A blunter approach to credit risk in a new set of global bank regulations will likely see an increase in the cost to banks of holding long-term project debt. Although the exact form the Basel III regulations will take is still unknown, three broad changes will likely make it more difficult for banks to make long-term debt commitments to infrastructure projects.

The three legs of Basel III

Firstly, banks will need greater tier one capital. Basel III will considerably tighten the definition of what constitutes tier one capital. This will be calculated as a percentage of a risk-adjusted asset base that will include all corporate loans, project financings and some ECA loans. The stress tests in the UK and US used a deterioration of 4% of tier one capital to test bank solvency, so banks will probably be asked to hold roughly 5-8% to provide them with a buffer against a deterioration in their asset base. Some banks have pre-empted the regulations with large capital raisings, particularly in the US, but several European and Japanese banks still need to raise substantial amounts of capital or greatly reduce their balance sheets to meet a more stringent tier one target.

Secondly, Basel III will concentrate on banks' short-term liquidity position to make banks more resilient to the closure of the short-term money markets. The new provisions introduce a liquidity coverage ratio, calculated as their stock of high quality liquid assets divided by their net cash outflows over a 30-day time period. The ratio would measure a bank's ability to convert assets into cash within a 30-day window, and would need to be a minimum of 100%. Project assets do not fare well, since banks cannot assume that there will be a liquid secondary market even for non-distressed project assets. A more stringent liquidity test will see banks favour more tradable assets such as government bonds.

And thirdly, and perhaps the most crucial point, banks will be asked to better match the tenors of their liabilities with their assets using a net stable funding ratio. This would be calculated as available amount of stable funding divided by the required amount of stable funding, and would need to be a minimum of 100%. The asset-liability mismatch in project and public lending brought down Depfa and its parent Hypo Real Estate and brought Dexia to the brink. The provision will promote increased use of medium- and longer-term funding. The numerator of this ratio is roughly equivalent to core funding and the denominator the weighted sum of the asset side of the balance sheet, weighted to the following schedule (courtesy of Credit Suisse):

• 0% weighting – cash, securities with a remaining maturity <1yr, interbank loans with a maturity <1yr, securities held with an offsetting reverse repo
• 5% weighting – unpledged high-quality liquid securities (similar definition to central bank eligible collateral)
• 20% weighting – corporate bonds and covered bonds with a proven record of liquidity (ie, no major increases in repo haircut in the last 10 years)
• 50% weighting – other corporate bonds, gold, equities, loans to corporates with a maturity less than one year
• 85% weighting – retail loans with a maturity less than one year
• 100% weighting – all loans with maturity > one year, all other assets.

Uncovered project loans would not flatter this ratio, but government-covered loans such as ECA facilities and loans covered by France's Dailly provision should be weighted 20% or less.

Basel II was focused on ratings, which have been shown to be cyclical, and has been criticised for amplifying the effects of the economic cycle, because lower ratings require greater capital provision. Basel III will tighten the rules that allowed banks to allocate capital according to their own credit ratings under the advanced internal ratings approach and build in an absolute margin of safety with an emphasis on stress testing and value at risk calculations. Quite how the Basel III risk-adjusted internal ratings model will differ from the Basel II regime is still unclear.

Whatever the outcome, there are likely to be wide divergences between national financial services authorities, so harmonisation among international banks could take ten years or more.

Coping with Basel becomes harder

In a more capital-constrained environment, project loans will increasingly compete for balance sheet capacity with the likes of shorter-term corporate financing, acquisition, leverage and commodity and trade financing. Although tightly-covenanted project financings have experienced low loss given default levels and usually have a low probability of default, project loans could fall foul of the new regulations, which try to ensure short-term liquidity by matching assets with liabilities within banks.

Before the credit crunch, project banks could free up regulatory capital using synthetic collateralised debt obligations that shifted credit risk from their balance sheets. This is now much more difficult because of the collapse of both the monolines and investors' appetite for CDOs, and because ratings agencies are more conservative. The change in agencies' methodologies has meant that that CDOs have become much less attractive, because they will require a far larger first loss tranche, which pays a higher margin but consumes more regulatory capital, to support the ratings of the more senior tranches.

