor’s consolidated debt. In fact, the agencies’ practices are not highly developed, and to the extent they exist, largely follow the sceptic’s position. This author had conversations with Standard & Poor’s and Moody’s in January 2013, which confirmed the following:

• The agencies’ primary focus is rating project debt, not assessing its impact on sponsor consolidated ratios. Indeed, until recently the company credit analyst did not always participate in the credit review of a project loan.
• No formal process exists for sponsors to disclose project loan terms in an effort to limit impacts on its credit rating. Few sponsors try, and in the rare instance where one does, the ratings process is ad hoc and case-specific.
• A Moody’s executive indicated that “if the sponsor puts the project loan on its accounting statements, we count it in full in their financial ratios.”

Whether this ratings agency stance is a cause or symptom of sponsor behavior is unclear, but it has not helped the development of project finance. High quality sponsors believe – and act as if – project loans cannot be separated from the sponsors’ debts. Sponsors that use project finance to expand credit capacity do so by hiding the debts from analysts and ratings agencies, and may understate contingent obligations even to their own managements, overstating their leveraged economics in the process.

Rebuilding robust reporting

A project loan reported in consolidated financials should be not automatically be equated with full recourse debt. Project loans have become exercises in partial recourse, offering sponsors different degrees of ability to walk away. Thus, sponsors need to find methods that can measure partial recourse on a case-by-case basis. This will allow sponsors to hold sensible discussions about their financial ratios with ratings agencies. It also will facilitate more effective disclosure of project loans in financial statements, and the appropriate inclusion of contingent liability capital charges in leveraged economics.

The best example of this approach involves the most common type of contingent support, the completion guarantee. Almost without exception, sponsors’ financial reports ignore completion guarantees. Ratings agencies do not focus on quantifying these exposures, and do not offer methodologies for doing so. In volunteering a quantification of their contingent supports, sponsors can take discussions with auditors and the agencies to a higher level. By acknowledging that contingent support has costs, sponsor efforts to segregate project loans from consolidated debt will gain credibility.

One simple method for handling completion guarantees would be to encourage auditors and ratings agencies to classify project debt as a consolidated obligation until completion tests are met and the loan turns non-recourse. While appealing in its simplicity, this method overburdens the sponsor during the guarantee period. Completion supports are largely assurances that cost overruns will be funded. To treat this type of guarantee as the equivalent of fully guaranteeing the debt is to ignore its contingent nature – the vast majority of completion support involves spending enough to complete a project, rather that paying off total project debt.

A better approach would be to record a probability-based capital charge that represents the likelihood that the sponsor may need to contribute additional funding to a project. One way to estimate this charge would be to use the project’s appropriation cost estimate (APC). Capital budgeting typically requires a developer to produce a 50/50 APC. This is a project cost that has an equal probability of going under or over budget. APCs imply that the probability distribution for project cost has a 50% zone, consisting of overruns of varying sizes, plus some risk that the project is abandoned.

With this APC in hand, a developer can ask cost consultants for two more data points – the mean probability cost overrun and the probability of abandonment. The former provides an overrun estimate with an equal chance of being larger or smaller. The latter will involve some combination of unforeseen technical, execution and political risks such that the project ends up being abandoned.

With this data in hand, a sponsor can produce a reserve capital charge estimate. This capital charge represents the average amount of financing capacity that a sponsor should reserve in order to service completion guarantees extended on a portfolio of similar projects. This capital charge can be calculated as follows:

Completion support capital charge =

(50% – % probability abandonment) x mean value overrun +

(% probability abandonment) x value of debt

Consider, for example a $500 million project funded with $300 million in project debt and characterized by a $50 million mean value overrun and a 1% chance of abandonment. Using the above formula, the project’s completion support charge would be:

(50% – 1%) x $50m + ($300m x 1%) = $24.5m + $3m = $27.5m

This calculation indicates that if a sponsor were to reserve financing capacity estimated in the above manner every time it extended completion support, it should be adequately reserved to make good on its guarantees. The sponsor can then volunteer these reserve estimates to ratings agencies when it requests that project debt be segregated from consolidated ratios. Ratings agencies can count these capital charges as debt equivalents when calculating debt/capital ratios. Each year the sponsor’s reserve levels could be adjusted as it offers additional completion support or projects meet completion tests. Sponsors can also use these charges as upfront capital costs to be added to their equity investment when calculating leveraged economics.

