THE FUTURE OF PROJECT BONDS IN LATIN AMERICA

THE FUTURE OF PROJECT BONDS IN LATIN AMERICA Emil Arca * There has been an increasing interest in the use of project bonds in Latin America. This article explains these debt instruments and charts their rapid rise in Latin America. It then goes on to explore the challenges surrounding the expanded use of these nascent means of financing, and further investigates how such challenges have been or may be overcome. This article aims to lay the foundations for better understanding a rapidly evolving instrument in the Latin American project finance arena. The Definition of a Project Bond At the outset it is worthwhile to define the term project bond by reference to its component parts: a project and a bond. A project is usually thought to involve infrastructure (e.g., roads, bridges and tunnels; ports and airports; water treatment plants; pipelines) or other types of operating assets (e.g., oil drill ships, power plants). In addition to the great variety of needs to which such a project may relate, the term is used broadly in this context to refer both to greenfield (entirely new construction) or brownfield (existing site) assets. In its purest form project finance is also usually thought of as nonrecourse financing, i.e., the source of repayment is the revenues generated by the project itself, which may be from customers of the project or concession payments from a government for which the project was built. This is in contrast to corporate finance, in which the source of payment is the general balance sheet of an operating company. In practice, however, some transactions may have diverse forms of support from the transaction sponsor or third parties. In contrast to the wide variety of projects that the foregoing matrix of attributes can produce, the term bond is seemingly straightforward. One might think that in this context it simply means the issuance of a debt security purchased by a large number of * Emil Arca is a partner in the New York City office of Hogan Lovells US LLP, where his practice encompasses a wide variety of capital markets and finance transactions, especially in emerging markets. Kristen Lam, an associate in Hogan Lovells s New York office, assisted in the preparation of this article. 187

HARVARD BUSINESS LAW REVIEW ONLINE 2013 institutional investors, such as insurance companies (compared to the more traditional means of financing projects via a loan from one or more commercial banks). Indeed, that taxonomic distinction between bonds/institutional investors and loans/commercial banks is a useful one that will form the basis for the discussion that follows below. However, in adopting this sharp distinction for heuristic purposes, we should not overlook converging characteristics between these two markets. For example, there are now mixed bank/institutional investor structures where bank loans and bonds appear in the same transaction pursuant to common terms or other inter-creditor arrangements. Also, bank loans, particularly on very large projects, may be syndicated to a large number of banks and thus, like bond offerings, have a large number of participants. Furthermore, commercial banks have increasingly become distributors of bonds, and funds (usually thought of only as bond buyers) have become investors in some types of bank loans. The Rise of Project Bonds in Latin America The impetus for the increased use of project bonds in Latin America and elsewhere has come from the confluence of three factors: (1) enormous unmet infrastructure and energy needs; (2) capital availability among fixed income investors; and (3) capital constraints among banks. We will pass over the first factor, as it is widely remarked upon. The second has both short-term and longer- term structural elements. The short-term impetus is that, with the U.S. Federal Reserve Bank pursuing a low-interest policy in the current low-growth economic environment (and, by its own estimates, continuing to do so for another year or more), the real return on risk-free assets such as U.S. Treasuries has hovered around zero. In this environment, fixed income investors have looked to riskier assets in their search for yield. As a result, instruments such as U.S. corporate junk bonds, also referred to as high-yield or speculative grade bonds (i.e., bonds rated non-investment grade), and bonds from Latin America corporate issuers have enjoyed record highs in issuance levels and record low spreads over U.S. Treasuries with comparable maturities. It remains to be seen whether such demand will exist in an interest rate environment more typical of the historic norm. The longer-term, structural impetus for the suitability of project bonds is that institutional investors, such as insurance companies, have longer-term liabilities than banks and a more natural ability to match a long-dated (e.g., 10 or 20 years) project s assets to those liabilities. Likewise, the constraints on bank lending have short-term and longer-term structural elements. Two short-term elements are: (1) the recent need for banks worldwide to shore up their balance sheets following the financial crisis that began in 2007; 1 and (2) the effect on European banks (a number of which were leading project 1 Even though in the public consciousness the financial crisis is usually thought of as beginning on September 15, 2008 with the bankruptcy of Lehman Brothers and the acquisition of Merrill Lynch by Bank of America, by the summer of 2007 the world had seen the failure of Bear Stearns, the freezing of much of the US asset-backed securities market, and disruption in the inter-bank lending market. 188

