The impact of post-crisis financial regulation on the shadow banking system : a panel data analysis

(1)

UNIVERSITY OF AMSTERDAM

The Impact of Post-Crisis Financial Regulation on the Shadow

Banking System: A Panel Data Analysis

Master Thesis

July 2017

MSc. Finance: Corporate Finance and Banking

Eveline van ’t Spijker 10400982

(2)

-Statement of Originality-

This document is written by Eveline van ’t Spijker who declares to take full responsibility for the contents of this document. I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it. The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.

(3)

Master Thesis

The Impact of Post-Crisis Financial Regulation on the

Shadow Banking System

Eveline van ‘t Spijker

July 2017

Abstract

This paper analyzes the impact of post-crisis financial regulation on the securitization activities of banks, and the impact of these regulations on the banks’ systemic risk contribution to the financial system as a whole. Besides that, we examine whether the securitization activity in our sample is driven by regulatory arbitrage. Specifically, the 5% risk-retention rate, the disclosure and due diligence requirements and the Basel III regulatory framework are examined. Using a panel data fixed effects model for a sample of 68 European and US banks over the period 2010- 2016, we do not find support for regulatory arbitrage. We find that the risk retention rate increases securitization activity and that the Basel III regulations negatively affect securitizations, albeit this finding is more pronounced for European banks. Finally, our estimates do not provide evidence for a significant change in the systemic risk contribution of these banks due to the new regulation, implying that shadow banking is not significantly different than before the 2007-2009 financial crisis.

(4)

List of figures and tables

Figure 1. Schematic overview of the securitization process………... 12

Figure 2. US and European historical securitization issuance……… 16

Figure 3. Basel II versus Basel III capital requirements………. 22

Table 1. Summary statistics for the sample of banks……… 34

Table 2. Summary statistics of securitization transactions……… 35

Table 3. Largest securitization transactions before and after post-crisis regulation…………. 36

Table 4. Securitization activities and post-crisis financial regulation………... 40

Table A.1 Sample of Banks……….. 51

Table A.2. Basel III phase-in arrangements……….. 52

Figure A.3. Securitization activity aggregated by year for the sample of banks……… 52

Table A.4. Pairwise correlation matrix of all variables included……….. 53

Table A.5. Systemic risk contribution over the period 2010-2016………... 54

Figure A.6. Cross-correlations of bank's risk in isolation and bank's systemic risk………... 54

Table A.7. Securitization activity and post-crisis financial regulation - Robustness Test……… 55

Table A.8. Regulatory arbitrage, securitization and post-crisis regulation - Robustness Test…. 56 Table A.9. Systemic risk and post-crisis financial regulation - Robustness Test………. 57

(6)

1. Introduction

he collapse of the shadow banking system in 2007-2008 played a significant role in the destabilization of the traditional banking system leading to the global financial crisis, and later on the Euro sovereign debt crisis. The benchmark of the Financial Stability Board (FSB) defines shadow banking as “credit intermediation involving entities and activities (fully or partially) outside the regular banking system” that raises systemic risk concerns in particular by maturity and liquidity transformation, the use of leverage, incomplete credit transfer and regulatory arbitrage”. Although this definition might be a useful benchmark, it defines shadow-banking activities as operating primarily outside the banking system, while many activities took place under the umbrella of large bank holding companies (Adrian, 2014). A majority of these activities involved securitizations, which is hereafter used as benchmark for shadow banking activities by banks. Because the traditional bank credit intermediation and non-bank credit intermediation are closely linked to each other, the crisis that started in the shadow banks soon spread to the regulated banking system. This interconnectedness and risk contribution of the shadow banking system to the traditional banking system has become the focus of post-crisis policy responses (Adrian, 2014).

In the decades prior to the 2007-2009 financial crisis, this shadow banking and the relative use of shadow banking short-term funding –e.g. commercial paper, securitized bank loans, and debt issued by non-bank entities - became increasingly important (Duca, 2016). There are several factors that contributed to the size of the shadow banking system, such as the economic growth, low interest rates, and the demand for safe assets by investors and banks.

This demand for safe assets was met through financial innovations such as securitization. In a typical securitization process, the issuer transfers financial assets to a Special-Purpose Vehicle (SPV), which sells asset-backed securities (ABS), -claims to future cash flow generated by these securitized assets- to outside investors (Ryan, Tucker, Zhou, 2016). Securitizations became increasingly popular, especially by financial institutions and other financial services firms that could now securitize their risky loans and create safe assets through credit, maturity and liquidity transformation (Gennaioli, Schleifer, Vishny, 2013). This securitization was also a way for banks to lower the capital requirements set by the Basel I and II accords. This led more capital-constrained banks to set up SPVs, also known as ABS conduits, to avoid the capital requirements (Acharya, Schnabl, Suarez, 2013).

A major downside to the securitization process is the exposure to systemic risk. Banks often provided liquidity guarantees to their SPVs, which created perceptions that shadow banking was stable. Banks, credit rating agencies and investors neglected this downside tail risk in the guarantees that could lead to a systemic breakdown. When the housing market started to realize losses and the aggregate tail risk of a systemic breakdown faced by the banking system became apparent, a run on the shadow banking system was inevitable (Gorton, Metrick, 2012). The liquidity crises that started in the

(7)

securitized assets market, called for liquidity support of the sponsoring banks, and soon spread to the traditional banking system leading to the global financial crisis. This suggests that while the credit risk transfer may have reduced banks’ individual risk, they posed greater systemic risk to the financial system, challenging the implementation of efficient financial regulation (Nijskens, Wagner, 2011).

