Banking and Trading. By Arnoud W.A. Boot and Lev Ratnovski
IMF Working Paper No. 12/238
Summary: We study the effects of a bank's engagement in trading. Traditional banking is relationship-based: not scalable, long-term oriented, with high implicit capital, and low risk (thanks to the law of large numbers). Trading is transactions-based: scalable, shortterm, capital constrained, and with the ability to generate risk from concentrated positions. When a bank engages in trading, it can use its ‘spare’ capital to profitablity expand the scale of trading. However, there are two inefficiencies. A bank may allocate too much capital to trading ex-post, compromising the incentives to build relationships ex-ante. And a bank may use trading for risk-shifting. Financial development augments the scalability of trading, which initially benefits conglomeration, but beyond some point inefficiencies dominate. The deepending of the financial markets in recent decades leads trading in banks to become increasingly risky, so that problems in managing and regulating trading in banks will persist for the foreseeable future. The analysis has implications for capital regulation, subsidiarization, and scope and scale restrictions in banking.
We study the effects of a bank’s engagement in trading. We use the term “banking” to describe business with repeated, long-term clients (also called relationship banking), and “trading” for operations that do not rely on repeated interactions. This definition of trading thus includes not just taking positions for a bank’s own account — proprietary trading — but also other short-term activities that do not rely on private and soft information, e.g. originating and selling standardized loans. Both commercial and investment banks over the last decade have increasingly engaged in short-term trading. We need to understand the rationale for that, and the challenges that it poses.
Such challenges clearly exist. They are perhaps most vivid in Europe, where some large universal banks seem to have over-allocated resources to trading prior to the crisis, with consequent losses affecting their stability (e.g., UBS, see UBS, 2008; an earlier example is the failure of the Barings Bank due to trading in Singapore in 1995). In the United States, the development of universal banks was until recently restricted by the Glass-Steagall Act. Yet there are many examples of a shift of institutions into shortterm activities, with similar negative consequences. Since early 1980-s, many New York investment banks have turned the focus from traditional underwriting to short-term market-making and proprietary investments; these have often backfired during the crisis (Bear Stearns, Lehman Brothers, Merrill Lynch). Also, in 2000-s, commercial banks have used their franchise to expand into short-term activities, such as wholesale loan origination and funding (Washington Mutual, Wachovia), exposing themselves to risk. And post-Glass-Steagall, there is evidence of trading being a drain on commercial bank activities in newly created universal banks, such as Bank of America-Merrill Lynch. A 2012 loss related to the market activities in JP Morgan is another example. The banks’ short-term activities, especially proprietary trading, have received significant regulatory attention: the Volcker Rule in the Dodd-Frank Act in the U.S., and the Report of the Independent Commission on Banking (the so-called Vickers report) in the UK.
The interaction between banking and trading is a novel topic. The existing literature on universal banks focuses primarily on the interaction between lending and underwriting. Such interaction is relatively well-understood, and also was not at the forefront during the recent crisis. Our paper downplays the distinction between lending and underwriting: for us both could possibly represent examples of long-term, relationshipbased banking. We contrast them to short-term, individual transactions-based activities. We see a shift of relative emphasis towards such “trading” as one of the major developments in the financial sector (for sure prior to the crisis).
The focus on trading as a possibly detrimental activity in banks, and its difference from underwriting in this regard, is supported by emerging empirical evidence. Brunnermeier et al. (2012) show that trading can lead to a persistent loss of bank income following a negative shock. In contrast, underwriting, while more volatile than commercial banking, is not associated with persistent losses of profitability.
The key to our analysis is the observation that the relationship business is usually profitable and hence generates implicit capital, yet is not readily scalable. The trading activity on the other hand can be capital constrained and benefit from the spare capital available in the bank. Accordingly, relationship banks might expand into trading in order to use ‘spare’ capital. This funding (liability-side) synergy is akin to the assertions of practitioners that one can “take advantage of the balance sheet of the bank”.
Opening up banking to trading, however, creates frictions. We highlight two of them. One friction is time inconsistency in the allocation of capital between the longterm relationship banking business and the short-term trading activity. Banks may be tempted to shift too much resources to trading in a way that undermines the relationship franchise. Another friction is risk-shifting: the incentives to use trading to boost risk and benefit shareholders as residual claimants. As a result of these two factors, a bank can overexpose itself to trading, compared to what is socially optimal, or ex ante optimal for its shareholders.
Both problems become more acute when financial markets are deeper, allowing larger trading positions. This increases the misallocation of capital and enables the gambles of scale necessary for risk-shifting. The problems also become more acute when bank returns are lower. Both factors have been in play in the last 10-20 years. Consequently, the costs of trading in banks may have started to outweight its benefits. These frictions are likely to persist for the foreseeable future, so a regulatory response might be necessary.
Full text: http://www.imf.org/external/pubs/cat/longres.aspx?sk=40031.0