A liquid bond market for infrastructure issuers is the next step in the evolution of the project finance market – and has been for several years. The demand exists, in the form of pension funds and life insurance companies with long-dated liabilities. The supply exists, because sponsors like fixed rates and longer-amortisation profiles, which usually allow for larger debt amounts. Banks, their arranging fees safely booked, are happy to have an exit from their commitments.

But both supply and demand have not lived up to expectations. Pension funds and life companies have become as comfortable with infrastructure equity as infrastructure debt. During the later years of the credit boom volumes of project bond issuance were low. In Europe, outside the UK PFI sector and scattered continental renewables deals – few with a flawless record – the product struggled to compete with banks offering multi-decade tenors.

Props kicked away

Even those markets with healthy issuance suffered from distortions. Banks were big players in European bond markets, but tended to prefer highly-rated paper, which attracted lower capital charges. They found that wrapped bonds or securitisations of their own loan books, offered more palatable exposure to infrastructure than the low-investment grade levels that unwrapped bonds tend to reach. So the monolines tended to dominate the European bond market. In the UK, wrapped bonds could be bought by banks' proprietary trading desks, hedged with the cheaper credit default swaps on the monolines, and the difference in spread captured upfront. Until the monolines collapsed, this was a profitable trade. "The demise of the monolines has ripped up the established capital markets template for European projects," says one banker specialising in capital markets in London. "We need to find a new participant that could assume some of the functions – particularly supervisory – that monolines once did."

Syncora (formerly XL) and FGIC have been told to suspend claims by the New York State Insurance Department while they work to restore their financial health. Ambac's regulator, the Wisconsin Office of the Commissioner of Insurance, has parked four Ambac infrastructure obligations in a segregated account for claims with a high probability of distress, alongside Ambac's distressed structured finance commitments. The four are its primary market wrap of the bankrupt Las Vegas monorail, and three credit default swaps, against Reliance Rail, the Sanef repackaging and the super-senior tranche of the EPIC project debt securitisation.

Now sellers of credit default protection on monolines are being asked to make good on their protection. An auction process is normally used to settle the amount outstanding on these swaps, because a buyer of protection must deliver a reference security, which has a residual value in exchange for payment under the contract, and these are not always easy to locate. In recent auctions for Syncora (formerly XL) and FGIC CDS, CDS sellers had to make good on 85% and 74% of face value insured, respectively.

In the UK, unwrapped debt has yet to stack up favourably against a resurgent bank market. A combination of UK-based (and in some instances UK government-owned) bank lenders and the European Investment Bank can match the tenors on offer from the bond market, and even with swap costs comes in 25-30bp cheaper.

Carillion and Lloyds, as sponsors, closed a £375 million 30-year loan with Lloyds and Royal Bank of Scotland to finance their £685 million Bristol Southmead hospital. The sponsors, and their financial adviser RBC, had been running financing options since 2006, and the discussion of a bond option never became advanced enough to lead them to get a rating.

The UK's water sector offers the largest potential issuer base for bonds, though these utilities are structured as whole business securitizations. While they benefit from project-type lender protections, they offer a credit profile that, though regulated, is closer to a corporate.

The traders dabbling in Latin American project bonds have so far had a happier experience. They tended to hedge their illiquid and low-rated bond deals, for issuers such as Peruvian and Panamanian toll roads, either with CDS or by shorting the sovereign bonds of the host country of the projects in question. So far, both sides of this pair have held up adequately, though appetite for such trades has been among the casualties of the crunch.

The US waits for liquidity to return

The US power market had an on-off relationship with project bonds, since in the aftermath of the Enron bankruptcy several bonds backed by tolling agreements with merchant power marketers struggled. Generating portfolios with outstanding bond debt also performed poorly, even when they did not join their owners in bankruptcy. Several active buyers of project bonds were forced to take losses, and some sold their holdings to distressed asset buyers like Beal Bank.

In the years between 2003 and 2008 sponsors began to target the lower ratings and higher leverage on offer in the B loan market. "The biggest thing keeping down bond issuance in those years wasn't the banks – they couldn't compete on terms or pricing – but the B loan buyers," says one managing director with a presence in all three markets.