In the aftermath of the financial crisis, policymakers started to assess the securitization process leading to several regulatory measures to address the flaws of the securitization process and off-balance sheet banking in general. The major measures affecting securitization in Europe and the US are the 5% retention rate, or “skin in the game” rule, and due diligence and disclosure requirements. In addition, the new capital and leverage requirements of the Basel III framework, although still recent and not fully phased in yet, affect securitization transactions through the risk-weights. This brings us to the empirical question this paper examines: “What is the impact of post-crisis regulation on the sensitivity of the traditional banking system to the shadow banking system?”

In particular, this paper examines the sensitivity in two dimensions; securitization activity of the banking sector, and the contribution to systemic risk of the originating banks. To answer this question, the paper develops three hypotheses. First, we hypothesize the regulatory arbitrage argument that capital constrained banks are more likely to engage in securitization. Acharya, Schnabl and Suarez (2013) found that this was a major driver for securitization in the pre-crisis period, and this paper test if this relationship still holds with the new regulation in place. Second, we hypothesize that the above- mentioned post-crisis regulatory measures negatively affect the securitization origination by banks. Although not meant to reduce securitization activity, the post-crisis regulatory measures might impede banks from securitizing or have taken away the incentives to securitize. The third hypothesis states that post-crisis regulation changes the form of shadow banking by banks. We will analyze this different form of shadow banking by examining the impact of post-crisis regulation on the systemic risk contribution of securitizing banks. It is argued that the focus on the particular risks of securitization could lead to other forms of shadow banking that could also endanger financial stability (WRR, 2016). We examine securitizations and no other structured finance products, since it is the most common type of structured- finance transactions, and although its use fell dramatically during the global crisis, it was gaining volume again when financial market started to recover (Ryan et al., 2016).

In order to conduct the empirical analysis, we use a novel panel data set consisting of quarterly securitization issuances of US and European banks from January 2010 to December 2016 that we call the post-crisis period. Besides that, we will use the systemic risk measures ⊗CoVaR and ⊗$

CoVaR, as presented by Adrian and Brunnermeier (2016), to estimate the effect of the implemented regulation on the systemic risk contribution of the banks. The main findings reveal that post-crisis securitization activity is not driven by the tendency to arbitrage capital requirements. The results further reveal that the publication of the retention requirements significantly increases securitization activity by banks, while the due diligence and Basel III leverage ratio decrease the

(8)

primarily driven by the European banks in our sample and it is most likely that the negative impact of the disclosure requirements is largely due to the general business cycle. We do not find supporting evidence for different shadow banking. Importantly, the post-crisis regulation does not significantly impact the systemic risk contribution, while the subsample of Globally Systemically Important Banks (G-SIBs) seems to contribute mostly to any changes in systemic risk.

This paper contributes to existing literature in several ways. First of all, there is little empirical evidence in general on regulation and prudential supervision and how it might affect the banking sector. The empirical assessment in terms of securitization activity and systemic risk contribution has to the best of our knowledge not been executed before. In addition to that, we assess the impact of regulation in terms of systemic risk and not on the banks’ risk in isolation, because neglecting excessive risk- taking from the systemic perspective is exactly what severed the financial crisis. Second, this study provides insights on a specific but major part of the shadow banking system, the origination of securitized assets by commercial banks. Notwithstanding it complements traditional banks in the creation of credit, its structure can also jeopardize systemic stability. Therefore the bigger picture that this paper aims to address is to assess the effectiveness of the implemented regulation, the reason being that shadow banking has macro economic implications, while regulatory frameworks often focus on individual bank characteristics. For regulators, it is important to monitor and assess the development of implemented regulation and to find the right balance between the benefits of securitized banking and simultaneously mitigating unintended side effects (IMF, 2014).

The remainder of the paper is organized as follows: section 2 presents a literature framework about shadow banking in general. It describes how the shadow banking process differs from the traditional banking system and how the two are interlinked. It further explains the securitization process and how it contributes to systemic risk. In addition it discusses the new regulatory framework affecting the securitized banking system. Section 3 discusses the data, methodology, chosen control variables, the results and their implications. Section 3 further performs additional robustness checks. In section 4, the limitations of this paper will be discussed and section 5 summarizes and concludes.

(9)

2. Literature Review

2.1 Shadow Banking

This section starts with describing how the traditional banking differs from the shadow banking system, and how they are linked to each other. Besides that, a description of the shadow banking securitization process and its instruments will be provided. Furthermore, this chapter will describe the rise and the major drivers of the shadow banking system and evaluate the regulatory responses to the fall of the shadow banking system in the aftermath of the 2007-2009 financial crisis.

2.1.1 From Traditional Banking to Shadow banking

In an environment of full available information and no regulation, the choice between traditional banking in the form of deposit funding, and shadow banking in the form of securitized funding should be irrelevant. However, this changes when banks have to consider the presence of deposit insurance, government intervention, capital requirements and asymmetric information (Bertay, Gong, Wagner, 2017). This section explains how these considerations can move traditional banking towards shadow banking and how the two systems are related to each other.

