The Routledge Companion to Banking Regulation and Reform

Ismail Ertürk, Daniela Gabor, Ismail Ertürk, Daniela Gabor

• Publisher: Routledge
• Year: 2016
• Print ISBN: 9781032242347
• eBook ISBN: 9781135007140
• Topic: Business
• Subtopic: Business General
• Index: Business

The Routledge Companion to Banking Regulation and Reform provides a prestigious cutting edge international reference work offering students, researchers and policy makers a comprehensive guide to the paradigm shift in banking studies since the historic financial crisis in 2007.

The transformation in banking over the last two decades has not been authoritatively and critically analysed by the mainstream academic literature. This unique collection brings together a multi-disciplinary group of leading authorities in the field to analyse and investigate post-crisis regulation and reform. Representing the wide spectrum of non-mainstream economics and finance, topics range widely from financial innovation to misconduct in banking, varieties of Eurozone banking to reforming dysfunctional global banking as well as topical issues such as off-shore financial centres, Libor fixing, corporate governance and the Dodd-Frank Act.

Bringing together an authoritative range of international experts and perspectives, this invaluable body of heterodox research work provides a comprehensive compendium for researchers and academics of banking and finance as well as regulators and policy makers concerned with the global impact of financial institutions.

Part I

Knowledges of credit risk and bank regulation

1

The credit crisis as a problem in the sociology of knowledge

Donald MacKenzie¹

This chapter analyzes the role in the credit crisis of the processes by which market participants produce knowledge about financial instruments. Employing documentary sources and 87 predominantly oral history interviews, the chapter presents a historical sociology of the clusters of evaluation practices surrounding ABSs (asset-backed securities, most importantly mortgage-backed securities) and CDOs (collateralized debt obligations). Despite the close structural similarity between ABSs and CDOs, these practices came to differ substantially and became the province (e.g. in the rating agencies) of organizationally separate groups. In consequence, when ABS CDOs (CDOs in which the underlying assets are ABSs) emerged, they were evaluated in two separate stages. This created a fatally attractive arbitrage opportunity, large-scale exploitation of which sidelined previously important gatekeepers (risk-sensitive investors in the lower tranches of mortgage-backed securities) and eventually magnified and concentrated the banking system’s calamitous mortgage-related losses.

Introduction

At the heart of the credit crisis that erupted in summer 2007 and culminated in the near collapse of the global banking system in the fall of 2008 were complex, esoteric financial instruments. At the peak of the crisis, in October 2008, the International Monetary Fund (IMF) categorized the estimated $1.4 trillion losses that, were it not for massive international government intervention, would most likely have caused an economic catastrophe on the scale of the Great Depression. More than half the total, $770 billion, was in mortgage-backed securities, asset-backed securities (ABSs) of other kinds, and collateralized debt obligations (CDOs).² The largest single category of loss, $290 billion, was in a class of instruments of which many outside the financial sector had simply been unaware prior to the crisis: ABS CDOs, in other words collateralized debt obligations whose underlying assets are tranches of asset-backed securities, most commonly mortgage-backed securities (IMF, 2008, table 1.1, p. 9). Not only were the sums lost on ABS CDOs very large, but (as discussed in this chapter’s fifth section) the losses were concentrated at the very core of the global financial system. ABS CDOs also had wider effects. The “assembly lines” via which they were constructed reshaped the underlying market for mortgage-backed securities in ways that facilitated ever looser mortgage underwriting. Those losses and these processes were by no means the only causes of the credit crisis, but to understand it fully we need to understand ABS CDOs, to grasp how they emerged from the world of mortgage-backed securities and the (cognitively quite different and organizationally largely separate) world of CDOs, and above all to develop a sociological analysis of how these complex financial instruments were evaluated by market participants. For example, differences between how market participants evaluated ABSs and evaluated CDOs, and the location of those evaluations in different groups or departments of credit rating agencies and banks, had a double effect. In a situation in which investment behavior was largely governed by credit ratings, they made the construction of ABS CDOs highly profitable. Simultaneously, however, they left the ABS CDO a kind of epistemic orphan, cognitively peripheral to both its parent worlds, ABSs and corporate CDOs.

In its emphasis on evaluation,³ this chapter contributes to a growing body of work in economic sociology that shows the importance and richness of what Beckert (2009, pp. 253–4) calls “the value problem,” in other words “the processes of classification and commensuration with which actors assign value to goods.” As Carruthers and Stinchcombe (1999, p. 353) point out, “buyers and sellers” need “to know the commodities they transact in,” and the ease with which those commodities are bought and sold is, therefore, “among other things, an issue in the sociology of knowledge.”

