New research by Rainer Haselmann, Christian Leuz, and Sebastian Schreiber finds evidence suggesting that German banks with commercial lending relationships improve their trading positions shortly before corporate events and announcements using private firm information.
New evidence suggests that German banks may benefit from information they collect from lending relationships. The findings confront a decades-old debate about whether regulation should prevent banks from both lending and investing activities within the same institution.
To investigate potential conflicts of interest between bank lending relationships and proprietary trading positions, Rainer Haselmann, Christian Leuz, and Sebastian Schreiber gathered data on 168 million trades around 39,994 corporate events at German banks and firms. Their results are striking: the authors find that universal banks whose commercial arm is the main lender to the firm – called ‘relationship banks’ in the paper– are likely to purchase significantly more shares ahead of positive unscheduled corporate events than banks that don’t have a lending relationship. Similarly, they build up negative trading positions ahead of unscheduled events with negative news.
According to James R. Barth, the Lowder Eminent Scholar in Finance at Auburn University, “this finding is significant and has major policy implications. In particular, the result of the authors’ analysis underscores the potential for conflicts of interest in universal banking and would seem to support those who supported the Volcker Rule and the proposed Liikanen Report.”
Universal banks—those with both commercial and investment operations—have a long history of suspicion about whether their structure allows them to make advantageous trades in capital markets based on privileged information from borrowers.
In 1933, the U.S. passed the Glass Steagall Act to separate commercial and investment banks and eliminate the possibility of an unfair advantage in the markets. Over time, concerns about this potential conflict dissolved, and the U.S. repealed the Glass Steagall Act in 1999. The 2008 financial crisis highlighted the issue yet again, as well as excessive risk-taking via proprietary trading by universal banks.
To guard against this, the Volcker Rule was included in the Dodd–Frank Wall Street Reform and Consumer Protection Act, which eliminated proprietary trading by American banks. The European Union considered similar proprietary trading bans around the same time with the Liikanen Report, but instead decided that universal banks must set up organizational structures to separate conflicts of interest from commercial and investment banking.
Regarding universal banks, Ata Can Bertay, Assistant Professor of Finance at Sabanci Business School in Turkey, said, “Documenting relationship banking-related information collection and its use for lucrative trading activities as universal banks has been extremely difficult as these information flows happen within banks, which tend to be secretive organizations.”
Under European Union regulations, trading on inside information is illegal. However, banks can make proprietary trades when their lending practice has private lending information, as long as their organizational structure prevents speculative traders on the investment banking side from accessing that same information. The paper’s aim is to understand how well this organizational structure mitigates sharing private lending information from the commercial side of the bank to the investment side.
An empirical obstacle for the authors’ investigation is to determine whether banks are making profitable trades due to their specialization within the industries or business models to which they lend, or whether they are using private information from their borrowers. The authors address the challenge of differentiating between specialization versus private information in a few ways.
First, they distinguish between known, publicly anticipated corporate events, and unscheduled corporate events that are harder to anticipate. The former might include previously scheduled earnings announcements, while the latter might include profit warnings and announcements of mergers and acquisitions. The results on profitable trading for relationship banks become even more pronounced when they focus on these unscheduled events.
Further, they explore whether banks still trade profitably even after the lending relationship has ended, as expertise should persist but private information flows should not. They find that relationship and non-relationship banks exhibit no differences in trading positions after the former’s lending relationship with a borrower has ended. This finding provides further evidence for the connection between private lending information and relationship banks’ profitable trading positions.
A relationship bank likely receives new, relevant information from its borrowers in certain situations– specifically when granting new loans or when helping with M&A transactions. Thus, the paper examines banks’ trading positions for these events. When borrowers have been granted a new loan in the previous quarter, the paper’s results for favorable trading positions prior to unscheduled corporate events become even stronger. Similarly, consistent with information flows between borrowers and their relationship banks before upcoming M&A transactions, the authors find that there is profitable trading around M&A events when the borrower is a seller, a target, or when the transaction is eventually cancelled.
In addition, the paper explores whether banks trade profitably around events that pertain to their borrowers and other firms, such as when the borrower and an unrelated firm have a joint event, which could include a legal dispute, joint venture or merger. Here, the authors find that banks also trade more successfully in the unrelated firm during the period surrounding the joint corporate event, but do not trade profitably in these unrelated firms for other events that do not involve the banks’ borrowers.
To avoid regulatory scrutiny, banks could shroud their trades. In this regard, the authors find that the evidence is concentrated in events with medium returns, defined as more than two but less than ten percent in absolute returns. In addition, relationship banks build up their seemingly informed positions through many small trades, as opposed to less frequent, larger trades, which are more likely to have price impact and catch the attention of a regulatory agency.
Barth added, “The supervisory authorities may be less concerned with modest returns from such behavior, as the authors seem to confirm. Moreover, modestly profitable trading by banks earned on potential proprietary trading may be less of a focus of supervisors as long as the risk of such behavior is not excessive.”
The question that arises from these results on favorable trade positions and lending relationships, naturally, is how does private firm information flow within these lending institutions?
The paper suggests that the centralization of information within banks regarding their exposures by the risk management department is a possible pathway. The underlying idea is that risk management could passively broadcast information, for instance, by adjusting trade limits, or through decisions on trading positions. Ironically, the authors conclude, the increased centralization of information on risk management that was introduced in response to the financial crisis of 2008 may be the same mechanism through which these suspicious trades take place.
Combining investment and lending activities within a bank may be more efficient, but as Haselmann, Leuz, and Schreiber show, it can lead to conflicts of interest. If governments want to mitigate these conflicts, they face a choice. They can ban proprietary trading or create organizational structures within banks to lower the potential for these conflicts of interest. While the United States implemented the Volcker Rule to eliminate proprietary trading by American commercial banks, the European Union took a different, softer route to preventing conflicts of interest by putting more of the responsibility on banks to create organizational structures and ethical walls. The paper’s results indicate that these internal barriers to information sharing within universal banks are imperfect, which provides useful evidence to policy makers weighing these choices.
Regarding policy implications, Bertay commented, “This unique empirical analysis should make policy makers rethink how they approach the organizational structures of universal banks. Given the ineffectiveness of ethical walls, new regulatory and supervisory approaches should be developed to prevent conflicts of interest in universal banking.”
Read more about our disclosure policy here.