影子银行

Forthcoming Oxford Handbook of Banking, 2nd edition August 21, 2013

Commercial Banking and Shadow Banking: The Accelerating Integration of Banks and

Markets and its Implications for Regulation

by

Arnoud W.A. Boot*

and

Anjan V. Thakor**

ABSTRACT

In this chapter we examine the implications of the increasing integration of banks and financial markets for the evolution of each as well the design of prudential regulation. The growth of the use of depository-banking-originated mortgage-backed securities sold in the market and used as collateral to back short-term funding transactions in the shadow banking sector illustrates this intertwining of banks and markets, and the potential systemic risk dangers of this, as is apparent from the recent crisis. An important implication of this integration is that it is becoming more and more difficult to isolate banking risks from financial market risks. Building on contemporary theories of financial intermediation, relationship banking and securitization, we show how these theories inform us about the underlying causes of the seismic shifts that have occurred in financial markets in recent years, and what they suggest about the kinds of regulatory reforms that may be most fruitful. We end with a list of open research questions suggested by our analysis.

KEYWORDS:

Relationship banking, financial markets, integration, shadow banking, regulation.

*University of Amsterdam and CEPR

**John E. Simon Professor of Finance, Director of the PhD Program, and Director of the WFA Center for Finance and Accounting Research, Olin School of Business, Washington University in St. Louis and ECGI

Commercial Banking and Shadow Banking: The Accelerating Integration

of Banks and Markets and its Implications for Regulation

Arnoud W. A. Boot

Anjan V. Thakor

Introduction

THE financial sector has evolved rapidly over the last decade, with the impetus for change provided by deregulation and advances in information technology. Competition has become more intense. Interbank competition within domestic markets as well as across national borders and competition from financial markets have gained importance. Both the institutional

structure of financial institutions and the boundary between financial institutions and financial markets have been transformed. At no stage has this blurring of boundaries been more evident than during the events leading up to the financial crisis, events that have highlighted how large the shadow banking sector has become.1Pozsar, Adrian and Ashcraft (2010) estimate the size of the shadow banking system in the U.S. $16 trillion in 2010, but estimates (and measures) vary greatly (see Claessens, Pozsar, Ratnovski and Singh, 2012). This chapter reviews the literature related to these developments and uses it to examine the importance of this changing landscape for the structure of the financial services industry and the design and organization of regulation.

As we will argue, the increasingly intertwined nature of banks and financial markets is not without costs. In particular, as the financial crisis of 2007–9 has illustrated, systemic risks may have become more prevalent. In this chapter, we seek to provide a fundamental analysis of the underlying forces that could explain the evolution of the banking industry. We begin by

discussing the key insights from the financial intermediation literature, including the potential complementarities and conflicts of interest between intermediated relationship banking activities and financial market activities (underwriting, securitization, etc.). While debt

contracts dominate the financial intermediation literature, the impressive growth of privateequity firms has turned the spotlight on equity. In a sense, one could interpret private equity (PE) as intermediation driven from the equity side. Given their economic functions as debt and equity intermediaries, respectively, how do banks and PE firms interact?

Our discussion reveals that the interaction between banks and PE firms is only one aspect of an increasing integration of banks and markets. Banks have a growing dependence on the financial markets not only as source of funding but also for hedging purposes and offloading risks via securitization, and possibly for engaging in proprietary trading. Financial market linkages often also imply that intra-financial sector linkages mushroom, e.g. the asset-backed securities created by securitization can serve as collateral that financial institutions use to fund themselves in the shadow banking system. The multiple dimensions of bank dependence on markets generate both risk reduction and risk elevation possibilities for banks. For example, while hedging may reduce risk, proprietary trading, liquidity guarantees for securitized debt, and positions in credit default swaps can increase risk. This raises potential regulatory

concerns. What do these developments imply for prudential regulation and supervision? Will the increasing interactions between banks and markets increase or decrease financial systemfragility? The financial crisis of 2007–9 suggests an increase in fragility, but how much can we generalize from this crisis? These questions have become particularly germane not only because of growing banks-markets integration, but also due to the (up to recently) growing cross-border footprint of financial institutions.

These developments have also focused attention on the role of ‘gatekeepers’ (Coffee, 2002), like credit rating agencies. While the financial intermediation literature has

acknowledged the role of credit rating agencies as information processors and sellers for some time now (e.g. Allen, 1990 and Ramakrishnan and Thakor, 1984), the literature has not discussed how rating agencies may affect the fragility of the financial sector through theimportant role they play as ‘spiders in the web of institutions and markets’. We take up this issue in our discussion.