In this telling, the revival of a B loan market – of which there are tentative signs – spells the end of sponsors, particularly buyers of operational assets, giving the bond market a second look. Barclays Capital has closed both of the B loan deals to clear the market since the crunch – the $290 million high-pressure refinancing of CIT's BRSP portfolio, which closed in June 2009 as CIT tottered on the brink of bankruptcy, and ArcLight's $220 million financing for the acquisition of the Great Point power portfolio from EIF, which has signed but waits on regulatory clearance. It also closed a $205 million high-yield junk-bond refinancing for IFM's North American Energy Alliance portfolio in September 2009.

Oil and gas infrastructure issuers were the earliest and most consistent borrowers to return to the bond market, as they were after Enron's collapse. Several Spectra-sponsored operational pipelines, including the Maritimes and Northeast Pipeline, and the South-East Supply Header pipeline, as well as FPL's Florida Gas Transmission asset, closed bond financings in 2009, although some of this issuance was left over from 2008, and the timing of others was dictated largely by expansion or construction schedules. Pipeline issuers have enjoyed the best pricing, with SESH's five-year issue pricing at less than 5%.

Non-US oil and gas issuers have benefited from the warmer welcome. The Dolphin pipeline, sponsored by Mubadala (51%), Total (24.5%) and Occidental (24.5%), closed a $1.25 billion ten-year bond in July 2009, at 337.5bp over the equivalent treasury, as part of a wider refinancing that included bank, sponsor and ECA debt. The bond priced 100bp lower than the bank debt, including swap costs. This followed on the heels of a well-received $2.23 billion bond issue by RasGas 2&3, sponsored by Qatar Petroleum (70%) and ExxonMobil (30%).

The two bond issues, which both had US sponsors, demonstrated the attraction of robust – and operational – Middle Eastern oil and gas credits to bond investors, of which US buyers accounted for the largest amount of orders. More unusual was the $270 million seven-year junk bond issue that Grupo R used to fund its acquisition of the PetroRig III drill-ship. Combined with a mooted bond issue from Brazilian operator Schahin's Lancer vessel, it has led to speculation that bond deals might take over from bank financings in the offshore sector.

The speculation is probably misplaced. The PetroRig bonds, led by Jefferies, carried a 12% coupon, and were used to supplement bank debt and allow Grupo R to fund the acquisition without selling project equity to outside investors. Bonds work well when issued just before a rig is delivered, for instance when a rig has find a new customer during construction, and bond carry costs make using the instruments on newbuild rigs uneconomical.

Is wind the next level?

From 2009's first stirrings in the oil and gas market, some bookrunners have been tempted to extrapolate a flowering of bond financings for assets in other sectors with greater levels of construction risk. Renewables projects, with short, simple construction phases, are most frequently mentioned. The midstream oil and gas sector suggests investor tolerance for construction risk is low, and that delays to construction on projects such as SESH justify this caution.

Sponsors continue, where possible, to finance either with equity or draws on corporate loans. Those without the financial resources have to turn to project banks. El Paso raised $700 million in equity and convertible debt from Global Infrastructure Partners, and infrastructure fund, and $1.5 billion in debt from ten banks for its Ruby pipeline.

Wind developers are, on the other hand, looking at using more long-dated debt, some of them because their financing requirements are larger than bank lenders alone could digest. Caithness Energy is said to be looking at a bond tranche for the financing of its $1.4 billion $845 million Shepherd's Flat wind farm. Coordinating bond, commercial bank, and Department of Energy-guaranteed tranches is likely to tax sponsor's development expertise. One developer, ArcLight's Terra-Gen, opted to split its Alta Wind project into separate phases, closing a $400 million construction loan for a 150MW first phase with Credit Agricole and Natixis in early March.

Canadian life company Manulife, which owns US power debt and equity investor John Hancock, has carved out a niche providing long-dated single asset project bonds to US and Canadian wind farms. This year it has provided a C$179 million tranche to Invenergy's Raleigh wind farm, and a C$88.5 million 21-year tranche to Boralex' thames River project. It also provided a $168 million long-dated tranche to RES' Hackberry wind project in December 2007, though in the US it shares the market with John Hancock.

