By Vinod Kothari

The failure of Superior Bank and the recent prosecution filed by the Federal regulators against accounting firm Ernst and Young must force accounting standard-setters to rethink on securitisation accounting. The current accounting standards are too much reliant on estimates – estimates which are accounted for as profits – and when estimates go wrong, and some estimates will always go wrong, we find someone to persecute. The following statement is not a defense of the accounting firm, but we need to realize that the problem is not really the wrong valuation of residual interests in a particular case, but the very basis of securitisation accounting, which is so dependant on guesswork.

The $ 2 billion prosecution:

1st Nov, 2002, FDIC in its corporate capacity filed a 50-page prosecution against accounting firm Ernst and Young for an over $ 2 billion claim on admitted "gross misstatement of Superior's assets". That Superior's assets were over-stated is an admitted fact, though E&Y has denied there being an accounting lapse and has claimed that the erosion in asset values was purely due to economic factors.

On the other hand, FDIC pleads that "as a direct result of E&Y's gross misstatement of Superior s assets, the bank became insolvent, which ultimately required the FDIC to pay out in excess of $750 million from the Federal Insurance Deposit Fund." Professional negligence apart, FDIC also rakes up the issue of consulting-auditing interface – "In clear violation of industry and statutory auditor independence rules, E&Y delayed disclosure of Superior s true financial condition. E&Y s consulting arm was engaged to "bless" the conclusions of E&Y's auditors during the time that disclosure of a $270 million error would have adversely affected an $11 billion sale of E&Y consulting."

Superior's failure:

The case of Superior Bank easily highlights the risks inherent in securitisation – that it might tempt a banker to originate assets that the bank might hate to keep on its own books. The bank was virtually romancing with subprime lending behind the securitisation facade. In 1993, it began to originate and securitize subprime home mortgages in large volumes and later, finding that there were investors who buy up what a banker itself would hate to keep on balance sheet, it expanded its activities to include subprime automobile loans as well. As would be usual, the bank was supporting its securitisation business with residual interests and over-collateralisation.

Superior 's residual interests represented approximately 100 percent of tier 1 capital on June 30, 1995. By June 30, 2000, residual interest represented 348 percent of tier 1 capital, which, put simply, would mean that that the risk on the asset side was 3 1/2 times the risk on the liability side. After all, the first loss risk retained by the originator in a securitisation transaction is comparable to equity in a corporation. If Tier 1 capital is the first loss support to the bank, the equity holders in Superior Bank agreed to absorb first loss risk of $1, and correspondingly, the bank went out in the market to bear first loss risk to the extent of $ 3.48. To a lay man, it would mean, I have $ 1 in my pocket and go to the casino and put a bet of $ 3.48 – however, the regulators did not see this for quite sometime. It is only from Jan 2002 that the Federal regulators have required dollar-for-dollar capital provision for retained interests, a regulatory provision which is yet to adopted by most other countries in the World.

Not only did the bank's financials hide this risk, it, on the contrary, continued to book profits on sale of subprime loans which is both allowed and required under US accounting standards. The A testimony before the Committee on Banking, Housing, and Urban Affairs, U.S. Senate stated: "Superior's practice of targeting subprime borrowers increased its risk. By targeting borrowers with low credit quality, Superior was able to originate loans with interest rates that were higher than market averages. The high interest rates reflected, at least in part, the relatively high credit risk associated with these loans. When these loans were then pooled and securitized, their high interest rates relative to the interest rates paid on the resulting securities, together with the high valuation of the retained interest, enabled Superior to record gains on the securitization transactions that drove its apparently high earnings and high capital. A significant amount of Superior 's revenue was from the sale of loans in these transactions, yet more cash was going out rather than coming in from these activities." 

The bubble burst when regulators required the bank to re-value its residual interests when the bank became undercapitalised and was ordered to be closed.

The FDIC complaint says: " The concentration of the bank in this high-risk, volatile arena was extraordinary. Virtually all of Superior s earnings were derived from the accounting related to Superior s issuance of bonds secured by its subprime mortgages ( securitization transactions ). Superior s reported recourse for its securitization transactions represented 163% of Superior s assets compared to .03% for banks like Superior. Moreover, Superior s return on assets, a basic yardstick of bank profitability, was 7.56%, which gave it the distinction of having the highest return on assets of any insured thrift in the nation — over twelve times higher than that of the average thrift operating in the United States (.62%)". 

The root of the problem:

The root of the problem is a yawning divergence between regulatory standards and accountings standards. While regulatory standards required a dollar-for-dollar provisioning on retained interests, accounting standards go by fair valuation of these very interests which may continue to show as assets on the originator's books. 

The retained interests mostly represent the credit enhancements required to make the securities acceptable to the investors. Assume Superior had assets of $ 100 that it wanted to securitise and the rating agencies reverted that a AAA-rating for the securities was not possible unless Superior gave a credit enhancement of 12%. This 12% may include trapped servicing income or excess interests, subordinated interests, etc. In plain words, this 12% is nothing but equity in the transaction that makes the debt rated AAA. In order to put the debt on a AAA-platform, it needs to climb up on the shoulders of the equity class – which in the instant case is 12%. The originator's retained interest is the originator's first loss risk, and very obviously, no originator would be prepared to take a higher first loss risk than is required to make the securities marketable. Of course, the credit enhancement is a several-times-multiple of the first loss, but that multiple itself is a function of the desired rating. 

In other words, if Superior wanted to export a less than AAA-rated paper off its balance sheet, the first-loss piece could have been lower, but at its own option, it opted to give a better rating to its securities, and hence increased the first loss risk. 

Present accounting standards value the same amount of first loss risk irrespective of the extent of credit enhancement, while it is a known fact that if the risk of what goes out of the balance sheet is lesser, the risk of what remains on the balance sheet must be higher. So, higher credit enhancement should call for a higher provision, which is not being done under FAS 140 currently.

There are two options: either the regulatory standards and the accounting standards could be harmonised and accounting standards could also require for a dollar-for-dollar capital provision for retained interests. Alternatively, the UK FRS approach, which does not provide for any off-balance sheet treatment could be followed.

The present approach is getting increasingly more complicated and more prone to miscalculations, either deliberate or otherwise.