United States: An Update On Covered Bonds

On February 4, 2009, Standard & Poor's ("S&P") issued a proposed revised covered bond rating methodology. On March 11, 2009, Fitch Ratings ("Fitch") followed suit. These proposals more closely link the rating of a covered bond to the rating of the issuer than did prior rating agency standards and increase required overcollateralization levels.

In 2007, development of the nascent U.S. covered bond market was put on hold as the financial crisis unfolded. Regulatory efforts in 2008 to encourage development of the covered bonds market were well received, but issuers were unable to act given the extreme dislocation of the capital and credit markets. Covered bonds have the potential to provide issuers with another alternative as part of a well-diversified liquidity management program and to provide investors with an asset-backed debt instrument that protects against many of the risks recently experienced in the originate-to-sell model. However, to date, federal government financial crisis recovery efforts have focused on restarting the securitization markets.

In January 2009, amendments to the FDIC's Temporary Liquidity Guarantee Program ("TLGP") were proposed that would give banks the option to issue a "quasi-covered bond product." Industry participants questioned the utility of this proposal in promoting the development of a covered bond market. This proposal to have secured FDIC-guaranteed debt, disclosed in January, has not been followed with rulemaking. It is not yet clear if the TLGP will be expanded to include such a product. On March 25, 2009, the first transactions of the Federal Reserve Board's Term Asset-Backed Securities Loan Facility ("TALF") program settled, with monthly transactions expected throughout 2009. Under the TALF, the Federal Reserve Bank of New York provides low-cost loans to purchasers of asset-backed securities. One of the goals of the TALF is to restart the securitization market to encourage consumer and small business lending.

Below, we provide a brief overview of covered bonds and discuss recent actions taken by the credit rating agencies. For additional information on the government's recent actions, covered bonds and the financial crisis generally, please see our Client Alerts and resources at Financial Crisis Legal Updates and News.

Background

Covered bonds are debt obligations with recourse either to the issuing entity or to an affiliated group to which the issuing entity belongs, or both. Upon an issuer default, covered bondholders also have recourse to a pool of collateral, known as the "cover pool." Covered bonds generally are fixed rate bonds with a maturity of no less than one year and no more than 30 years. The bonds are low risk yield-bearing products having long maturities. Unlike securitization, where assets are removed from an issuer's balance sheet, in a covered bond structure, the cover pool remains on the issuer's balance sheet but is 'ring fenced' from the issuer's other assets. The cover pool usually consists of high quality assets, including residential mortgages, public debt, or ship loans. Cash, or cash equivalents, also may serve as cover pool collateral. The cover pool is subject to ongoing eligibility criteria, and cover pool assets must be replaced if the pool fails to meet such criteria. By requiring ongoing maintenance of the cover pool, a covered bond protects investors from asset degradation and ensures ongoing management by the issuer. Typically, the issuer is required to overcollateralize the cover pool.

Many European countries have covered bond legislation establishing a framework for the structure and treatment of covered bond programs. Legislation provides exceptions for covered bonds in the event of bankruptcy. Specifically, in the event of an issuer insolvency, covered bond holders have priority rights over the cover pool assets.

U.S. Funding Alternatives

Historically, U.S. depository institutions have had a number of available alternatives to finance mortgage originations, including selling loans to government-sponsored entities ("GSEs") and borrowing from the Federal Home Loan Banks (the "FHLB").

The GSEs, particularly the Federal National Mortgage Association ("Fannie Mae") and the Federal Home Loan Mortgage Corporation ("Freddie Mac"), facilitated the development of a strong secondary market for mortgages by integrating the mortgage markets with the capital markets. Securitization of residential mortgages, made mainstream by the GSEs, became an affordable and widely-used funding tool for mortgage assets not satisfying the GSE program's origination criteria, i.e., loans that did not conform to GSE standards due to loan size (jumbo loans) or the origination standards (no documentation loans).

Depository institutions also have had access to FHLB borrowing. The FHLB are 12 banks set up under a government charter in the early 1930s to provide support to the housing market by advancing funds to over 8,000 member banks that originate mortgages. The FHLB banks, which benefit from an "AAA" credit rating and GSE status, are able to raise debt at rates only slightly higher than Treasury securities. The FHLB banks advance funds to their member financial institutions at interest rates that are lower than those in the commercial market, particularly on longer-term funds.

An Optimal Time for Covered Bonds

GSEs are facing liquidity crises of their own, and the future of the GSEs seems uncertain. Although FHLB members can borrow at attractive rates, depository institutions seeking such funding must buy equity in the FHLB equal to 5% of their borrowings and post 120-130% over-collateralization. There is no equity buy-in cost for covered bonds, and the recommended over-collateralization is only 105%. With the securitization market closed and other sources of lending either scarce or more expensive than in prior periods, covered bonds may assist in reviving the lending market.

