Finding Systemically Important Financial Institutions ... ... Finding Systemically Important...

download Finding Systemically Important Financial Institutions ... ... Finding Systemically Important Financial

of 54

  • date post

  • Category


  • view

  • download


Embed Size (px)

Transcript of Finding Systemically Important Financial Institutions ... ... Finding Systemically Important...

  • Finding Systemically Important Financial Institutions around the Global Credit Crisis: Evidence from Credit Default Swaps

    Jian Yang

    The Business School, University of Colorado Denver, Denver, CO 80217

    Yinggang Zhou The Faculty of Business Administration, Chinese University of Hong Kong, Hong Kong

    This draft: September 16, 2010

    Abstract With an international dataset of credit default spreads as a credit risk measure, we propose a novel empirical framework to identify the structure of credit risk network across major financial institutions around the recent 2007-2008 global credit crisis. The findings directly shed light on credit risk transmission in a financial network and help find systemically important financial institutions from the perspective of interconnectedness. Specifically, Lehman Brothers, Morgan Stanley, Sefeco, Chubb, and possibly AIG in the US and BNP Paribs, Dresdner bank, and UBS in the Europe are primary senders of credit risk information. Goldman Sachs, Bear Sterns, Bank of America, and Metlife in the US and Barclays, RBS, Commerzbank, and HVB in the Europe play the role of the exchange center on the credit market by intensively receiving from some financial institutions and then transferring credit risk information to others. Finally, Citigroup, Wachovia, JPMorgan and Hartford in the US, and ABN AMRO, ING, Rabobank, and Deutsche Bank in the Europe appear to be prime receivers of credit risk information. Further analysis shows that leverage ratios and certain aspect of corporate governance (i.e., CEO duality) may be significant determinants of identified different roles of financial institutions in credit risk transfer, while no such evidence is found for other factors including size, liquidity and asset write-downs.

    Key Words: credit risk; financial network; directed acyclic graphs; structural VAR JEL Classifications: G01, G15, G32 We gratefully acknowledge helpful comments from Warren Bailey, Andrew Karolyi, David Ng, Meijun Qian (the discussant), Hao Zhou, Haibin Zhu, and seminar/conference participants at Central Bank of China, Chinese Academy of Science, University of Colorado Denver, and National University of Singapore 4th Annual Risk Management conference. *Corresponding author. Email: Tel: (303) 315-8423; Fax: (303) 315- 8084

  • 1

    Finding Systemically Important Financial Institutions around the Global

    Credit Crisis: Evidence from Credit Default Swaps

    1. Introduction

    During the 2007-2008 global credit crisis, the comovement of financial institutions’

    assets and liabilities increased dramatically. Such an increase of comovement gives rise to the

    systemic risk that institutional distress may spread widely and distort the supply of credit and

    capital to the real economy (Adrian and Brunnermeier, 2009). Understanding the nature of the

    systemic risk is the key to understanding the occurrence and propagation of financial crises

    (Allen, Babus, and Carletti, 2009). According to Allen, Babus, and Carletti (2009), there are at

    least three types of systemic risk. The first type of systemic risk is a common asset shock such as

    a fall in real estate or stock market prices. The second is the danger of contagion where the

    failure of one financial institution leads to the failure of another due to investor panics or other

    psychological factors. A third type of systemic risk is the failure of one financial institution

    which likely coincides with the failure of many others due to the more correlated portfolios and

    enhanced financial connections of individual financial institutions resulting from their individual

    original incentive to diversify.

    The recent global credit crisis has underscored the importance of systemic risk and

    exposed critical weakness in the financial regulatory system due to serious deficiency of our

    understanding in this regard. As a result, a top-down system-wide marco-prudential approach has

    been proposed to supplement the traditional bottom-up micro-prudential approach focusing on

    soundness of individual banks. While the literature (Allen, Babus, and Carletti, 2009; Cossin and

    Schellhorn, 2007; Elsinger, Lehar, and Summer, 2006) theoretically demonstrate the credit risk

  • 2

    transfer in a network structure due to tangible (rather than psychological) financial connections

    (e.g., interbank loans) between individual firms, other types of systemic risk mentioned above

    can also be reflected by the comovement of credit risk or other asset prices of these financial

    institutions. For example, Elsinger, Lehar, and Summer (2006) provide the evidence that

    correlation in banks’ asset portfolios dominates contagion as the main source of systemic risk,

    while contagion is rare but can nonetheless wipe out a major part of the banking system.

