Print
E-MailRainer Haselmann is Professor of Finance, Accounting, and Taxation at Goethe University Frankfurt, Christian Leuz is Charles F. Pohl Distinguished Service Professor of Accounting and Finance at the University of Chicago Booth School of Business, and Sebastian Schreiber is Professor of Finance , Accounting and Taxation at Goethe University Frankfurt. This post is based on their recent paper.
Related research from the Program on Corporate Governance includes Insider Trading Via the Corporation (discussed on the Forum here) by Jesse M. Fried.
Since the implementation of the Glass-Steagall Banking Act of 1933 concerns about combined banking operations, i.e., commercial and investment banking within the same universal bank, faded. Notwithstanding, the Global Financial Crisis in 2008 reinvigorated the debate about separating commercial and investment banking, resulting in the Volcker Rule and general requirements for ethical walls. In particular, universal banks could use borrowers’ confidential information when selling securities to investors or trading in capital markets. Despite the persisting political and scholarly debate, it is surprising how little we know about banks’ use of private information and banks’ proprietary trading in general. One reason is that internal information flows cannot be directly observed, and proprietary trading data is not readily available.
This study aims to demystify information transmission within universal banking systems. We combine detailed German data on banks’ proprietary trading with lending data, both provided by German supervisory agencies. Due to the granularity of our data, we are for the first time able to identify instances when banks likely make informed trades in their borrowers’ stocks. Focusing on banks’ trades in stocks of their clients around important corporate events, we find that relationship banks build up positive (negative) trading positions in the two weeks before events with positive (negative) news. This is particularly true when these events are unscheduled, and banks unwind positions shortly after the event. This finding is remarkable, as it should be more difficult to build positions in the ‘right’ direction ahead of unscheduled events.
One challenge for the analysis is that banks may specialize in dealing with certain industries, business models or firms. Such specialisation, and the expertise that comes with it, could manifest in profitable trading, even without any direct information flow from the lending side to the trading desk. To overcome this challenge, we identify corporate events (e.g., legal disputes or mergers) that involve two firms, a borrower and an unrelated third party, with whom the bank has no lending relationship. We then analyze bank trading in the unrelated firm around the joint corporate event. We find that the probability of suspicious trades around the joint event increases. However, relationship banks do not exhibit such suspicious trade patterns around other events of the same unrelated firms when these events do not also involve their borrowers, suggesting that banks have no special expertise in these unrelated firms generally.
Furthermore, we investigate banks’ trading patterns after a bank-firm lending relationship ended. Building up bank expertise for a client takes time and does not disappear immediately when the lending relationship ends. Thus, if bank specialization is the source of a bank’s superior information, profitable trading should not coincide exactly with the duration of the lending relationship. We find that banks have profitable positions around corporate events only when they concurrently have a lending relationship. Even more tellingly, relationship banks’ trading during ’non-relationship periods’ and after a relationship ended does not look significantly different from the trading of non-relationship banks.
A potential channel through which private information can travel within banks is the risk management function. As it is centralised, it requires information on all bank exposures, whether they are from lending or trading. In particular, risk management monitors larger trading exposures as well as short positions. Existing exposures can pose a conflict when the bank receives information about impending corporate events with news in the ’opposite’ direction. Exploiting this mechanism, we find that relationship banks are more likely to unwind (reduce) an existing short (long) position before unscheduled positive (negative) news events. Interestingly, we also find that informed trading results vanish for events with absolute returns greater 10%. These significant events are expected to hit the supervisory radar, which incentivizes banks to shroud their informed trading to avoid supervisory scrutiny. Furthermore, we find that relationship banks usually build up profitable imbalances around corporate events using many small trades rather than a few large ones. These trading patterns arguably indicate that banks actively aim to remain undetected by supervisory institutions when conducting trades around unscheduled abnormal return events.
Related to the shrouding of informed trades, we show that relationship banks obtain worse prices for borrower stocks in OTC trades when the identities of the trading parties are known, as compared to exchange trades, consistent with price protection. Relationship banks in turn respond by building up their suspicious positions mostly on exchanges. This trend highlights that relationship banks persistently obtain worse prices for borrower stocks in OTC trades, when the identities of the trading parties are known, indicating active price protection.
To conclude, our novel evidence suggests that banks engage in profitable proprietary trading ahead of corporate events when they are a firm’s main lender. Using unique micro-level data, we find that relationship banks possessing private information build up significant imbalances before events, even when these events are unscheduled, and unwind positions shortly after. Interestingly, we find evidence suggesting that other market participants are aware of the information advantages of relationship banks. Furthermore, evidence suggests that relationship banks shroud their trades to fly below the supervisory radar and to avoid price protection by other market participants. These findings underscore the potential for conflicts of interest in universal banking, arising for instance through private information leakage in centralized risk management functions. Following the Global Financial Crisis, organizational structures that collect information centrally within banks (i.e.: risk management) have been strengthened globally. Intriguingly, it could be that precisely these organizational structures play an important role in explaining banks’ informed trading patterns.
The complete paper is available for download here.
