Asset securitization can be described as a technology by which certain classes of contractually homogenous corporate assets, typically loans, leases, and receivables, are pooled and legally partitioned from the firm that originated the assets (the originator firm). This is accomplished by a legal transference of title from the originator firm to a special purpose vehicle (SPV). The SPV is typically an irrevocable trust but can also be a corporation, partnership, or association. Securities are then issued by the SPV, hacked by the future cash flows of the partitioned assets. The resulting financing alternative has produced a number of real economic benefits for originator firms including improved corporate liquidity, lower out-of-pocket costs of debt financing, and better matching of financing and investment cash flows. Further, because current banking regulatory requirements (RAP) are sufficiently flexible to allow many securitizations to be treated as sales, the technology has helped the banking industry reduce capital funding requirements. Not surprisingly, given all these benefits, the market for asset-backed securities has grown from $1.237 billion in 1985 to $130.659 billion by 1993. The accompanying exhibit graphically illustrates that growth is proceeding exponentially. An entire financial industry and a distinctly unique class of investment grade security has emerged.

Risks Associated with Securitization

Risks associated with securitization arise from four sources: 1) interest rate or market risk associated with changes in yield required for resale in the secondary market; 2) credit risk associated with failure of credit-card account holders to pay off balances; 3) prepayment risk which may be associated with the business cycle and/or interest rate movements; and 4) securitization risk that the trust structure will not function as contractually agreed. Securitization structures often have provisions to mitigate or shift risk components among the various parties. This is particularly true for credit and prepayment risk. For example, the originator firm might agree to absorb any credit losses up to an agreed percentage of the transferred amount of assets. Other possibilities include third-party guarantees, "excess" accounts that accumulate the SPV's excess cash flows for use in absorbing credit losses, letters of credit, and overcollateralization of asset-backed debt.

Interest-rate risk is a function of the difference in terms between securitized assets and asset-hacked debt issued by the SPV. In credit- card securitizations, for example, the underlying receivables contain essentially a fixed rate across similarly originated accounts. If variable-rate, asset-backed debt is sold by the SPV, increases in interest rate will adversely affect the value of the securitized assets, thus giving rise to interest-rate risk.

Current Reporting Standards

Securitizations are currently reported for as either sales or collateralized loans, in accordance with SFAS No. 77, entitled Reporting by Transferors for Transfers of Receivables with Recourse, and Technical Bulletin 85-2, Accounting for Collateralized Mortgage Obligations (CMOs), respectively. This accounting is based on "predominant characteristic approach." In the case of SFAS No. 77, if a transaction "purports" to be a bona fide sale, the transferor's obligation under the recourse provisions can be reasonably estimated, and if certain other qualifying criteria are met (i.e., no calls or puts on transferred assets, control of assets has been relinquished), the transaction is treated as a sale. Similarly, if the transaction takes the legal form of debt, Technical Bulletin 852 applies. While allowing for some exceptions, the bulletin generally requires such transactions to be reported as collateralized loans.

Both pronouncements have been criticized because they use the legal form of a transaction to determine whether sales or loan treatment applies. Legal form is arguably weak as a basis for reporting standards because it is arbitrary, can be manipulated, and can diverge markedly from economic substance. To see this, consider the securitization of credit- card receivables. Credit-card securitizations are unique because closed pools of credit-card receivables amortize quickly (usually within eight months). To maintain principal, the rights to future receivables are conveyed by the originator firm to the SPV trust. In addition to proceeds from the "sales" of receivables to the SPV, the originator firm receives a service fee and "excess fees," representing cash flows received in excess of those needed to provide credit enhancements, service the receivable accounts, absorb credit losses from defaults, and pay interest payments on the asset-backed securities.

Take a case where a pool of credit-card receivables yields 20% annually. If, as is typical, service fees are around 3%, debt securities are issued at 6%, and other nominal charges such as fees for guarantee letters, interchange, etc., equal 1%, the originator firm retains a significant economic interest through excess fees. In this case 10%, or 50% of gross future cash flows, will ultimately return to the originator firm. Most of these cash flows are subordinate, however, in the sense they will first bear any credit, interest rate, and prepayment risks associated with the securitized assets. Yet, because the transaction "purports" to be a sale in legal form, the assets would be removed from the originator firm's balance sheet in accordance with the provisions of SFAS No. 77.

