• No results found

Leveraging the leverage ratio : does the unweighted leverage ratio as additional capital requirement to Basel III incentivise ‘constraint’ banks to increase their risk appetite

N/A
N/A
Protected

Academic year: 2021

Share "Leveraging the leverage ratio : does the unweighted leverage ratio as additional capital requirement to Basel III incentivise ‘constraint’ banks to increase their risk appetite"

Copied!
59
0
0

Bezig met laden.... (Bekijk nu de volledige tekst)

Hele tekst

(1)

Executive master of finance and control

Thesis: Leveraging the leverage ratio:

“Does the unweighted leverage ratio as additional capital requirement to Basel III incentivise ‘constraint’ banks to increase their risk appetite“

Abstract:

Agency problems and incentives for banks result in high leverage for banks and causes regulation. Previous regulation was based on risk weights, however this has several shortcomings. To address these shortcomings the Basel Committee on Banking Supervision introduced a risk

unweighted leverage ratio as a backstop for risk based capital ratios net to the CET 1 capital requirements, to become effective as per 1 January 2018. There are various valid reasons to add the leverage ratio as a supplement to obviate the shortcoming of the RWA approach, however it may also

incentivise constraint banks to re-risk their activities in order to meet return requirements on the additional capital that should be held.

This paper tests empirically whether constraint banks actually shift their risk appetite based on a dataset of Eurostoxx-600 banks and banks that were identified to have a capital shortfall in the 2014 AQR. First of all the empirical data does not show a relationship between risk and return, measured as ROA and RWA / Total Assets. So based on the dataset used there would be no incentive for banks to re-risk.

Non-constraint banks operate their banks with a statistically significant higher balance sheet risk. In terms of re-risking, a statistically significant increase in the overall riskiness of the balance sheet (RWA/Total Assets) was found, increasing with 1.3% for the constraint banks compared to the non-constraint banks that reduced the ratio with -0.9% points. The latter underpins the expectation that

constraint banks have re-risked their balance sheets.

Banks have in total raised c. 103bn in new capital, constraint banks represented c. 70% of total new issuance in straight equity and AT1 capital combined, addressing leverage shortfalls.

Author:

Drs. M.H. Lamers (10680721) Supervisor:

Prof Dr. A.W.A. Boot

(2)

1

Table of Contents

1. Introduction 2

2. Regulatory framework, technical description and problem 4

3. Why RWA approach and leverage ratio? Costs and benefits 13

4. Description of the dataset 25

5. Methodology and hypotheses for the statistical tests 30

6. Hypotheses 35

7. Presentation of the results 40

8. Conclusions 48

9. Literature list 49

(3)

2

1. Introduction:

The banking industry is in constant change and has become increasingly complex and interlinked in the run-up to the latest crisis. According to Blanchard (2009) a global credit crisis was triggered due to I) underestimation of risk and disaster myopia II) Difficulties to value categories of assets and increasingly complex products and III) increased interconnections between financial institutions1. Financial institutions increased their leverage to increase their return on capital with less equity. Encouraged by low interest rates and regulation lagging financial innovation, banks have been able to boost their balance sheets with less capital. The main drivers of leverage have been real estate and structured finance and more generally trading book activities. Increasing leverage was a common strategy to improve returns on capital, especially within an increasingly competitive environment.

Increased risk levels for banks do not automatically lead to higher costs for shareholders of a bank, due to some distorting factors including agency problems and incentives including the reluctance of banks to raise equity in order to bolster their capital position. This result to risk shifting from banks to society causing a need for supervision and regulation of banks.

Highly levered balance sheets provided minimal cushion when asset prices decreased, this led to significant and accelerated depletion of capital. Post credit crisis, banks have been actively deleveraging and have focussed on rebuilding capital levels, seeking for more conservative levels. The role of regulators has been pivotal in this process. As a response to the most recent crises the Basel Committee on Banking Supervision developed Basel III regulation. As part of the revised framework, a leverage ratio is included as a minimum capital level for banks in addition to the risk weighted Core Equity Tier-I ratio (CET1 ratio). This should address the shortcoming of Basel II and the RWA approach including regulatory arbitrage, information asymmetry, procyclicality, off-balancesheet banking, differences in local standards and incentives for staff.

This thesis will test whether a potential conflict between introduction of the leverage ratio next to the risk-weighted requirements occurs in practice and whether banks, for which the leverage ratio is basically their ‘first stop’ rather than a ‘back stop’ tend to increase risks in their operations. Data collected over the period 2013 - 2015 will be used to test whether increased risk taking has taken place over time by ‘constraint’ banks.

1

(4)

3 Similar regulations were imposed in the US where the FDICIA introduced Prompt Corrective Actions (PCA) for banks depending on several capital requirements. The impact of such regulation has repeatedly been subject of research. Especially the impact on risk appetite and behaviour of banks. This study will focus on the effects of the introduction of an unweighted leverage ratio, versus a risk-weighted capital ratio. Two groups of banks have been created based on constituents of the 48 Eurostoxx-600 banks index and 25 Banks that were identified to have a capital shortage in the 2014 Asset Quality Review (’AQR’). Any duplicates in the two groups have been removed. The remaining group of 63 banks was split in two, the first group of 29 banks are the ‘constraint banks’ with a leverage ratio of <3% as per FY 2013 if reported, or a bookvalue of equity / totals assets of <6% as per FY 2013 as a proxy for the leverage ratio. The second group consisting of 34 banks are the ‘non-constraint banks’, meeting both the CET1 and leverage requirement based on the above thresholds. Data over the period 2013-2015 has been collected from SNL.

From a Controlling perspective it is crucial to monitor a potential shift in risk behaviour. The risk management organisation of a bank should acknowledge the potential behavioural change that regulation could unintendedly have on people within the organisation. Managing risk is the core business of a bank, and a potential shift in risk appetite on the back of changing regulation could therefore have a significant impact on the organisation. Banks are adapting to new regulation and are acting based on potentially wrong incentives resulting in excessive risk taking. Therefore it is crucial for risk-managers and controllers in banking to understand the impact of changing regulation and the consequences for people’s behaviour in banks. Banks operate with high leverage, inherent to the business model, so a small shift in risk-appetite could have disastrous consequences for a specific bank and/or the banking system at large.

The introduction of a leverage ratio next to the CET-1 ratio also raises other interesting controlling questions with respect to capital allocation and subsequent pricing implications. Capital deployment is costly for a bank, the cost of holding additional capital will need to be allocated to the different departments and products, this could have a significant impact on profitability and the bank’s competitive position vis-à-vis other players in the market. The costs of additional capital that a ‘constraint’ bank will need to attract can be allocated to departments and products for example based on risk (RWAs) or exposure (Assets). The choices that banks make to allocate these costs, also in comparison with their competitors, are crucial for its competitive position and profitability in certain areas of banking.

(5)

4

2. Regulatory background, technical description and problem

In the next section the need for regulation for banks is described, followed by the history of banking regulation in Europe. The Basel Committee on Banking Supervision (‘BCBS’) developed Basel III regulation for Banks. In the sections 2.2 – 2.4 a historical overview is provided describing the various Basel Accords, their shortcomings and the amendments made over time.