THE FUTURE OF PROJECT BONDS VOLUME 3 finance lenders in Latin America) of the Eurozone crisis. The foregoing factors may be transitory but there are also structural, and therefore long-term, constraints to the amount of project lending the bank market in general may take, including: (1) for individual banks, internal or regulatory limits on particular project sponsors, industries, countries or obligors (i.e., off-takers on the project); (2) the increased cost of capital under Basel III rules that are in effect or coming into effect for banks; and (3) the obverse of the match funding issue noted above in respect of institutional investors, namely: banks fund themselves from a variety of sources, but many sources, such as deposits, are short-term. Early Project Bonds and the Role of Monoline Insurance Many early project bond transactions all over the world benefitted from financial guaranty insurance provided by monolines (i.e., insurers who are only in the business of providing such insurance). In Latin America, these transactions appeared in the late 1990s and included such high-profile projects as the Autopista Central in Chile. Such transactions typically had to have a shadow investment grade rating (i.e., a rating of BBB- or Baa3 or above prior to giving effect to the insurance) from the rating agencies rating that insurer. In return for a premium paid over the life of the transaction, the insurer would wrap the bonds. A wrap is the promise to pay on demand by the indenture trustee, without any defense to payment upon submission of the proper form (as in a simple letter of credit), any shortfall in the relevant period s debt service. As the insurers typically had the highest possible ratings (AAA or Aaa), the bonds they insured likewise carried those AAA/Aaa ratings. Such insurance provided a number of benefits: (1) the higher ratings on insured securities allowed the sponsor to access a deeper and more liquid market than it could with an unwrapped securities offering; (2) as the insurer typically retained voting and control rights on virtually all matters except those that would change the underlying payment characteristics of the bonds and a few other key provisions, the sponsor only had to go to one party to seek waivers and consent; (3) the guaranty effectively provided liquidity as well, as it not only covered ultimate credit risk but it also insured each scheduled payment; and (4) the monoline s staff had the time that many institutional investors did not to study the project in-depth, to visit the site, and to monitor subsequent reports. As will be seen below, the latter three attributes mitigated important challenges now facing the unwrapped project bond market. Unfortunately like virtually all other market participants and regulators, and despite stellar track records in other credit assessments the monolines and the rating agencies misjudged the severity of credit losses on one product (and derivative products based on such underlying product): securitized subprime U.S. residential mortgage loans, especially those that originated during the 2005-2007 period. The amount of capital the rating agencies required the monolines to have depended on the ratings of their portfolio risk being at least investment-grade (below investment-grade risk increased the capital needs of the monolines considerably), so the ratings on the monolines plummeted when 189

HARVARD BUSINESS LAW REVIEW ONLINE 2013 the ratings on their underlying portfolios likewise plummeted as a result of losses in excess of the expected investment grade loss scenarios. It seems unlikely, at least in the immediate future, that monolines will return not only due to the existing companies having disappeared altogether or restricting themselves only to insuring U.S. municipal bond risk (the sector in which the business began), but also due to the lack of market confidence in the business model described above. In their absence, to the extent that project bonds are getting done, investors have had to take on the full credit risk of projects themselves. As discussed in the following section on challenges to project bond execution, the liquidity support and administration convenience of the monolines role have been or can be dealt with in a variety of ways in project bonds. Recently, there has been speculation that some of the monolines role might be taken up by MLAs (multi-lateral development agencies) or ECAs (export-credit agencies), as they are government funded, and project bonds often come within their respective mission statements to promote social development in emerging markets such as Latin America or exports to Latin America. Challenges of Project Bond Execution Each project bond poses its own challenges and there is not enough space here to discuss the sui generis ones. There are however certain recurring, interrelated issues that can be discussed in general terms. The first are common to project bonds globally, and the last couple are of particular relevance in emerging markets such as Latin America: Industry Expertise. Banks active in project lending usually have teams of bankers familiar with issues specific to the asset type, the country, etc., and if they do not, the manner of underwriting the credit on a bank loan and documenting it may leave them with more time to learn about it than bond execution does. While there may be some institutional investors with specific expertise, in general they do not have the ability to replicate the expertise of banks consistently active in project lending, as fixed income investors typically need to look at a wide variety of investments. Of course, to the extent relative market share between lenders and bond buyers shifts in favor of the latter, a corresponding shift in experience could be expected to occur as well. In smaller transactions executed in the traditional private placement market (i.e., transactions where the investor buys the debt directly from the sponsor and the arranging bank acts only as a placement agent), such education opportunities for bond investors may exist, especially if the investor is involved in the transaction at the outset. In these transactions, the investor has the opportunity to do its own diligence and negotiate documentation. However, in larger deals that are commonly executed in the 144A market (i.e., deals where the end investors buy from the bank which has underwritten the bond), there is little or no opportunity to negotiate terms or documentation, and the investor has only a short period following the presentation of an offering memorandum and a live or virtual roadshow to make an investment decision. In the latter circumstance, the 190