In the traditional banking system with deposit insurance, depositors transfer money to the bank, and receive insured savings in return. The bank uses these funds to provide e.g. a mortgage to a borrower, whose house serves as collateral. The bank holds this mortgage on the balance sheet along with other loans to retail and commercial customers. Hence, in the traditional banking system, loans are originated and held by the bank, and the main source of funding is the insured demand deposits. This deposit insurance only applies to retail-depositors, and reduces the incentive for these depositors to withdraw their money. In addition, public funds (e.g. government safety net) serve as a backstop in the traditional banking system (Gorton, Metrick, 2012). Backstops such as credit guarantees by deposits and contingent liquidity by lending of last resort, were created to ensure stability of the traditional commercial banking system. However, these guarantees also incentivized banks to engage into excessive risk-taking and maturity transformation, motivating the implementation of supervision and prudential regulation (Adrian, 2014). Such regulation involved for instance capital requirements since individual banks do not internalize the costs of their risk-taking on other banks or the economy as a whole. However, the presence of explicit or implicit government guarantees and capital requirements did actually incentivize risk-taking and the arbitraging of regulation, especially in the form of securitized banking (Acharya et al., 2013).

Consequently, in the decades prior to the crisis, this traditional form of commercial banking was accompanied by securitized credit intermediation. The securitized banking system differed from the traditional banking system in the sense that it had no deposit insurance, no public backstop, and designed to transfer credit risk. Although this securitized banking system was not subject to the same regulation as the traditional banking system, it can mimic deposit insurance similar to one in the

(10)

regulated system, for non-retail depositors such as sovereigns and mutual funds. Investors transfer large deposits to the bank and receive collateral in return. These deposit-collateral transactions are in the form of a repo-agreement, where the investor buys an asset from the bank and the bank agrees to repurchase this asset in the future. The %age difference between this repurchase price and purchase price is the repo rate and is similar to the interest rate on a bank deposit. The banks have to hold a repo haircut; a fraction of their assets in reserve when they borrow money in the repo market. In case the bank defaults on this repo, the investor has the right to end the agreement and sell the asset.

Moreover, the shadow banking entities, known as SPVs or conduits, operating under the umbrella of bank holding companies often had liquidity guarantees (Adrian, 2014). This is because most of the regulated commercial banks provided explicit liquidity guarantees to their SPVs, the commitment to repurchase maturing securitized assets in case that the SPV could not roll over the securities. This liquidity support or recourse is not legally required, however originators often provided support because of reputational risk, since no support might worsen their future access to capital markets (Gorton, Souleles, 2005). As we will see in the following section, the loans originated by the bank are no longer held on the balance sheet, but distributed. Therefore, it runs in the shadow of the regulated banking system because it is off-balance sheet and not subject to bank regulation (Acharya, Schnabl, Suarez, 2013).

Because of the above-mentioned implicit guarantees of commercial banks through their conduits to investors, recourse risk was created from the conduits back to the balance sheets of the commercial banks. Effectively, the risk was not transferred to outside investors, but instead concentrated on the bank’s balance sheet, which is in contrast to what was perceived as a risk transfer in the securitization process before the crisis (Acharya et al., 2013). Furthermore, the guarantees provided to the SPVs could eventually lead to government intervention enjoyed by the regulated banking system, since the execution of guarantees caused financial problems for the banks, which in turn increases the use of public funds (Górnicka, 2016). This means that even though shadow banking is perceived as prevailing outside the regulated banking system, it is interlinked to the traditional banking system by the repo-agreements and liquidity guarantees

2.1.2 The Shadow Banking Securitization Process and its Instruments

This section explains how the securitized banking system transforms illiquid assets into safe, short-term and liquid assets by the means of SPVs, and how it mimics the traditional banking system in the credit, maturity and liquidity transformation. It further discusses the most common securitization instruments.

As displayed in figure 1, the securitization process starts at the originator, also referred to as sponsor, which is the bank that originates the underlying loan. The originator then transfers the loans to the issuer. The issuer is usually in the form of a conduit or Special Purpose Vehicle (SPV), a company that has been set up to carry out the securitization transactions. The transfer of securitizations from the originator to the SPV can be accounted for as either a ‘true sale’, or ‘secured borrowing’ (synthetic) as

(11)

set out in the Financial Accounting Standard No. 140. In practice, the majority of the securitizations transactions prior to the crisis were accounted for as ‘true sale’, and hence receive off-balance sheet treatment (Cheng et al., 2011).1 One of the criteria of a true sale is that the underlying pool of assets is held separately in the SPV, so the assets are removed from the banks balance sheet. The reason being, that in case of default of the originator, the assets held in the SPV will not be affected. More specifically, SPVs have no bankruptcy costs since they are structured in such a way they cannot go bankrupt. This feature is known as ‘bankruptcy remoteness’. To achieve the structuring of an SPV as bankruptcy remote, its business activities are constrained and the issuance of debt is limited (Gorton, Souleles, 2005). Because of this bankruptcy remoteness, the financial position and credit rating of the originator become irrelevant to the investors, leading to a higher credit rating for the SPV than the originating institution. This is called credit enhancement, which means that credit quality is obtained from a third-party guarantee and the securitized assets are rated at investment grade up to AAA-grade, usually resulting in a higher rating than the average of the underlying pool. The SPV then issues the securities that are backed by the underlying cash flows of the assets. The last step involves the partitioning of the notes into different classes of seniority to reflect the underlying credit risk. The lowest rate is assigned to the most junior tranche and often referred to as the ‘first-loss tranche’, since it absorbs the initial losses. European originators mostly retained this first-loss peace on their balance sheet (Baig and Choudry, 2013). These notes are then sold to outside investors such as insurers and hedge funds. When securitizations are not isolated from the balance sheet and accounted as secured borrowings, the assets remain on the bank’s balance sheet and the proceeds from the transfer are recognized as a liability (Cheng et al. 2011).

