Asset securitization and executive compensation are frequently cited as leading causes of the 2008 financial crisis. Is there a connection between these two purported causes? The legal literature has failed to explore this question, but in recent research we have found a clear linkage between the two that we believe calls for better monitoring by bank regulators.
In securitization, a firm, called an “originator,” sells financial assets, such as mortgages or accounts receivable, to a “special purpose entity.” This special purpose entity then directly or indirectly issues securities into the secondary market. Because the originator receives the cash from the sale of financial assets to the special purpose entity but is largely unconstrained contractually in how it uses it, there have long been concerns about agency costs of securitization.
Agency costs can arise whenever a party (termed an agent) who is hired to act on behalf of another party (termed a principal) acts adversely to the interests of the principal. In securitization, observers have typically worried that originators would misspend the proceeds of the securitization on ill-conceived projects, leaving a judgment-proof corporate shell from which small and involuntary (e.g., tort) creditors could not collect.
Yet, little attention was paid to the most obvious agency cost: excessive executive compensation. This is problematic because understanding the relationship between securitization and executive compensation sheds light on underlying incentives to securitize in the first place. If we want to understand why securitization exploded as it did in the run up to the 2008 financial crisis, we need to develop a better understanding of incentives to securitize.
While there are many legitimate reasons to securitize, we find that securitization by banks bears a strong relationship to above-average compensation for bank CEOs. We created a unique dataset of over 20,000 firm-year observations from 1993 to 2009. Holding other things constant, if we take two commercial banks of roughly the same size, the one that securitizes pays its CEO over $400,000 more than the one that does not, about 20% more than the median compensation for all banks in our sample. Moreover, this is the opposite of the pattern we see among non-bank (industrial) firms: industrial securitizers actually pay their CEOs less than securitizing banks.
One might think $400,000 is not enough to worry about. In a world of CEOs like Angelo Mozilo (former CEO of Countrywide Financial) and Lloyd Blankfein (CEO of The Goldman Sachs Group, Inc.), who earned hundreds of millions of dollars through securitization leading up to the crisis, this may seem like a rounding error. But we focus specifically on commercial banks—not non-bank financial firms, such as Countrywide and Goldman—because they invented securitization, were its predominant practitioners until the late 1990s, in theory were subject to special “safety and soundness regulation,” and have not been carefully studied in this regard.
Our findings lead to several important questions: Why does this difference exist? Is this difference a problem? If it is, what should regulators do about it?
We think the difference exists because securitization improves a bank’s return on assets, and banks were first-movers in securitization, developing early expertise in these complex transactions. Securitization began in the early 1970s in part as a response to regulators’ desire to see banks lend more without relaxing bank regulation. By securitizing home mortgages, banks were able to remove these mortgages from their balance sheets and stay within regulatory parameters, even as they were lending more to accommodate baby-boomers’ growing demand for housing.
Because securitization increases the return on a bank’s assets, it created the appearance of value. But the complexity of the transactions, involving several steps and parties, produced barriers to entry for non-banks or industrial firms. The complexity and liquidity of securitizations—coupled with the absence of common lending covenants—may have made it more difficult for those who ordinarily monitor compensation (such as shareholders, directors, and regulators) to observe and act on the relationship between securitization and compensation.
We think this relationship between bank securitization and CEO compensation is problematic for two reasons. First, we find that the shares of securitizing banks perform no better than the shares of non-securitizers. Thus, CEOs of securitizers may be paid better than non-securitizers, but the banks’ shareholders are not doing any better. Second, securitization created or contributed to significant social costs in the 2008 financial crisis, including distorted property values, needlessly complex foreclosure proceedings, and all the consequences of the federal bank bailout.
Although we find the linkage between bank securitization and the compensation of bank executives to be problematic, we do not think our findings warrant significant restrictions on either securitization or compensation. Rather, we think that bank regulators should simply pay more careful attention to the relationship we have identified between securitization and compensation, giving heightened scrutiny to banks exhibiting this relationship.
To date, the chief regulatory response to the financial crisis has been the 2010 Dodd-Frank Act, which adopts neither a straightforward nor intuitive strategy with respect to securitization and compensation. The Dodd-Frank Act requires originators of all sorts—banks and industrials—to retain a certain amount of risk associated with securitized assets and to “claw back”—that is, demand the return of—excessive compensation paid by failed banks or any firm that restates its earnings.
While these might have been politically appealing tactics, neither is likely to deter excessive securitization and the destructive instability that it can create. Originators have long kept “skin in the game” contractually. Typically, they would do this by purchasing the riskiest tranche of securities issued in the securitization or by promising to repurchase securitized financial assets (e.g., mortgages) if the underlying obligor defaulted—or both. Our data show that it was not the originator’s skin in the game that mattered, but instead that of the bank originator’s CEO. Therefore, Dodd-Frank’s additional regulation requiring originators to retain some risk associated with securitized assets creates needless complexity.
The threat of a clawback might deter some excessive compensation. However, we worry that if the policy goal is to reduce system-threatening transactions, a clawback is too late and too drastic to be effective. Further, under Dodd-Frank, a clawback is available only after a financial institution has failed or has restated earnings. Regulators will likely be reluctant to commence formal resolution proceedings against troubled banks, and the rules on earnings restatements are under-enforced and thus unlikely to have a deterrent effect.
We think a better solution for regulators is a simpler one: monitor bank CEO compensation before a bank fails. Such monitoring should be sensitive to changes in compensation that associate with highly complex (and potentially risky) transactions. Given the structure of securitization, it is improbable that other bank stakeholders are in a position or have incentives to monitor this relationship as effectively as regulators. Thus, while monitoring by regulators is certainly not without its costs, it would appear to be a better alternative than other options currently in play.
Jonathan C. Lipson is the Harold E. Kohn Professor of Law at Temple University James E. Beasley School of Law.
Emre Unlu is an Associate Professor of Finance at the University of Nebraska-Lincoln College of Business Administration.
Ella Mae Matsumura is the Robert and Monica Beyer Professor of Accounting in the Department of Accounting and Information Systems in the Wisconsin School of Business at the University of Wisconsin-Madison.
Rachel Martin is a doctoral candidate in the Department of Accounting and Information Systems in the Wisconsin School of Business at the University of Wisconsin-Madison.