Excerpt
Table of Contents
List of Abbreviations
List of Figures
List of Tables
1 Introduction
2 To Bail-in or Not to Bail-in
2.1 The General Approach
2.2 The Regulator’s Decision Process
2.3 Effect of too-big-to-fail
2.4 Addressing the Problem
3 Event Study
3.1 Previous Studies
3.2 Context
3.3 Hypotheses
3.4 Methodology
3.5 Robustness Testing
3.6 Interpretation of Results
3.7 Shortcomings and Further Research
4 Conclusion
References
Appendix
List of Abbreviations
Abbildung in dieser Leseprobe nicht enthalten
List of Figures
Figure 1. Bail-in waterfall
Figure 2. Time line for the event study.
Figure 3. Index returns
Figure 4. Index returns for the entire time frame
List of Tables
Table I. Summary for treatment group
Table II. Regression results for the market model
Table III. One-sample t-test
Table IV. OLS regression results
Table V. Summary of the robustness checks
Table VI. Herfindahl-Hirschmann-Index
Table VII. Individual results for banks in the treatment group
Table VIII. Standard errors for the OLS regression
Table IX. T-tests for robustness checks
1 Introduction
During the recent financial crisis, resolution mechanisms proved to be insufficient to wind down systemically important financial institutions. Especially the bankruptcy of Lehman Brothers led to severe macroeconomic instability (Fleming & Sarkar, 2014). As a result, governments had to step in and bail out unsecured creditors with the help of guarantees, liquidity support or even nationalization (Laeven & Valencia, 2012). Thereafter, politicians univocally promised that taxpayers would never have to suffer for the mistakes of the financial industry again. National and supranational authorities implemented new regulations for preventing and managing bank failures, most noteworthy the Dodd-Frank Act in the U.S. in 2010 and the Single Resolution Mechanism (SRM) in the European Banking Union in 2016. Commissioner Lord Hill commented on the implementation of the SRM in a press release on 31 December 2015:
“This means that we now have a system for resolving banks and of paying for resolution so that taxpayers will be protected from having to bail out banks if they go bust. No longer will the mistakes of banks have to be borne on the shoulders of the many”
The new rules are designed to protect the taxpayer. Losses are now supposed to be primarily borne by stockholders and unsecured creditors. Exceeding shortages are covered by funds provided by the banking industry itself. If these improvements were credible to efficiently resolve any bank in any condition, the implicit government insurance for too-big-to-fail (TBTF) banks should vanish (Schäfer et al., 2016). However, regulatory improvements often lack credibility for large, systemic banks, as their resolution can still create much turmoil in the industry (Ignatovski & Korte, 2014). The source of the TBTF problem is that creditors can anticipate their bail-out and thus lose the incentive to monitor and discipline the respective bank (Sironi, 2003). For this reason, bail-out expectations can lead to disproportionate levels of risk-taking (Dam & Koetter, 2012).
The failure of Banco Popular in June 2017 put the SRM to the first test. After a deteriorating situation of the bank, depositors started a bank run. The Single Resolution Board (SRB), endowed with the mandate to handle distressed banks, had to step in and draw up a solution. Stockholders and junior bondholders were ultimately wiped out and Banco Popular’s competitor Banco Santander announced the purchase of Banco Popular for the symbolic price of €1 (Single Resolution Board, 2017a).
Schäfer et al. (2016) proved that the introduction of the SRM did not invoke a strong market reaction due to the lack of credibility for systemic banks. As the resolution mechanism apparently worked smoothly and did not contaminate other banks or even the non-banking industry, the question arises whether the resolution mechanism has consequently become more credible for bank investors. By analyzing market reactions to the Banco Popular bail-in, I primarily aim to answer this question. Hypothetically, if investors perceive the bail-in as a credible commitment by the SRB, the implicit insurance should cease. For this reason, investors would demand higher, risk-reflecting interest rates, impairing the bank’s profitability. Therefore, in the case that this is the dominant effect, a bank’s share price should fall.
The remainder of this thesis is structured as follows: In the following chapter, I will elaborate the regulator’s decision process and review the literature on how bail-out expectations create moral hazard, what the consequences are and how to address the problem. In chapter 3, I will present previous event studies on bail-out expectations and then conduct an event study to detect the market reaction to the Banco Popular bail-in. Thereafter, I will present possible shortcomings of my study and further open research questions. Chapter 4 concludes the thesis with a summary and stimulates further discussions.