THE FUTURE OF PROJECT BONDS VOLUME 3 investor has to rely, to some extent, on the expertise of other parties who may have been working on the transaction for months or even a year or more. In reviewing the offering memorandum for the bonds, these investors are dependent on: (a) the underwriting bank s team (often comprised of members with experience in projects from the bank s lending work) who, along with the issuer and their respective lawyers, structured the transaction; (b) the rating agencies who may have rated the bonds; and (c) various experts who have provided their opinions on certain issues, such as the projected demand for the project s output, its engineering, its budget and the adequacy of its insurance (the opinions of these experts are often included in the offering memorandum). Construction Risk. Not all project bonds have construction risk, for as noted at the outset of this article the term is used broadly to refer to non-recourse financing of large or single operating assets already in existence. Such project bonds without construction risk are sometimes referred to as the take-out phase of a project financing or operating asset securitizations. Recent examples of such transactions in Latin America include offerings backed by drill ships operating in the oilfields off the Brazilian coast (and their charters and service contracts with Petrobras) and bonds secured by toll road revenues from the northern and southern toll roads in Panama. However, bonds with construction risk, whether in Latin America or elsewhere, are usually deemed the most difficult for the bond market to handle. Recent examples of such bonds in Latin America include those relating to water treatment plants in Peru. Bonds may also involve partial construction risk or improvements to existing facilities, such as to choose a couple of examples from Latin America another Brazilian drillship transaction or a bond relating to the financing of the Lima, Peru airport. The level of difficulty will vary depending on the construction s duration, novelty, complexity, susceptibility to delay, and the experience and resources of the sponsor. From a bondholder s point of view, there are two principal elements to construction risk: (a) credit risk (i.e., the risk that the bond will never be paid back or the unfinished collateral will need to be sold at a loss if the project is not completed and therefore will never generate revenues); and (b) timing risk (i.e., the risk of delayed revenue generation and payments based on those revenues due to construction delays). (a) Credit Risk. If there is no recourse to a creditworthy sponsor (or another creditworthy counterparty, e.g., a financial guarantor as described in the preceding section or, in a more limited capacity in one of the Brazilian drillship transaction, Kexim (a Korean ECA), which provided a completion guaranty related to construction at a Korean shipyard), the only way to avoid these issues altogether is not to have bond investors take this risk. Thus, a bank may provide a loan to take the project through the construction phase, with incentives such as cash sweeps and increasingly higher margins over time to provide the sponsor with an incentive to finance the operating or take-out phase elsewhere (such loans are sometimes referred to as miniperms ). As the Latin American project bond market develops, there may 191