Claessens et al. (2012) describe the steps in figure 1 as different stages of risk transformation. The first step of the process involves the transfer of maturity risk. This means selling the long-term AAA security to the SPV that is funded in short-term money markets, transforming it into a short-term asset, also known as structured money market instrument (SMMI). The second step is the transfer of liquidity risk. The asset is transformed into a safe, short-term and liquid asset through the use of puts, which is the obligation of banks to give liquidity support to the vehicle, and implicitly the issuing of stable net asset value (NAV) claims to the investors in case the SPV could not rollover the securities. Finally, the third step involves the transfer credit risk: the tranching of a pool of loans into an equity, mezzanine and safe long-term AAA security.

In summary, the loans follow a chain of special purpose vehicles and transformations and are repackaged into safe, short-term, liquid assets. Even though the securitized banking system as displayed in figure 1, was traditionally meant to transfer risk from the single bank’s balance sheet, along a chain of balance sheets, it was eventually securitization without risk transfer. The assets that were sold of balance sheet appeared to flow back in case of distress, thereby retaining all the risk on the balance

(12)

sheet of the bank (Acharya et al., 2013). The securitization process created safe assets in way that was risky and unstable.

Figure 1. Schematic overview of the securitization process

In the securitization process, several terms relate to the different securitized instruments and underlying pool of assets (Jobst, 2002). In general, mortgages are mostly used as underlying assets in the securitization process, but often consumer loans, loan receivables, short-term commercial paper, credit card receivables, auto loans, lease receivables, student loans or second mortgages are also used. The development of the structured finance market created two prime asset classes that serve as underlying collateral; Asset-backed securities (ABS) and collateralized debt obligations (CDO) (Jobst, 2002). ABS is the term used for the securitization of corporate and sovereign loans, auto loans, credit card receivables, student loans etc. Asset-backed commercial paper (ABCP) has the same characteristics as the true sale of ABS, but commercial paper is a short-term financing instrument, used for short-term funding needs. Mortgage-backed securities (MBS) are also a subcategory of ABS and include residential-mortgage backed securities (RMBS) and commercial-mortgage backed securities (CMBS). The second asset class involves the CDOs. The main difference from the ABS and MBS class is that it

(13)

is backed by debt obligations instead of consumer products. The two main subcategories of CDO are collateralized loan obligation (CLO) and collateralized bond obligation (CBO), but several other categories exist such as collateralized mortgage obligation (CMO), CDO of ABS and CDO of CDO (Jobst, 2002). We will use the origination of these instruments in the empirical analysis.

2.1.3 The Economic Benefits of Securitization

When properly structured, securitization has several benefits to banks and investors, and is a beneficial complement to the traditional banking system by increasing access to capital markets, supporting liquidity, reducing maturity mismatch and better risk sharing (IMF, 2014). First of all, securitization could lower the funding costs for banks. For example, banks that have difficulties funding their existing loans can use securitization of loans to get access to more liquid capital markets. This might in turn increase the credit creation and lending activities of banks, resulting in lower borrowing rates. Besides that, the higher credit rating of the SPV compared to the originating banks results in lower funding cost of issuing securitized notes than issuing assets by the bank itself (Baig, Choudhry, 2013). Furthermore, securitization helps banks to more efficiently manage their asset-liability duration mismatch. This mismatch arises because long-term loans are funded with short-term deposits. As explained in section 2.1.2, securitization allows for maturity transformation reducing this mismatch (SEC, 2014).

A potential benefit shown by Boot and Thakor (1993) is that under asymmetric information the pooling of assets and the partitioning of claims could enhance the expected revenue of the issuing bank, compared to issuing a composite security. More specifically, by splitting a composite security in a senior security with no informational asymmetry, and in another informative sensitive security, the more senior security is made informative insensitive and can thus be priced closely to its true value. Since this made the other security more informative sensitive than the composite security, informed trading becomes more profitable and increases the issuers expected return from selling these informative sensitive claims.

Another benefit of securitization under Basel I and II was regulatory capital relief. Clasessens et

al. (2012) explain two ways through which the capital requirements could be reduced. First, banks could

sell the asset off-balance sheet to the SPV, allowing banks to reduce their regulatory capital since the risk was meant to be transferred to outside investors. Also, banks could just hold the highly rated

securitized assets on their balance sheet instead of the lower rated non-securitized securities. The capital

requirement on the retained part of the issued securities was significantly lower than the 8% required under Basel II. For instance, assets consolidated on the balance sheet from conduits, did not needed to be included in the risk-based capital measure but instead had a 10% conversion factor for the amount covered by a liquidity guarantee, meaning that the risk-weight attached to the senior tranches of the securitized loan (e.g. 20%), was lower than the original loan (e.g. 50%) (Acharya et al., (2013). By reducing the amount of capital that needs to be held against the assets, the bank can in turn improve the

(14)

return-on-equity (ROE) (Baig, Choudhry, 2013). In the proceedings of the paper we will see that this relief of regulatory capital played an important role in the rise of the securitized banking system.