2 To Bail-in or Not to Bail-in
When a bank enters a crisis, the regulator, with his mandate as a crisis manager, has primarily two options: He can either resolve the bank and wind the creditors down or he can rescue the bank with public funds. In this chapter, I will briefly explain the approach of the SRM, before examining the regulator’s motivation for choosing either of the alternatives or bailing-in or bailing-out a distressed bank. I will highlight the rationale behind the fact that the regulator predominately prefers and opts for a bank bail-out. It is essential to be aware of the regulator’s trade-off when evaluating any regulatory decision. After considering the trade-offs behind both policies, I will outline the subsequent moral hazard problem and additional costs on the economy arising from bail-out expectations. Finally, I will examine the proposed measures to tackle these problems.
2.1 The General Approach
The SRB was adopted in 2014 as an independent EU agency to ensure the orderly resolution of distressed banks within SRM jurisdiction. If a bank’s situation becomes critical, the European Central Bank, as supervisor, notifies the SRB that this bank “fails or is likely to fail” (European Banking Authority, 2015). Thereafter, the SRB examines the bank’s assets and liabilities and the possibility to bail the bank in, that is to wind the shareholders and subordinate creditors down in the order of the bail-in waterfall (see Figure 1). First Common Equity Tier 1 (CET1) capital, consisting of a bank’s core capital, then Additional Tier 1 (AT1) capital, that is convertible securities, and finally the Tier 2 (T2) capital such as subordinated debt is wound down (Basel Committee on Banking Supervision, 2011). Moreover, the SRB seeks either a private solution through mergers and acquisitions or liquidates the bank’s assets. If the point of non-viability (PoNV), defined by the Basel Committee of Banking Supervision (BCBS), is reached, the SRB has to decide to take further measures.
Figure 1. Bail-in waterfall
Figure 1 specifies the order in which the respective liabilities of banks are subject to bail-in.
Abbildung in dieser Leseprobe nicht enthalten
The Single Resolution Fund, financed by the banking industry itself, is only incorporated if the previous steps are found to be insufficient. Finally, if a bail-in is expected to cause too much turmoil, Art.44 (3) BRRD enables the resolution board to grant the distressed bank government support in exceptional cases and to bail specific creditors out.
The U.S. counterpart for resolutions is the Orderly Liquidation Authority (OLA). As they face similar decisions, I will refer to “the regulator” or “the decision maker” as the respective body responsible for financial stability from now on in this thesis.
2.2 The Regulator’s Decision Process
The regulator, in case of a bank’s distress, faces the decision whether to resolve or support the bank. Moreover, it is a trade-off between preserving market discipline with a bail-in and market liquidity with a bail-out in a multi-period game (DeYoung et al., 2013). For this reason, long term benefits of bail-ins need to be weighed against the short-term benefits of bail-outs.
Historically, central banks held the role of managing crisis (Gorton, 1985), but over time, their responsibility has expanded to preventing crisis as a regulator . This strengthens the time-inconsistency problem of their actions (Acharya & Yorulmazer, 2007): Ex ante, it is optimal to commit to strict bail-ins of a distressed bank in order to preserve market discipline, but ex post, it is predominantly[1] decided to bail the bank out.
2.2.1 Rationale Behind Bail-ins
Bagehot (1873) already recognized the moral hazard problems arising from the rescue of a failing bank and therefore advocated a laissez-fare policy. Credible commitment to resolutions that might mitigate the moral hazard effect, can only be achieved when the time-inconsistency problem is solved. In order to be time-consistent, the regulator needs to commit to his ex-ante pledge and to resolve the distressed bank ex-post. In this way, the regulator sacrifices the short-term liquidity for a long-term market discipline (Acharya & Yorulmazer, 2007).
Demonstrated by Schäfer et al. (2016), the actual execution of a large-scale bail-in, despite its possible negative effect on liquidity in the market, is a powerful instrument to tackle bail-out expectations and the moral hazard problem. Winding down unsecured creditors sends a credible signal that the resolution board might also be willing to resolve distressed banks in the future. Especially, in a stable political and economic environment, with an efficient resolution tool at hand, a resolution would be considered the better option (DeYoung et al., 2013). Moreover, by resolving a bank, the regulator can spare billions of taxpayer’s money (Honohan & Klingebiel, 2003). The Deutsche Bundesbank quantified the costs for supporting domestic banks after the financial crisis from 2008 at €236 billion (Deutsche Bundesbank, 2015).