2.1 Why capital regulation?

Before the evolution of banking regulation is described, one need to understand fundamental principles in banking driving the need for regulation. Many rationales for bank regulation have been advanced over time, including protecting safety and soundness of an individual institutions, the stability of the financial system, fostering competition, consumer protection, credit allocation and monetary control2. Stability of the financial system is of crucial importance for well-functioning of the real economy. Several financial crises have shown how costly failures of the banking systems are to the real economy. The protection of banks, for example in the form of LLR facilities (Lender of Last Resort), deposit insurance, protection of the payment system and too ‘big to fail’ policies, all shift costs and risk from the banks to the public. Similar as to private contracting, for example in insurance taking, restrictions on bank behaviours are designed to limit exploitation and moral hazard.

After the recent crisis, the prudential regulation of banks has emerged as a pivotal issue. The key question was asked: What is the socially optimum amount of capital that banks should be required to hold on their balance sheets? Underlying this question is the premise that privately-optimal bank capital levels may fall below the social optimum, due to disturbing factors such as agency problems and counter effective incentives, and thus necessitate regulation3. The remainder of this section describes these agency problems and incentives including the reluctance of banks to raise equity in order to bolster their capital position, causing a need for supervision and regulation of banks.

Optimal capital levels for banks are delicate, Acharya et. (2010) find that the privately-optimal capital structure of a bank is like a ship navigating carefully between the mythological sea monsters Scylla (rent-seeking) and Charybdis (asset substitution)4, meaning that if bank leverage is too low, debt becomes so safe that there is no longer an incentive for creditors to impose the discipline that induces bank

2

Greenbaum S., Thakor A., Boot A. (2016), “Contemporary Financial Intermediation”, 3rd edition, Elsevier, Academi Press Publisher

3,4 Acharya V., Mehran H., Thakor A., (2010), ‘Caught between Scylla and Charybdis? Regulating Bank Leverage when there is rent seeking and risk shifting’, Federal reserve bank of New York Staff reports

(6)

5 monitoring. In case leverage is too high, the debt overhang problem occurs resulting in asset substitution, managers are taking excessive risks at the costs of the creditors. Asset substitution is often even correlated ((e.g. Reinhart and Rogoff (2008)). They find that this phenomenon is attributable to fiscal stimulus provided by governments (the tax benefits arising from tax deductibility of debt, whereas dividends paid to shareholders are not) or central banks as lenders of last resort. When bank failures are correlated all bank’s creditors may be protected as the social costs of collective failure may become too high.

In theory, banks that have capital ‘available’ should not choose to engage in risky activities, just because they have certain headroom. What is meant with availability or headroom is that more asset risk can be taken without violating capital requirements, believing that capital has a (high) fixed price5. Exploiting the bank’s ‘available capital’, by banks that have the leverage ratio as constraint should elevate the cost of capital. Based on the MM propositions, one would indeed expect that an increase in risk should lead to a higher required return. However, in the 1995 paper ‘Do the M&M propositions apply to banks”, Merton Miller came to a dual conclusion: ‘Yes and No’. The cost of equity is not a fixed number, but a function of both the risk the banks earning assets and the degree of leverage. However, for banks, there are a couple of factors that complicate the relationship between risk and return. First of all there are too big to fail safety nets that create an incentive for banks to hold inadequate capital levels and make equity disproportionally expensive. These protection mechanisms create moral hazards that shift risk from the banks to the public. Government backstops represent an implicit asset on the bank’s balance sheets, this creates reluctance for banks to raise equity because of a windfall gain to debtholders. An increase in equity will cause a disproportional advantage for debt holders as the value of the guarantee will be lowered at the expense of the (new and existing) equity holders. If a bank’s capital must be boosted through issuing new shares, it generally signals to investors the adverse news that retained earnings are unlikely to be enough to meet the capital requirements (Mayers and Majluf 1984)6, and the new equity injections will dilute the value of existing shares (Myers 1977). Jensen Meckling (1976) case up with the agency problem of asset substitution or risk-shifting by borrowers, where a borrower after raising debt, has incentives to transfer wealth away from lenders by switching to riskier assets unless the expected profits from safer assets are sufficiently high.

5 Boot (2012), ‘Financial Sector in Flux’, Journal of Money, Credit and Banking, Supplement to Vol 46, No1 . 6

Myers S., Majluf N. (1984), ‘Corporate finance and investment decisions when firms have information that investors don’t have’, Journal of Financial Economics 13, p187-221

(7)

6 Low leverage levels furthermore create debt overhang issues. This creates an incentive for companies to increase their risk appetite, even to an extent that they would invest in negative NPV projects. This effect is even stronger for banks, that have an implied bail-out guarantee. By increasing the riskiness of their activities they can increase the value of the implicit guarantee at the expense of the debt holders and society ultimately in case of a bail out. As creditors anticipate to be bailed out, the downside is ‘socialized’, the increase in bank leverage is not reflected in a higher cost of debt financing. Rent seeking relates to behaviour of an organisation to obtain economic gain from others without reciprocating any benefits to society through wealth creation, the bill of the excess risk taking is taken up by society. Furthermore, deposits which typically represent a significant part of the bank liabilities are a key production factor for banks to perform their operations. It is an essential part of the intermediation services that banks provide.

Other arguments supporting that MM doesn’t hold for bank. For example it is argued that capital that banks have to put aside to meet capital requirements is not available for lending. So an increase in bank capital will reduce bank capital. However a capital requirement is entirely different, it related to the equity portion of a bank’s total funding mix. And unlike cash capital reserves, equity can be directly invested in risky assets, similar to any other form of financing of the bank. Therefore equity does not reduce lending capacity for banks.

Finally, the need for regulation is applicable to incentivisation of staff. Acharya and Richardson (2009) state that the genesis of the financial crisis was partly caused by the desire of employees at highly leveraged institutions to take high risks, generating high short term ‘profits’. They did so by getting around the capital requirements imposed by regulators. Top traders in investment banks received giant bonuses in the years in which risk-taking generated high revenue and profits. However, the positions were not closed and the bill came somewhat later. In 2006 Goldman Sachs had a bonus pool of c. USD 16bn. In spite of the investment bank disasters of the second half of 2007, which saw Wall Street investment banks lose over USD11 billion, the average bonus fell only 4.7%. In 2008 losses skyrocketed causing the five largest independent investment banks to lose their independence: two failed, one was taken over by a conglomerate, and two became bank holding companies to qualify them for bailout money. Yet Wall Street bonuses were over USD18 billion—about what they were in the boom year of 2004 (DiNapoli, 2009). Another interesting example is AIG, the bailed out insurer that took on significant positions in CDSs, lost 40.5bn in 2008. Though the US government owns 80% of AIG's shares and

(8)

7 invested USD180 billion in the corporation, AIG nevertheless paid the 377 members of the division a total of USD220 million in bonuses for 2008, an average of over USD 500,000 per employee.7

Summarizing, implicit risk shifting from banks to society causes moral hazard, if regulators would allow banks to operate at too low capital levels, terrible incentives are created for banks with respect to risk taking. Fundamentally higher capital requirements would be better, as it internalises risk taking, and no moral hazard exists anymore. In a debt overhang world, in which we probably are, distortions still apply and therefore banks need to be regulated. Imposing additional capital measures may come at a cost for the banks and its shareholders, however the overall costs, including those for regulators and society, would decrease in case of sufficiently adequate capital levels. In the next sections an overview of Basel regulation is provided.

7

(9)

8 2.2 The Basel I Accord

Since 1978 regulators in Europe started to focus on bank capital. Due to the increased international competition among banks in global markets, regulators recognised the need to coordinate capital requirements for banks. Negotiations took long, but in 1993 the first Basel Accord was implemented in 12 countries for all insured banks.