    Arguably, as a modest first step towards the understanding of the systemic risk, it may be worthy

    to objectively identify the structure of risk transmission across major financial institutions before

    we investigate different causes or channels of generating the systemic risk.

    This paper aims to use a dataset of international credit default swaps to identify the

    structure of credit risk transmission across major financial institutions on the eve of Lehman

    Brothers’ failure. The paper contributes to the literature in the following aspects. First, this study

    is perhaps the first to provide a data-determined identification of structure of credit risk

    transmission. Such an investigation is important itself, as it is directly motivated by the earlier

    theoretical work on credit risk in a network economy (e.g., Allen, Babus, and Carletti, 2009;

    Cossin and Schellhorn, 2007; Elsinger, Lehar, and Summer, 2006) and furthers our

    understanding of credit risk transfer (e.g., Allen and Carletti, 2006). It should also be informative

    to investors for their investment decisions on international equity and credit markets.

    Equally important, it can also shed some light on recent important efforts of identifying

    systemically important financial institutions (SIFIs) and designing and deploying macro-

    prudential regulation (e.g., Adrian and Brunnermeier (2009). More specifically, we focus on a

    particular major criterion of identifying SIFIs, which is their connectedness with other financial

  • 3

    institutions.1 Interconnectedness captures situations when financial distress in one institution

    materially raises the likelihood of financial distress in other institutions. In this context, SIFIs

    arguably can be those financial institutions which are primary senders or even exchange centers,

    but not primary receivers of credit risk information. 2 Lastly, it extends the literature on

    international asset return spillover (e.g., Eun and Shim, 1989; Malliaris and Urrutia, 1992;

    Gagnon and Karolyi, 2009) (see Gagnon and Karolyi (2009) for more related studies) by

    investigating the credit market (rather than the stock market) and focusing on the firm-level

    (rather than the national/market level) (credit) risk measure (rather than any other asset return) as

    represented by a particular type of asset price changes (i.e., CDS spreads).3

    Second, we propose an innovative empirical framework of combining the cluster

    analysis, the principal component analysis (PCA), the direct acyclic graph (DAG) and structural

    vector autoregression (VAR) analysis to facilitate an in-depth search for credit risk transmission

    network. Specifically, we classify international financial institutions into several clusters in the

    first step, then extract the major driving force behind the changes of CDS spreads in each cluster

    into principal components, and finally apply DAG-based structural VAR analysis to identify the

    structure of credit risk spillover within each cluster while controlling the influence of the other

    clusters. Noteworthy, our empirical framework corresponds well to the theoretical discussion on

    (clustered and unclustered) credit risk network (e.g., Allen, Babus, and Carletti, 2009; Cossin

    1 As a useful analytical device to structure the assessment of systemic important, three primary criteria are proposed: size, substitutability, and interconnectedness. Typically, the magnitude of the direct impact of an institution on the financial system relates to size and the degree of substitutability while the magnitude of the indirect impact depends on the strength of interconnectedness. 2 This is an important difference to alternative systemic importance measures, such as CoVAR by Adrian and Brunnermeier (2009) or MES by Acharya, Pederson, Philippon, and Richardson (2010), which address how an institution contributes to the financial system's overall risk. Our approach also allows for the transmission of credit risk with time lags. 3 As mentioned earlier, the structure of credit risk spillover could be due to either contagion or the link by fundamentals. It is beyond the scope of this paper to distinguish these different channels. See Forbes and Rigobon (2002) and Bae, Karolyi, and Stulz (2003), for recent works on financial contagion.

  • 4

    and Schellhorn, 2007). In particular, the DAG analysis (Pearl, 2000; Spirtes et al., 200