Do Banks Always Protect Their Reputation?

A firm’s reputation is a valuable asset. Arguably, conventional wisdom suggests that a reputable firm will always act in the best interest of their clients to preserve the firm’s reputation. For example, in his testimony/defense of Goldman Sachs before Congress, the Chairman and CEO Lloyd Blankfein states, “We have been a client-centered firm for 140 years and if our clients believe that we don’t deserve their trust, we cannot survive.” In our forthcoming Review of Financial Studies article entitled Complex Securities and Underwriter Reputation: Do Reputable Underwriters Produce Better Securities?, we examine the extent to which this conventional wisdom holds with complex securities.

We start by developing a model with asymmetric information where a financial firm (i.e., an underwriter) produces securities whose quality is later observed by investors. With simple securities such as most stocks and bonds, the conventional wisdom holds true: a better reputation will lead underwriters to more accurately represent the value of the securities they sell to their clients. The result is intuitive; if an underwriter were to misrepresent a simple security, it would become evident to market participants as soon as the next financial statements are released, and the underwriter’s loss of reputation would result in the loss of a large amount of future business revenue. As long as these future revenues are large in proportion to the gains from short-term misrepresentation, firms act in their customer’s best interest because it is also in their own best interest.

With complex securities the incentives change. The quality of complex securities such as MBS and CDOs is not directly observable. In fact, it is typically not until the next market downturn that the stress-test scenario occurs where investors can observe whether the security delivers the cash flow promised in the extreme state. Because of the long lag between when the security is typically sold and when the quality is revealed, high reputation firms often have an incentive to cash in on the firm’s reputation by bundling low quality collateral and selling it as purportedly high quality securities. Such a scenario can be amplified by agency conflicts within the firm, although those are not necessary to produce the equilibrium that we describe in the paper as the profits from misrepresentation can be quite large. Although outside of our model, it is interesting to note that the reputation loss and legal liability of the bank will be reduced to the extent that an underwriter can at least partially blame the poor asset performance on unforeseen market conditions.

We empirically evaluate the relation between reputation and performance using a large dataset that covers most of the structured finance universe. Our sample includes over 14,000 deals issued between January 2000 and December 2010, with $10.1 trillion USD in underlying collateral. We do not find any evidence to support the conventional wisdom that more reputable banks issue better performing securities. In fact, we find that investment banks with the highest perceived reputation issue securities that consistently underperform securities issued by low reputation underwriters. This result holds true for non-agency mortgage-backed securities (MBS), asset-backed securities (ABS), and collateralized debt obligations (CDO), across many years of security issuance, and with a barrage of control specifications. A possible alternative explanation for our findings is that the best-informed banks were simply terribly unlucky. Contrary to this possibility, we find that issuance volume by high reputation banks increased after warning signals available to insiders were revealed in early 2007, thus indicating that these banks may have known that the market was likely to face difficulty going forward. They may have done so because it was much more profitable to unload their balance sheet to clients rather than to take the losses themselves. An informal interpretation of our model would suggest that at that point in time the incentives to maintain a reputation are no longer present; underwriters can anticipate that the market might be close to an end and all past actions (and misrepresentation) will finally be revealed. Short-term profit maximization would then be the best possible choice.

In sum, our paper shows that firms of highly perceived reputation might not always produce the best products, nor act in the best interest of their customers. Consistent with Partnoy (1997), our model shows how complexity can be used to misrepresent securities. From a legal perspective, our work suggests that transparency and disclosure are important to counteract complexity and to promote well-functioning markets.

The full paper is available for download here.