The organization of the chapter is as follows. In the next section, we focus on the economic role of financial intermediaries. The primary focus here is on the banks' role in lending and how this compares to non-intermediated finance directly from the financial market. We will also analyze the effects of competition on the banks' lending relationships. Does competition harm relationships and reduce their value and hence induce more transaction-oriented banking, or does competition augment the value of relationships? This discussion will summarize the key insights from the modern literature of financial intermediation. In the next section we discuss the increasingly interconnected nature of banks and financial markets, with a focus on securitization. This ‘technology’ has been at the center of the 2007–9 financial crisis. What are the future prospects for securitization? The proliferation of non-banking financial institutions, and particularly private equity firms, is discussed in the following section. We will argue that much of this activity is complementary to the role of banks, rather than threatening their raisond?être. Subsequently, we focus on the role of credit rating agencies. These agencies have been indispensable for the explosive growth (and temporary demise) of securitization. How will their role develop? We then discusses regulatory implications. Here we link the role of banks in lending (as emphasized in our earlier discussions) to their role as providers of liquidity. This brings in the issue of fragility which is at the heart of the current regulatory debate.

Understanding banks as information-processing intermediaries

In this section we discuss two issues: (1) what is the key role of banks vis-à-vis markets? and (2) how does competition impinge on this role?

The economic role of banks

We first discuss the role of banks in qualitative asset transformation—i.e., the process by which banks absorb risk to transform both the liquidity and credit risk characteristics of assets (see Bhattacharya and Thakor, 1993). For example, banks invest in risky loans but finance them with riskless deposits (e.g. Diamond, 1984, Millon and Thakor, 1985, and Ramakrishnan and Thakor, 1984). They also invest in illiquid loans and finance them with liquid demandable deposits (e.g. Diamond and Dybvig, 1983). The theory of financial intermediation has placed special emphasis on the role of banks in monitoring and screening borrowers in the process of lending. Bank lending is typically contrasted with direct funding from the financial markets.

What are the comparative advantages of bank loans over public capital market bond financing? The most striking insight of the contemporary theory of financial intermediation is that banks are better than markets at resolving informational problems. The possession of better information about their borrowers allows banks to get closer to their borrowers. Interestingly, a feedback loop is generated as this proximity between the financier and the borrowing firm in bank lending arrangements may also help mitigate the information asymmetries that typically plague arm's length arrangements in market transactions. This has several aspects. A borrower might be prepared to reveal proprietary information to its bank that it may have been reluctant to reveal to the financial markets (Bhattacharya and Chiesa, 1995). A bank might also have better incentives to invest in information acquisition. While costly, the substantial stake that it has in the funding of the borrower and the enduring nature of its relationship with the borrower—with the possibility of information reusability over time—increase the marginal benefit of information acquisition to the bank.2 Boot and Thakor (2000) analyze the economic surplus that relationship banking can generate.

Such borrower-lender proximity may also have a dark side. An important one is the hold-

up problem that stems from the information monopoly the bank may develop due to thespontaneous generation of proprietary information on borrowers. Such an informational monopoly may permit the bank to charge higher loan interest rates ex post (see Sharpe, 1990; Rajan, 1992; and Boot, 2000, for a review). The threat of being ‘locked in’, or informationally captured by the bank, may dampen loan demand ex ante, causing a loss of potentially valuable investment opportunities. Alternatively, firms may opt for multiple bank relationships (see Carletti, Cerasi, and Daltung, 2007). This may reduce the informational monopoly of any individual bank, but possibly at a cost. Ongena and Smith (2000) show that multiple bank relationships indeed reduce the hold-up problem, but can worsen the availability of credit (see Thakor, 1996 for a theoretical rationale).

Another aspect is that relationship banking could accommodate an intertemporal smoothing of contract terms (see Allen and Gale, 1995, Allen and Gale, 1997, and Boot and Thakor, 1994) that would entail losses for the bank in the short term that are recouped later in the relationship. Petersen and Rajan (1995) show that credit subsidies to young or ‘de novo’ companies may reduce the moral hazard problem and informational frictions that banks face in lending to such borrowers. Banks may be willing to provide such subsidized funding if they can expect to offset the initial losses through the long-term rents generated by these borrowers. The point is that, without access to subsidized credit early in their lives, ‘de novo’ borrowers would pose such serious adverse selection and moral hazard problems that no bank would lend to them. Relationship lending makes these loans feasible because the proprietary information generated during the relationship produces ‘competition-immune’ rents for the bank later in the

relationship and permits the early losses to be offset. The importance of intertemporal transfers in loan pricing is also present in Berlin and Mester (1999). They show that rate-insensitive coredeposits allow for intertemporal smoothing in lending rates. This suggests a complementarity between deposit taking and lending. Moreover, the loan commitment literature has emphasized the importance of intertemporal tax subsidy schemes in pricing to resolve moral hazard (seeBoot, Thakor, and Udell, 1991 and Shockley and Thakor, 1997) and also the complementarity between deposit taking and commitment lending (see Kashyap, Rajan, and Stein, 1999).

