| Search |
03 October 2007
Banks which arrange and participate in project finance debt are building significant risk concentrations around this asset class. As margins fall and maturities extend, it is expected that there will be fewer participant banks active in the syndicated loan market. Mark Gordon, Head of Securitisation, Structuring and Placement at SMBC Europe, examines how project finance debt arrangers have turned to the securitisation market to manage their loan portfolios more efficiently
Global demand for infrastructure and the project financing that supports it is at an all time high, with investors financing US$129bn worth of infrastructure projects in H1 2007, US$114bn in H2 2006, US$111bn in H1 2006 and US$104bn and US$66bn in the second and first halves of 2005 (these figures do not include so-called 'Alternative Investments')1.
However the knock-on effects of the US sub-prime crisis have adversely impacted the wider securitisation market. Issuers, arrangers and rating agencies continue to structure project finance securitisations in anticipation of the market returning gradually over the next few months.
Issuer Motivations
For the banks that have issued synthetic securitisations of portfolios of project finance loans and bonds, the prime motivation to date has been regulatory capital release. Under Basle I, drawn project finance loans are risk weighted at 100%. Under Basel II, these loans (which are a category of Specialised Lending) are risk weighted at 70% at best under Foundation IRB. Under Advanced IRB, this risk weighting may be reduced to a much lower level. Most of the structures we have seen are Basel I compliant and incorporate call features on the super senior swap, reflecting the low risk weighting for retained super senior under the new capital accord. The release of regulatory and economic capital around a specific portfolio improves both return on equity and return on risk asset.
These deals are generally replenishable and as such are used as ongoing portfolio management tools. During the pre-agreed replenishment period, as loans amortize, prepay or cancel, the originating bank may replenish with new loans. These new loans must meet the replenishment criteria set out in the offering circular. The replenishment criteria are set to protect the interests of the note investors.
A secondary motivation for issuing banks is the arbitrage between the margin income on the loan portfolio and the cost of buying protection via the securitisation. Despite falling margins in the loan market, the arbitrage case is quite compelling given the alternatives i.e asset sales or monoline wraps, both of which can have a negative impact on a lending bank's relationship with project borrowers/sponsors. The arbitrage is improved by the expected maturity mismatch between the underlying assets and the credit-linked notes.
For an originator to achieve full regulatory capital release from the securitisation, the legal maturity of the hedge must equal or exceed the legal maturity of the underlying assets. The final legal maturity of SMBCE's SMART PFI 2007 transaction is March 2044. This reflects the maturity of the longest dated asset in the reference portfolio. As you can imagine, ABS investors have little interest in 37 year PFI risk.
Therefore synthetic securitisations are normally structured to incorporate a call feature at the originator's option. In the case of SMART PFI 2007 the call is at year 7. The note investors are comfortable that the originator will call the deal at this time, as the securitisation starts to become less efficient once the replenishment period has ended. This is due to the effect of scheduled amortisation and prepayments on the reference portfolio when set against certain fixed costs of securitisation.
Typical Structures / Deal History
Between 1999 and 2001 there were a small number of cash CLOs of project finance loans. Since 2004, there have been five bank balance sheet project finance synthetic CLOs.
Originating banks tend to prefer synthetic structures over cash structures (where a portfolio of loans is transferred by means of a true sale to an SPV). The synthetic transfers part of the risk associated with a portfolio of loans, but not the loans themselves. As such the originating bank continues to service the loans in the normal way and preserves it relationship with the borrower.
As far as the project borrowers and sponsors are concerned, the fact that its loan is included in the reference pool of a synthetic securitisation does not change the operation of that loan in any way or under any circumstances. Given the additional liquidity that securitisation brings to the project finance market and the effect it can have on margins (this is most notable in the UK PFI market) most project borrowers and sponsors are active supporters of the synthetic securitisation process.
Given the long term nature of project finance debt, cash deals do provide a funding advantage. We understand that at least one project finance cash deal is currently at the structuring stage, however we expect the market to remain predominantly synthetic for the foreseeable future.
A summary of the main features of the synthetic deals is given in Figure 1
You will note from Figure 1 that four out of the five deals closed since November 2004 have involved the intermediation of KfW. KfW benefits from a guarantee from the Federal Republic of Germany and as such is zero risk weighted under Basel I and II. KfW will intermediate only on certain asset classes. PFI/PPP loans and bonds originated by European banks are one such asset class (the others being SMEs and certain property related assets). The fact that banks can achieve a zero risk weighting on the whole reference portfolio above the first loss/threshold amount provides a significant advantage in terms of risk asset release. This is one of the main reasons that PFI/PPP has featured so heavily in recent issuance. A typical KfW intermediated Basle I structure is given in Figure 2
In the example depicted above, the asset originating bank (the "originator") buys credit default swap protection from KfW for realised losses that occur in the reference portfolio which exceed the threshold amount. KfW in turn buys protection from a super senior investor (typically an unfunded CDS from a AAA rated monoline). The SPV (the "Issuer") issues rated notes, linked to the performance of the reference portfolio. The Issuer uses the proceeds of the notes to purchase KfW certificates ("Collateral") which are credit linked to the reference loans. KfW is then fully hedged via the super senior swap and the cash collateral from the sale of the credit linked certificates of indebtedness to the SPV.