Completion support is only the most common type of contingent obligation. Others include operating cash deficiency agreements and take-or-pay offtake contracts. For these the approach should be similar – estimate a capital charge that reflects the probability of an obligation being called and the mean size of the financial contribution it would need to make to the project. The sponsor would then reserve financing capacity against these obligations, report these reserves to auditors and ratings agencies, and include them in leveraged economics.

More difficulties arise when dealing with informal sponsor assurances and voluntary support. Informal assurances are fundamentally a disclosure problem. It is not wrong or illegal for borrowers to advise lenders that voluntarily support is likely. However, sponsors must avoid dissembling or remaining silent about these assurances with other stakeholders, auditors and ratings agencies. Too often these assurances are not carefully scrutinised before they are given, and are sometimes offered without the full knowledge of sponsor management.

Sponsors can use full disclosure to their advantage, building up a reputation for not giving informal guarantees. This can be accomplished in several ways, one of which follows. Sponsor disclosures could include a senior management representation in loan documents that a sponsor has not provided any informal assurances. Sponsors can repeat this representation in a footnote to their audited financial statements, and would be liable for misrepresenting its financial statements if this turned out not to be the case. By giving such representations in the face of potential liability, sponsors should be able to convince the ratings agencies that informal assurances have not been given and should not be factored into the agencies’ treatment of the sponsor’s project loans.

The remaining challenge comes from sponsors’ willingness to voluntarily rescue a project, even if they have not provided a formal or informal commitment to do so. Its difficulties arise because many sponsors have economic incentives to remedy project debt defaults, and sponsors will not know whether they want to intervene until it becomes necessary.

Sponsors should start by acknowledging that project finance offers them the option of walking away, and go on to describe the circumstances under which they could foresee exercising that option. For some sponsors, that description will indicate that walking away is likely, and the case for project debt segregation is strong. Other sponsors will walk away in specific scenarios with relatively low probabilities. Their case for debt segregation is correspondingly limited.

To use this approach, sponsors should provide ratings agencies with their decision framework for walking away. This framework could, in theory, contain one or more of the following elements:

• Financial capacity to intervene is limited, so voluntary rescue is resource constrained
• The sponsor has more attractive projects, resources are limited, and the opportunity cost of voluntary rescue is high, so voluntary rescue is unlikely
• Since the project is highly leveraged, the sponsor will see more incentive to walk-away rather than support the debt if the cost of a remedy surpasses a fairly low threshold
• The sponsor’s business model relies upon a reputation for not intervening to support stand-alone debt. Such sponsors, for instance a private equity firm, will not expend good money after bad. Collectively these reputational concerns make voluntary rescue unlikely
• The sponsor intends to hedge risks by limiting its equity stake, and will not want to convey to the host or other governments that the hedge is for appearances only.

Once the sponsor makes its decision framework for walking away explicit, ratings agencies will have to decide which default scenarios will result in a sponsor walking away and how likely these are. Agencies already make similar probabilistic judgments on management turnaround strategies or borrowers’ contingency plans for rectifying financial ratios when these are under pressure.

Sponsors can help their cause by volunteering risk assessments and decision criteria that ratings agencies can apply to their case. This involves an analysis of a real option to walk away in default, and is complicated by the fact that circumstances may look different to the sponsor at the time the option would be exercised. These complexities can best be depicted with the decision tree in Chart 1 below.