2.2.2 Rationale Behind Bail-outs
Even though the policy makers are aware of the long-term disciplining effects of a resolution, they still often prefer the short-term advantages of bail-outs. To understand how to address the lack of credibility, it is essential to examine the reasons why the regulator might be unable to maintain his ex-ante commitment. Policymakers act rationally, and they have to weigh the costs against the benefits of a bail-out during their decision-making. Three major sets of incentives for bail-outs are prevalent in literature: Macroeconomic stability, personal reward and direct credit allocation (Stern & Feldmann, 2004).
2.2.2.1 Macroeconomic Stability
Firstly and most importantly in the discussion, policymakers fear spillovers from the distressed institution to other banks and the real economy. The failure of a bank creates the hazard of possible contagion to other banks (Diamond & Rajan, 2005), mainly due to the interconnectedness of banks. If banks are resolved, creditors either lose their investment, or have to wait until the bank’s assets are liquidated. Both negatively affect real spending activity in the economy and can therefore slow down economic growth (Ashcraft, 2005). Furthermore, banks hold vital functions for non-bank companies, for example by granting short-term liquidity (Saidenberg & Strahan, 1999). The substitutability of a bank is therefore dependent on whether the financial system is market-based or bank-based, that is whether another bank can undertake the function (Barth & Schnabel, 2013). DeYoung et al. (2013) emphasize the risk of a bank failure on the liquidity situation in a country and develop the following logic: Regulators tend to prefer bail-outs if the resolution mechanism is inefficient and the respective applied discount rate of long-term effects is high. In times of political and economic pressure, positive short-term effects are thus more highly valued in the liquidity-market discipline trade-off.
The “single most important factor leading to TBTF protection” (Stern & Feldmann, 2004, p.59), concerns regarding spillovers, is further specified by Barth and Schnabel (2013) with four levels of contagion within the financial sector. Firstly, the failure of one bank can lead to the perception for creditors, that other banks may also get into financial difficulty. This could lead to a bank run (information contagion, Acharya & Yorulmazer, 2008). Secondly, as a result of the interconnection of banks through the interbank market, the failure of one bank can also generate significant write-offs in open claims at other banks (Interbank market, Allen & Gale, 2000). Thirdly, failed banks need to liquidate their assets as a matter of urgency and this results in an excess supply of financial assets and often to an asset price downturn (Macroeconomic feedback, Brunnermeier & Pedersen, 2009). Finally, the failure of one bank creates expectations for remaining banks, especially in times of instability in the financial system. (self-fulfilling expectations, Barth & Schnabel, 2013).
Therefore, policymakers act in the public interest by preventing macroeconomic instability by bailing out a failing bank. Even though Stern and Feldmann (2004) state that the possibility for a spillover is often exaggerated by decision makers, they admit the magnitude of the possible consequences.
2.2.2.2 Personal Reward
Additionally, a principal-agent problem emerges in the regulator’s decision (Kane, 1990). This is because the public, more precisely the taxpayer, must bear the cost of a bail-out, and the regulator can benefit personally from a bank rescue. The decision maker might want to progress his career and can therefore proactively improve his ties with the private sector (Mailath & Mester, 1994). A bail-out is also less stressful and can lead to lower personal costs for the decision maker. Alternatively, it is assumed that the regulator’s mandate is judged on evident events and a bank failure might suggest incompetency (Kane, 2001). Dam and Koetter (2012) incorporate regional political factors to explain bail-out expectations
Even though accepting the existence of this rationale, Stern and Feldmann (2004) credit low reliance on personal motivation as an explanatory phenomenon for TBTF bail-outs.
2.2.2.3 Direct Credit Allocation
Finally, bailing out a bank can grant the government codetermination in the bank’s credit allocation. Hence, the political policymakers can influence the decision whom to award credit. These might either be companies, which are not sufficiently served by the private lending market, but considered by the government to be important for the economy (Stiglitz, 1993), or companies, which belong to a policymaker’s cronies, implying nepotism (Pomerleano, 1999). According to Stern and Feldmann (2004), this incentive is of lower relevance for countries with efficient central planning and independent governments.
2.3 Effect of too-big-to-fail
When the failure of a bank is considered to cause severe macroeconomic instability, this bank is deemed too-big-to-fail. In the following paragraphs, I will outline which bank criteria are detected in literature that make a bank’s failure undesirable. I will then elaborate on the effects of being TBTF on a bank’s risk-level. Finally, I will expand the scope of TBTF to further consequences on the economy.