2.1.4 Systemic Risk of Shadow Banking

In this section we examine the systemic risk concerns of securitization. Boot and Thakor (2013) examine some implications of the intertwining of banks and markets, as for instance illustrated by the offloading of risk through securitization. As a consequence, the integration of banks and markets complicates the isolation of bank risk from systemic risk, raising regulatory concerns. For instance securitization was meant to offload risk, while the liquidity guarantees for securitized debt actually increased risk (Boot, Thakor, 2013). The recourse that was created, provides little risk transfer during a run on the shadow banking system, increasing systemic risk of the supporting banks without holding much capital against this risk (Acharya et al., 2013). It is therefore important to regulators to be aware of how securitizations could increase the interconnectedness between banks mutually and with the financial system in general.

The paper of Coval, Jurek and Stafford (2009) demonstrates how securitization, through the means of pooling and tranching, substitutes the diversifiable risk for highly systemic risk. In large diversified underlying pools of assets, the default losses are entirely attributable to the systemic risk exposure. As a result, securitized assets are far less likely to survive an economic bust than non- securitized securities of similar credit rating. Before the crisis, this vulnerability was not recognized by the credit rating agencies setting credit rating that only reflected the expected payoffs of the securities. Accordingly, the credit rating was imprecise in reflecting its dependence on systemic risk.

In their paper, Gennailoli et al. (2013) present a model of shadow banking that explains how the shadow banking system is stable and increases welfare under rational expectations, but contributes to systemic risk and is vulnerable to liquidity crises when investors ignore the downside tail risk of negative macroeconomic shocks that quickly propagate through the financial system as a whole. The additional credit of securitization can be efficient and welfare improving. For example, a bank can issue both safe and risky loans, but when investors have limited wealth available, they prefer the riskless debt. The problem then is that the banks cannot create enough collateral with their safe projects, and the risky projects cannot serve as collateral due to their exposure to idiosyncratic risk. To be able to meet the demand for safe debt, banks pool the loans as to diversify away the idiosyncratic risks and issue the debt that is perceived to be riskless now. This might reduce the probability of distress due to an idiosyncratic shock, but it increases the systemic risk of their portfolio since the banking system is now exposed to common risks. This is because shadow bank activities and traditional banking activities are exposed to common risks like the interest rates and the housing market (Diamond, Rajan, 2002). This securitization by banks therefore increases the exposure to tail co-movement between the securitized banking system and regulated banking system.

(15)

tail risk of the securitization activities. They belief that the collateral they receive in the worst scenario, is higher than it actually is, which lead them to buy more debt that they perceive riskless, this further increases the banks balance sheet and the exposure to systemic risk. When the extreme scenario appears to be worse than anticipated, the bank faces an enormous exposure to this downside risk, since they sold the ‘risk free’ debt to investors. This risk becomes then systemic since the exposure to macroeconomic risk causes the banks to fail together. Hence, the problem is how securitization enables the banks to increase their risk taking that exposes the banks to common risks and not per se in the misperception of risk in the first place (Gennailoli et al., 2013).

2.1.5 Conclusion

Even though shadow banking is said to differ from traditional banking in the form of risk transfer, deposit insurance and government intervention, section 2.1 describes how the shadow banking mimics the traditional banking system, and how the two are intertwined. Furthermore, the securitization process is complex and involves several steps, participants and instruments. Although securitization can be a beneficial complement to the traditional banking system by increasing liquidity and lowering borrowing costs, it also contributes to systemic risk that could cause failure for the financial system as a whole, and even though shadow banking activities are not publicly regulated, it is very unlikely to be solved without public support.

2.2 Development of the Shadow Banking System

The shadow banking system has grown remarkably in the decades preceding the global financial crisis, with securitization gradually becoming an important feature in the financial markets. Prior to the 1970s, banks typically held loans on their balance sheet, but since then, securitization markets in the US and Europe grew steadily, until the beginning of the 2000s, which are characterized by a boom in the global securitization markets (Bertay, 2017). Different securitization industry trends prevailed during the growth of the system, with auto loans and credit cards as mostly used asset classes in the 90s in the US, followed by RMBS that mostly dominated the securitization markets in the US and Europe thereafter up to the financial crisis.

According to the AFME and as can be seen in figure 3, structured finance issuances in the beginning of the 2000s grew above $2500 billion in the US and reached a peak of $800 billion in Europe. The rise in the shadow banking was driven by financial innovation and technology, such as securitization and improved technological systems to originate and distribute, and coincided with the fairly favorable economic environment of low interest rates, economic growth and few defaults

(16)

in the Eurozone in the beginning of the 2000s, where key drivers in the ‘search for yield’ by investors in structured finance products (IMF, 2014). Also, the demand for safe, short-term and liquid assets from large corporations and banks, especially European banks, drove the growth of off-balance sheet banking. Banks used the securitized assets for repo funding and to boost their leverage. At the end of 2006, European and US banks held over one third of the total AAA-tranche issuance (Claessens, Poszar,

Ratnovski, Singh, 2012).

