Evidence on the Naive-Investors Theory
The evidence on securitizer losses also calls into question a more fundamental issue: whether moral hazard caused by selling MBS to naive investors is much of a problem in the first place. This version of moral hazard predicts that most of the risk from securitization would be passed on to investors. But, as we have shown, the aggregate evidence overwhelmingly shows that securitization concentrated risk in the same financial intermediaries that organized the securitization chain. (174)
In the standard rational choice model of moral hazard, the high level of trade in mortgages during the boom could only have been achieved if market participants had devised effective solutions to mitigate this incentive problem. (175) Nonetheless, a series of articles based on a particular "natural experiment" research design has convinced many economists and policymakers that moral hazard caused by securitization did indeed play an important role. (176) This body of research interprets jumps in the default rate of mortgages at particular credit score thresholds commonly used to screen mortgage applicants as evidence that securitization led to lax screening. (177) This research is frequently cited as justifying the risk retention requirement. (178)
However, an article one of us recently published in the Journal of Monetary Economics shows that this research design has a fundamental flaw, and, in fact, provides no evidence that securitization resulted in lax screening. (179) Although the details are somewhat technical, (180) the intuition for this flaw is simple. Credit score thresholds are ubiquitous in lending for reasons that have nothing to do with securitization. As a result, there are jumps in the default rate of mortgages at these thresholds even when securitization rates do not change. Consequently, this research in fact provides no evidence that moral hazard caused by securitization was a significant contributor to the financial crisis. (181)
And there are good reasons to think it was not. The risk retention requirement is premised not only on the moral hazard problem posed by securitization but also on a failure of private responses to that incentive problem. While it is difficult to rule out conclusively all versions of the naive-investor theory, there is substantial evidence that the organizational and contractual structure of the securitization market was designed to mitigate moral hazard. The originate-to-distribute model of mortgage lending dates back at least to the 1950s, when independent mortgage companies that sold their loans to outside investors grew to become an important part of the housing finance system. (182) MBS sponsored by Fannie Mae and Freddie Mac became the dominant source of funding for mortgages beginning in the 1980s. (183) Private-label securitization was born in 1977, roughly thirty years prior to the onset of the recent financial crisis. (184) Over these decades of experience with securitization, many institutional mechanisms were developed to mitigate the incentive problems it poses.
First, originators and securitizers reduced information asymmetries by making verifiable disclosures about the loans to investors. (185) This information formed the basis of lending decisions both for originators who retained almost all their loans as well as pure originate-to-distribute lenders. Beginning in the mid-1990s, a quantitative revolution automated much of the underwriting process, as lending decisions were increasingly based on hard information, such as credit scores and debt-to-income ratios, that were fed into computer-based underwriting systems rather than on subjective judgments of creditworthiness by loan officers in long-term relationships with borrowers. (186) The result was that underwriting was based on information that could be--and, in fact, was--disclosed to loan buyers. Because of these disclosures, investors were well aware of the general changes in the quality of mortgages during the boom period. (187)
These representations about the characteristics of the mortgages were backed up by warranties that obligated the securitizer or originator to repurchase the mortgages at par in the event that the representations were false. (188) Investors thus protected themselves contractually--further evidence against the naive-investors theory. Following the crisis, some large securitizers made sizeable payments to settle claims based on false representations. (189)
Second, a range of mechanisms resulted in substantial risk retention by both originators and securitizers. We have already discussed securitizer risk retention. (190) Originators also faced substantial losses if the mortgages they originated performed poorly or were of a different quality than the quality disclosed. Originators faced the risk of buying back loans that defaulted within a given, warrantied period after sale or for which their representations were inaccurate; (191) the risk that securitizers would "kick out" specific loans following due diligence review, requiring the originator to sell the loans at a substantial discount; (192) the risk that the originator would be unable to sell loans still in the pipeline for securitization should the market cease to accept its loans more generally; (193) and the risk of losing valuable servicing rights retained by the originator. (194) As we discuss in more detail in the next Part, these arrangements caused originators to suffer catastrophic losses when the housing bubble burst. (195) These mechanisms suggest that market participants were quite sophisticated about the incentive problems posed by securitization.
Finally, the moral hazard story does a poor job explaining the timing of the decline in mortgage underwriting standards or why similar problems did not emerge in other securitized asset classes. A vibrant secondary market for private-label MBS was in place long before the subprime boom without apparent incident. (196) If there was an inherent flaw in the incentive structure of mortgage securitization, why did it not manifest earlier? Similarly, many other types of assets are securitized and make use of sophisticated devices designed to mitigate incentive problems. (197) While markets for other types of asset-backed securities (ABS) suffered a loss of liquidity from the general financial crisis, there has been no indication of a breakdown in incentives. Why would the same investors understand the incentive problem and contract for protection for some asset classes but not others? In this sense, the underwriting dynamics and performance of MBS in the early-to-mid 2000s was unique, both relative to the MBS market in the past and to the contemporaneous ABS market. The naive-investors theory cannot explain this difference. In contrast, irrational exuberance in a housing bubble provides a parsimonious explanation for why mortgage underwriting standards specifically eroded during the run-up to the crisis, and why participants on all sides of the market suffered large losses as a result.
To be clear, we are not arguing that securitization played no role in the crisis. To the contrary, securitization was a key factor in the collapse of the financial system as the wave of mortgage defaults burst upon its shores. Securitization ended up concentrating mortgage risk on the balance sheets of large, systemically important financial institutions and creating opacity as to the ultimate bearer of this risk. (198) But the case for securitization causing moral hazard in mortgage underwriting is both theoretically and empirically weak.
In sum, the high levels of risk retention by securitizers during the bubble strongly imply that reforms should focus on reducing securitizers' exposure to the MBS they sponsor, not increasing it. And there is little compelling evidence for the moral hazard problem that the Dodd-Frank Act's risk retention requirement was intended to solve. Accordingly, as a regulatory tool, risk retention should be abandoned.
ABILITY TO REPAY
We turn now to the second pillar of the Dodd-Frank Act's reforms to the mortgage market: the ability-to-repay rule. While at first glance the requirement that lenders make a reasonable determination of the borrower's ability to repay may seem to have little relation to the risk retention requirement, we show that there are important parallels. Both the risk retention requirement and the ability-to-repay rule aim to improve mortgage underwriting through incentives for supposedly sophisticated market participants. As such, in the face of a bubble, the ability-to-repay rule suffers from the same basic problem as the risk retention requirement.
Section 1411 of the Dodd-Frank Act creates a duty for creditors making a residential mortgage loan to make "a reasonable and good faith determination based on verified and documented information that, at the time the loan is consummated, the consumer has a reasonable ability to repay the loan, according to its terms." (199) The statute and its implementing rule provide some guidance on the contours of this duty. Lenders must consider the borrower's credit history, employment status, income, debt-to-income ratio, and assets. (200) More concretely, lenders must verify the borrower's income and assets using third party documentation, such as tax returns, payroll receipts, and bank statements. (201) In calculating the monthly debt payment on the mortgage to determine its affordability, lenders must use a payment schedule of substantially equal payments that fully pays off the loan over its term (202) based on the greater of the "go-to" fully indexed interest rate and any initial introductory "teaser" interest rate. (203) However, the statute and implementing regulations do not mandate any specific standards for how lenders use this information to determine a borrower's ability to repay. They simply impose a duty to make a "reasonable"...