2.3.1 Really too-big-to-fail?
In 1984, the Comptroller of the Currency created a new type of bank after insuring all creditors and depositors of the Continental Illinois Bank and announcing that eleven further banks in the state were important enough to save. These banks were considered TBTF and a failure would lead to undesirable turmoil in the system (O'Hara & Shaw, 1990). Regulatory improvements of resolution regimes thus often lacked credibility for the largest and most important banks, as shown by Ignatovski and Korte (2014).
Nevertheless, size is often merely heuristically used as a proxy for systemic relevance and has, conditional on systemic risk, little explanatory power to justify bail-out expectations (Barth & Schnabel, 2013). Zhou (2009) incorporates three measures to explain systemic importance, of which none is correlated with size. Banks are therefore rather too-systemic-to-fail (Rajan, 2009). Regardless of systemic risk, DeYoung et al. (2013) describe the effect of a bank’s complexity on bail-out expectations. When a bank’s business model is complex with regards to branches and activities, the resolution cannot be executed promptly and can therefore cause more disruption. Thus, the authors credit high importance on the too-complex-to-resolve rationale. Acharya and Yorulmazer (2007) argue that banks have an incentive to herd, for example by lending to a specific sector, so that they fail simultaneously. The more banks that fail at the same time, the less feasible an alternative for a private solution becomes. The failed banks are simply too numerous for a private solution and therefore too-many-to-fail. To revive the contagion risk (see chapter 2.2.2.1), banks can also be deemed too-interconnected-to-fail (Bernanke, 2009).
Furthermore, Barth and Schnabel (2013) proved that banks can also be too-big-to-save, that is the government cannot afford to bail the distressed institution out. A prominent example of this was the Icelandic Bank Kauphting. The bank had a balance sheet of 228% of the Icelandic GDP in January 2006 and was therefore simply too-big-to-save for the Icelandic government. Hence, size can also decrease bail-out expectations when exceeding a country’s financial capability (Völz & Wedow, 2011).
As demonstrated above, the label TBTF can be misleading as size is not the main variable of interest to measure bail-out expectations. Nevertheless, due to the prominence of the phrase and for the sake of overview, I will continue to use the term TBTF.
2.3.2 Effect on Market Discipline
In theory, investors discriminate between safe and risky investments by adjusting the respective price of unsecured securities (Chen & Hasan, 2011). This ideally provides an incentive to bank managers to choose an adequate risk level. For this reason, firstly investors need to understand a bank’s risk level and then be capable to affect the manager’s decision via price and quantity signals (Bliss & Flannery, 2002). As a result, managers are influenced by the markets via two channels: Directly through capital costs and indirectly through signalling for bank supervisors (Sironi, 2003).
However, when investors do not see that their full investment is at stake because of bail-out expectations, their incentive to monitor and influence the management’s risk level vanishes (Sironi, 2003). Consequently, banks do not have to pay a premium for taking higher risks. Price and quantity signals, that would have influenced the management’s decisions, are dampened by the government’s implicit insurance. Furthermore, the safety net subsidizes the bank through lower capital costs ceteris paribus, resulting in a wealth effect for covered banks (O'Hara & Shaw, 1990).
2.3.3 Creating Moral Hazard Through Implicit Insurance
As funding costs are not further directly connected to a protected bank’s risk engagement, managers are incentivized, with the goal of expected profit maximization, to engage in high risk - high return investments (Black & Hazelwood, 2013). A bank can benefit from the upside potential, but does not have to bear the consequences of the downside potential of risky investments, as it is protected by a safety net. Shavell (1979) argued that moral hazard leads to the tendency that insurance protection decreases the insured individual’s incentive to avoid losses. This is just as well as the insurance for TBTF banks induces excessive risk-taking through moral hazard.
There is abundant empirical evidence that covered banks take higher risks. Dam and Koetter (2012) show that a bank’s risk level is dependent on the estimated bail-out probability. Black and Hazelwood (2013) and Duchin and Sosyura (2014) prove that banks that have previously been supported by state aid, expect government support again in the future and thus increase their ex post risk level on average. Duchin and Sosyura (2014) also prove that banks close to distress have an incentive to “gamble for resurrection” with large upside potential. Gropp et al. (2013) consistently demonstrate that discontinued guarantees for German saving banks helped to decrease the risk level.
