home‎ > ‎

Institutional Risk Analytics: Covered Bonds

Covered Bonds and the Reform of Structured Finance

March 31, 2010



On March 18, 2010, Representative Scott Garrett (R-N.J.) introduced the United States Covered Bond Act of 2010 (the “Act”).  His goal was to create an alternative to securitization for financing mortgages and other types of consumer debt.


What are Covered Bonds


Covered bonds are the original structured finance security.  They share several features with all structured finance securities.  First, covered bonds use the cash-flow from a pool of assets to make the bond’s principal and interest payments.  Second, they give the investors in the bond a perfected security interest in the pool of assets that isolates the assets and insulates the pool should the issuer becoming insolvent.  Third, they use over-collateralization in the asset pool as a credit enhancement.  These features are where the similarities with other structured finance securities stop.


Historically, covered bonds have been fixed rate instruments issued with maturities ranging from two (2) to ten (10) years.  They have been structured with periodic interest payments and the repayment of principal at maturity.


By design, covered bonds have a recourse requirement.  The issuer retains ownership of the assets in the pool.  During the life of the bond, if the collateral does not perform as expected whether from prepayments or defaults, the issuer remains on the hook to replace all “non-performing” assets in the collateral pool with performing assets or other acceptable collateral.  For some covered bonds, if the issuer becomes insolvent, an investment contract covers any gap between the cash flow on the pool assets and payments on the covered bonds.   


Unlike securitizations, there is no ambiguity about the loans remaining on the issuer’s balance sheet and the need for the issuer to retain capital to support them. 


Experience with Covered Bonds


Covered bonds originated several decades ago in Europe.  The European experience has been mixed.  As would be expected given its call on the issuer for replacement of non-performing assets, there have been very few, if any, defaults.  Unfortunately, this same structural reliance on the issuer to replace non-performing assets results in covered bonds trading based on the perceived solvency of the issuer.  This reliance also leads to the market freezing, except for purchases by the European Central Bank, as it did from late 2007 through early 2009 when the solvency of the issuers was called into question.


The market for covered bonds has not taken off in the United States.  Rep. Garrett tries to address this through the creation of a regulatory structure.  The Act that he proposed creates a covered bond regulator in the U.S. Treasury.  This regulator has numerous responsibilities including setting over-collateralization levels for specific asset types, maintaining a registry of covered bond programs with information on individual outstanding issues and creating a periodic asset-coverage test to be disclosed to market participants showing if the pool backing the bond exceeds a minimum level of over-collateralization.  The Act also subjects covered bonds to securities regulations and requires issuers to submit to the trustee on at least a monthly basis a schedule showing all assets pledged.


Obstacles to Market Development and Solution


While Rep. Garrett’s proposals are helpful, they don’t go far enough to address the obstacles preventing the widespread use of covered bonds.  From an issuer’s perspective, one obstacle is the ongoing retention of credit and prepayment risk without receiving any offsetting benefit in the form of liquidity in times of financial stress.  From the investor’s perspective, one obstacle is an almost total lack of information about the performance of the assets in the pool.  Without an ability to assess the risk of the assets in the pool, the investor cannot know how dependent the interest and principal payments are on the issuer’s replacing non-performing assets.  Therefore, the investor cannot give the issuer any credit for the assets in the pool and the issuer’s access to funds reflects the issuer’s perceived solvency.


As currently proposed, the Act solves these problems for the U.S. covered bond market by basing the market on enhanced moral hazard ala Fannie Mae and Freddie Mac.  Under this solution, the attractiveness of covered bonds will be driven by investors assuming the government is on the hook for large bank solvency. 


The preferred solution to overcome these obstacles is to disclose the loan-level performance of the assets supporting the covered bonds on a daily basis.  With the performance data, investors in both the primary and secondary markets can value the underlying assets and assess how dependent the interest and principal payments are to replacement of the non-performing assets.  The less dependent, the more the covered bond is priced off the asset quality and not the issuer’s perceived solvency.  The more the covered bond is priced off asset quality, the more attractive this product becomes as a source of funding because there will be investor demand.  The more the covered bond is priced off asset quality, the more robust the primary and secondary markets are and the less likely they are to freeze.