Changes in Rating Agency Methodology

Securitization credit ratings are linked exclusively to the assets in the securitization vehicle and the servicer credit rating. Unsecured debt credit ratings are linked exclusively to the issuer's credit rating. Covered bond credit ratings have been linked primarily to the asset quality and characteristics of the program, and only tangentially to the issuer credit rating. Proposed changes in S&P's and Fitch's covered bond rating methodologies change that.

Standard & Poor's

S&P's proposed changes in covered bond rating methodology have created controversy in the covered bond industry, causing S&P to extend its initial five-week public comment period by an additional two weeks. The key change in the methodology is the creation of an "explicit 'soft link' between certain covered bond ratings and the issuer's ratings."¹ The proposed methodology will lead to a tighter relationship between issuer jurisdiction and default rating, and the covered bond program rating. As many as 60% of global covered bond programs could experience a ratings downgrade under S&P's proposed rating methodology.² The proposed methodology divides covered bond programs into three categories based on the risks described below.

Asset-Liability Mismatch. This factor weighs market value risk posed by liquidity needs that investors face upon an issuer's insolvency. The proposed rating methodology determines the magnitude of this risk by taking into account the maturity mismatch between the assets in the cover pool and the covered bonds and the maximum stressed net liquidity need on a rolling basis based on the amount of covered bonds outstanding.

Depending on the liquidity exposure, a program will be viewed as having minimal or heightened asset-liability mismatch risk. The asset-liability mismatch is minimal when the weighted-average maturity gap is below 18 months and the stressed maximum liquidity need is below 15% of the then outstanding assets. The asset-liability mismatch is heightened when the weighted-average maturity gap is above 18 months, or the stressed maximum liquidity need is above 15% of the then outstanding assets, or the amount of assets that can be pledged to a lender under a repurchase facility is less than the stressed maximum liquidity need after applying the applicable 'haircut.'

Segmenting Jurisdictions. The proposed rating methodology considers the jurisdiction in which the covered bond is issued. S&P considers whether covered bond legislation exists, the systemic importance of covered bonds for that market, and the historical track record for covered bonds in that market. Based on these considerations, S&P rates programs issued in Germany, France, and Denmark the highest. Programs coming from German or French banks that issue covered bonds out of Ireland and Luxembourg are also viewed favorably. The current proposal assigns the lowest rating to all remaining countries.

Market Value of the Assets. This factor considers assumptions relating to refinancing costs depending on the exit strategy of the covered bond program. These assumptions are based on asset-liability mismatches and the alternatives available to programs needing to access third-party liquidity, sell assets, or both. For programs that are able to borrow from third-party liquidity sources, such as the European Central Bank ("ECB") or National Central Banks ("NCB"), the liquidity analysis continues to assume that the refinancing costs are based on central bank refinancing rates and banking rates. However, an additional test will be performed to verify that the available assets will at all times be sufficient to meet the stressed liquidity shortfalls.

For programs that rely to a greater degree on the option to sell assets to meet liquidity needs, collateral levels will need to be increased to account for the risk of market value fluctuations or increased refinancing costs) as the result of increased stresses in refinancing costs. The new framework proposes to calculate this risk as the lower of: the estimated proceeds from selling the assets based on the maturity of the covered bonds and stressed refinancing costs, or the proceeds from a tailor-made residential mortgage-backed security after applying ECB or NCB haircuts to allow for an alternative funding source. Additional calculations are performed for programs where the liquidity needs are greater than the amount of available liquid assets.

Based on the above factors, each covered bond program would be assigned a category. There are three categories, and the potential covered bond ratings uplift above the issuer's credit rating is different for each category.

Category One. All programs in this category could theoretically be assigned a rating of "AAA" regardless of the issuer's credit rating. A category one program would either: include no mismatch between the maturities of the assets and the liabilities, as in pass-through structures, or provide a structural mitigant to address this risk, such as a committed liquidity facility from an appropriately rated counterparty.

Category Two. The maximum uplift from the issuer's credit rating to the covered bond rating for covered bond programs in this category would be six notches above the issuer's long-term credit rating and up to a maximum of five notches with respect to an "AAA" covered bond rating. Covered bond programs in this category could theoretically be rated "AAA" provided the long-term rating on the issuing bank is at least "A", or "A-1" if it has no long-term rating. The covered bond programs placed in this category should be able to pledge their cover pool assets through repurchase facilities with the NCBs or the ECB and should have a minimal asset-liability mismatch.