Another major driver of the rise of the shadow banking activities by banks was regulatory

arbitrage. Acharya et al. (2013) find in their paper that the majority of banks provided liquidity

guarantees in order to reduce regulatory capital. More specifically, they find that especially commercial banks with low economic capital ratios used securitizations transactions, while they find a much weaker relationship between the Tier 1 regulatory capital ratio and the issuances of ABCP. Hence, this suggests that the banks were allowed to lower their economic capital, while maintaining a stable regulatory capital ratio when setting up conduits. The minimum capital requirements that were set by Basel I and Basel II have played an important role in the increase of the ‘safe’ securitized assets.

3500,0 3000,0 2500,0 2000,0 1500,0 1000,0 500,0 0,0 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 Year

Figure 2. US and EU historical securitization issuance. Source: own reproduction based on AFME data.

Because of few defaults in the structured finance markets, investors had little reason to worry about the

riskiness of the underlying assets. Likewise, credit rating agencies evaluated these assets-backed securities and collateralized debt obligations as safe. Moreover, risk was perceived to be transferred away from the bank’s balance sheet. This lack of understanding by the credit agencies also influenced the regulators that tied the bank capital requirements to these ratings and investors in the financial markets that outsourced their due diligence to these same rating agencies. The combination of these low capital requirements and the perception in the financial markets that these securities where safe,

EU US € Bil lio ns

(17)

increased the relative use of securitization dramatically (Coval, Jurek, Stafford, 2009).

These perceptions changed rapidly in the summer of 2007 when U.S. housing market defaults were increasing and mortgage-backed securities started to experience losses. This in turn created concerns about the structured finance market in general. The way the securitization process created ‘safe’ assets appeared to be risky and unstable. The investors of the AAA-rated structured finance products suddenly stopped buying when it became apparent the securitization process had ignored the downside aggregate risk of a systemic breakdown (Claessens, Poszar, Ratnovski, Singh, 2012). This led to a liquidity crisis in the securitized asset market and was followed by a securitized-banking run that was driven by the exercising of guarantees by investors, leading to unexpected exposures for the originating banks, withdrawal of the repurchase agreements but also other short-term securities like asset-backed commercial paper, structured investment vehicles (SIVs) and the run on money market funds (Gorton, Metrick, 2012).

The run created a breakdown of the securitization process described above, the reason being that supporting banks started to suffer losses or had to take the securitized assets back on the balance sheet. Moreover because the haircuts on repos rose significantly during this period, the commercial banks had to issue new securities to fund this increasing haircut. However, fire sales decreased asset prices and therefore collateral sharply, so investors get even more worried about the solvency of banks, increasing the haircuts further, reinforcing the cycle and eventually driving the banking system to insolvency. The securitization activities by banks and other financial institutions decreased to levels similar to before the 2000s, and stabilized thereafter (Bertay, 2017).

After the breakdown of the securitized banking system a discussion emerged among policymakers and regulators about the future of securitization. Some policy-makers have argued against the revival of securitization, wile others (ECB, BoE) have spoken in favor of the revival of the securitization markets, focusing on high quality securitizations (Bertay, 2017). Regulatory reforms should help in restoring confidence and enabling securitizations that are less vulnerable to liquidity risk.

2.2.1 Conclusion

Securitization already became a common source of funding in the 70s, leading to a boom in the 2000s. The favorable economic environment, positive attitude of investors and credit rating agencies towards securitization and arbitraging regulation are considered some of the key drivers of this growth. The 2007-2009 crisis was special in the sense that it did not occur in the traditional banking system, but instead occurred in the securitized banking system. When the markets dried up banks had to provide liquidity guarantees to their SPVs, or take the securitized assets back on their balance sheet. This collapse in the shadow banking system played a critical role in the risk posed to the traditional banking system, and the global financial crisis that followed this collapse.

(18)

2.3 Post-Crisis Assessment of the Shadow Banking System

The shadow banking run revealed the procyclicality of off-balance sheet banking, the widespread regulatory arbitrage, and the spillover effects to the real economy. It further revealed that the task of bank regulators to ensure the financial well-being and soundness of the banking system, should actually also encompass the participants in the financial markets outside the regulated banking system (Boot, Thakor, 2013). In response to this, regulators started to assess the real risk transfer of securitization transactions, the complexity of the securitization process and its contribution to systemic risk. Several regulatory developments affecting securitization transactions in both the United States and Europe have been implemented aiming to improve the securitization process and address the vulnerabilities that intensified the global financial crisis. The major legislation affecting securitization transactions in the US is the Dodd-Frank Wall Street Reform and Consumer Protection Act (Dodd-Frank Act). In Europe, several regulatory reforms affect the securitization transactions, in particular the Capital Requirements Directive and Capital Requirements Regulation (together CRD). Further, the Basel III accord is designed for macro-prudential supervision, but also affects securitization transactions (Arca et al., 2015). This section provides the post-crisis assessment of the securitization process and a description of the new regulations affecting securitization transactions in both the United States and Europe in the areas of risk retention, due diligence and disclosure, and the Basel III regulatory framework.

2.3.1 Assessment of the Securitization Process

First of all, the securitization process suffers from moral hazard and adverse selection. The moral hazard problem is created by the originate-to-distribute model in the securitization process. Since the loans are created through multiple balance sheets, a distance is created between the originator of the loan and the eventual investor that bears the default risk. The securitization process therefore reduces the incentives of the securitizer to carefully scrutinize the borrowers and ensuring the quality of the assets underlying the securitization (Keys, Mukherjee, Seru, Vig, 2010). Likewise, adverse selection is created by the information asymmetry and misaligned incentives between the investors of the asset-backed securities and the banks that have private information about their securitized loans. Since investors do not obtain this information, they require a ‘lemon discount’, which lowers the securities’ prices below their book value. Investors require this discount because opaque banks might be incentivized to securitize mostly low quality loans (Buchanan, 2016).