Compounding this reduced efficiency is a widening in spreads. Many top-tier project banks have tens of millions in loans with margins of below 100bp and tenors of beyond 20 years sitting on their books. It would be impossible to sell these loans in the secondary market without offering a big discount and a CDO would be loss making because the super senior tranches of those deals offer spreads of around 100bp. Most project banks – notably Lloyds and RBS, which are part-owned by the UK government – are carrying this debt at par even though it is priced at below their long-term funding costs. In weighing the merits of project financing versus other types of banking activity, the paltry return of boom-year project debt weighs heavily.

So, while CDOs do not, and may not ever, offer the efficiency banks desire, the requirements of Basel III and the agencies may force them to use this market anyway. Unlike the project CDOs of the boom years, some banks such as Santander and SocGen are passing through regulatory capital to pension funds as private placements. SocGen closed an open-ended Eu940 million first-loss synthetic credit default swap, Comsos 1, in August 2008 with a pension fund. And Santander closed a Eu300 million managed synthetic CBO with two pension funds in late 2009.

Both deals, along with Lloyds' Eu1 billion Blue Exeter synthetic CLO, were done on a bilateral basis, with investors asked to take a view on the bank's management team, its risk assessment practices, and its origination capabilities. SG avoided the inefficiency of tranching by adopting a flat structure, hiving off the first loss risk using a credit default swap, whereas Santander used a standard tranched synthetic structure but adopted what is believed to be the first for any project bank – a managed structure where it can withdraw and place assets subject to certain criteria. Santander continues to hold 50% of any loans contributed to the CDO to align its interests with those of the investor, and managed to place the entire first loss tranche.

Using banks as a conduit to the capital markets makes sense. Banks, unlike most pension funds, have the personnel to structure, originate and credit approve bespoke project financings and the pension funds and life companies are natural homes for long term debt.

One drawback giving bond investors access to the loan books of project banks is that credit committee approvals and rating agency methodology do not perfectly align. Bond buyers reference similar credits, looking for pricing differentials and are largely reliant on ratings. Rating agencies have largely avoided criticism for their single-asset ratings but came unstuck when applying actuarial approaches to pools of assets. Having the capital markets investing directly into projects is a cleaner way of substituting or recycling bank commitments to projects. As one debt capital markets specialist says, "Banks tend to get booted around by clients much more than rating agencies and banks are notoriously bad at pricing risk."

Are the there primary bond market options?

Basel III could be the jolt that sends more projects, by whatever route, to the capital markets. However, the clarion call of capital markets has been sounded before and the associated bond activity has failed to materialize (see feature, p.45 this issue). A big hurdle for the primary issuance of project bonds is the reluctance of bond investors to dedicate the resources to the credit process for a bespoke project financing. "It is a chicken and egg scenario," says an adviser. "Bond buyers want to see some dealflow before they invest in credit teams, but need credit teams to invest in bespoke project financings. Nevertheless, I know a number of large bond buyers that are investing in credit teams."

The need for specialist credit teams at pension finds and life companies is much more acute in the project finance market now that the monoline option has disappeared. The only viable monoline left is Assured Guaranty, and while it may wrap the occasional private placement it will quickly reach its exposure limits and so will offer scant liquidity. Since monolines issued an irrevocable and unconditional guarantee to lenders they effectively acted as an administrative agent, monitoring the project and stepping in when things turned sour. This agency function post-monoline will likely be replaced by a combination of greater investor knowledge and an enhanced role for banks acting as agents and trustees.

The benefit of a monoline over a third party agent was that the monolines' interests were directly aligned with bondholders. Hadrian's Wall Capital, a UK-based debt fund founded by Marc Bajer, formerly a managing director at Assured Guaranty, has come up with a new solution to this agency issue. The fund has teamed up with insurer Aviva to fund an initial £500 million vehicle, rising to £1 billion split equally between Euros and sterling to invest in infrastructure projects. The vehicle will provide mezzanine financing to infrastructure projects alongside senior debt bond issuance. The mezzanine lender would effectively act as a super-agent for the bondholders, similar to the traditional role of the monoline. But the mezzanine lender has less to lose in the event of default, and so control passes back to the bondholders in a stress scenario.