The first Basel Accord was primarily focussed on credit risk and risk weighting of assets. The Accord distinguished five groups of assets with risk weighting ranging from 0-100%. The framework was developed by representatives of the G10 supervisors in order to serve the need for a multinational accord to strengthen the stability of the international banking system and to remove competitive inequality arising from different national capital requirements, creating ‘a level playing field’. Basel I required banks to hold 8% to their risk-weighted assets of which maximum 50% could consist of Tier-II to be implemented as per 1992, this includes credit equivalent amounts of any off-balance sheet exposures. The Basel I accord was just a first step in banking regulation, and was always intended to evolve over time. Criticism on the first accord were numerous. First of all the risk classes were crude, mortgages for example required c. half of the capital compared to business loans, whereas one could imagine situations were mortgages are far more risky than certain corporate loans. Furthermore, classifications could be manipulated via financial structures. Technically the instrument/position would qualify as a low capital requirement asset, whilst the underlying position would require other capital treatment. Others argued that interest rate risk was not taken into account in Basel I whilst this represents an important source of risk for bank capital. Special interests were also accommodated to safeguard local interests of participating countries. For example Japanese Banks typically operate as Keiretsus, implying that they hold equity stakes in other companies to diversify risk. 45% of unrealised equity gains could be counted as Tier-II capital. Basel I furthermore assumes that banking risk is the same across countries, not taking into account local market dynamics and economies. Other criticism on the first Basel accord focusses on the fact that capital requirements are set linearly, and do not take into account co-variabilities of returns affecting diversification and portfolios risks. Finally, the ratios are prescribed based on book values, not taking into account actual return profiles based on market values8. Many of these topics were addressed in the subsequent Basel II and Basel III regulations.

8

Greenbaum S., Thakor A., Boot A. (2016), “Contemporary Financial Intermediation”, 3rd edition, Elsevier, Academi Press Publisher

(10)

9 2.3 Basel II Accord

In June 1999, the Basel Committee issued a new capital adequacy framework to replace the Basel I 1988 framework to become effective in 2008. The framework was developed to better reflect underlying risks and financial innovation that was developed in the years before. The framework aimed to prevent a situation in which the framework could act as a source of disadvantage. Furthermore much more emphasis was put on coverage of a more comprehensive set of risks that banks are facing, including credit risk, interest rate risk and operational risk. A more risk-sensitive approach was chosen for particular asset classes making greater use of bank’s own internal risk models and assessments. Finally, the Committee wanted to make market discipline and regulatory monitoring part of regulation, intensifying the dialogue between banks and regulators and not solely rely on capital requirements. The updated regulation was crafted along three Pillars. The First Pillar did still focus on capital requirements. Risk weighted assets now included market risk and operational risk as well. Under Basel II banks are allowed to either use an Internal Ratings-Based (IRB) model or apply a Standard Model (SM). The risk components in the IRB include the probability of default, loss given default, exposure at default and effective maturity. Basel II also includes capital requirements for market and operational risk. Operational risk is defined as the risk of loss resulting from inadequate or failed internal processes, people or systems or from external events. Market risk is based on Value at Risk (VaR) models and stress testing, including the concentration risk and illiquidity in stressful market circumstances.

The second Pillar focussed on the supervisory review process. Supervisors are expected to evaluate bank’s own assessment of their capital needs relative to their risks and take appropriate actions if needed. The third Pillar focusses on market disclosure to stimulate banks to provide more transparency and disclosure. However exact details on how to comply with the increased disclosure request are not given. As described the Basel II mechanism provided a much more tailored and comprehensive approach to bank risk. However, this framework was still subject to criticism. Criticism is mainly focussed on the fact that banks can use IRB models which could be subject to manipulation, the complexity of the revised framework and the fact that liquidity risk is not adequately addressed. Masera (2010) stated that Basel II allowed systemically important institutions to gain an advantage with respect to diversification, resulting in the lower required capital buffers. Finally, it is argued that the risk-based minimum requirements would increase cyclicality, banks may estimate risk levels too low during periods of economic growth and can easily attract capital to fuel further growth. In a downturn however, the statistics change and actual risk may become more adverse than expected.

(11)

10 2.4 Basel III Accord

On the back of the criticism and the 2007-2009 financial crisis the Basel Committee proposed the new Basel III Accord on June 1, 2011. The updated framework will become effective as of 1 January 2018 and aims to increase the quality of bank’s capital. Basel III differs basically in five areas compared to Basel II, capital definition, capital requirement changes, risk coverage, leverage ratio and liquidity management. Liquidity management was non-existent in the Basel II framework. The Committee acknowledged that the importance of liquidity was underestimated and that well capitalised banks faced trouble during the crisis due to liquidity issues. Therefore it introduced two liquidity ratios, namely the Liquidity Coverage Ratio (LCR) and the Net Stable Funding Ratio (NSFR). The first ratio should safeguard that the bank has sufficient liquidity on its balance sheet to service its 30 days net outflows in a stress scenario. The second focusses on the funding of medium and long-term assets. Different types of funding instruments receive weights ranging from 100%, 90%, 80%, 50% or 0%. Available Stable Funding (ASF) sources are expected not to leave the bank within a year and can include equity, preferred shares, obligations with a maturity exceeding one year or deposits without a defined maturity. Required Stable Funding (RSF) is defined as the sum of assets that cannot be easily converted into money or used as collateral in a stress period exceeding one year. Different percentages are again allocated to the different asset classes. The ratio of the AFS/RFS should exceed 100%.

In order to address the cyclicality issue raised in the previous chapter, Basel III proposed several additional measures. A Capital Conservation Buffer of 2.5% over Risk Weighted Assets is prescribed that can be build up in times of normal operations and that can serve as an additional cushion in a stress scenario. In case the buffer is partially depleted, the bank needs to restore it first before it can distribute earnings to its shareholders in the form of dividend, buybacks and bonus payments to its employees. Basel III also prescribes a Countercyclical buffer of 0-2.5% of its RWAs at the discretion of the local regulator. Such buffer will be implemented if there are signs of excessive credit growth in a specific country increasing systemic risk. Finally, the Basel III Committee introduced a particular requirement for Systematically Important Financial Institutions (SIFIs) to prevent a too-big-to-fail status for individual banks potentially triggering a systemic crisis. An additional top-up ranging from 1.0-3,5% is added to the risk-weighted capital requirements depending on size, scale, complexity, interconnectivity etc. So with all surcharges activated, capital requirements could be as high as 15.5% of risk weighted assets.

The Committee also limited the bank’s ability to pay significant bonuses that could trigger excessive risk taking, and introduced ‘ claw-back’ mechanisms to reclaim previously paid bonuses in case of bad-times.

(12)

11 Most importantly in light of this paper, the committee introduced a leverage ratio to address the concerns of regulatory arbitrage in the model-based approach. The leverage ratio was intended to restrict the build-up of leverage that could destabilise the broader financial system and the economy, and re-inforce the risk-based metrics with a simple ‘ backstop’ measure. The committee was of the view that a simple leverage ratio framework is critical and complementary to the risk-based capital

framework and that a credible leverage ratio is one that ensures broad and adequate capture of both the on- and off-balance sheet sources of banks’ leverage9.