Using Economic Substance to Distinguish

Current literature uses legal form as the principal discriminating criterion as to whether a transaction should be treated as a sale or collateralized loan. Many have suggested that economic substance should be used to make this distinction. A transaction could be classified as a sale when most of the economic benefits and rights, as defined by prescribed rules, have been relinquished. This approach has been used in other areas, such as lease accounting via SFAS No. 13. Such an approach is objective, thus promoting comparability. The attest function is also well structured under this approach. An auditor has only to determine if 1) the triggering prescribed conditions exist and 2) if the applied accounting treatment is dictated by the prescription.

There are, however, a number of problems with such an approach. Consider the case where a firm transfers $100,000 of credit-card loans, paying an average of 20%, to a SPV. The SPV issues, at par, $100,000 of asset- backed 7% securities, maturing in five years under a fixed repayment schedule. The SPV pays the $100,000 it receives back to the originator firm and agrees to let this firm service the assets for a fixed fee of 3%. Another 2% is used to cover the costs of letters of credit and credit fines, interchange, fiduciary fees, and other minor servicing costs. Any residual cash flows belong to, and revert immediately to, the originator firm. The originator firm provides recourse for all credit losses up to 10% and for all losses from prepayments that must be reinvested at lower rates of interest. The question is whether, in substance, a collateralized loan transaction or sale has taken place. In this case, the answer appears to be "loan." The originator firm carries almost all risk, mainly credit and prepayment, associated with the assets. As discussed earlier, the originator firm also retains a significant portion of the expected future cash flows of the assets.

What if, however, credit-card rates, as a result of increased competition, fall to 16%? Sales treatment, in this case, may capture more of the economic substance of the transaction since little of the securitized assets' expected future cash flows will now pass to the originator firm. Sales treatment, however, may still not reflect sufficient substance since investors still hold only a senior security interest in the securitized assets, and the originator firm bears most of the credit risk through the recourse provision.

Requiring No Recourse

Now suppose FASB issues a new predominant characteristics standard to partially address this concern. The standard allows sales treatment, but only if there is no recourse provision included in the transaction. Unfortunately, the new standard, though very conservative, will still not mitigate the problem. The SPV can be structured in many other ways to channel credit risk through other mechanisms. In this case, future excess fees can be accumulated and used to offset credit losses. This kind of structure, in fact, is prevalent in credit-card securitizations that are often originated by banks subject to RAP rules. Under RAP rules, sales treatment applies only when no recourse is provided.

No recourse is, of course, only one possible way a standard could be written. Far more than lease contracts, which do not create separate legal entities, trusts are very flexible devices used to partition, mix, and shift various risk and benefit components among numerous parties. Trusts, after all, serve a myriad of personal and business purposes, particularly in estate and tax planning. They have very limited statutory requirements as to form and thus can support broad and complex transaction structures. Numerous mechanisms could be developed to comply with the prescriptions of almost any standard. Many exist now, including third-party holdings of nominal residual interests, letters of credit to the trust, and cash collateral accounts and credit lines made available to the trust. How can a standard be written to cover the many contingencies, structures, and forms that are possible within securitization structures? How can a prescriptive standard remain relevant in the face of the ongoing innovation that has characterized securitization and many other forms of derivative financing? It appears that predominant characteristics, however operationalized, may not be able to fully and consistently capture the economic substance of securitization transactions. The extent to which it can will be a function of the securitization structures employed in the marketplace, the agreement of the standards with these structures, and the extent of true risk shifting and mixing within the SPV.

Other Problems

There is also the risk that a conservative, prescription-based standard may dictate collateralized loan treatment when, in substance, most of the economic benefits and risks have been relinquished by the originator firm. What if, in the previous example, the securitization had no residual "seller's interest," little excess fees, and no prepayment guarantees, but the firm did provide a 3% recourse for credit losses? If expected losses were 3% or higher, the situation would be similar to the sale of any asset with a warranty guarantee. Yet a no-recourse clause in the standard would dictate collateralized loan treatment, thus overstating the economic impact of securitized assets.

Another argument against standards based on predominant characteristics is the lack of influence the originator firm has over securitized assets. The representation of securitized assets within a balance sheet suggests firms have influence over such assets. Securitized assets, however, fall under the auspices of a trustee who is constrained strictly by 1) the provisions of the trust agreement, and 2) fiduciary responsibilities as dictated generally by the domicile of the originator firm. Residual interests in trust have no voting power. SPVs are usually "irrevocable" in nature, with exceptions, and do not explicitly provide any power to the grantor, although they could do so.