2.3.4 Charter Value
An opposing effect of the implicit insurance of TBTF banks is created by the increase of charter value. Due to lower funding costs, the firm value of a covered bank increases ceteris paribus. Hence, managers would jeopardize more value by engaging in riskier activities and are incentivized to decrease a bank’s risk level (Keeley, 1990). Hakeness and Schnabel (2010) argue that the net effect on a protected bank’s risk level can be lowering, in the case of low transparency and an elastic deposit supply. Cordella and Yeyati (2003) demonstrate that bail-out guarantees for adverse macroeconomic shocks can incentivize banks to lower levels of risk. When only failures, that are independent of the manager’s decision-making scope, are rescued, the moral hazard effect can be muted, but the charter value effect increases. Nevertheless, there is more empirical evidence on an increased risk engagement, that is, a subordinated charter value effect (Black & Hazelwood, 2013; Duchin & Sosyura, 2014).
2.3.5 Further Effects
The media and politicians are mainly concerned with fiscal bail-out costs borne by the taxpayer. Nevertheless, the TBTF problem induces further costs on the economy and these are all a waste of resources. Stern and Feldmann (2004) highlight this by estimating the fiscal costs from the 1980’s savings and loan crisis to $150 billion and the lost output associated with the subsequent moral hazard problems to $500 billion.
As elaborated in chapter 2.3.2, banks generate advantages of being TBTF, mainly through lower funding costs. For this reason, they wish to be considered too-big, either through organic growth or through mergers and acquisitions (Kane, 2000). In this way, disproportionally large banks are created, which cannot be justified by increased cost-efficiency or revenue generation.
TBTF banks have a comparative advantage against non-covered banks through lower funding costs. Comparable to the inefficiency of public-sector firms (Bartel & Harrison, 1999), covered banks have less pressure to operate cost-efficiently. Taking the finding from Shleifer (1998) into account, namely that government owned firms are less innovative, Stern and Feldmann (2004) speculate that also government protected firms show a lower level of innovation.
Moreover, bail-out guarantees do not only affect the risk level of protected banks, but also of their unprotected competitors (Gropp et al., 2011). Insured banks have lower funding costs and can therefore price their products more aggressively by granting higher deposit rates and lower credit rates. In order to compensate for this competitive disadvantage, uninsured banks are urged to increase their risk level. This is not only “unfair” (Fukao, 2004), but also increases the instability of the system (Hakeness & Schnabel, 2010).
2.4 Addressing the Problem
Since managers and investors can estimate the underlying rationale (see chapter 2.2), they are able to anticipate a bail-out. The primary goal of regulatory improvements should be to tackle these expectations credibly and thus the subsequent moral hazard effect. However, the largest effect on credibility is achieved through actual, successful bail-ins (Schäfer et al., 2016). If managers and owners fear to lose their position and investment, and if creditors and depositors decrease their expectations tobe bailed-out, banks can be disciplined towards less risk-taking. As bankruptcy is a natural resource reallocation process, the regulator should not aim at prohibiting bank failure, but should aim to decrease the subsequent spillover effect (DeYoung et al., 2013).
Significant improvements of the resolution regime can decrease the fear of spillovers and therefore the incentive for the regulator to renege on his pledge (Stern & Feldmann, 2004). This regime needs to incorporate the special role of banks and should therefore be more applicable than the general corporate bankruptcy law. The resolution board must to be endowed with both sufficient legal and financial resources to resolve a bank promptly and efficiently (Ignatovski & Korte, 2014). Götz and Tröger (2016) argue that bail-in able debt should only be held by sophisticated, non- bank investors, who are able bear the possible losses. Hence, the resolution of a bank generates less turmoil in the economy ex post and is therefore more credible ex ante. Additionally, all banks need to be covered by the resolution regime comprehensively, with special attention to TBTF banks, in order to avoid risk shifting from unprotected to protected banks (Ignatovski & Korte, 2014).
To continue this thought consistently, Stern and Feldmann (2004) argue to employ a regulator who has proved strictness in the past and is therefore more credible to enforce the resolution mechanism rigorously. Such a regulator would not be prone to the temptations of a bail-out (see chapter 2.2.2.2).
[...]
[1] Goodhart and Schoenmaker (1995) analyzed 104 bank failures across 20 countries, where the decision maker chose to bail-out 73 times.