We understand the rating agencies are currently working on six new project finance securitisations and that a number of these are not PFI/PPP related.
For example, SMBCE will securitise a portfolio of project finance loans totalling $1bn that it has originated in the 6 countries of the GCC (Bahrain, Kuwait, Oman, Qatar, Saudi Arabia and UAE). These assets will be representative of the sort of project finance that is prevalent in the region, i.e. LNG, petrochemical, refinery, power/water, etc. This will be the first securitisation of such assets and one of the very few non-property related securitisations of Middle East assets. This type of portfolio would not be eligible for KfW intermediation. As such, and as SMBCE is now operating under Basle II, the structure of this and other non intermediated transactions is as per Figure 3.
You will note that in this example the originator retains the first loss and also the super senior portion of the capital structure. This reflects a Basle II scenario where banks can risk weight retained super senior at 6/7 per cent (depending on regulatory jurisdiction).
Rating Agency Approach
All three major rating agencies have invested considerable resources in understanding project finance; they have recruited heavily from the banking sector and developed well published rating methodologies around project finance. That said, significant differences of opinion can exist between agencies with regard to the credit quality of a particular loan.
Typically, each project within the reference pool will be separately rated on a "shadow" (non-public) basis. Each asset will be assigned a recovery rate and a correlation matrix (if one loan defaults, the probability that other loans in he pool will also default) will be established for the pool. In general terms, the recovery assumptions around project finance are relatively high. The pools that have been securitised in recent years have been predominantly investment grade and correlation has generally been assessed as low.
S&P published its "Project Finance Default and Recovery Study" in December 20063.The report highlights key findings of the aggregated data experience of Standard & Poor's analysis of the historical project finance loan default and recovery rates. 32 banks participated in the study. The data comprises 4,029 projects from the period January 1st, 1990 to December 31st 2005. The data showed that project finance loans had a ten year cumulative average historical default rate of 11.34 per cent. The average recovery rate based on widely defined defaults is 72.1%. This recovery rate varies from region to region. The combined effect of strong ratings, high recovery and low correlation assumptions means that from an originators perspective, project finance debt, in general terms, lends itself to securitisation.
Practical Issues
Of course, if the rating agencies are to analyse each project separately, they require detailed information on each project. To complete the shadow rating process, the rating agencies require the level of information that would be released to a potential participant in a syndicated loan, e.g. project legal bible, due diligence reports, financial close model and depending upon the status of the project, updated project information e.g. construction and compliance certificates, construction or operation reports, updated financial model, waiver requests, etc.
Most project finance loan facility agreements prohibit the disclosure of such information to third parties. As such, we at SMBCE seek specific consent to release information to the rating agencies from each project company and where necessary, the project sponsors.
The second category of information that will be released around a securitisation is that which is released in investors. The majority of this information relates to the characteristics of the pool as a whole rather than specific data on each project e.g. portfolio breakdown by sector, geography and shadow rating, average DSCR/LLCR etc.
As a large number of project finance borrowers and sponsors are unfamiliar with securitisation and/or the rating agencies, it is worth investing some time to get all parties comfortable with the process. From the perspective of the sponsors, synthetic securitisation is preferable to a sale of the loan as it preserves the lending relationship with a bank and keeps bank syndicate size at a minimum. It also introduces a new pocket of liquidity to the project finance market. In our experience, those institutions which purchase the CLNs that come out of project finance securitisations are generally non-bank investors.
In the Middle East alone, the predicted requirement for project financing over the next 3 to 5 years is considered to be far in excess of what can be absorbed by the bank market. It is essential for borrowers and lenders alike, that originating banks develop the synthetic risk market to allow them to efficiently manage their portfolios. With regard to the PFI sector in the UK, securitisation is frequently credited with allowing certain debt providers to lend at lower margins than would have otherwise been possible.
It is worth noting that the existence of a small number of project finance securitisation programmes can not in itself cause a reduction in debt pricing across the board. The main market for the distribution of project debt is still the syndicated loan market. As such, margins can only fall to a level which is acceptable to the participant banks.
Other issues to consider from an originators perspective are the not insignificant up front and on going costs associated with securitisation e.g. legal fees, rating agency fees, trustees, accountants, arrangement and distribution fees etc. The upfront component of these costs can usually be amortised over the expected life of the deal.
Given the cost component and the requirement for a certain level of diversity in the reference pool, a portfolio of a certain type and size is necessary for the securitisation to be efficient. You will note from Figure 3 that the minimum amounts are around £400m and 24 assets. In addition to those assets included in the original reference pool it is prudent to hold an amount of eligible assets for replenishment purposes.
For these reasons a number of banks have attempted to pool portfolios and securitise on a joint basis. Given structural and regulatory issues this is very difficult to achieve.
Future
World-wide demand for infrastructure projects is growing at a tremendous rate. Our expectation is that the number of active arrangers and participants in the project finance debt market will decrease due to the effects of Basle II, margin reduction and longer tenors on the underlying loans.
Those banks which remain will need to manage their risk synthetically and recycle capital through securitisation. Investors are drawn to this asset class by the strong cash flows, low defaults and high recoveries of the underlying assets. Market permitting, we expect project finance to become an established asset class within the securitisation market within the next 12 months.
No comments:
Post a Comment