Both Invenergy, with a 2009 $318 million holding company issue, and NextEra (formerly FPL), with a $305 million BNP-led 20-year placement for its Mountain Prairie portfolio, have recently completed portfolio private placements, for operational assets. FPL remains the only sponsor to close a public bond issue for a wind portfolio. It closed two deals in the middle of the last decade, for its $465 million National Wind portfolio, and its $480 million American Wind portfolio, each with a concurrent holding company issue.

But sponsors will need to tread carefully to reopen the market. Ratings agencies were sceptical that portfolios would achieve the production levels – and turned out to be right. FPL's portfolio has performed adequately, while the European Breeze transactions have suffered multiple downgrades. Any improvement in agencies' perceptions of construction risk must be set against this harsher view of wind forecasts.

Canada is lifeco country

The most consistent project bond market – at least for greenfield projects – is in Canada. Canada's life insurance companies, and by extension the Canadian banks that control access to these investors, have established a dominant position in Canadian PPP. Since the crunch, only deals with particularly long construction periods, or elements of traffic risk, have been funded by bank debt alone.

The life companies, with copious appetite for single-A bonds, took time to warm up to PPP. They, and the Canadian banks, were unwilling to hold spreads for prolonged bidding periods, and foreign lenders with appetite for interest risk such as Deutsche and ABN Amro took market share from Canadian banks before selling bonds they warehoused following project completion, capturing the construction risk premium in the process. Both have now left the market.

European banks competed in the Canadian market by funding long-dated tranches with wraps – booked, for regulatory reasons, outside Canada – from monoline bond insurers. The bond insurers, together with the most enthusiastic wrapped loan providers, such as Fortis and Depfa, are no longer in business.

Bank lenders now have a limited role funding the position of the costs of Canadian PPP concessions that will be repaid at completion with acceptance payments from the country's provinces. The acceptance payments were introduced to compensate for lower amounts of project debt that projects could support in an era of higher pricing.

Canadian banks are happy to provide these short-term tranches, and continue to enjoy a small advantage in funding costs. In any event, these acceptance payments could shrink, or even disappear, as Canada's provinces begin to reduce their commitments in the face of larger budget deficits. For instance, SNC-Lavalin and Acciona are believed to be using a bond tranche alone for their Stoney Trail SE concession in Alberta, which was due to close as Project Finance went to press. Bond bankers caution, however, that European lenders may yet be able to offer attractive mini-perm loans and market rumour suggests that a 30-year bank bid was recently submitted in Ontario.

Banks still hope that ratings agencies will stop giving PPP project bonds the crucial single-A ratings, pointing out that similar assets in Europe tend to garner triple-B ratings. Sponsors see it differently. "There's room for pricing on PPP bonds to get lower. We're still paying a premium of almost 150bp over similarly-rated corporate debt. And it's a much nicer experience putting together a bond deal than dealing with banks."

Still hazy after all those deals

The traditional limitations of the bond market have not disappeared. Outside the private placement market staggered drawdowns remain rare, and sponsors therefore have to guard against negative carry. The gap between what sponsors must pay in interest and what they would get by reinvesting proceeds in a money market account is wide, and the guaranteed investment contracts that the monolines and AIG offered are a thing of the past.

The growth in the product offerings from the major trust services firms promises to erode some of the advantages that banks claim in dealing with borrowers on routine waivers and stepping in if projects have difficulties. Banks, and until recently monolines, claimed that having one point of contact was important to sponsors, because sponsors would find it difficult to deal with a large and diverse group of bondholders on unwrapped deals. Having trustees with experienced staff and the ability to communicate quickly and rapidly with bondholders might give bonds an advantage over bank debt. To date, however, trustees are preoccupied with unraveling distressed structured finance, and a small number of infrastructure finance, deals.

Prepayment penalties on project bonds remain stiff, and issuers had little ability to pressure institutional lenders to reduce these during the boom years. Banks' swap breakage costs are not as steep, and bank tenors are creeping towards 18 years in wind, because banks view renewables as a core market to be defended at all costs.

Most importantly, there is little evidence that the universe of bond investors is broadening. Even with the imprimatur of a rating, a small, single-digit number of life insurance companies, and a similar number of bond fund managers in the US participate in the project bond market. The cartel is not as pronounced as in Canada, but it shows few signs of opening up. Without a pronounced change in the way banks book their project finance debt, and a corresponding loosening in the way bond investors price risk, the long-awaited capital markets takeover – outside Canada – is still a long way off.