The banking crisis appears remarkable for so much depending on just one asset class? If you are puzzled as to why US legislators and EU politicians are especially angry about the central role of the ratings agencies (Standard & Poor's, Moody's Investors Service and Fitch Group) in the financial crisis, read on. Ratings agencies are arguably more vital to the financial system than even accountants! Why? Because the ratings agencies value banking assets more succinctly or persuasively than accountancy firms. As a theological reminder, it is important to bear in mind that risk grade ratings (using Letters A,B,C,D with + or – and numbers 1.2.3 etc.) are essentially the aggregate of re-payment default and final loss after recoveries. The potential to default and loss given at default are estimated historically over a full economic or credit cycle pertaining to many thousands of examples of asset classes, obligors, counterparties, and very large pools of millions of retail loans, credit cards, mortgages etc. A credit risk grade expresses the known likely % within large populations of borrowers that some will default on loan repayment. Risk grades can also reflect market risks of tradable financial instruments if they are suitably long-established with a reliable history of price volatility. It is on the basis of the risk grades that financial assets have value and prices (or spreads). These risk estimates have to adjust however depending upon the maturity of a loan. That is to say if a loan (asset) has a maturity of months or a few years only, say 1-3 years, then if repayment of principal and interest can be forecast to occur in a downturn or other predictable shock period, or cycle turning point, then the average default rate over a whole cycle is not the true representative risk-price. This is the problem especially with SME loans, credit card and car loan debts which are short term, while mortgage debts are much longer term and are likely to experience one or two economic cycles. Corporate loans are also rated according to the company's type and size since a large % of small firms (about 10% close or fail each year) and firms that are heavily borrowed, such as property developers, do not generally survive well over economic cycles, similarly for certain classes of household borrowings such as overdrafts and general consumer credit.
The credit crunch story of ratings begins with Moody's and Fitch the first bond ratings agencies, both formed in 1913. S&P followed in 1941. Only Moody's is publicly quoted (as of 2000). Its revenues grew fast, as did the other two global raters, reporting the highest profit margins of any company in the S&P 500 index for five years running; shares up 500% in 4 years, earnings up 900 % in 7 years, with nearly half of its revenue coming from its collateralized debt obligations (CDO) ratings service, until in the past year it lost over 60% share value to a P/E of 5 ($2bn income, $1bn profit, $5bn mkt.cap.) Two years ago, by the winter of 2006 it was clear something was going wrong with the risk ratings of CDOs most of the contents of which (about 90-95%) were originally rated triple-A by the three ratings agencies. This was hardly unexpected, however, as US real estate began falling in certain areas in 2005 and in the US there is always a risk of mortgagees abandoning their properties when the remaining equity is below the remaining debt; the only recourse of the lenders is to the value of the property, not the borrower, unlike in Europe. In areas between San Francisco and Los Angeles a large % of new house buyers have negative equity today of approx. 66% compared to the outstanding loans! Before the end of 2006, about 750 CDO bonds had been downgraded in the previous 2 years. They fell in value by about $25bn (or 2% of the total mortgage backed CDOs) because of falls in repayment rates by sub-prime U.S. mortgage holders, according to Lehman Brothers Holdings who claimed to maintain the most thorough analysis. CDOs (a broad term for unregulated fixed income asset-backed securities, mainly bonds supported by mortgages or credit card debt) were designed to sell chunks of banks assets to free up capital for faster lending growth (fueling 'credit boom' economies). CDOs could average 5-8% yields and given their high credit-rating appeared to be very good value to investors; Cdos were easy to sell. The first CDOs were packaged by Drexel Burnham Lambert in '87. In those days (since '78) the term 'sub-prime' meant high credit-worthy borrowers just below the highest grade. In late 1992 the term suddenly changed meaning. It now meant low credit-worthy borrowers? In 1992 at the end of the outgoing George H. W. Bush administration regulation had changed allowing financial brokers, not just commercial banks to offer mortgages. It was then when financial engineers at JP Morgan figured out how large amounts of sub-prime mortgages could be included in investment-grade CDOs. The sub-prime mortgage market grew dramatically at an annual rate of 25% between 1994 and 2005, a tenfold increase in a decade. By early 2007 CDOs were worth over $2tn (almost as much as all