Risk weights are determined by the following factors: Credit risk, Market risk, Operational risk and Counterparty risk. The Core Equity Tier 1 (‘CET 1’) capital ratio is the key metric under Basel III, and is calculated by dividing the available Tier I Capital over the Risk Weighted Assets (‘RWAs’). The quality of bank Capital is defined according to the following tiering:

9

Basel Committee on Banking Supervision. ‘Revisions to the Basel III leverage ratio framework’ , 6 July 2016 Figure 1: Components of bank Capital under Basel III

Source: BCBS: ‘Basel III: A global regulatory framework for more resilient banks and banking systems’ (Dec 2010)

Tier 1 Capital Tier 2 Capital Common Equity Tier 1 (CET1) Additional Tier 1  Equity

 Non controlling interests

 Limited to amount required for subsidiary  Less deductions, examples

 Intangibles and goodwill  Shortfall in provisions

 Gains on change in own credit risk

 100% investments in other financial institutions  Net deferred tax

 Hybrid Tier 1 instruments

 Subordinate to subordinate of bank  Perpetual

 Callable

 Discretionary dividend  No step ups

 Loss absorbing through write off or conversion  General provisions

 Tier 2 instruments

 Subordinate debt of bank  Maturity > 5 years  No accelerated payment Tier 2

(13)

12 The minimum required capital levels under phase-in Basel III (CRD IV) are shown in the below graph:

A bank’s risk-weighted capital requirements, based on different tranches of capital (Core Equity, Tier 1 capital and total capital) should meet specific minimum hurdles, that are phasing-in in the years ahead.

Figure 2: Phase-in capital requirements according to Basel III (CRD IV)

Source: Basel III Handbook Accenture

2.0% 3.5% 4.0% 4.5% 4.5% 4.5% 4.5% 4.5% 2.0% 1.0% 1.5% 1.5% 1.5% 1.5% 1.5% 1.5% 4.0% 3.5% 2.5% 2.0% 2.0% 2.0% 2.0% 2.0% 0.625% 1.25% 1.875% 2.5% 0.625% 1.25% 1.875% 2.5% 0.0% 2.0% 4.0% 6.0% 8.0% 10.0% 12.0% 14.0% Until 2012 2013 2014 2015 2016 2017 2018 From 2019

(14)

13

3. Why RWA approach and leverage ratio? Costs and benefits

Banking regulation is in place for 200 years and keep evolving and developing to address contemporary issues in the global banking landscape. Some valuable parallels and lessons can be drawn from previous regulations. In order to get a understanding of the benefit of the leverage ratio and Basel III it is crucial to understand the good things of Basel I first. Section 3.1 describes Prompt Corrective actions in the US, which is very consistent with Basel I with exception of the leverage ratio. Both regulations were an efficient first step in capital regulation, leading to increased capitalisation levels for banks. Section 3.2 describes the limitations of the RWA approach and why adding a leverage ratio might help. Potential conflicts between both measures are described in section 3.3.

3.1 An evaluation of Basel I and ‘PCA’

As described in section 2.1 Basel I called for a 8% capital requirement over risk weighted assets, with at least 50% in the form of Tier-I capital. In the US, banking regulators imposed additional ‘PCA’ requirements. Prompt Corrective Action (‘PCA’) is very consistent with Basel I, except from the Tier I leverage ratio that was part of the regulation.

PCA was the legislative response to the insurance fund losses generated by the US savings and the loan banking crises of the 1980s10. In December 1991, the US Congress passed the Federal Deposit Insurance Corporation Improvement Act (FDICIA), addressing the importance of capital ratios in addressing the problems that led to the problems in the 1980s. The Act comprised of two components, first FDICIA contained an early closure policy for institutions which had no positive level of capital. Secondly PCA involved early intervention for problem banks by bank regulators.

The PCA standards distinguished 5 categories of banks, depending on how well they met either of three capital standards: i) Total risk-based capital standard, ii) Tier-I risk-based capital standard and iii) Tier-I leverage ratio. The five categories were ‘Well capitalised’, ’Adequately capitalised’, ‘Undercapitalised’, ‘Significantly undercapitalised’ and ‘Critically undercapitalised’. If a bank falls into one of the three undercapitalised categories, mandatory restrictions apply to its activities that become increasingly severe in case the bank’s capital further deteriorates. The Act became effective as per December 1992, in the run up to the effective date, banks significantly improved their capital positions. In 1991 43.3% of

10

(15)

14 the US banks were classified as ‘well capitalised’, whilst one year later 87.5% of the US banks qualified as ‘well capitalised’11.

One criticism against the imposed capital standards however, is that it may lead to increasing levels of portfolio risk. Kahane (1977), Koehn and Santomero (1980) and Kim and Santomero (1988) have shown that regulatory capital standards cause leverage and risk to become substitutes and that as regulators require banks to meet more stringent capital standards, banks responded by choosing assets with greater risk. Whilst the primary purpose of early intervention/closure is to prevent banks from taking increased risk when approaching insolvency, research by Levonian (1991) and Davies and McManus (1991) demonstrates that early closure of banks, may fail to protect deposit insurance funds from losses as it creates and incentive for banks to take increased asset risk, a sort of debt overhang problem where the capital of the bank is de-facto a call option. Portfolio risk is typically measured in two ways, using both the total Risk Weighted Assets (RWA) as a percentage over Total Assets (TA) and nonperforming loans as a percentage of Total Assets. Avery and Berger (1991) have shown that RWA/TA correlates with risky behaviour, other studies by Berger (1995) and Shrieves and Dahl (1992) use non-performing loans. Baer and McElravey (1992) used the leverage ratio as a measure for capital, as it was more binding than the risk-based capital standard over the period under review12. With respect to the speed of adjusting capital ratios, Aggarwal and Jacques (1998) find that undercapitalised banks adjusted their capital ratios at much faster pace than the group of well capitalised banks, not surprisingly because undercapitalised banks faced severe restrictions on their actions once PCA would become effective.

Basel I and PCA had many similarities and proved to be a relatively efficient first step in international banking regulation. Introduction of PCA led to quick recapitalisation, however as a consequence of the higher capital levels, banks started to increase risk levels. In Europe the leverage ratio as a backstop, was only introduced as part of Basel III. Before the potential conflict between the leverage ratio and the risk-weighted capital requirements are discussed in section 3.3, the next section first describes the limitations of the RWA approach in isolation and why adding a leverage ratio might help.

11

FRBNY Economic Policy Review / October 1998 12

(16)