Carruthers and Stinchcombe focus on a particular set of knowledge-generating arrangements, to be found, for example, in the trading of the shares of large corporations, that one might call the “canonical mechanism.” This involves the standardization of the financial claims or other commodities being traded, continuous auctions coordinated either by an exchange or by dealers who act as “market makers,”⁴ and wide dissemination of the resultant prices. These arrangements are, as Carruthers and Stinchcombe show, powerful generators of public knowledge, but they are also limited in their scope, even in their primary domain, the financial markets. The ABS and CDO tranches discussed here were not, in general, traded in canonical-mechanism markets. They were usually bought directly from those who had constructed them, who frequently were dealers based at major international banks, and in many cases then simply retained by the purchasers. Secondary trading of them was on a limited scale and was always “over-the-counter” (conducted by direct institution-to-institution negotiation) rather than on an organized exchange. Even in the limited cases in which some of these instruments were made sufficiently standard that canonical-mechanism trading was possible, there was an undercurrent of dissent, touched on in the penultimate section below, about whether the publicly quoted prices of them were fully reliable and legitimate.

In consequence, this is a case in which the analysis of the “social processes behind the constitution of value” (Beckert, 2009, p. 254) needs to look beyond the canonical mechanism. There is a substantial body of work by economic sociologists on these processes, mainly concerning contexts outside the financial markets and often – though not always – goods and services that are “singular” (Karpik, 2010): not straightforwardly commensurable. The situations on which this literature has focused include those in which the legitimacy of a product or of monetary valuation is contested (see, e.g. Zelizer, 1979 on life insurance and Zelizer, 1994 on children); incommensurable forms of evaluation or “orders of worth” contend (Boltanski and Thévenot, 2006; Stark, 2009): perceptions of value interact with aesthetic judgments (e.g. Velthuis 2005; Aspers 2005); the quality of a product is inferred from the status of its producer (e.g. Podolny 1993, 2005; see Aspers 2009); or the value of a commodity to one buyer depends directly on anticipation of its value to other buyers (as in the case of dot.com stocks or houses bought in the anticipation of selling them to others at a higher price).⁵

ABSs and CDOs are not valued for their aesthetic properties, and the moral legitimacy of monetary valuation of them has never been challenged. With those exceptions, however, all the phenomena listed in the previous paragraph can be found in respect to ABSs and CDOs, and I return to two of them in the conclusion. However, the main way in which the evaluation of ABSs, CDOs, and ABS CDOs contributed to the crisis concerns the apparently “technical” core of evaluation. ABSs, CDOs, and ABS CDOs are debt instruments. They normally entitle investors (a) to defined “coupons” (interest payments), set either as a fixed percentage or as a fixed margin or “spread” over a benchmark interest rate such as Libor (London Interbank Offered Rate) and (b) to eventual repayment of principal (their initial capital investment). The monetary worth of an investment in an ABS or CDO is thus the aggregate present value of those future payments. If the payments were entirely certain, the valuation of an ABS or CDO would be a matter simply of arithmetic, but they are not. There are two main risks: default (in other words that the payments are not made or not made in full) and prepayment (i.e. principal is repaid earlier than anticipated, in a situation in which it can be reinvested only at a lower rate of interest). This chapter’s focus is on whether and how those risks were taken into account in the evaluation of ABSs, CDOs, and ABS CDOs.

How might “technical” processes of evaluation of this kind be analyzed sociologically? This chapter draws its inspiration from studies of scientific practice. Historians and sociologists have found that practice to be far less uniform than traditional notions of a unitary “scientific method” might suggest (see, e.g. Galison and Stump 1996) and have sought to capture distinctive clusters of practice in notions such as the “local scientific cultures” of Barnes, Bloor, and Henry (1996), the “subcultures” and “competing traditions” of Galison (1997), the “experimental cultures” of Rheinberger (1997), “epistemic cultures” of Knorr Cetina (1999), “epistemological cultures” of Fox Keller (2002), and “evidential cultures” of Collins (2004).

Can similar patterned differences in evaluation practices be found in financial markets?⁶ This chapter suggests that they can,⁷ using as its main evidence differences between the evaluation of ABSs and of CDOs, which are structurally very similar instruments (indeed sometimes simply lumped together, as, e.g. by McDonald and Robinson, 2009). In evaluation, as in scientific practices, one can find “aggregate patterns and dynamics that are on display in expert practice and that vary in different settings of expertise … patterns on which various actions converge and which they instantiate and dynamically extend” (Knorr Cetina 1999, pp. 8–9). Let me call these patterns “clusters of evaluation practices.” (Following the literature on science and calling them “evaluation cultures” might be taken to imply greater homogeneity and “bounded-offness” of their practitioners than is the case.⁸ It could also be taken wrongly as implying a theory of action as based solely on “belief ” and “habit” – for which see Camic, 1986 – rather than self-interested, reflexive rational choice. As discussed in the conclusion, belief and habit were present, but by no means exclusively so.)⁹