banks' own capital), of which a quarter was sub-prime mortgages. At first this was considered a good thing. Sub-prime mortgages were an important social policy, growing the property-owning democracy and appropriately underwritten by federal government agencies Fannie Mae and Freddie Mac. In 1994, less than 5% of mortgages in the US were sub-prime. Home ownership increased 1.94% annually during the Clinton administrations, reaching 67.7% by 2000. Today, of about $11.6tn in outstanding mortgages, one-quarter are sub-prime and Alt-A loans. In the '80s S&L crisis, by comparison only about one tenth as much at today's prices was sub-prime and yet even at that level was partly blamed for the '90-'91 recession! A major problem for investors in CDO securities was the absence of a liquid secondary trading market with price transparency. Investors or their brokers had no screen service they could look up to check daily prices. Consequently, investors (unless they were market insiders) did not know when values dropped as defaults increased. There was no easy way then or since for investors to find out what their CDOs are worth in the market, which is not an uncommon problem for many bank bonds, but became alarming for a market that grew to become huge in scale. Though, from a regulator's viewpoint, however, CDOs represented only about 2% of the total of world financial assets (totaling $118tn by end of 2005). Even today, the delinquency in the underlying US mortgages is estimated at about 3.5%, but 3-4 times this in sub-prime mortgages (say 13% of $700bn) and rising. Indicators did emerge of the market value of the bonds, but mainly the valuations (ABX index etc.) are based on risk gradings. CDOs are variable in their contract details and not easy to compare. To date write-offs have totaled about $505bn, or 4.3% of the total mortgage market.
The difficult-to-chart slide in CDO values exposed the little understood role played by rating companies in assessing risk and acting as the only de facto regulators in a market that lacked official watchdogs. Again, from a regulator's perspective the market was considered to be a professional interbank and institutional investor market. Most of the world's CDOs are owned by banks and insurance companies, and the people who regulate those firms rely on the ratings to provide a sound basis for valuations of CDOs. But, the market changed once regulations were relaxed to allow Alternative Investments to be included in retail investment funds, pension and insurance funds. Rating agencies facilitated this greatly by maintaining high risk grades for new issues, surprisingly so when signs of possible recession were on the horizon and obvious at least to economists. The ratings agencies did more than value CDOs by giving them letter grades; they helped banks and other financial firms create CDOs (e.g. wrapping together 100 or more bonds and other securities, including debt investments backed by home loans) and then slice and dice them into tranches, each with a separate grade. The raters told CDO originators how to get most profit by maximizing the tranches with the highest ratings so that typically 80%-95% will be AAA and the rest AA or A.
All pretence that financial engineering could dispel expected risks was lost on August 16 2007 when, after an internal revision of its ratings practices, Moody announced new corrected ratings models. These ignited the credit crunch like a dirty bomb devastating $trillions of assets and derivatives and in turn most the world's biggest banks, and as yet untold numbers of investment funds worldwide! If a company's bond is given a triple-A rating, it means Moody (or its nearest competitors ,S&P and Fitch) believes unsecured loans to the company (its bonds etc.) are as safe as Government debt, an extremely high probability of lenders being repaid the principal plus interest, bonds that could become collateral for at least 80% of face value. A great deal more in loans and funding was teetering on the values of CDOs. The lower the rating - from Triple A to Double A to Single A - and on down to Baa, below which the bond or loan is rated junk, certainly not investment grade i.e. high-risk! rapidly $100s of billions of CDOs dropped 4 to 17 notches, from highest grade to lowest grade and over a year were discounted by 60% to 80% to over 90%!
It has been essential for nearly a century that all bonds by publicly quoted issuers receive a credit rating. Ratings services are paid for by issuers, and only in some cases by investors, through fees or subscriptions. As debt markets grew, investors sought two ratings, Moody's and Standard & Poor's, a duopoly and sometimes as a third check, Fitch, which with Moody's and S&P, is internationally recognised by all regulators. Basel II, for example, requires 3 independent ratings from which banks may take only the middle or lowest value for risk assessment if the ratings differ. Until 3 ratings were required and not merely 2, Banks could avoid Moody's, which was known for giving lower ratings than its competitors until recent years when some of their strict internal rules were abandoned.