While negative carry, involving paying interest on bond proceeds before they are required for construction, and the prepayment penalties attached to refinancing bonds will continue to disadvantage a bond issuance in comparison with bank debt, sponsors and their advisers are moving toward greater real-time disclosure, which makes it easier to get waivers from bondholders if, say, construction goes over budget or schedule. Ideally an open web-based platform would provide a summary of the ongoing performance of a project, including whether it is meeting its coverage ratio covenants and fulfilling its rating criteria. Such transparency would help the trading of paper in the secondary market, and provide more liquidity for new issues.

Mitigating construction and refinancing issues can be achieved with sequential bank, then bond, financings. Bonds have been best suited to projects that are large enough to be considered corporate issuers in their own right. Just as UK utilities visit the capital markets using tightly-structured whole business securitizations, Qatar Petroleum has successfully employed a multi-sourced funding strategy for its stable of large LNG projects and has been able to tap the capital markets at opportune moments.

On a single-asset greenfield basis bond investors need convincing. Without monolines, are there enough buyers of long dated BBB to AA-rated paper for greenfield projects? The likes of Mubadala hope so. Mubadala, the Abu Dhabi infrastructure investment vehicle, has a slate of 50 projects worth some $200 billion in the pipeline, many of which will be greenfield PPPs. To overcome investor inertia it is considering upfront payments to bond investors to entice them into projects. At the smaller end of the spectrum the European solar developer SunRay is hoping to tap the capital markets for a Eu200 million issue in the coming months. The issue will be one of the first ever publicly-rated bonds for a solar PV asset, as well as being the first project bond in Italy to feature construction risk.

A wholesale shift to project bonds in the short term is unlikely, and the Basel III fears for projects may be overdone, as most banks are moving to build in an equity buffer to sweeten market perceptions in advance of new regulation. Basel III changes will come into force in 2012 and will be introduced gradually. Meanwhile, the bank market has already returned to tenors of over 20 years, pricing is coming down and there are clear signs that the syndication market is returning. For example, at least one bank in the recent Eu3.9 billion Nord Stream gas pipeline has approached the borrower for consent to sell down in the secondary market.

And the Bristol Southmead hospital PPP in the UK was financed with 30-year bank debt, though RBS, Lloyds and the European Investment Bank provided most of it, despite a shortage of long-dated sterling investment-grade paper. In late March, Wales & West Utilities Finance closed a £515 million bond offering with the £300 million, A- (S&P/Fitch), 20-year tranche attracting orders of £1.25 billion and priced at 130bp over gilts, which narrowed slightly in immediate trading.

According to one market participant the Southmead deal, which had a debt to equity split of 85/15 in preparation of a BBB rating, could have been placed in the capital markets but for Lloyds, RBS and the EIB. The UK market was also close to a bank-bond solution for the Gatwick Airport acquisition. The losing Citi Infrastructure Investors-led consortium had planned to use multi-sourced financing.

Project banks with healthy balance sheets, such as Credit Agricole, Santander, BNP Paribas, and Societe Generale, have taken market share. They are highly motivated to book project assets because senior management believes that project finance debt is remunerative in the long term and will offer risk-adjusted returns in excess of their long-term funding costs. Basis risk is also diminished at higher margins, that is, the risk that future funding costs could swamp the return on previously booked debt has fallen.

The traditional avenues of managing project debt – syndication and CDOs – are remerging. But while in the past making a long-term commitment of balance capacity sheet was a business decision, in the future it will be guided more by regulation. Bonds cannot compete with relationship deals, personified most vividly by Total-Aramco's Jubail refinery deal, which is priced below 200bp despite refinery margin risk. But Basel III should make risk-priced bank debt more expensive and bonds more attractive, particularly to sponsors that do not want to deal with increases in margins or cash sweeps at years seven to 10. Perhaps a more probable driver for the capital markets is not bank regulation but the sheer size and scope of the infrastructure pipeline – particularly as international banks hit country limits in the GCC, for example – or more likely still, a combination of the two.