Beyond common equity: The influence of secondary capital on bank insolvency risk

https://doi.org/10.1016/j.jfs.2020.100732Get rights and content

Abstract

Banks must adhere to strict rules regarding the quantity of regulatory capital held but have some flexibility as to its composition. In this paper, we examine if bank insolvency (distance to default) is sensitive to capital other than common equity for a sample of listed North American and European banks. Decomposing tier 1 capital into tangible equity and non-core components reveals a series of heretofore unidentified non-linear links with insolvency risk. We assess the influence of binding capital requirements, finding that low regulatory capital buffers are associated with increased insolvency risk for banks holding greater quantities of non-core tier 1 and tier 2 capital. The links between insolvency and capital, evident when the latter is denominated relative to tangible assets or total regulatory capital, are found to be expunged when defined relative to risk-weighted assets.

Introduction

This paper seeks to understand a complex and relatively unexplored question, namely, beyond common equity is bank insolvency risk sensitive to the quantity and mix of other secondary regulatory capital held? At the core of banking regulation is the concept of minimum capital requirements, but capital regulation has often struggled to keep pace with the evolution of bank financial and operational sophistication. Bank management, using innovative capital instruments and strategic risk modelling, contribute to the complexity of the nexus between capital and risk. Requiring a bank to hold greater quantities of capital is expected to be associated with reduced risk, but the ability of management to shift between capital of differing quality and to manipulate standardized risk-weights obfuscates the relationship.

The literature examining the relationship between capital and risk has largely focussed on equity capital or more comprehensive tier 1 capital.1 In contrast, bank management have a feast of capital ingredients available to satisfy regulatory capital requirements. Tier 1 capital may be decomposed into high-quality tangible common equity and a lower quality residual component, termed non-core tier 1 capital (NCT1). Subordinated to tier 1 capital, tier 2 capital comprises undisclosed and asset revaluation reserves, loan-loss reserves, hybrid capital and subordinated debt. In quantifying regulatory capital, each of these is adjusted for asset risk using risk-weighted assets. Given the variety of moderating influences relevant to regulatory capital, we ask an important research question which has received only limited attention. That is, to what extent and under what circumstances is bank insolvency risk sensitive to low-quality capital, specifically NCT1 and tier 2 capital?

We provide an outline of the empirical contributions of the paper. Links between low-quality capital and insolvency risk are examined for a sample of European and North American banks. While the literature has presented mixed views regarding the risk-reduction capacity of the specific components that make up low-quality capital, our contribution relates to the assessment of such secondary capital from the regulatory perspective and in identifying the specific circumstances under which risk-reduction benefits exist.2 We first explore the significance of low-quality capital for the full set of banks over all periods without explicitly accounting for any moderating influences. The baseline findings emanating from this analysis indicate that low-quality NCT1 and tier 2 capital are not linked with insolvency risk. In contrast, a strong relationship between risk reduction and greater quantities of tangible equity and tier 1 capital is documented. The remaining analyses assess the impact of moderating features, such as non-linearities and capital constraints, upon the links between capital and insolvency risk. We initially uncover a series of non-linear links between capital and risk. In keeping with a hypothesis of Calem and Rob (1999), we find a U-shaped link between insolvency risk and both tangible equity and tier 1 capital. A clear distinction emerges for NCT1, however, where an n-shaped relationship between capital and insolvency risk is uncovered. For banks holding low levels of NCT1, holding further capital is associated with increased insolvency risk, but this effect tapers as capital increases. In order to explain these non-linear findings, we examine the influence of binding capital constraints.

Banks may be constrained by formal capital requirements, with the implication that some banks may hold regulatory capital above and beyond that prescribed by market forces. We consider the moderating impact of such constraints on the effectiveness of low-quality capital, focussing on banks with low capital buffers, where constraints may be greatest.3 For constrained banks with low capital buffers, we find evidence of increased insolvency risk for greater quantities of NCT1 or tier 2 capital. For banks unconstrained by capital requirements, NCT1 is found to have some risk-reduction potential. Capital constraints are also found to have a moderating influence upon the relationship between insolvency risk and each of tangible equity and tier 1 capital. For banks constrained by regulatory capital thresholds, such capital is linked with lower insolvency risk. While risk reduction from tangible equity and tier 1 capital is evident for unconstrained banks, the strength of the relationship is curtailed. We verify that our findings are not a consequence of capital substitution, by examining capital composition as a proportion of total regulatory capital. These findings reinforce a negative relationship with distance to default for capital constrained banks where regulatory capital consists of greater proportions of NCT1 and tier 2 capital.

