CONCLUSIONES Y RECOMENDACIONES
5.4. Fundamentación Científica
This study investigates the impact of securitisation and CBs on the credit risk taking behaviour of European banks in the short and the long run. The comparison of securitising banks with CBs issuing banks and the use of the CBPS methodology are distinguishing features of this
30The analysis was performed using ‘twang’ package inR. 31Results are reported for the outcome model only.
study. Some studies suggest that CBs issuing banks might have a lower inclination towards credit risk taking because of the so-called ‘skin in the game’, whereas securitising banks tend to issue greater amount of risky assets because of the opposite reasons. Even though securitising banks maintain enough ‘skin in the game’ in one form or the other, regulators emphasise upon on-balance sheet risk retention. The results reported in this study suggest that the absence of ‘skin in the game’ in the form of on-balance sheet retention may not be the driving factor for the banks to take the elevated risk. The results of multinomial model did not confirm that securitisation is used for the credit risk transfer by European banks. These results are in line with Cardone-Riportella et al. (2010) and Martín-Oliver and Saurina (2007). However, this study evaluates the ex post effects of securitisation on credit risk taking behaviour of banks, particularly focusing on the role of on-balance sheet risk retention. The use of CBs for comparison with ABS serve this purpose, as CBs are on balance sheet instruments.
The results, although, suggest a positive relationship of securitisation with LLP in the short run but this relationship becomes negative in the long run. The negative relation of securitisation with LLP in the long run cannot be translated as risk under-pricing when securitisation does not lead to an increase in risk taking. These empirical findings are attributed to the retention of the first loss position in the securitisation transaction to generate a signal about the quality of securitised products, as explained by Gorton and Pennacchi (1995), and implicit/explicit recourse provided to investors (Casu et al., 2011; W. Chen, Liu, & Ryan, 2008; Cheng, Dhaliwal, & Neamtiu, 2011).
There can be many other plausible explanations for the decline in underwriting standards in the loan origination process during the pre-crisis years. The optimism of lenders and investors because of the housing prices boom is a potential reason for the underestimation of the risk (Hellwig, 2008). According to Ruckes (2004), the changes in lending standards are linked with the economic cycle. At the time of an economic boom, banks screen their borrowers superficially. Banks tend to lend aggressively because of the intensifying competition and they want to save the resources spent on screening (Cerasi & Rochet, 2014). A bad shock in the economy may cause the failure of many projects. The widespread losses during the GFC are consistent with the belief about an unusual bad shock as explained by Chiesa (2008)
The deterioration of lending standards and lax screening can also be linked with com- pensation practices. The compensation in the banking system is often linked to the volume of loan origination and sale of high fee products (Segoviano et al., 2015). As a result, the suitability of products sold to borrowers is often ignored. This results in predatory lending that may deteriorate the performance of banks’ loan portfolios (Surti, 2010). Therefore, lax
lending standards cannot necessarily be linked with the issuance of a specific security, but can be caused by increased competition during the economic boom and compensation practices.
Summarising, this study suggests that European banks use securitisation and CBs as funding tools and not as a risk management tool. The changes in the risk level of banks can be attributed to the increase in liquidity but securitisation is not the only tool used to generate liquidity. CBs and many other similar instruments also serve the purpose of increasing bank liquidity. When banks are flushed with liquidity, they might get the motivation to take higher risk because of excess capacity. However, admonishing securitisation solely may not be the right approach.
The regulations enacted after the GFC which emphasise on risk retention seem unneces- sary. Risk retention may not help reduce the risk level as shown by the positive impact of CBs in the long run. Hence, CB is not the silver bullet it is flaunted. Regulators should identify the loopholes and distortions in the process of securitisation. These distortions should be controlled, instead of penalising the entire process and the securitisation market.
Chapter 5
Effects of Securitisation and Covered
Bonds on Bank Stability
Banking is a very treacherous business because you don’t realize it is risky until it is too late. It is like calm waters that deliver huge storms.
Nassim Nicholas Taleb
5.1
Introduction
Funding sources of banks were put under extensive scrutiny after the Global Financial Crisis (GFC). Banks were not able to use securitisation to raise funds, as investors lost confidence in this market. Banks heavily involved in securitisation faced liquidity and funding pressures during the GFC that turned into a full blown systemic crisis (Blommestein et al., 2011). Amid this situation, a shift from the preference for returns to the preference for safety was observed among investors. Covered Bond (CB) — a close counterpart of Asset-Backed Securities (ABS) and adual recourseinstrument1— emerged as an important alternative to ABS to meet investors’ preference for safety and to provide funding to banks.2 Regulators also provided
1Covered bonds holders enjoy a dual recourse. First, they have a preferential claim over the assets held in the cover pool backing the issuance of these bonds. This pool remains unaffected in post-bankruptcy situation. Second, CB holders have an unsecured claim on bankruptcy state of issuers if the cover pool does not generate enough cash or its value drops. In such a condition CB holders are treatedpari passuwith unsecured creditors of the credit institution and in some cases they also get a preferential treatment over unsecured creditors.