HARVARD BUSINESS LAW REVIEW ONLINE 2013 also be more structures where such bank and bond financing exist together at the outset of the transaction. (b) Timing Risk. Payments on project bonds are often rated by rating agencies, either corporate-style (for timely interest and timely principal on scheduled payment dates) or structured finance-style (for timely interest on scheduled payment dates and ultimate payment of principal on the legal final payment date, although the structure itself may provide for payment of principal on various payments along the way). Such ratings are indicative of the fact that fixed income investors typically have a greater need than banks for predictable payments on designated dates. In project bonds, timing issues are often addressed with the use of reserve accounts. Depending on the stage of the transaction, funding for these reserve accounts may come from the following sources: (a) at the outset of the transaction, cash from bond proceeds; (b) in the later stages, the project revenue source; or (c) where there is a reimbursement right to a creditworthy counterparty or, in later stages, the revenue source, letters of credit (which can be drawn upon without any conditions and thus essentially function as a cash substitute). Reserve accounts have a number of uses: they can be drawn upon (a) to pay debt service (i.e., to take any delay risk, including risks related to construction delays) or (b) in other cases, to fund key expenses (e.g., insurance and, once the asset is in existence, further capital expenditures and operating expenses). Sponsors may also be allowed to provide an equity cure (i.e., fund a liquidity account themselves) a limited number of times during a transaction to prevent it from going into default. However, such cures are voluntary, so they are more for the sponsor s benefit than something investors can rely upon at the outset of the bond. In order to be cost-efficient, debt service reserve accounts have to be sized appropriately that is, large enough to absorb likely delays but small enough to make the transaction cost-efficient so as not to absorb an undue amount of the bond offering s proceeds. As it is, the entire construction phase results in a project being subject to negative carry (i.e., absorbing the cost of having fully disbursed funds from the bond offering to fund construction accruing interest at the long-term bond interest rate, while such idle cash is invested in short term and, therefore, almost always lower yielding eligible cash equivalent investments). Such negative carry is one of the reasons project finance was traditionally thought to be a more natural bank product, as its funds typically only accrued interest when they were disbursed (although banks were able to compensate themselves through commitment fees or interest rates for the implicit cost of stand-by credit). Thus, while reserve funds may mitigate delay risk, they will add to the negative carry issue for the project s economics if they are excessively sized. Bond Administration, Waivers and Consents. Project bonds typically involve a large number of investors, all of whom have consent rights (in contrast to a bank loan 192

THE FUTURE OF PROJECT BONDS VOLUME 3 which may have fewer banks and/or more authority invested in an administrative agent or other lead bank or banks). On most issues, waivers require the consent of a majority in interest of investors. Other issues may require other percentages of approval, ranging from zero (for noncontroversial matters such as correcting typographical errors in the bond documents or adding collateral) to 25% (if such a corporate-style threshold is used for declaring a default on a bond rather than the majority-in-interest standard more common in structured finance offerings) to 100% (for changes in the payment characteristics of a bond or to release collateral). Obtaining such consent may be difficult and/or time-consuming for several reasons: (a) the large number of investors that may have to consent to reach a majority in interest; (b) contacting beneficial owners with voting rights, involves making communications through the book-entry systems, as the bonds are often held in street name through brokers and other intermediaries; (c) the lack of expertise the investor may have to evaluate the request; and (d) the lack of time the investor may have to evaluate the proposal (project finance is unlikely to be the investor s full-time focus). In most corporate bonds, or structured finance transactions involving relatively passive pools of financial assets, such requests for waivers, consent or amendments are rare, but the complexities of operating assets (particularly if construction risk is present) make such requests more likely in a project bond. There are a variety of mechanisms that have been used in or proposed for project bonds or similar transactions to move such decisions away from investors to other parties. The mechanisms that are employed will depend on the nature of the issue, but may include: (a) affording greater flexibility (than in a bank loan) to the sponsor to make decisions, or including less stringent covenants (making it less likely that the sponsor will breach a document requirement in the first place); (b) requiring rating agency confirmations (i.e., allowing the sponsor to take a particular action if the agencies rating the bond confirm that it will not result in a downgrade or withdrawal of the rating on the bond); (c) placing greater reliance (than in a bank loan) on an independent engineer s judgment of the permissibility of a change in plan; (d) allowing a subordinated investor (i.e., the party in a first loss position below the bulk of the debt) to make certain decisions (this is analogous to tranched commercial mortgage-backed securities in which the lowest tier of notes often has the authority to direct the actions of a special servicer charged with foreclosures and similar matters); (e) in structures with both banks and bonds, allowing bank lenders to make certain decisions when their exposure is of the same type as or prior to that of bond investors. Ratings Ceilings. The deepest part of the international bond market is for issues rated at least investment grade (BBB- or better by Standard & Poor s Ratings Services or 193