Finally, we consider how the choice of denominator impacts upon the links between forms of capital and insolvency risk. While risk-weighting of assets appears a sophisticated way to link capital held to asset risk, empirical evidence for its effectiveness is weak (Barakova and Palvia, 2014, Mariathasan and Merrouche, 2014, Acharya et al., 2014). We provide fresh evidence on the questions surrounding risk-weighting of assets. Links between capital and insolvency risk, which are evident when denominated by tangible assets, are regularly expunged when considered relative to risk-weighted assets.

Our findings relate to and augment the literature in a variety of ways, placing the paper amongst the debate on the benefits of banks holding further capital and the literature which highlights the potential for regulatory arbitrage. Although various studies highlight the importance of high-quality equity capital for bank performance, little attention has been paid to the components which comprise total regulatory capital. Closest to this paper is Demirgüç-Kunt et al. (2013), who examine the relationship between capital components and bank stock returns over the financial crisis. Our paper also builds upon the findings of Berger and Bouwman (2013), which analysed links between equity capital and a survival probability dummy variable. In contrast to these papers, our focus is on insolvency risk throughout the business cycle, of greater importance to financial stability, and the decomposition into capital components, especially NCT1, which sheds fresh light on the overall capital picture. Furthermore, our assessment of the non-linear links between the various tiers of capital and insolvency risk extends previous analyses, illustrating the marginal risk reduction capacity of each capital form.

The regulatory requirement for banks to hold capital is related to the safety net provided by government guarantees and the potential for systemic risk resulting from bank failure. Capital restrictions attempt to prevent equity holders from decreasing the size of their limited liability stake and creating risk-shifting opportunities (John et al., 1991). The implication is that banks may hold capital above and beyond that prescribed by market forces. For a bank constrained by capital requirements, two possibilities exist; first, by changing capital composition they might be able to reduce the quantity of equity capital held, while maintaining a constant level of regulatory capital (Boyson et al., 2016). Second, they might choose to arbitrage risk-weights associated with the regulatory capital denominator, using, for instance, asset-backed commercial paper to cosmetically transfer risk off the balance sheet (Acharya et al., 2013, Jones, 2000). Our findings are linked with both sides of the regulatory arbitrage debate. We augment the findings of Boyson et al. (2016) where focus was confined to trust preferred securities. We show that the more expansive NCT1 capital is associated with increased risk, only providing risk reduction capacity for banks with above median regulatory capital. In other words, banks constrained by capital requirements use NCT1 as a substitute to meet regulatory capital ratios, with the adverse consequence of increased risk. Furthermore, the limited links between risk-weighted capital and insolvency risk found may be explained by the arbitrage of risk-weights proposed by Acharya et al. (2013) or through the mechanism of model manipulation suggested by Mariathasan and Merrouche (2014).

Ostensibly, many of the findings outlined in this paper are congruent with the renewed focus on high-quality capital under the Basel III framework. This framework will distinguish between non-core tier 1 capital and high-quality common equity, requiring greater quantities of the latter. Banks are required to maintain a minimum 4.5% of risk-weighted assets in common equity, while remaining tier 1 capital will consist of securities (such as contingent convertible instruments) designed to provide loss absorbing capacity on an ongoing basis. This focus on greater quantities of equity capital is generally aligned with the loss reduction capacity of such capital outlined. As highlighted here by the non-linear analysis of insolvency risk and tangible equity, links are attenuated for banks holding large quantities of tangible equity, indicating a potential limitation to the focus on greater core equity capital under the Basel III framework. Also, in keeping with the limitations documented here, alternative tier 1 components will be strictly comprised of high-quality, subordinated perpetual instruments, thus limiting the potential for regulatory capital arbitrage. Finally, tier 2 capital will be restricted to 25% of the total minimum risk-weighted capital and treated as gone-concern capital. Reduction in the relative importance of such capital is appropriate, given the limited role in reducing going-concern banking risk highlighted here.

The remainder of the paper is organized as follows. In Section 2, the empirical model and data sample are described and summary statistics provided. Section 3 provides empirical results and sensitivity tests, while Section 5 concludes.

Section snippets

Empirical model and data

To empirically test the relationship between components of regulatory capital and bank risk, we estimate variants of the following model (Demirgüç-Kunt et al., 2013, Beltratti and Stulz, 2012),DDi,j,t=α+δDDi,j,t1+β1Ci,j,t1k+β2Xi,j,t1+β3dj,t+ui,j,twhere subscript i corresponds to individual banks, j to countries and t to the year of measurement. In our main specifications, DDi,j,t is distance to default at time t, Ci,j,t1k is either a single capital metric vector or a matrix of k capital

Empirical results

Eq. (1) suffers from possible identification issues, primarily due to the potential endogeneity of capital metrics in risk equations and the persistence of bank risk. The latter issue is addressed through estimation of a dynamic panel model, accounting for risk persistence. To this end, we follow Beck et al. (2000) and Levine et al. (2000) and estimate Eq. (1) using the two-step system GMM approach proposed by Blundell and Bond (1998).11

Robustness analysis

In this section, the robustness of findings detailed are examined for alternative periods, bank sizes and methodologies. Relevant tables are available in the online appendix provided with the paper.