To mitigate this potential adverse selection, the banks could retain the first-loss piece of the securitization or offer recourse. Nevertheless, this recourse created another adverse section problem, since the disclosure of the transactions was too opaque, assessment of the true risk transfer was hard to interpret. Cheng, Dhaliwal and Neamtiu (2011) emphasize that information disclosure of securitization transactions is insufficient and, and that more transparency is needed to assess the risk transfer or recourse provided with securitized assets. It is argued that financial reporting requirements substantially

(19)

simplified the economics underlying the securitizations, leading to inadequate descriptions of the complex structures. Also, the extent to which recourse was provided differed dramatically between the securitizations transactions. This opacity in the risk transfer of securitization transactions does lead to information asymmetry between banks and market participants, consequently causing lack of understanding and uncertainty about the external shocks that trigger recourse, and its impact on the balance sheets of the originating banks (Cheng et al. 2011).

At last, pre-crisis financial regulation, such as the Basel I and II regulatory frameworks, was designed to limit the risk of a financial institution in isolation but did not address their contribution to systemic risk (Acharya et al., 2016). The system itself and asset exposures of the banks remained still vulnerable to large negative macroeconomic shocks Therefore, future regulation should focus more on the systemic risk contribution and correlations of the particular assets of financial institutions and require capital charges according to their contribution to systemic risk and hence be designed to mitigate risk within the financial system as a whole (Acharya, Pedersen, Philipon, Richardson, 2009). Besides that, new capital requirements should deal with the opportunity for regulatory arbitrage.

Clearly, the securitization activities appeared to be vulnerable because of lack of appropriate regulation leading to regulatory arbitrage and additionally the systemic nature of the securitization activities causing spillovers to the real economy (Adrian, 2014).

2.3.2 Regulatory Improvements to the Securitization Process

In light of the above-mentioned flaws of the securitization process, three major regulations were developed, aimed at improving the securitization process affecting both Europe and the US. These regulations include the risk retention rate, the disclosure requirements and the Basel III framework.

2.3.2.1 The Risk Retention Rate

The risk retention rate requires the securitizing banks to keep an interest in their own off-balance sheet securitized assets, which is generally 5% of the credit risk of the assets collateralizing the asset-backed securities. This rules applies to all asset-backed securities collateralized by residential mortgages and to all other asset-backed securities classes. There are some exemptions, for instance if ABS are collateralized by residential mortgages that are ‘qualified’ residential mortgages (SEC, 2014). Also, the SEC states that liquidity guarantees or repurchase agreements are not eligible forms of risk retention since these forms of credit support are in general not funded when closed, and may therefore not be able to absorb the losses when they occur, as experienced in the recent financial crisis. This ‘skin in the game’ should result into better screening and monitoring of the borrowers by the banks, hence mitigating the moral hazard problem that prevailed in the pre-crisis period. Moreover, it should correct the misalignment of incentives between originators and investors (Sarkisyan and Casu, 2013).

(20)

-European Union-

In the EU this rule was formerly adopted in 2011 in Article 122a of the Capital Requirements Directive (CRD II) but on 1 January 2014, the securitization risk retention rule was replaced by Articles 404-410 of the Capital Requirements Regulation (CRR). The new rule reduces the differences in the implemented retention rules across member states of the EU, but the provisions in the CRR are broadly similar to the articles contained in the CRR. The CRR articles state that EU credit institutions must retain a material net economic interest of at least 5% of securitization. This 5% retention must be held as long as the life of the securitized asset and hedging the retention is not permitted (Arca et al. 2015).

-United States-

In the United States a very similar retention rule is enforced. The retention rate is incorporated in section 941 of the Dodd-Franck Wall Street Reform and Consumer Protection Act published in July 2010 and requires originators to retain a material portion of at least 5% of the credit risk of the securitized exposures, and the originator is not allowed to hedge the retention (Sarkisyan and Casu 2013). Although, the credit risk retention rule was proposed already in 2010 it was under consideration for several years before the final rule became effective, which was as of July 2014 (SEC 2014).

2.3.2.2 Due diligence and Disclosure

In the pre-crisis period, although disclosure requirements did exist, they appeared to be insufficient and too opaque (Cheng et al. 2011). Therefore, originating banks could have had the incentive to only securitize the relative low quality and risky loans compared to the loans that were retained on the banks’ balance sheet. This led to adverse selection between the investors and originators. Due diligence came into force to provide access to the prospective investors to all available and materially relevant information and data on the retention, credit quality and performance of the underlying pool of assets used in the securitization transactions (Arca et al., 2015). This reduces the information asymmetry and misaligned incentives between the investors and originators, and potential uncertainty about the provided recourse.