HARVARD BUSINESS LAW REVIEW ONLINE 2013 Fitch Ratings or Baa3 or better by Moody s Investors Service). Historically, few countries in emerging markets were rated investment grade on a foreign currency basis as a result of various, and related, combinations of political and economic system risks, external debt profiles, and transfer and convertibility risk on foreign currency payments originating from within the country. In fact, over the past quarter century in Latin America, Chile was the only country consistently rated in this category. In recent years, however, other major countries in Latin America such as Brazil, Colombia, Mexico, Panama, and Peru have maintained investment grade ratings at various points from one or more of such rating agencies. As such sovereign ratings usually provide a ceiling for foreign currency-denominated bonds originating from payment sources within the country, the maintenance of such sovereign ratings at or above investment grade can expand the project bond market, and, conversely, their decline below such threshold can be expected to contract international demand for such bonds. Currency Issues. If it is true that the greatest demand in the international bond market is for investment grade bonds, then it is even more true that in the international bond market the deepest demand is for such bonds based in one of the major international currencies, typically US dollars. While there are pockets of international interest in local currency bond issues, these have typically existed for corporate issuers without the additional complexity of analyzing a complex nonrecourse project. Likewise, while it is true that there has been a number of project bonds (defined broadly) in local capital markets, only some countries in Latin America (notably Brazil, Chile, and Mexico) have sufficiently developed capital markets (in size and sophistication) to handle such issuances. Even there, capacity is limited compared to the overall social needs for project spending. The issue of currency risk has typically been handled in one of several ways: 1. Having a dollar-denominated source of payments. In such circumstance, there is no mismatch between the asset (the contract or market supporting the project) and the liability (the issued bonds). Such contracts may be present in, for example, the offshore oil drilling market in Brazil. However, unless the payor in US Dollars likewise has a dollar-based balance sheet, the investor is simply shifting the asset-liability mismatch risk from the project to the off-taker or market making the dollar denomination payments to the project. 2. Having a legally and factually dollarized economy. Currently in Latin America, there are only two countries with such dollar denominated economies: Panama and El Salvador. The reason for the addition of the adjective factually in the foregoing caption is that one can have a legally dollarized economy as Argentina did in the 1990s when it set a one-to-one peg between the peso and dollar but if the underlying economy is not sufficiently dollarized, the maintenance of that peg (or other form of dollar denomination) is pure sovereign risk against the day the sovereign decides that it needs greater 194

THE FUTURE OF PROJECT BONDS VOLUME 3 flexibility to have a local currency-based economy. 3. A currency hedge from a credit-worthy counterparty. Hedges can come in a variety of forms depending on the degree of protection that is sought, the simplest being a standard cross-currency hedge whereby the project entity exchanges local currency at a scheduled exchange rate against delivery of the required amount of dollars for that period s debt service. Such hedges are typically provided by banks. One issue of special concern for bonds is where in the priority of mark to market termination payments upon breakage of such hedges are paid in the transaction s payment waterfall. The hedge provider will typically want to be pari passu with the debt and investors and rating agencies will due to the theoretically infinite amount of such termination payments want it to be subordinated to debt. One compromise is to have a designated amount of such payments made pari passu with principal (covering what the hedge provider believes is a likely payment scenario, but which investors and rating agencies then can size in running their transaction cash flow stress tests) with the remainder paid on a subordinated basis. 4. Overcollateralization. If the transaction s cash flows are sufficiently robust and the local currency is viewed as sufficiently stable it may be possible to cover currency risk from the transaction s collateral structure. However, project bonds present special challenges to making such assessments, both because demand (absent a contract with a dedicated offtaker) may be uncertain and because of their longer tenors. As the foregoing suggests, bonds will not be the sole source of financing for projects across Latin America. However, it is hoped that they can be part of the mixture of funding which encompasses public and private, bank and capital markets, as well as local and international, funding that together can supply the capital to meet Latin America s growing and unmet needs for infrastructure, energy, and economic development. 195