Lessons for the Broader Securitization Market


            The disclosure solution for the covered bond market is also applicable to the broader securitization market.  Investors in the securitization market also face the issue of a lack of information on the performance of the assets backing these securities.  As recognized by the SEC in its Proposed Rules for Asset-Backed Securities, “information asymmetry” exists between the information that issuers report to investors in asset-backed securities (“ABS”) and the information that is needed in order to assess the risk, value and monitor the underlying assets. 


In order to solve the problem of information asymmetry in the ABS and covered bond markets, it is necessary to address not only what information is to be provided but also the timing of the disclosure of this information.  The section of the Proposed Rules on the specific data points for each asset type backing a structured finance transaction addresses the issues of what information should be disclosed.  The Proposed Rules do not however specifically address the timing of the disclosure of this information.


In the ABS market, a select few have access to loan-level performance information on a daily basis.  Firms such as Goldman Sachs and Morgan Stanley have subsidiaries involved in originating, billing and collecting loans backing ABS.  They receive “fresh” loan-level performance data on a daily basis that they can use for trading days, even weeks, before most other market participants receive the information. 


By contrast, other ABS investors and market participants currently have to wait to receive the “stale” accumulation of daily loan-level performance data in a single periodic report.  The periodic report is distributed to investors by trustees, third party data vendors or on a website on a once-per-month or less frequent basis.


It follows from the Nobel Prize winning work of Columbia University Professor Joseph Stiglitz that once markets with information asymmetry freeze, they don’t unfreeze until the information asymmetry is eliminated. 


To eliminate information asymmetry in and unfreeze the ABS markets requires that each ABS provide all market participants with fresh performance data on a daily basis on the individual loans that support the ABS.  If all market participants receive equal and full information on a daily basis, they can evaluate the risk and return of the ABS in both the primary and secondary markets.


Benefit of Receiving Loan-Level Performance Data Daily


What would have happened if there had been access to loan-level deal specific data daily and its forced recognition of the deterioration in loan underwriting and performance in the years leading up to the current financial crisis?  A significant amount of losses could have been avoided if investors had access to loan-level deal specific data daily and had been able to accurately assess the risk in securitizations. 


It has been reported that securitization investors like Goldman Sachs and Morgan Stanley who had this data daily recognized that risk was mispriced, stopped buying new securities by late 2006 and in fact went further and shorted the subprime market.  According to the Securities Industry and Financial Market Association, over $1.75 trillion in non-agency mortgage backed securities, home equity loan-backed securities and CDOs were issued globally between the time Goldman Sachs and Morgan Stanley decided to stop buying such securities and the beginning of the credit crisis in 2007.  Analysts and traders estimate that there were several hundred billion dollars of losses on these thinly traded securities. 


These losses would have been avoidable if either the other investors had exerted market discipline (based on more up-to-date data) by not providing liquidity for unsustainable origination practices or regulators had noticed that securities firms (such as Goldman Sachs and Morgan Stanley) with access to loan-level performance data daily were placing massive shorts on the market and intervened.  In addition, there were other avoidable losses in the financial system as there were loans made during this time period that ended up on the balance sheets of financial institutions both to replace the loans sold into the capital markets and to grow the financial institutions’ internal loan portfolios.  If market discipline had been exerted in late 2006, these loans might not have been made.  These loans have also incurred a significant amount of losses.  At a minimum, the benefit to the financial system from providing loan-level deal specific data on a daily basis would be several hundred billion dollars of losses avoided.


Cost of Providing Loan-Level Performance Data Daily


The annual cost of providing loan-level deal specific data daily for securitization and covered bond transactions will be much lower than the losses described above.  In order to provide this data, a new data-handling infrastructure will be needed to collect, store and distribute this information.  Based on the cost for comparable information services for securitizations, the on-going annual cost of the infrastructure for loan-level deal specific performance data daily would be approximately 5 basis points (0.05%) of the principal amount of the loans that are supporting a particular securitization or covered bond.  


Benefit Outweighs Cost


By spending 0.05% per year of the amount of the loan collateral, the covered bond and securitization market can avoid repeating the several hundred billion dollars of losses from not being able to accurately assess and price the risk of securitizations.  Spending such amount will also restore confidence in and restart the securitization markets.