Conclusions

Prudential regulation requires banks to hold capital as a buffer in the event of losses and, among other reasons, as a means to mitigate risk shifting by shareholders. Under capital regulation, a large menu of securities are permitted as part of regulatory capital. While previous studies have predominantly concentrated on relationships between bank risk and either tier 1 or equity capital, we expand this analysis to regulatory capital other than equity. This places the work among the literature

References (70)

  • A. Beltratti et al.

    The credit crisis around the globe: Why did some banks perform better?

    J. Financ. Econ.

    (2012)
  • A.N. Berger et al.

    How does capital affect bank performance during financial crises?

    J. Financ. Econ.

    (2013)
  • D. Besanko et al.

    The regulation of bank capital: do capital standards promote bank safety?

    J. Financ. Intermed.

    (1996)
  • L.K. Black et al.

    The effect of TARP on bank risk-taking

    J. Financ. Stab.

    (2013)
  • R. Blundell et al.

    Initial conditions and moment restrictions in dynamic panel data models

    J. Econom.

    (1998)
  • P. Calem et al.

    The impact of capital-based regulation on bank risk-taking

    J. Financ. Intermed.

    (1999)
  • T. Conlon et al.

    Anatomy of a bail-in

    J. Financ. Stab.

    (2014)
  • A. Demirgüç-Kunt et al.

    Bank activity and funding strategies: the impact on risk and returns

    J. Financ. Econ.

    (2010)
  • R. DeYoung et al.

    Nontraditional banking activities and bank failures during the financial crisis

    J. Financ. Intermed.

    (2013)
  • M.A. Duran et al.

    Risk shifting in the US banking system: an empirical analysis

    J. Financ. Stab.

    (2014)
  • F. Fiordelisi et al.

    Efficiency and risk in European banking

    J. Bank. Finance

    (2011)
  • F.T. Furlong et al.

    Capital regulation and bank risk-taking: a note

    J. Bank. Finance

    (1989)
  • V.K. Goyal

    Market discipline of bank risk: evidence from subordinated debt contracts

    J. Financ. Intermed.

    (2005)
  • G. Iannotta et al.

    Ownership structure, risk and performance in the European banking industry

    J. Bank. Finance

    (2007)
  • K. Jacques et al.

    Risk-based capital, portfolio risk, and bank capital: a simultaneous equations approach

    J. Econ. Bus.

    (1997)
  • T.D. Jeitschko et al.

    Incentives for risk-taking in banking – a unified approach

    J. Bank. Finance

    (2005)
  • K. John et al.

    Risk-shifting incentives of depository institutions: a new perpspective on federal deposit insurance reform

    J. Bank. Finance

    (1991)
  • T. Jokipii et al.

    Bank capital buffer and risk adjustment decisions

    J. Financ. Stab.

    (2011)
  • D. Jones

    Emerging problems with the basel capital accord: regulatory capital arbitrage and related issues

    J. Bank. Finance

    (2000)
  • M.C. Keeley et al.

    A reexamination of mean-variance analysis of bank capital regulation

    J. Bank. Finance

    (1990)
  • L. Laeven et al.

    Bank size, capital, and systemic risk: some international evidence

    J. Bank. Finance

    (2016)
  • L. Lepetit et al.

    Bank income structure and risk: an empirical analysis of European banks

    J. Bank. Finance

    (2008)
  • R. Levine et al.

    Financial intermediation and growth: causality and causes

    J. Monet. Econ.

    (2000)
  • K.G. Lindquist

    Banks’ buffer capital: how important is risk

    J. Int. Money Finance

    (2004)
  • K. Mannasoo et al.

    Explaining bank distress in Eastern European transition economies

    J. Bank. Finance

    (2009)
  • Cited by (0)

    We would like to thank the Editor, Iftekhar Hasan, two anonymous referees, Bo Becker, Allen Berger, Santiago Carbo, Bob DeYoung, Iftekhar Hasan, Kose John, Steven Ongena, George Pennacchi, Valerio Potì, Amine Tarazi, Anjan Thakor, Karin Thorburn, Farzad Saidi, Klaus Schaeck, René Stulz, Greg Udell and participants at the Marstrand Finance Conference 2017, FINEST Spring Workshop 2015 (University of Limoges, France), Wolpertinger Conference 2015 (Granada, Spain) and Cardiff Business School for helpful comments and suggestions. Conlon and Cotter would like to acknowledge the financial support of Science Foundation Ireland under Grant Number 16/SPP/3347 and 17/SP/5447. Conlon would like to acknowledge the financial support of The Irish Research Council under Grant REPRO/2015/109.

    View full text