15 Savings & Loans Crisis – lessons learned from previous crisis

The Savings & Loans crisis of 1980’s has been studied and analyzed by many, as there are valuable lessons to be learned from this crisis. In the early 1980’s there were c. 4,000 federally chartered savings & Loan institutions with c. USD 604bn of assets. Legislation at the time was driven by the public policy goal to stimulate home ownership. However regulation and supervision of S&L associations was less stringent compared to commercial Banks. Legislative actions in the early 80’s were aimed to help the S&L, but in fact contributed to the eventual cost of the crisis. Initially, prior to the crisis, most legislative, political and regulatory decisions were led by the spirit of deregulation. And, when interest rates increased, S&L’s, like mutual savings banks were losing money due to asset/liability mismatches, the policy makers had insufficient tools to deal with the crises and to close for example insolvent S&Ls. In the run up to the crises, capital requirements were lowered and eased and lax accounting rules were accepted. One of the most impactful regulatory changes related to the treatment of goodwill, in July 1982 the Bank Board eliminated the 10-year amortisation restriction of goodwill, allowing S&Ls to use the general GAAP standard of ‘no more than 40 years’ in effect at the time. This policy was aimed to let healthy S&Ls to take-over insolvent institutions, to ultimately prevent the guarantors to compensate the acquirer for the entire negative net worth. This led to an increase in goodwill from c. USD 7.9m to USD 22bn over a period of nearly 1.5 years. Goodwill represented c. 67% of total Regulatory Accounting Principles (‘RAP’) capital at the time. The deregulation led to strong growth of the S&L industry, between 1982 – 1985 total S&L assets grew with c. 52%, more than twice the growth in the saving and commercial banking segment. This growth was largely fueled by money inflows from deposits. Deposits guarantees were also increased from USD 40,00 to 100,000 in the early 1980s. Another effect of the de-regulation of the industry was the shift from traditional home mortgage financing into new more risky activities like equity securities, service corporations and real estate investments. In 1982, after a steep decline in interest rates, many S&Ls turned to profitability again, however 10% of the industry was still insolvent based on a GAAP basis and 35% based on a tangible basis. Regulators were trapped by their own policies as they could only act based on the relatively lax Regulatory Accounting Principles (‘RAP’ reporting). The deregulation also had significant spill-over effects to the general banking industry. The bidding up of deposit interest rates by opportunistic S&Ls also increased the cost of funding for commercials banks. Also on the asset side competitive pressure from S&Ls had an impact as much of the invested assets were concentrated in commercial real estate lending.

(17)

16 This led to erosion of underwriting standards for commercial banks, as a more prudent pricing of risk would leave them out of business. The channelling of large volumes of deposits into high risk investments did create market distortions. The estimated costs of the crisis are c. USD 160bn, including USD 132bn for the tax payer. Regulator’s belief that the markets would develop and adhere to its own discipline, drove deregulation and loosening protection of depositors, proved to be a miscalculation. The Government guarantee of insured deposits exposed US tax payers to the risk of loss, whilst the profits go to the equity holders of the banks.

This crises showed the clear need of effective supervision of insured deposit institutions, especially if they are given additional tools and experience fast growth. Furthermore the S&L crises tough that regulators should be well trained and equipped and need to have the tools to act/close insolvent institutions based on the right criteria.

(18)

17 3.2 Limitations of Basel II and the RWA approach

Capital requirements prior to Basel III were primarily based on risk weights, however the 2007-2009 global financial crisis has highlighted the limitations of the risk-weighted capital ratios. Despite numerous refinements and revisions over the past 2 decades, the weights applied to asset categories seem to have failed to fully reflect bank’s portfolio risk causing an increase in systematic risk. In this chapter some of these shortcomings are discussed providing rationale for inclusion of an additional un-weighed leverage ratio.

First of all, Basel II allowed banks to use their own internal models. Rugemintwari (2012) shows that in the presence of asymmetric information between banks and their supervisor, banks have an incentive to understate its risk taking which could be prevented by the introduction of a simple leverage ratio as included in Basel III. Banks entirely control the information they communicate to the regulator concerning their risk positions and corresponding risk capital. Blundell-Wignall and Atkinson (2008) quote in their paper an investment banker that states: ‘ It amazes me that regulators asked us to set our capital regulation weights, given the way the incentives are. But good luck to any supervisor who wants to find out what really happens inside a business, that is difficult for insiders to know and impossible for outsiders. Based on the model of Rugemintwari, it seems difficult, if not impossible for the supervisor to devise appropriate sanctions to deter a bank from misreporting its risk assessment when the regulator can only detect potential gaming with a very low probability. As a potential solution the model introduces a simple leverage ratio is addition to the IRB based risk-weights.

Secondly, Basel II was criticised for amounting little more than lightweight pro-cyclical bank capital regulation with ample possibilities for regulatory arbitrage, and as such contributed to the banks being undercapitalised and unable to resist a serious economic storm. Not only model based approaches were criticised being easily manipulated and producing overoptimistic assessment of risk in good times (leading to too low capital buffers), but also the quality of capital was criticised.

Thirdly, securitisation transactions were widely used to take risk ‘off-balance sheet’ to lower capital requirements. Acharya V. and Richardson M., (2009) describe the process where banks use securitisation transactions to circumvent capital requirements. They did so by temporarily placing assets such as securitized mortgages in off-balance sheet entities, so that the banks did not have to hold significant capital buffers against them. Second, the regulations allowed banks to reduce the amount of capital they held against assets that remained on their balance sheet for AAA-rated tranches of securitised mortgages. Repackaging of mortgages reduced the amount of capital that banks had to maintain for potential loan losses. Finally structured securitised mortgages were classified as AAA by

(19)

18 rating agencies due to modelling failures and possibly conflicts of interest. The growth in securities from 2002 – 2007 was staggering, securitisation worldwide went from USD 767bn at the end of 2001 to USD 1,400bn in 2004 to c. USD 2,700bn at the peak of the ‘bubble’. But instead of acting as intermediaries between borrowers and investors and transferring the risk from mortgage lenders to the capital markets, banks became primary investors and kept concentrated risk at greatly magnified level. The main purpose of securitisation was not to share risk with investors, but to make and end run around capital regulations13.

Another shortfall of Basel II and the RWA approach comes from different local standards with respect to implementation and model failures. As a reaction to the ECB stress test in which a shortfall of EUR 25bn was identified, Acharya and Steffen (2014) provide benchmark stress tests suggesting that regulatory stress test outcomes are potentially affected by the discretion of the national regulators. Furthermore they state that the use of the CET-1 ratio is problematic, as both the nominator and the denominator have shortcomings. The use of risk-weights for the denominator is based on Bank’s Internal ratings based models. Even in the standardised approach, risk-weights are not necessarily reflecting the true risk of bank’s assets. For example zero risk weight is applied to government bonds. With respect to the numerator, common Tier-I capital is a new measure of regulatory capital that incorporates a substantial number of transitional arrangements until it is fully implemented. Many deductions from capital are phased-in over time, which treats the composition of capital for banks differently. For example the impact of goodwill and other intangibles and Deferred Tax Assets (DTAs) can differ between banks, as national competent authorities can decide on recognising these items at their discretion. These differences between countries do not add to a level playing field for banks that operate internationally. A simple leverage ratio, based on unified standards would address this issue.

Finally, a general shortcoming of banking regulation is found by Kupiec (2016), he states that it is based on lagging indicators not reflecting the true market value of a bank’s equity capital. Under amortized cost accounting, many bank assets continue to be recorded at full value until a bank books a reserve for loan losses or an impairment for a security held to maturity. There is ample evidence showing that banks are slow to book provisions, especially when the bank has insufficient operating earnings to offset the provisions. Inadequate provisioning is an important cause of why failed bank resolution costs are that high. Chernykh and Cole (2015) found that the Nonperforming Asset Coverage Ratio or NACR14 provides an informative forward looking measure to asses a bank’s solvency condition that shows accurate

13

Acharya, V. and M. Richardson, (2009), ‘Causes of the financial crisis’. Critical review, Vol. 21, pp. 195-210 14

(20)

19 warning signals well before regulatory capital measures show weaknesses. Cole and White (2015), find in their paper that PCA interventions based on NACR would have closed banks sooner in the recent financial crisis and reduced Deposit Insurance Fund losses with c. 25%. Would this be the next step in banking regulation?

Why adding a leverage ratio might help.