The research on which this chapter is based, which is outlined at the end of this introduction, supports six postulates about these clusters.¹⁰ First, clusters of evaluation practices are the path-dependent outcomes of historical contingencies.¹¹ For example, while the evaluation practices surrounding CDOs always had default risk as their primary object, those surrounding mortgage-backed securities were concerned primarily with prepayment. As the following section will show, that latter focus originally arose because of features of the political economy of mortgage lending in the United States that can be traced back to the 1930s. The focus on prepayment remained in place even in the very different circumstances of the past decade: it formed a criterion on which that decade’s subprime mortgage-backed securities were judged superior to their prime counterparts. In emphasizing long-lasting effects such as this, I do not want to suggest that evaluation practices never change. They do – change in them is a major focus of this chapter – but the way in which they change is path-dependent: it is easier, for example, to modify an existing practice than to develop an entirely new one.

Second, the more elaborate of evaluation practices give rise to, and are informed by, distinctive ontologies: distinctive presuppositions about the nature and properties of the features and processes of the economic world. Thus the third section of the chapter will show that the evaluation practices surrounding CDOs came to be oriented heavily to one such feature, “credit correlation” (a term that will be explained in that section), which was a notion entirely absent, at least in any explicit form, in the evaluation of ABSs. Like many scientific objects, correlation was neither simply “real” nor simply “fictional” (Knorr Cetina, 1999, pp. 248–52). It was not observable in any straightforward sense: to invoke it was to invoke the unseen. Yet, like the scientific objects analyzed by Daston (2000), it had the potential to become “more real,” as specific markets (the tradable credit indices described below) were created in which its effects were more easily traced. Indeed, some of those involved with CDOs came to hold that in those markets correlation was not just real but tradable. For others, though, the frustrating difficulties of measuring correlation indicated that it was a misconception, an artifact of inadequate models.

Third, evaluation practices become organizational routines, and when different practices are pursued in the same organization, they frequently are the province of separate parts of it.¹² For instance, the evaluation of ABSs on the one hand and CDOs on the other typically became the responsibility of different sections of banks, of the specialist “monoline” insurers, and of credit rating agencies. In the case of the rating agencies, for example, both ABSs and CDOs fell within the remit of their structured finance departments, but the latter had separate groups dealing with each. When ABS CDOs (which are CDOs with ABSs nested within them, so to speak) came into being, the decision as to how to evaluate them was thus also a decision about how their evaluation should be mapped onto the organizational structure of rating agencies. All the three main agencies – Moody’s, Standard & Poor’s (S&P), and Fitch – found the same solution: they relied on the existing ratings, by ABS groups, of the component ABSs, and assigned the analysis of the higher-level structure to CDO groups. Those groups analyzed that structure largely as if it was simply another variant of a CDO, for which existing practices were therefore appropriate, rather than treating an ABS CDO as a radically different instrument that demanded new evaluation techniques.

Fourth, in modern debt markets (in which I include the markets for bonds, tradable loans, and structured instruments such as ABSs and CDOs) evaluation practices regulate actions and become means of governance via the process of credit rating.¹³ Ratings (see Figure 1.1) encode rating agencies’ conclusions about either the likelihood of default on debt instruments (in the case of S&P and Fitch) or, in the case of Moody’s, the expected loss on them (the likelihood of loss multiplied by its severity). For institutional investors such as banks, insurance companies, and pension funds (private individuals were never major participants in the ABS and CDO markets discussed here), ratings frequently become rules. Cantor, ap Gwilym, and Thomas (2007, p. 14) note that in the United States “there are currently over 100 federal laws and 50 regulations incorporating credit ratings,” and they report that the purchases of 74 percent of their sample of US investment fund managers (and 78 percent of European managers) were subject to a minimum-rating requirement: if an instrument’s rating was below the minimum, they were not allowed to buy it. Especially toward the end of the period discussed here, banking regulation in particular relied heavily on ratings, with banks able to hold much smaller capital reserves in respect to instruments with high ratings, a factor that greatly enhanced the attractiveness of the most senior tranches of the instruments discussed here.

Fifth, evaluation practices crystallized in ratings reduce a difficult problem of evaluation (assessing complex, novel financial instruments that involve potentially uncertain payments stretching years into the future) to a simple one, by establishing a rough equivalence among debt instruments of different kinds and with different particularities. Though some buyers of ABSs and CDOs had a good understanding of the detail of evaluation pr...

Table of contents

• Cover
• Half Title
• Title Page
• Copyright Page
• Table of Contents
• List of figures
• List of tables
• List of contributors
• Acknowledgements
• Introduction
• PART I Knowledges of credit risk and bank regulation
• PART II Critical perspectives on financial innovation
• PART III New approaches to banking, risk and central bank role in the Eurozone
• PART IV Regulation of misconduct in banking
• PART V Limits of post-crisis bank regulation
• PART VI Dysfunctional global finance and banking reform
• Index