Securitized assets would be categorized together with unencumbered assets and loans of the same classification. While disclosure provisions could provide details about all securitization transactions, this approach would be far more cumbersome as securitization transactions grow in number. Further, the balance sheet numbers would lose comparative utility and meaning since categories would now be qualified and conditioned on disclosure data.

In a similar vein, collateralized loan treatment does not recognize the innovation in securitization technology: the capacity of the SPV to fully partition the obligations of the trust and the originator firm. Without this innovation, a number of the benefits of securitization would be absent, including the ability to obtain AAA/Aaa investment grade ratings on asset-backed debt. In addition, if a senior interest in securitized assets has been sold, via the SPV, it follows that what remains must be of lower average credit quality. By including securitized assets and asset-backed claims in the balance sheet of the originator firm, however, investors remain uninformed about the benefits and risks that securitization has added. Again, disclosure could mitigate this problem, but only at a cost, namely the loss of comparative meaning within the financial reports and the inconvenience to investors of utilizing disaggregated data of varying content, presentation, and meaning.

The Financial Components Approach

The FASB is now considering a new approach to securitization known as the financial components approach. In this approach, all securitization transactions, irrespective of substance, would be recognized as sales. In addition, however, all component interests and obligations retained by the originator firm would be recognized at the time of sale. The model has been partially applied via SFAS No. 77, which requires recognition of any explicit recourse provisions when sales treatment is accorded to receivable transfers. Financial components, however, would go much further by 1) treating all securitizations as sales, irrespective of form, and 2) recognizing numerous "components" of securitization transactions, including call and put options on securitized assets, forward contracts for future receivable transfers, service rights, guarantees and letters of credit, and excess fee interests in addition to recourse provisions. The model is promising because it appears to allow individual recognition of the numerous trust mechanisms that can shift asset benefits and risks during securitization. As such, misspecifications of economic substance and or corporate domain arising from the "all or none" approach inherent in predominant characteristics would likely disappear under standards based on this approach.

There are, however, a number of new conceptual, recognition, and measurement problems that may arise under the model, perhaps explaining why the approach has not already been adopted.

Matching Costs with Revenues. Sales treatment recognizes immediately the expected future cash flows of securitized assets. These cash flows, however, require extensive service performance before they are "earned." Sales treatment under financial components, therefore, may weaken the information content attainable from accrual accounting: namely, the process of recognizing sales revenues, as earned, and matching these to periodically incurred costs.

What Parts of the Contract Should Be Recognized. Securitizations at any point up to completion are essentially a partially executed, connected group of contractual provisions. How much of these nonexecuted contractual provisions should be recognized? For an extreme case, consider the credit-card receivable securitization transaction described earlier. The originator firm agrees, throughout the revolving period, to transfer additional receivables to the SPV to maintain a stable balance of securitized assets, even though a pool of credit-card receivables will paydown at rates of 12% per month or better. If the transaction is viewed from an overall point of view, these receivables have been "sold" at the point where the securitization transaction is initiated. Given an average transaction life of five years, the amount of future receivables to be transferred will exceed, by several times, the initial assets transferred. Yet the receivables do not actually exist yet. Should gains on their "sale" be recognized immediately upon inception of the transaction? Or should a value be assigned to the forward contract itself. If so, what value will be assigned, given that no recognized secondary market exists for this component?

Valuation of Components. For many securitization components, as with the forward contracts mentioned above, no valuation procedure is clearly evident. For certain components, such as put and call options, theoretical valuation models do exist. For other components, such as recourse, guarantees, and letters of credit, historical credit risks can be determined and an expected value derived. For still other components, an active secondary market may exist, such as, for example, some actively traded "seller" certificates. For many components, however, no clear valuation approach appears evident. If components are to be valued at market value, this presents an obvious obstacle to recognition. If, however, historical cost values are used and allocated to components, the problem does not go away. Any reasonable allocation would stiff have to be based on proportionate market values at time of inception. Again, a valuation model for all recognized components would be necessary.

Recognition of Synergy. A final recognition problem that arises with the financial components approach concerns the lack of recognition of the securitization innovation itself. Securitization ties together, in a complex structural manner, various provisions, rights, agreements, and obligations among potentially numerous parties. By recognizing each individual component separately, however, the whole transaction is not reported, thus leaving unrecognized any synergistic information. Yet the value of the synergy attainable through the SPV trust is precisely what may drive the occurrence of securitizations in the first place.