Trading by banks was a major factor in the recent crisis. Market-based activities – trading in, or holding securitised debt instruments – led to the failures of major universal banks in Europe (RBS and UBS among the largest) and of both investment and commercial banks in the US (Bear Stearns, Lehman Brothers, Merrill Lynch, Washington Mutual, Wachovia).
Since the crisis, trading remains a major source of instability. It caused repeated losses in banks (including $6.8 billion in JP Morgan in 2012). And commentators argue that trading is a drain on resources in universal banks, such as Bank of America-Merrill Lynch (which moved risky trading exposures to the commercial bank unit). Emerging empirical evidence (Brunnermeier et al. 2012) confirms the significant risks of trading by banks.
Trading can destabilise banks: Our conclusions
Do the recent trading-related bank failures represent one-off phenomena (maybe related to the crisis), or are they a sign of deeper structural problems in the financial system? An answer to this question would determine the optimal policy response.
Our recent paper (Boot and Ratnovski 2012) suggests that it is the latter. We argue that the deepening of financial markets in the last 10 to 15 years has fundamentally destabilised banks by inducing trading. Specifically, banks have incentives to use their franchise value to trade on a large scale. This gives rise to two key negative effects:
a) Banks trade too much. Points at a misallocation of capital, in part at the expense of lending. This is detrimental for real economic activity.
b) Bank trading is too risky. This is intrinsically linked to point (a): large trading positions encourage risk-shifting. This leads to failures and financial instability.
These (reinforcing) negative effects were not present historically when financial markets were not as deep, but will remain so in the foreseeable future. Without policy action, crises associated with trading by banks are bound to recur. Even strong supervision will not be able to prevent them. Consequently, it appears necessary to restrict trading by banks.
Towards a comprehensive policy response
Given the implications for systemic risk, trading by banks has received significant regulatory attention. Most noteworthy initiatives are the Volcker Rule in the Dodd-Frank Act in the US, the recommendations of the Independent Commission on Banking (the so-called Vickers report) in the UK, and the recent Liikanen report to the European Commission.
Despite these initiatives, the policy response appears slow and inconclusive. Implementation details are not yet worked out; compliance timelines, if at all present, are very long. Also, while the problem is common to many countries, there is a large international heterogeneity in response, which can compromise the effectiveness of national measures.
The lack of a comprehensive policy response is in part surprising. Some can be attributed to lobbying and legislative frictions. But there is also a deeper, conceptual problem. Economists and policymakers lack a good understanding of the economic forces at play. This makes it very difficult to formulate an unambiguously effective, let alone optimal, policy.
Our paper (Boot and Ratnovski 2012) attempts to fill this gap. We study the economics of trading by banks, and particularly the interaction between banks’ trading and relationship-based activities, and highlight some key market failures. The analysis allows unique insights into the optimal structural policy in banking.
There are three fundamental questions. First, why do banks engage in trading? Second, what are the possible market failures? Third, why has trading by banks become such a significant problem recently?
Traditional banking is a long-term relationship-based business, focused on repeated interactions with customers. Trading by banks can be defined broadly as any short-term (not based on repeated interactions) activities. Thus, fundamentally, trading includes not only taking positions for a bank’s own account (proprietary trading), but also, for example, originating, selling, or holding standardised loans.
The key to our analysis is the observation that the traditional banking business is usually profitable, yet not readily scalable. The trading activity, on the other hand, is often capital constrained so can benefit from the capital of the bank, and is scalable. Accordingly, banks can expand into trading in order to use their “spare” capital. This synergy is akin to the assertions of practitioners that one can 'take advantage of the balance sheet of a bank”.
Opening up banks to trading, however, creates frictions (market failures). One such friction is the misallocation of capital. Banks may opportunistically shift too many resources to trading in a way that undermines their relationship franchise. This happens, for example, when banks chase short-term opportunities in financial markets and end up using the risk bearing capacity necessary for their core business. Another friction is risk-shifting: banks may use trading to boost risk to benefit shareholders. As a result, banks trade too much, and in a too risky a fashion, compared to what is socially optimal.