Previously discussed factors show some of the limitations of Basel II including the RWA approach. During the credit crisis, banks were forced to reduce leverage, which put significant pressure on asset prices. This process of dismantling of leverage resulted in a vicious circle of losses, weakening capital positions and capital markets becoming illiquid. To prevent this situation from happening again, the BCBS first introduced a leverage ratio in 2010 that should serve as a back-stop for risk based capital ratios. With the introduction of an additional leverage ratio, regulators were trying to constrain the innovative reactions to banking regulation and slow-down the race between the banking sector and the regulatory bodies. A mandatory minimum leverage ratio of 3% would be implemented beginning of 2018. In a response, the Dutch Minister Dijsselbloem even considered to raise the minimum threshold for Dutch banks to 4% as per that same date. The leverage ratio simply measures a bank’s capital in relation to all its assets, on- and off-balance sheet (equivalents), irrespective of their riskiness. The ratio is calculated as the unweighted average of the monthly ratio’s during the quarter. The numerator of the ratio stands for the Tier 1 capital, that is at permanent disposal of the bank, no repayment obligations should apply. Examples are Equity, retained earnings, perpetuals, hybrids, CoCos etc. The contribution of hybrid capital instruments with re-payment obligations for calculation of the leverage ratio is capped at 15% of the overall Tier 1 capital .

The leverage ratio will set a ceiling to non-risk weighted exposures, a non-risk weighted ‘back-stop’ measure. Another reason for introducing the leverage ratio is described by C. Rugemintwari (2011), he states that the introduction of a leverage ratio next to the CET-1 ratio limits the effects of asymmetric information between the bank and its supervisors. Banks use internal models for RWA calculations and could therefore have an incentive to understate their risk taking. This could be mitigated by the introduction of a simple leverage ratio as suggested in Basel III. It also appears as a necessary remedy to the supervisor’s imperfection. The leverage ratio should counter balance the build-up of systematic risk by limiting the effects of risk weight compression during booms. Gambacorta and Karmakar (2016) find that the leverage ratio is more counter-cyclical than the risk-weighted capital ratio: it is a tight constraint during a boom and a soft constraint in a bust, further more it is concluded that the benefits of

(21)

20 introducing the leverage requirement appear to be substantially higher than the associated costs. The paper furthermore states that new capital regulation consists of three complementary components: i) risk-weighted regulation, ii) stress-testing framework which assesses banks’ resilience to tail risk (BCBS, 2009) and iii) leverage regulation independent of risk assessment15. The leverage ratio should not be seen in isolation, when the information on individual asset risk would not be taken into account when assessing capital adequacy, banks may be incentivised to shift investments from low-risk to high-risk assets. Turner (2010) and Goodhart (2010) state that the most important step regulator should take to achieve the broader macro prudential goal of protecting the banking sector from excess aggregate credit growth is a significant increase in equity requirements. Goodhart (2010), Acharya et al (2011) and Acharya et al. (2015) go even one step further and propose that regulators should also impose restrictions on dividend and equity pay-outs as part of prudential regulation.

Others have come up with more complex proposals to regulate banking capital including Flanery’s (2005) contingent capital certificates, forced equity issuances by banks during periods of deteriorating performance (Hart and Zingales (2009) and Duffie (2010), expanding the limited liability of equity (Admati and Pfliederer (2009), ‘capital insurance’ (Kashyap, Rajan and Stein (2008) and taxing the systematic risk of financial institutions (Acharya, Pedersen, Phyllipon, and Richardson (2010). The leverage ratio is expected to address most of the identified shortcomings of the RWA approach. However for a specific group of banks, the once that adhere to the risk-weighted requirements but fall short on the leverage ratio (‘constraint banks’), it may give an incentive to re-risk their activities. This phenomenon is described in section 3.3. and tested in the remainder of this paper.

15

Gambacorta L., Karmakar S., (2016), ‘Leverage and Risk Weighted Capital Requirements’, BIS Working Paper, No 586

(22)

21 3.3 Potential conflict Leverage ratio versus the RWA approach

The implementation of the leverage ratio, next to the CET-1 ratio has caused a wave of reactions. In a working paper written by PWC, it is raised that both measures can work opposing/conflicting. Härle P., Lüders E., Pepanides T. et al. (2010) find that ‘closing of the leverage capital shortfalls’ will impact ROE negatively. The capital shortfall of banks that comply with the CET-1 capital requirements, but fall short on the leverage ratio is estimated at multiple billions. In order to make a decent return on the additional capital that a bank should hold from a leverage perspective, banks may seek to optimise risk-return within the new boundaries set. According to BIS (2012), European banks offered for sale a significant amount of assets, mostly those with relative high risk-weights, including distressed bonds, commercial property and low-rated securitised assets.

Originally the leverage ratio was considered ‘supplementary’ to the risk-based ratio, and regulators continue to describe it this way, however banks with a low risk profile may fear that this leverage ratio will become the binding capital requirement. Combining both metrics may not be compatible with the objective to adopt a low-risk approach that drives savings and retail banks’ business models. Banks with a low-risk profile could be sufficiently capitalised based on their risk-weights, however could be obliged to hold a minimum leverage which increases their cost of capital and has a negative effect on their Return on Equity. This could have an impact on the competitive positioning of the bank and could make the bank less attractive to investors, as they could make a superior return on banks that use their balance sheet ‘more effectively’. The combination of both metrics may incentivise bank managers to adopt a higher risk approach and align risk weighted capital with minimum leverage requirements. The graph below visualises the potential shift in risk behaviour to compensate for additional leverage capital requirements.

(23)

22 The Financial Stability Review of November 2015 discusses the concerns raised that the non-risk-based nature of the leverage ratio could incentivise banks to increased risk taking. The paper presents a model which explicitly takes into account the fact that banks are subject to the maximum of two capital requirements. The model is briefly explained below:

The model assumes that banks can invest in safe or risky assets. ω denotes the investments in safe assets and (1-ω) represents the investment in risky assets. The risky asset has a higher results volatility. It could result in a higher profit, but could also lead to a loss. The model assumed two states of nature, s1 with a probability μ to occur and s2 is assumed as a bad state with probability (1-μ) to occur. The safe

asset returns R1≥1 if state s1 occurs and (1-λ1) ϵ(0,1) if state s2 occurs. The risky assets returns R2>R1 in

state s1 with probability π and (1-λ2) ϵ(0,1) with probability (1-π). In state s2 the risky assets returns

(1-λ3) ϵ(0,1) with probability (1-π) and zero otherwise. In the case of both a risk-weighted capital requirement and a leverage ratio requirement, and hence banks are subject to the maximum of the two capital charges. The risk-weighted requirement k(ω) depends on the risk choice of the bank. The risky asset requires a higher capital charge, whilst under the leverage ratio the capital requirement klev is

independent of the riskiness of the portfolio. This leads to a kinked nature in the capital requirement as shown in figure 4.

The risk-based requirement increases when a bank increases it holdings in risky assets, until the level (1-ωcrit)

where the risk-weighted capital requirement starts to exceed the leverage ratio. At low-risk holding the required capital lies below the leverage ratio requirement (dotted line). The model yields two results. First, imposing a leverage ratio will incentivise banks to increase risk-taking. This can be shown by comparing the first order condition (FOC) when the model is solved under a solely risk-based capital requirement, and when

Figure 4: Kinked capital requirements in case of both a risk-weighted capital requirement and independent leverage ratio

Source: Grill, Lang and Smith (2015)

krisky

klev

(1-ωcrit) 1 Investment in risky asset

(24)

23 the leverage ratio constraint is added to the model.