2The market for covered bonds remained resilient to the shocks of GFC in 2007/8 when ABS market faced a decline of 51%. However, the CB market grew by 5% in 2009. CBs outperformed government bonds as well during the sovereign debt crisis in Europe in 2009.
a favourable treatment to CBs that incited the interest in this class of instruments.3 On the contrary, securitisation faces stringent regulations after the GFC. Many market participants are concerned about this discriminatory treatment of ABS vs CBs, as they think it may result in a distortion of the funding base diversification of Financial Institutions (FIs). It may have severe implications for the individual bank stability and the financial stability.4 This study contributes to the literature by examining how the focus on a particular instrument affects the bank stability.
The proponents of CBs do not support the concerns about the increasing use of CBs, mainly for two reasons. First, issuance of CBs can never go beyond a certain limit because of their on-balance sheet nature and strict requirements for the cover pool. The on-balance sheet nature of CBs requires banks to hold capital against these bonds. The issuance of new capital can be difficult, hence CB issuance beyond the limits of current capital can be challenging for a FI. The strict requirements for the cover pool make it harder for the banks to have enough assets to be used against CB issuance. Second, the strict requirements for the cover pool and its continuous monitoring by the regulatory authorities make it harder for risks to materialize to the extent of threatening the bank stability (ECBC, 2013).
Most of previous studies have been evaluating ABS and CBs (hereinafter collectively referred as Securities and Bonds (SB)) from the investors’ perspective — especially CBs (e.g. Buchholst, Gyntelberg, & Sangill, 2010; Prokopczuk, Siewert, & Vonhoff, 2013). The regulations devised after the GFC are also meant to provide greater protection to investors. However, the effects of CBs on issuing institutions are not yet explored. This study evaluates the issuance of ABS and CBs from issuers’ perspective. Despite having important differences and being affected differently during the GFC, both instruments share many similarities.
This study helps in developing an understanding of the effects of ABS and CB issuance on bank stability that is assessed in terms of systemic risk, and is measured with the SRISK and the Long-Run Marginal Expected Shortfall (LRMES) of Brownlees and Engle (2015). The literature is widely divided between the “securitisation-fragility” and the “securitisation- stability” views. This study presents a different view, denoted “securitisation-scalability”. CBs are also analysed along similar lines. The study argues that the relationship of ABS or
3See Regulation (EU) No. 575/2013 (Capital Requirement Regulations (CRR)) Articles 129(5), 395, 400(1a), 416(2a), 509(3a), Credit Rating Agency (CRA) Regulation (EU) No. 462/2013 Article 8c (1), Liquidity Coverage Ratio (LCR) and the securitisation frameworks of Basel III and the haircut regulations of European Central Bank (ECB).
4CBs may generate greater amount of risk for issuers as they must actively manage the underlying pool of collaterals. Therefore, from a financial stability perspective, CBs might not prove to be risk-free and an over-reliance on them may result in risk concentration in the banking system.
CBs with bank stability is not totally distinct; rather, it varies with the issuance level of these instruments. Their respective volumes determine their implications on bank stability. The potential benefits of these funding sources might not be accessible either beyond or below a certain level. For these reasons, it is often argued that limits should be imposed on the issuance of these instruments. Such regulations have been imposed on CBs in some countries like Australia and Belgium.5 The empirical analysis performed here is likely to provide a foundation to evaluate these regulations.
The data is taken from 46 ABS and CB issuing banks from Europe for the time period between 2000-2014. The study performs an extended analysis of both instruments, aiming at evaluating the link between the scale of the issuance of these two instruments and the systemic risk of banks. The analysis focuses on the variation in the relationship of ABS and CBs with bank stability with respect to changes in the issuance scale of these two instruments. The study also tests the impact of size on the relationship between SB issuance and the systemic risk of banks. The analysis in the study starts with a quadratic model and extends to a partially linear setting that uses Generalised Additive Model (GAM).
The empirical results show the presence of a U-Shaped relationship between ABS and systemic risk. The issuance of ABS initially helps banks in controlling their systemic risk, but the impact is reversed when a bank keeps on issuing ABS. This is what is dubbed by some as ‘securitisation beyond limits’, and by this study as “securitisation-scalability”. For covered bonds, initial estimations showed that this relationship is of the opposite nature. Small issuance of CBs leads to an increase in systemic risk, but large scale issuance of CBs decreases systemic risk. I attribute this relationship to bank size and jumbo CBs. Further investigation revealed that smaller banks face a higher systemic risk when they issue CBs but larger banks remain unaffected. These findings do not support the proposal of putting a uniform limit on CB issuance. Such a proposal should be linked with the bank’s size. Moreover, the regulators should think about devising a framework that can control the limitless issuance of ABS.
The remainder of the chapter is organized as follows. Section 5.2 provides some insights from the previous literature. There are very few studies on CBs, but ABS have been studied widely. Section 5.3 explains the research methodology. This section provides economic intuitions and then an empirical model to test the suggested relationship and explains the characteristics of the data including the sample details and descriptive statistics. Section 5.4 provides the empirical results of the model tested herein. This section provides the results of estimations of a quadratic model and generalized additive specification of a partially linear
model. Section 5.5 discusses the results and draws conclusions.