Rating a corporate bond, involves rating a company, and companies have assets, business models, balance sheets, management qualities, shares if public, and, most importantly, credit history. A rating is based on a quantitative model plus qualitative human judgment, just as banks do internally. Increasingly banks have become almost wholly dependent on external ratings and their internal analytics have therefore suffered. Rating a structured-finance instrument like a CDO is different if it is an asset-backed security vested in a 'failsafe' or 'bankruptcy remote' Structured Investment Vehicle (SIV) or Special Purpose Entity (SPE), which more often than not is a registered company in an off-shore tax-haven managed by lawyers and/or an accountant. Rather than having an established corporation or bank as direct issuer, the securitized (structured) bonds are backed by pools of debts (either mortgages, or auto loans, or credit card loans or other loan types, not a mix of these (usually from a single national jurisdiction) but can be from more than one bank or mortgage broker , building society etc.) These are collected, packaged into tranches, rated per tranche, and sold on by banks' arms-length SIV/SPEs or others (while the banks or agencies that originated the assets continue to manage them for a fee and who may contribute commercial paper and/or insurance against defaults above certain % levels in the issued bonds). Higher ratings were believed to be ensured (justified) by various credit enhancements including over-collateralization (pledging collateral in excess of the debt issued), credit default insurance, and equity tranche investors willing to bear the first 5% losses and mezzanine investors the next 5%. The high ratings achieved allowed banks to sell on mortgage assets paying them 8% or higher in the form of bonds that paid the senior tranche (80-95% of the total) only say 5-6%. The banks called the difference 'the spread' and booked it as profit. Such a wide spread should have indicated to experienced secondary market traders that the bonds were much riskier than they appeared. Banks could use the spread to cover the insurance and CP cost of securing higher grades, but also borrowed to do so thereby keeping the spread as profitable additions to own capital i.e. they replaced the funds loaned (enabling accelerated loan growth), sold of much or most of the risk, and still generated a profit equivalent to what they would have earned had the loans remained on their balance sheets – or so they thought that is, until Moody's discovered a bug error in its rating model software!
After August 16 2007 when Moody's announced new improved ratings models, rating agencies lowered the credit ratings on $1.9tn in mortgage backed securities by July 2008. The downgrades may have averaged about 50% of the face value a year earlier, but that became academic when Merril-Lynch sold $30bn of CDOs for 22 cents on the dollar, which now looks actually a good deal?! About 6m sub-prime mortgages were securitized in the US (of over 7.5 million first-lien sub-prime mortgages outstanding) and one quarter of these resulted in possible foreclosures by 2008! Big banks and other financial firms around the world reported losses of approx. $435bn by mid-July 2008 and $505bn by end of September, equal to about one fifth of all banks' capital reserves. Approx. 16% of sub-prime loans with adjustable rate mortgages (ARM) were 90-days delinquent or in foreclosure proceedings as of October 2007, roughly triple the rate of 2005. By January 2008, the delinquency rate had risen to 21%and by May 2008 it was 25%.
The U.S. mortgage market is just under $12tn with approx. 9.2% of loans delinquent or in foreclosure by August 2008. Sub-prime ARMs are 6.8% of the loans outstanding in the US, but 43% of foreclosures in 2007 when 1.3m properties were subject to foreclosure filings (up 79% compared to 2006).
HSBC was the first bank to announce a large write-off in mortgage backed assets of $10.5bn in January 2007. Such announcements by other banks followed in a steady stream ever since.
As we all now know the knock-on multiplier effect of these risk down-grades, write-offs, interbank credit crisis etc. is enormous. The question is how different would this have been if the risk gradings had been realistic from the start? The market would not have grown so fast. The credit boom economies would have been curtailed to slower growth. Property values would not have ballooned so much and the recession therefore less severe. Why were the ratings wrong? Most structured bonds are collections of mortgage or credit card loans originally packaged by banks and evaluated and underwritten by other banks based on rating agencies risk grades. Credit card securitizations are many times riskier assets (higher default risk) than mortgages because credit card debt is unsecured, yet as much credit card debt has been securitized and sold off banks balance sheets just as sub-prime mortgages have been (both just over $900bn in the US, but credit card bonds have a much shorter maturity, typically 4-5 months!). The insurance cover of now $54tn in CDS also poses tens of $billions in losses. $16 trillions or so have been written off world stock market equity capitalisations! Much of this would have happened anyway whenever recessions hit. But, could a much more manageable situation have arisen if $ trillions had not been over-rated? The measure of manageable has to be within the Basel capital reserve ratios of the banks at 8% of net risk-weighted assets.The actual impact looks like being 2-3 times this! (For more on this see the attached comment.)