Category two includes those covered bond programs from banks in the historically active covered bond issuing nations, such as Germany, France, and Denmark, as well as covered bonds issued by their subsidiaries in Luxembourg and Ireland. These countries all have a detailed framework for covered bonds. Although Sweden has a long track record for the covered bond market, it is not included in this category due to the inability of covered bond programs to access third-party liquidity under the Swedish legislative framework.

For programs in category two, the new framework proposes to calculate the overcollateralization based on a stressed liquidity need. Assumptions for borrowing costs from the ECB or NCB will be applied.

Category Three. Under the proposed rating methodology, the maximum uplift from the issuer's credit rating to the covered bond rating for covered bond programs in this category would be four notches above the issuer's longterm credit rating and up to a maximum of three notches with respect to an "AAA" covered bond rating. Therefore, covered bond programs in this category could be rated "AAA" provided the rating on the issuing bank is at least "AA-", or "A-1+" if it has no long-term rating. Programs that have a heightened asset-liability mismatch will be put in category three. It is possible that programs from those countries classified in category two would be placed in this category if the asset-liability mismatch is greater than the guidelines for minimal liquidity exposure. Category three will include those programs that have no access to the NCBs and ECB, or programs governed by a covered bond legislation regime that does not permit liquidity borrowings, such as Sweden, or programs structured so that liquidity could only be accessed by selling the assets. Programs may be placed within this category where the uplift from the issuer's credit rating could be further restricted by the type of assets within the program or the conditions for asset sale.

Below is a chart provided by S&P illustrating the new categories, and the factors considered in determining the relevant category for a covered bond program.³

Fitch Ratings

The current process that Fitch uses for rating covered bond programs is a three-step analysis. The factors are similar to those used by S&P, with a focus on liquidity gaps and overcollateralization. Although jurisdiction plays a factor in Fitch's rating methodology, Fitch does not weigh jurisdiction in the same way as S&P. The proposed rating methodology will lead to a tighter relationship between the issuer's default rating (the "IDR") and its covered bond program rating. Fitch concludes that if all issuers increase overcollateralization in order to obtain the highest achievable rating with no other changes made to the covered bond program, only 5% of public sector covered bond ratings would experience a rating downgrade, and only 10% of mortgage covered bond ratings would experience a rating downgrade.⁴

Probability of Issuer Default. Step one determines the maximum achievable rating for a covered bond program based principally on the probability of issuer default with consideration given to recovery expectations. Fitch determines the IDR and then determines the program's discontinuity factor (the "D-Factor") (a measure of the likelihood of an interruption of payment on the covered bonds as a direct result of issuer default). Factors considered in determining the D-Factor include segregation of the covered pool, liquidity gap solutions available upon default, alternative cover pool management plans, and the role of a supervisory authority, if any. By combining the IDR with the D-Factor, Fitch determines the maximum rating for the covered bond program, provided that overcollateralization between the cover assets and the covered bonds is sufficient to sustain the corresponding stress scenario.

The proposed rating methodology will change this step by increasing the weighting of liquidity gap scores in determining the D-Factor, and reducing in parallel the weighting of asset segregation. This change reflects Fitch's concern that limited asset liquidity will result in less certainty regarding whether liquidity gaps can be addressed within the timeframe allowed by the program's structure.

Overcollateralization Stress Testing. Fitch performs overcollateralization stress testing by simulating the wind down of a cover pool following an issuer default, and compares the stressed asset cash flows to payments due on covered bonds while factoring in credit losses, costs associated with bridging maturity mismatches, adverse interest and currency rates, and third party wind down expenses. The proposed rating methodology focuses on an assessment of the liquidity gap risk when a program must wind down within a stressed liquidation timeframe in an illiquid market. Fitch assumes that if a stressed sale occurs, the stressed sale proceeds will be less than expected before the credit crisis, which will require higher overcollateralization to be consistent with a given rating scenario. In most cases, Fitch expects that overcollateralization may need to be increased by 10%.

Calculating the Recovery Uplift. Current rating methodology also reflects stressed recoveries from the cover pool in the event the covered bonds default. The recovery prospects are assessed, at the time of the supposed default, by calculating the stressed net present value of the cash flows expected from the cover pool and comparing them with the net present value of the privileged liabilities (outstanding covered bonds and, if any, privileged swap agreements).

The assets' stressed net present value is calculated based on the following principles: scheduled cash flows are modified to reflect stressed defaults in the cover pool and recovery patterns on defaulted assets; they are further reduced by the worst-case effects of interest rate and foreign exchange fluctuations; and then discounted based on a stressed interest rate. Further deductions are made to reflect haircut suffered in large volume liquidations.