Under the risk-based capital requirement, the FOC reflecting the optimal risk choice is as follows: Equation 1: μ[πR2+(1-π) (1-λ2) – R1] = -ρk’(ω) - μ k’(ω) – c’(ω)

This equation shows that banks increase return until the marginal return from increased exposure to risky assets equals the marginal costs, i.e. the need to increase capital levels.

In case the leverage ratio becomes the binding constraint, the terms related to the risk-weighted capital requirement disappear from the equation, since increasing risk no longer requires the bank to hold additional capital. Sufficient capital is already held by means of the minimum leverage ratio. The FOC now becomes:

Equation 2: μ[πR2+(1-π) (1-λ2) – R1] + (1-μ)Y = c’(ω)

Removing the risk dependent capital requirement from the equation implies that the bank can shift its risk profile without having the obligation to hold more costly capital, the marginal cost of risk-taking declines. As a consequence of having a higher cushion to absorb economic shocks, the shareholders of the bank have more ‘skin in the game’. This is reflected by the Y in the formula. This effect opposes to the first and incentives to reduce risk. This effect is however small according to the paper and is significantly out weighted by the first incentive. The paper of the Financial Stability Review (2015) concludes its model by stating that although the banks with the leverage ratio as binding capital requirement may take on greater risk, at the same time they hold greater capital buffers which means that they can absorb bigger losses. The model suggests that if the leverage ratio constraint is not set excessively high, it will weakly decrease the banks’ probability of failure and if not all banks stick to the minimum hurdle of the leverage ratio, strictly decrease expected losses. The caveat in the model on setting a too high leverage ratio comes from the outside option that equity holders have. Since investors require a higher return than the debt holders, this could oblige the banks to go beyond their optimal risk choice to meet the required returns on equity. As the leverage ratio starts to bind, both expected losses and the probability of default declines (as additional capital is held). The increased risk-appetite is finally not unbound, as discussed the ‘skin-in-the-game’ effect reduces the bank’s incentive to increase risk exposure. Furthermore the risk-based capital framework still underlies the capital framework, so if a bank takes on too much additional risk, it will simply move back into the risk-based framework as binding constraint. Hence as both requirements apply alongside, it acts to constrain the risk-incentive.

(25)

24 The empirical analysis presented in the paper is based on the data of 500 EU banks covering years 2005-2014. Since the data for the Basel III leverage ratio is not available over the period, Tier-1 capital over total assets is used as a proxy. The analysis finds that the leverage ratio is a very important indicator for determining a bank’s distress probability, the model suggests that a 1% increase in the leverage ratio results in a 35-39% decline in the probability of distress to non-distress. This impact is much larger than the sensitivity of an increase in risk-weighted capital. Increasing the risk-weighted assets ratio by one percent point, the distress probability decreases only marginally with c. 1-3.5%. Assuming non-linearity of the distress curve, increasing the leverage ratio from low levels has even a more beneficial effect. The impact of the leverage ratio on banks below the leverage ratio requirement was also tested on a subgroup, using data since 2010. The model finds that EU banks with low leverage ratios have slightly increased their risk-taking, expressed as RWA/Total Assets. Banks bound by the leverage ratio increased their risk-weighted assets ratio by around 1.5-2.0% points more than they otherwise would have, in addition these banks have increased their leverage ratio with 0.5-1.0% point more than they would have done otherwise, in anticipation of the new regulation. Finally the paper demonstrates that bank distress probabilities should significantly decline with a leverage ratio requirement, even when considering a steep increase in risk-taking by the banks. The paper supports the introduction of a leverage ratio alongside the risk-weighted measures.

(26)

25

4 Description of the dataset

4.1 Compilation of the dataset

An Excel dataset has been compiled, consisting of data of 48 banks that are part of the Eurostoxx-600 Banks index. A capitalisation weighted index of European listed banks. In total 48 banks (8 Italian, 7 Spanish, 6 UK, 4 French, 4 Swedish, 3 Suisse, 3 Danish , 2 Dutch, 2 German, 2 Greek, 2 Austrian and 1 bank from respectively Belgium, Norway, Ireland, Portugal and Czech.) are included in the index. In order to broaden the sample a second group of 25 banks was added consisting of banks that were identified to have a capital shortfall as a result of the 2014 Asset Quality Review (AQR)16. In October 2014, the European Central Bank released the outcomes of the Asset Quality Review. This operation has been performed in the run up to the operational start of the Single Supervisory Mechanism (SSM). The report identified shortfalls for 25 banks in the review sample consisting of 130 European banks in total. The total identified capital shortfall for these bank equalled EUR 25bn. The AQR was based on the Capital Requirements Regulation and Directive (CRDIV). Under the AQR, banks were required to have a CET-1 ratio of at least 8%. In the stress-test scenario banks were required to hold a minimum level of 5.5%. The AQR resulted in an aggregate adjustment of EUR 47.5bn to bank’s carrying asset values as per 31 December 2013. In the stress test scenario, the bank’s capital is projected to be depleted by EUR 215.5bn, representing 22% of the total capital held by the participating banks. The risk weighted assets were expected to increase with EUR 860bn by 2016. This would lead to a decrease in the CET1 ratio of 4% from 12.4% to 8.3% in 2016. Since the start of the financial crisis in 2008, and the ARQ review date (31 December 2013), more than EUR 200bn in capital was raised by banks participating in the AQR. A further EUR 57.1bn was raised in between the review date and the date of publication of the AQR report. Net of these capital raisings EUR 9.5bn remained as identified capital shortfalls.

Data is derived from SNL Financial Business Intelligence Service. Data has been collected based on five moments in time (FY 2013, H1 2014, FY 2014, H1 2015 and FY 2015) including Total Assets, Risk Weighted Assets, CET1 ratio, leverage ratio, net earnings, bookvalue of equity and ROA. Missing data has been gathered from company reports, when available. Missing data that could be derived from other data in the file has been manually calculated, figures that were not published in half year reports, have been extrapolated. Data on AT1 capital issues has been derived from Bloomberg. Straight Equity issuance data has been retrieved from Dealogic.

16

Oesterreichischer Volksbanken-Verbund, Jyske Bank, C.R.H. -Caisse de Refinancement de l’Habitat, Dexia and Eurobank Ergasias have been omitted from the sample as well as duplicates between both groups.

(27)

26

4.2 Variables used for the descriptive statistics

The below table provides an overview of the SNL data per variable. These statistics are based on the complete sample including both constraint and non-constraint banks. In chapter 7 descriptive statistics per sub sample will be presented to describe potential differences versus the hypotheses as presented in chapter 6. Average assets for the 63 banks over the period 2013 – 2015 equal 429bn. As many banks have taken actions during the period under review, it is difficult to draw conclusions on the quality of the banks over the period. With an average CET-1 ratio of 12.62, banks seem to be well capitalised, however the minimum and maximum values in the table provide a much better picture of the difficulties some of the banks faced during the period under review. The CET-1 ratio of 5.09% for Carige per FY 2013 for example shows the weak capitalization of some banks, also underpinned by the 0.27% solvency of AXA Bank Europe. Over the review period the banks collectively raised 75.2bn in straight equity, in addition 45.8bn in AT1 capital was raised to bolster banks’ balance sheets. For the purpose of this paper the sample will be split in two to compare relevant metrics between constraint and non-constraint banks.