Importantly, trading by banks becomes more distortionary in deeper financial markets, which allow larger trading positions (increasing the misallocation of capital and enabling larger-scale gambles). Trading also becomes more distortionary when returns in the traditional relationship-based banking business are lower.
Implications: The dynamics of trading in banks
The simple observations above offer a very fundamental implication. In the last 10 to 15 years, financial markets have deepened substantially and traditional banking has become less profitable. The two trends had the same driver: information technology has increased the availability of hard information, expanding the universe of tradable claims and making banking more contestable. This means that while trading in banks was benign and contained before, it has irreversibly become more distortionary now.
To put it in starker terms: because of financial development, the business model where a bank combines core relationship operations with a transactional activity – be it in traditional European universal banks, in commercial banks that hold securitised products, or in investment banks – is no longer sustainable.
Trading in modern banks opens the door to risk-shifting and hence will lead to bank failures. Trading also leads to a misallocation of resources from lending; this makes banks unable to maintain relationships and leads to a reduced supply of customer-oriented services, such as SME lending. Trading then also compromises the role of banks as providers of liquidity during economic slowdowns (Kashyap at al. 2002).
Knowing the market failures – the misallocation of resources to trading and the potential for risk shifting – helps inform policy design. Specifically, we suggest the following:
- Which activities to restrict? The study suggests that risks are posed by transaction-based activities of banks. This offers two implications.
(a) Restrictions may need to cover more than proprietary trading (i.e. be wider than the core part of the Volcker rule). Other transactional activities, such as buying and holding securitised debt, pose similar threats (cf. Washington Mutual) and may need to be restricted.
(b) There is little justification for restricting customer-oriented investment banking activities, such as underwriting (so the restrictions can be narrower than the Vickers proposals). In fact, ample empirical evidence points to synergies between lending and underwriting.
- Segregate or prohibit? Segregation (as proposed by Vickers) can discourage overly risky trading, and is a necessary first step. But the study suggests that, even then, banks may still be able to allocate too much capital to their trading subsidiaries, leaving lending constrained. So it is important to protect capital and risk bearing capacity of bank lending operations. For this, trading within bank groups may have to be limited or prohibited altogether (as suggested by Volcker).
- What about hedging? The study suggests that trading at low scale does not create negative effects. At low scale, there is little misallocation of capital, and it is impossible to use small trading positions for risk-shifting. Hence the approach of allowing a limited (and sufficiently small) percentage of bank capital to be put to trading risk (as in the Liikanen report) might be appropriate.
- Can trading move to the shadow and become even riskier? Our analysis suggests that this is unlikely. Trading by banks is particularly risky because available implicit bank capital (i.e. rents coming from activities other than trading) enables and induces trading at large scale. If trading is removed from banks, it will resemble that in hedge funds – at smaller scale and with lower threats to financial stability.
- What about capital regulation and cyclicality? The study suggests that restricting trading by banks will free up capital and mitigate the pro-cyclical effects of capital regulation. So restrictions on trading are complimentary to the Basel III initiatives. Also, trading by banks, when not fully restricted, can be charged punitive capital to discourage it, or at the very least, have it internalise its risk.
Editors' note: The views expressed in this column are those of the authors and should not be attributed to the IMF, its Executive Board, or its management.
Boot, AWA and L Ratnovski (2012), “Banking and Trading”, IMF Working Paper 12/238.
Brunnermeier, M, G Dong, and D Palia (2012), “Banks' Non-Interest Income and Systemic Risk”, Working Paper.
Kashyap, AK, R Rajan, and J Stein (2002), “Banks as Liquidity Providers: An Explanation for the Coexistence of Lending and Deposit-Taking”, Journal of Finance, 57(1):33-74.