Jurisdictional Considerations. Fitch does not consider issuer jurisdiction as an independent factor in determining ratings for a covered bond program, but it does consider jurisdiction when reviewing asset liquidity. Fitch also may give a limited amount of credit to legislative covered bond programs because of the likelihood that the applicable regulator will exercise pressure on peers to provide liquidity in the event of an issuer default, or it will intervene directly through support measures.

The TLGP and TALF Programs

TLGP Program

On March 17, 2009, staff of the FDIC met with market participants to discuss proposed amendments to the TLGP. Proposals under consideration included an FDIC guarantee for "quasi-covered bond products." Eligible debt would be collateralized by a pool consisting of mortgages and "AAA" tranches of consumer debt, with half of all loans in the pool required to be issued within six months prior to a guarantee. Requests to include other types of collateral will be considered on a case-by-case basis. The collateral under the program is held for the benefit of the FDIC, making the bond more similar to government-guaranteed debt than an actual covered bond. In a covered bond program, collateral is held for the benefit of covered bond holders. The FDIC did not take action at the meeting to add the collateralized FDIC-guaranteed debt to the TLGP, although it is still considering doing so. Another open question is, if the FDIC does add this product to the TLGP, whether or not it will create a dual security interest in the collateral for the benefit of covered bond holders.

Given recent announcements of limited expansion of the TLGP and an effort to transition issuers away from reliance on the TLGP it is unclear whether the FDIC will expand the program to cover secured debt.

TALF Program

In November 2008, the Federal Reserve Board announced the TALF and authorized the Federal Reserve Bank of New York to provide up to $1 trillion of low-cost loans to purchasers of asset-backed securities. One of the goals of the TALF is to restart the securitization market to encourage consumer and small business lending. Initially, the TALF included asset-backed securities secured with credit card receivables, loans and leases, student loans and loans guaranteed by the Small Business Administration. In March 2009, the program was extended to include asset-backed securities secured by mortgage servicing advances, auto loans or leases relating to business equipment, leases of vehicle fleets and floorplan loans other than auto floorplan.

As noted, securitization provides off balance sheet financing of primarily discrete and unmanaged pools of assets. During the current financial crisis the off balance sheet securitization model has been blamed for disincentivizing originators from exercising due care during the underwriting process. Many market observers believe that if originators had retained a greater long-term interest in the performance of underwritten assets, particularly in the mortgage industry, more stringent underwriting procedures would have been applied and the scope of the mortgage meltdown would have been dramatically reduced. The covered bond structure directly addresses these concerns by retaining with the originator the long-term performance risk of assets within the cover pool. However, given the recent importance of securitization to the consumer lending market, federal regulators have found it necessary to focus their efforts on restarting securitization. Longer term, the benefits of covered bonds to investors, including the retained risk by the originator, may lend support to further development of the market.

What Does this Mean for the U.S. Covered Bond Market?

If the proposed rating agency methodologies are adopted, covered bond programs may become more expensive, especially based on the S&P rating criteria that automatically puts U.S. covered bond programs in its lowest category limiting the number of credit rating notches a covered bond could achieve above the issuer's own credit rating.

On September 25, 2008, the FDIC's treatment of covered bond holders in the event of an issuer insolvency was tested, and the outcome was consistent with the published FDIC guidance. Washington Mutual was closed by the Office of Thrift Supervision, and the FDIC was named receiver. The covered bond liabilities of Washington Mutual, the failed sponsor, were immediately transferred by the FDIC to a successor bank, JPMorgan Chase Bank, N.A. As a result of the acquisition, JPMorgan Chase Bank, N.A. effectively replaced Washington Mutual as the sponsor of the covered bond program. This swift action should provide comfort to investors that the FDIC will respect the contractual arrangements under a covered bond program in the event of an issuer insolvency.

Although there is no U.S. covered bond regulation and alternative funding sources exist, diversification, transparency and investor interest may ultimately drive the covered bond market.

Footnotes

^{1 Standard & Poor's Request for Comment: Covered Bonds Rating Methodology (February 4, 2009), available at http://www2.standardandpoors.com/spf/pdf/japanArticles/20090204CoveredBond.RFC.pdf .}

^{2 Id.}

^{3 Id.}

^{4 Fitch Ratings, Exposure Draft on Assessment of Liquidity Risks in Covered Bonds (March 11, 2009), available at www.fitchratings.com.}

Because of the generality of this update, the information provided herein may not be applicable in all situations and should not be acted upon without specific legal advice based on particular situations.

© Morrison & Foerster LLP. All rights reserved