Table 1: Variables used for the descriptive statistics

Source: SNL, Company reports Total Assets (EURbn) Total RWAs (EURbn) RWA/Total Assets CET-1 ratio Equity / Total Assets Net Earnings (EURbn) Bookvalue of Equity (EURbn) ROA Common Equity Raised (EURbn) AT1 Capital Raised (EURbn) Mean 429 152 44.08% 12.62% 6.61% 809 24,836 0.90% 1.19 727 Median 131 60 43.44% 12.00% 6.42% 309 10,364 0.91% 0 0 Maximum 3,644 1,691 86.34% 25.13% 14.67% 21,407 285,349 2.49% 9.25 8.58 Minimum 4 2 12.48% 5.09% 0.27% -13,583 113 -2.50% 0 0 Std. Dev. 618 233 16.9% 3.14% 2.37% 3,482 38,892 0.62% 2.26 1,612 Observations 315 315 315 189 315 315 315 315 63 63

(28)

27

4.3 Variables used in the statistical analyses

To test the relationship between the Risk of a banks balancesheet expressed as RWA/Total assets on the Return on Assets (‘ROA’) a statistical analysis will be performed. The will be explained in the next chapter. Tabel 2 presents the summary statistics of the main dependent, independent and control variables used in the regression. The sample consists of 63 banks, and data is collected based on five moments in time, resulting in 315 observations per variable. The dependent variable ROA is on average 0.91%. The operating return on assets (pre-provisioning) is taken, to exclude leverage effects in banks’ balance sheets and timing differences in the recognition of provisions by banks. The returns vary between a max of 2.5% (Helenic Bank, H1 2014) and a minimum of 2.5% (Nova Ljubljanskabanka, FY 2013). The mean and median are close with respectively 90 ad 91 basis points. The main explanatory variable, a bank’s balance sheet risk shows a maximum value of 86.34% (Banca Popolare di Milano, FY 2013), and a minimum value of 12.48% (AXA Bank Europe, FY 2014). The mean and median for this sample are 44.08% and 43.44% respectively.

Instrumental variables

The instrumental variables used for the 2SLS model are Corporate loans as percentage of total assets (‘Corporate’) and retail exposure as percentage of total assets (‘Retail’). The average Corporate exposure for banks in the sample equals 30.0%, with a maximum of 74.9% (Banco Commercial Portugues, FY 2014) and a minimum of 1.90% (Cooperative Central Bank, H1 2015). Retail exposure is on average 30.0%, with a median of 28.0%. Values range from a maximum of 85.6% (Permanent TSB, FY 2015) to a minimum of 2.8% (Natixis, FY 2014).

Table 2: Descriptive statistics of the regression variables

Source: SNL, Company reports

ROA CORPORATE RETAIL RWA_TA NIM FEES C_I

Mean 0.90% 30.08% 30.12% 44.08% 1.58% 25.25% 61.44% Median 0.91% 28.05% 28.09% 43.44% 1.48% 24.30% 59.10% Maximum 2.49% 74.87% 85.63% 86.34% 3.66% 60.61% 127.72% Minimum -2.50% 1.90% 2.81% 12.48% -0.04% -3.12% 20.85% Std. Dev. 0.006 0.166 0.149 0.169 0.007 0.113 0.167 Observations 315 315 315 315 315 315 315

(29)

28

Control variables

In order to optimise explanatory power of the model additional control variables are added to the model, so called control variables that could also explain the return on assets for banks. The NIM for the banks in the sample is on average 1.58% over the five time periods, with a median of 1.48%. The maximum equals 3.66% (Raiffeissen Bank International, FY 2014) and the minimum was a 4 bps negative spread for Dexia (FY 2013). The second control variable is Fee & Commission income as percentage of total income (‘Fees’). This is income that banks generate without directly using the bank’s balance sheet, and hence no capital requirement apply. This type of income has become increasingly important with the new capital regulations for banks, as it can boost returns. On average fee & commission income represents 25.3% of total income, with a median of 24.3%. Julius Baer (H1 2014) has the highest contribution of 60.6% and the lowest contribution was reported by Dexia (FY 2013) when the bank reported a negative operating income. Finally Cost/income ratios varied from a maximum of 127% (Permanent TSB, FY 2014) to a minimum of 20.9% for Bank of Cyprus (FY 2014). The average Cost/Income ratio for the bank in the sample equals 61.4% with a median of 59.1%. The relationship with the return on assets is expected to be inverse as described in the hypotheses section.

Dummy variables

Additionally Dummy variables running from 1-5 corresponding to the reporting date (per half year running from FY 2013 – FY 2015) are included to test whether the reporting period is of significant influence to explain the return on assets. The group dummy correspond to the group in which the bank falls, being either a constraint (1) or non-constraint bank (2). The methodology on which the split is based is described in section 5.2.

(30)

29 Table 3 provides a correlation matrix between the dependent variables and the explanatory variables in the model. As an indication of potential endogeneity in the model, correlations between the explanatory variables should not exceed 80%, which is the case. Correlation between balance sheet risk and return on assets equals 37.5%, which could indicate that there is indeed a relationship between both variables. The Group Dummy, NIM and C/I ratio are other variables that report relatively high correlations with the dependent variable (ROA). The Cost/Income ratio has a negative correlation with the return on assets as expected.

4.3 Other data gathered

For the hypotheses that will be tested based on descriptive stat only, details can be found in appendices 1 and 2. These variables include total Equity raisings, total AT1 capital raisings, CET-1 ratios and leverage ratios based on the bookvalue of equity as percentage of total assets (as a proxy for the leverage ratio, which is only reported for a few banks during the review period).

Table 3: Cross correlations of the dependent and independent variables

Source: Eviews, SNL data, Annual reports

1 2 3 4 5 6 7 8 9

1. Return on Assets 1

2. Time period dummy 0.063 1

3. Group dummy: (non-) constraint 0.416 0.000 1

4. Corporate Exposure 0.123 -0.034 0.298 1

5. Retail exposure -0.093 0.028 0.110 0.074 1

6. Risk Weighted Assets / Total Assets 0.375 0.000 0.575 0.200 0.056 1

7. Net Interest Margin 0.561 0.025 0.569 0.093 0.005 0.542 1

8. Fee Income -0.059 -0.004 0.000 0.000 -0.459 -0.285 -0.199 1

Referenties

GERELATEERDE DOCUMENTEN

In het kort geeft dit boek als antwoord op deze dreiging: ga voor ‘real-time’ evaluatie; verleg uw aandacht van ex-post evaluatie ten behoeve van verantwoording achteraf, naar

When we sort residential areas by the share of loans that was given in excess of the DTI policy in 2014, and subtract each area ’s local exemption quota, we see in Fig.. 11.5 that

Noteworthy, systems vaccinology has demonstrated that responses measured early after vaccination in peripheral blood mononuclear cells (PBMCs) are predictive of

the state, above all the culturalisation of citizenship, with a particular focus on educational policies and practices. The interest in this specific subject originated from

This research focused on both the impact of Basel III, the new capital requirements for banks, on the amount of risk banks take as well as the relationship between the leverage

The results on capital adequacy show that banks from countries with high uncertainty avoidance, high power distance, and banks from French code law countries hold significantly

Dit onderzoek is specifiek gericht op het verstrekken van krediet door een bank, wat voor de risk response betekent dat deze dient te worden geoperationaliseerd door

Voor het uitwisselen van decentrale informatie gelden nieuwe werkprocessen met als belangrijke kenmerken: Een goede webomgeving is cruciaal Ondersteun de praktijk door een