Chart 1: Project finance walk-away option

To help ratings agencies use this approach, sponsors will have to provide:

• Their estimate of the probability of default for any reason
• Their division of default scenarios into between cases in which they will walk away and in which they will not
• Estimated mean rescue costs in each of these cases
• Their own estimate of the likelihood they will behave as expected should a default arise. Such estimates are not self-assessments of one’s capacity for dissembling, but rather allowances for the potential impact of unforeseen circumstances.

The ratings agencies will test these sponsor estimates and make their own adjustments

This option assessment can be illustrated using the hypothetical case of a multinational energy company concerned about political risk in Venezuela. Its Venezuelan project has very strong economics and its management has credibility with ratings agencies. Management says that it will walk away from its project only if the host government arbitrarily revises its contracts, tax regime and ownership. It estimates that default for any reason is a 20% probability and that 90% of defaults would be politically driven cases where it would walk away. Any voluntary rescue has a mean cost of $100 million. Finally, management expects that it will behave as represented in 80% of the cases in which it anticipates walking away and 90% of the cases in which it does not.

Assume for illustrative purposes that the ratings agencies accept management’s estimates. The financial reserve associated with this sponsor’s potential for voluntary rescue is represented by the decision tree in Chart 2.

Chart 2: Project finance walk-away option
Possible Venezuela $100m voluntary rescue requires $5.4m financial reserve

Calculating the combined expected value of the branches ending in voluntary rescue would unfold as follows, with P representing the compound probability of a branch:

[P(Default in walk-away scenario,

and sponsor does not behave as expected) +

P (Default in non-walkaway scenario and sponsor

behaves as expected)] * Rescue Cost

= [(20% x 90% x 20%) + (20% x 10% x 90%)] x $100m = $5.4m

The provision for voluntary rescue is low in this case because the circumstances under which the sponsor will actually ride to the rescue are quite limited. This in turn suggests that the rating agencies should not heavily penalise the sponsor’s request to segregate its project loans on the grounds that it will voluntarily stave off default.

This methodology can be further refined. For example, in most cases there would be different rescue cost estimates for the walk away and non walk-away scenarios. What the above illustrates is the need to capture both the compound probabilities built upon sponsors specifying the conditions under which they would exercise their option, and the possibilities that forecast and actual sponsor behaviours may diverge.

It is important to stress that this option analysis estimates the amount of money a sponsor is likely to spend voluntarily to rescue a specific project loan. It is not an estimate of the amount of debt that the sponsor is likely to voluntarily stand behind. This voluntary rescue cost estimate can then be recognised as a charge against reserve financing capacity that a sponsor sets aside, and added as a debt equivalent to its debt/total capital ratio.

One last benefit of this methodology is the clarity it brings to leveraged economics. The fundamental premise of such economics is that the sponsor only risks its equity. Voluntary rescue, born of a decision to forego the option to walk away, undermines the foundations of levered economics. By applying the above methodology, sponsors disposed to use leveraged economics can adjust such economics for their strategic intentions, financing capacity and risk tolerances. The ultimate result should be more legitimate usage of leveraged economics, with returns on equity somewhat closer to a project’s return on total capital.

Making disclosure pay

A process that allows sponsors disposed towards transparency to segregate some project debt from the ratios that determine their debt ratings would be desirable. It would also facilitate a disciplined use of leveraged economics, one more suited to a world where limited recourse loans are the norm. Disclosure standards should reward sponsors for integrity, and stigmatise those whose project financing activities seem not to register on their financial statements. Quality sponsors and the capital markets would both gain. Accounting gimmickry would be devalued.

The methodologies above will not be the last word on this subject. However, they should facilitate initial discussions of contingent obligations with auditors and ratings agencies. Undoubtedly these methodologies can be improved with trial and error. Sponsors will also need to weigh the benefits and costs of increased disclosure of their decision-making processes. Many will find it is better to make a quantitative estimate of contingent events than to use their uncertainties as reasons not to address the issues surrounding project debt segregation and valid leveraged economics.

The author is an executive professor of finance at the University of Houston’s CT Bauer College of Business, and a former treasurer of the ExxonMobil Chemical Company.