Comments are due today to the Securities and Exchange Commission (SEC) regarding a rule that it proposed last year which prohibits sellers of risk in synthetic securitizations (underwriters, sponsors, etc.) from taking short positions in the securitization. I summarized the rule here.
If implemented as proposed, the rule may decrease the ability of banks to reduce risk. It may also increase regulatory arbitrage.
Prohibiting Risk-Reducing Securitizations
The proposed rule, mandated by the Dodd-Frank Act of 2010, targets the types of synthetic securitizations made infamous by an array of hedge funds and investment banks including John Paulson's fund, Magnetar, Goldman Sachs, JP Morgan, and Citigroup. The rule seeks to prohibit certain types of conflicts of interest in synthetic securitizations, including in those securitizations that are used by a firm
to reduce its long investment exposure to the relevant assets because it has come to believe that the assets will perform poorly. If the firm accomplishes this result by transferring the risk of its long positions to [asset-backed securities] ABS investors through a synthetic ABS—while marketing the ABS securities to investors as a good investment opportunity—it could be viewed as benefiting from a decline in the ABS at the expense of the ABS investors, who now have the exposure to the underlying assets. [proposed rule (page 60,338)]
Unfortuantely, this aspect of the proposal may have the effect of prohibiting synthetic securitizations that are used by banks reduce risk, so-called "balance sheet" transactions.
An example of such a transaction is the 2002 Deutsche Bank London Wall synthetic collateralized debt obligation (CLO). It transferred the risk of a "reference portfolio" of 264 corporate loans to investors, and is depicted in the following diagram from Standard and Poor's.
CLOs have in fact been effective in helping banks reduce their exposures to corporate loans. Investors have also benefited from their CLO investments by earning their expected returns and seeing minimal defaults, even for CLOs created before the 2008 financial crisis (see this Federal Reserve report, page 62). And contrary to what is implied by the proposed rule's discussion, investors have indeed found "good" investments in "poorly" performing assets. Due to pooling, tranching, and other credit risk transfer governance mechanisms, CLO investors experienced minimal losses even though corporate loan defaults spiked in early 2010.
It is unclear, therefore, where the actual conflict in these types of securitizations exist. True, all derivatives transactions are a zero-sum in the sense that one party's gain is the other's loss; and vice versa. Indeed, the very nature of a synthetic transaction is such that the risk transferring firm must "benefit[] from a decline in the ABS at the expense of the ABS investors."
But there is nothing inherently suspicious (or conflicted) about the balance sheet securtization the SEC describes, any more than there would be in a credit default swap purchased by a bank for hedging purposes. If adequate disclosures are made to investors about the underlying long positions whose risk is being transferred, and about the structure of the securitization itself, the transaction is no different from a conflicts perspective than a wide range of routine and healthy transactions that transfer risk, including a loan guaranty or loan participation agreement.
Increasing Regulatory Arbitrage
The potential for the proposed rule to limit the ability of firms to manage their credit risk exposures was also noted in a comment filed in response to the rule by a representative of banking institutions (the International Association of Credit Portfolio Managers).
That comment letter makes the additional point that the rule as proposed would nonetheless permit economically equivalent transactions:
[P]rohibiting synthetic securitizations that hedge balance sheet risk while permitting traditional securitizations that hedge balance sheet risk is not logically sustainable because it would have the consequence of treating economically identical transactions differently merely because of the means through which they are implemented. Notably, neither DFA § 621 nor the Proposal prohibits a financial institution that sells or grants participations in financial assets to an SPE in a traditional securitization from underwriting or having a subsidiary or affiliate underwrite the SPE’s ABS. Nor would DFA § 621 or the Proposal prohibit a lender from purchasing credit protection or insurance on corporate loans in bilateral transactions with counterparties other than securitization vehicles. All of these transactions – the traditional securitization of assets that are sold or participated to the SPE and the bilateral purchase of credit protection without an SPE – could have identical outcomes to the transactions that Example 3B identifies as involving a material conflict of interest.
A potential problem here is regulatory arbitrage. The proposed rule would incentivize parties to transfer risk through cash securitizations, participations, or bilateral credit default swaps; and it is not obvious that these transactions are any less "conflicted" or better promote investor protection than synthetic securitizations. Perhaps more importantly, because synthetic securitization is more efficient than other methods of transferring risk, it is unlikely that there will be any viable alternative available for banks to transfer certain loan exposures if the proposed rule is implemented.
Balance sheet securitizations should accordingly not be prohibited on the basis of being unduly conflicted. Full disclosure about the transaction's assets and structure is adequate to protect investors. This aspect of the SEC's proposed rule should not be adopted as proposed.
Comments