The bank-borrower relationship also displays greater contractual flexibility than that

normally encountered in the financial market. This flexibility inheres in the generation of hard and soft proprietary information during a banking relationship. This information gives the bank the ability to adjust contractual terms to the arrival of new information and hence encourages it to write ‘discretionary contracts’ ex ante that leave room for such ex post adjustments. This is in line with the important ongoing discussion in economic theory on rules versus discretion, where discretion allows for decision-making based on more subtle—potentially non-

contractible—information. See, for example, Simon (1936), and Boot, Greenbaum, and Thakor(1993).

The papers by Stein (2002), and Berger, Miller, Petersen, Rajan and Stein (2005) highlight the value of ‘soft information’ in lending. This could be an example of this more subtle and non-contractible information. On this issue, two dimensions can be identified. One dimension is related to the nature of the bank-borrower relationship, which is typically long-term, with accompanying reinforcing incentives for both the bank and the borrower to enhance the durability of the relationship. This allows for implicit—non-enforceable—long-term

contracting. An optimal information flow is crucial for sustaining these ‘contracts’. Information

asymmetries in the financial market, and the non-contractibility of various pieces of

information, would rule out long-term alternative capital market funding sources as well as explicit long-term commitments by banks. Therefore, both the bank and the borrower may realize the added value of their relationship, and have an incentive to foster the relationship.3

The other dimension is related to the structure of the explicit contracts that banks can write. Because banks write more discretionary contracts, bank loans are generally easier to

renegotiate than bond issues or other public capital market funding vehicles (see Berlin andMester, 1992). Such renegotiability may be a mixed blessing because banks may suffer from a ‘soft-budget constraint’ problem: borrowers may realize that they can renegotiate ex post, which could give them perverse ex ante incentives (see Bolton and Scharfstein, 1996 and Dewatripont and Maskin, 1995). The soft-budget-constraint problem is related to the potential lack of toughness in enforcing contracts due to the ex post distribution of ‘bargaining power’ linked with relationship banking proximity (see Boot, 2000). In practice, one way that banks can deal with this issue is through the priority structure of their loan contracts. If the bank has priority/seniority over other lenders, it could strengthen the bank's bargaining position and allow it to become tougher. These issues are examined in Diamond (1993), Berglöf and von Thadden (1993), and Gorton and Kahn (1993).

The bank could then credibly intervene in the decision process of the borrower when it believes that its long-term interests are in jeopardy. For example, the bank might believe that the firm's strategy is flawed, or a restructuring is long overdue. Could the bank push for the restructuring? If the bank has no priority, the borrower may choose to ignore the bank's wishes. The bank could threaten to call the loan, but such a threat may lack credibility because the benefits of liquidating the borrower's assets are larger for higher-priority lenders, and the costs from the termination of the borrower's business are higher for lower-priority lenders. When the bank loan has sufficiently high priority, the bank could credibly threaten to call back the loan, and this may offset the deleterious effect of the soft-budget constraint. This identifies a

potential advantage of bank financing: timely intervention. Of course, one could ask whether bondholders could be given priority and allocated the task of timely intervention. Note that bondholders are subject to more severe information asymmetries and are generally more

dispersed (i.e., have smaller stakes). Both characteristics make them ill-suited for an ‘earlyintervention’ task.

Intermediation and competition

Since relationship banking is an integral part of the economic services provided by banks and generates rents for banks, it also potentially invites multiple bank entry, which then generates interbank competition. An interesting question this raises is how competition might affect the incentives for relationship banking. While this may ultimately be an empirical question, two diametrically opposite points of view have emerged theoretically. One is that competition among financiers encourages borrowers to switch to other banks or to the financial market. The consequent shortening of the expected ‘life-span’ of bank-borrower relationships may induce banks to reduce their relationship-specific investments, thereby inhibiting the reusability of information and diminishing the value of information (Chan, Greenbaum, and Thakor, 1986). Banks may then experience weaker incentives to acquire (costly) proprietary information, and relationships may suffer. There is empirical evidence that an increase in relationship length benefits the borrower. Brick and Palia (2007) document a twenty-one-basis point impact on the