-European Union-

Under Article 406 and 409 of the CRR, credit institutions are obliged to provide due diligence and disclosure to investors so that they have a thorough understanding of the securitized transaction, and the risk and structural characteristics. Furthermore, they are obliged to obtain information from the issuer or originator and an explicit statement from the originator wherein the originator declares to fulfill the retention requirement. For instance, disclosure of the retention involves the (i) identity of the retainer, and whether it retains as an originator, sponsor or original lender, (ii) the form of the retention (e.g. on balance sheet, first loss piece, multiple originators), (iii) any changes to the retaining method, (iv) the level of retention at origination and commitment to retain over the life of the securitization. Besides that, all the prospective investors should have readily and easily access to information concerning the quality

(21)

and performance of the underlying exposures of the securitization transaction, which allow the investors to perform their own stress test. The publication of the due diligence rules for different asset classes started in April 2011, and became effective for the major asset classes like RMBS from January 2013 onwards (Arca et al. 2015).

-United States-

Under the Dodd-Frank Act Section 945 Rule 193, several rules regarding ABS due diligence and disclosure were adopted. The first disclosure rules were already announced in the last quarter of 2010, and became effective as of 20 January 2011 (SEC 2014). All ABS issuers are required to perform a review of the nature of assets underlying and ABS such that disclosure regarding the pool of underlying assets is accurate. Among others, asset-level data includes a unique identity number, collateral of each asset, cash flows generated by each assets, timing and amount of payments, the underwriting of the assets and several more items. Also, third parties may conduct parts of the due diligence, and if the third party contributes to the findings this should be stated in the prospectus. If it is the case that assets in the pool deviate, there must be disclosure about how the assets deviate, which entity included the assets in the pool and if compensating factors were used to determine to include the assets (Arca et al. 2015).

2.3.2.3 Basel III

The severity of the 2007-2009 financial crisis was in large part due to the built up of excessive on and off-balance sheet leverage. Simultaneously, the level and quality of the banks’ capital bases was gradually eroding, and the financial institutions were holding insufficient liquidity buffers. Consequently, the banking system could not absorb the systemic and credit losses, nor cope with the withdrawal of off-balance sheet exposures that had built up in the shadow banking system. This was further amplified by the interconnectedness of the financial institutions through several complex transactions (BCBS 2011). Therefore, the BIS together with the BCBS implemented the Basel III global regulatory framework that aims at a more resilient banking system and reducing spillover risks to the real economy.

One such reform presented in the Basel III framework is the enhancing of risk coverage by raising the quality and quantity of capital. A key lesson from the recent financial crisis is that regulation failed to capture major risks of on- and off-balance sheet risks as well as the risk involved in complex securitization exposures. The reform raised the standard capital requirements, relatively to Basel II,2 but also introduced higher capital requirements for securitizations in both the banking and trading book, and requires the banks to conduct more thorough credit analysis of securitization exposures. Figure 3, shows the differences between the capital requirements in Basel II and Basel III. The Common Equity Tier 1 (CET1) is now 4.5% compared to 2% in Basel II, with additional Tier 1 capital of 1.5% compared to 2%. The maximum of Tier 2 capital is reduced from 4% to 2%. Although the overall requirement is still 8% of risk-weighted assets, the change in composition increased the quality of capital. As well as the

(22)

quality, the Basel III framework increases the quantity with the introduction of several capital buffers. These include the capital conversation buffer, the countercyclical capital buffer and the systemic buffer. The first one is to build up capital buffers in good times that can be withdrawn in times of distress. The second one aims to build up capital in times of economic growth to lean against excessive credit growth, which can be used when credit is constraint. The latter one is to address the systemic risk and interconnectedness of G-SIBs. The G-SIBs are further subject to additional loss absorbency requirements.

In addition to raising the quality and level of capital, several reforms to address counterparty credit risk (CCR) were made. Both the failure to capture significant on-and off-balance sheet risks, as well as derivative exposures intensified the global financial crisis. Reforms regarding the CCR include default risk capital requirements and a charge for losses known as credit value adjustments (CVA). These capital requirements already started as off 1 January 2013 and are gradually phased in (BSBC 2011).

Figure 3. Basel II versus Basel III capital requirements. Source: own reproduction based on Citi Research

The risk-based capital requirements are supplemented with a leverage ratio. Before the crisis, banks built up excessive leverage while maintaining strong risk-weighted capital ratios. Therefore, Basel III introduces this ratio, which is simpler and more transparent than the risk-based measures and also more countercyclical. It is stricter in booms, while looser in busts compared to the risk-weighted requirements. The leverage ratio is said to serve as backstop to the risk-based capital requirements and mitigate excess leverage in the banking system and the risk of rapid deleveraging. The leverage ratio is a non-risk-based leverage ratio including both on- and off- balance sheet exposures such as securitized transactions, and is measured as Tier 1 capital over average consolidated assets and should be at least

(23)

3%. This leverage ratio is expected to become effective by 1 January 2018 but disclosure started from 1 January 2015 (BSBC 2011). A full overview of the phase-in arrangements for the different capital requirements is provided in the appendix A.2.

There are exemptions possible to the stricter capital requirements for securitization transactions. In particular, article 243(2) and 244(2) of the CRR allows originating banks to address the regulatory capital requirements applicable under Basel III, when they apply to structure their securitization transactions as significant risk transfer (SRT). More specifically, SRT applies when the transactions is structured in such a way that it transfers credit risk from the balance sheet of the originating bank to third parties that will bear potential losses. Recognition of a SRT requires several conditions that need to be met on a continuous basis over the life of the securitization transaction (ECB, 2016). This SRT structure might be beneficial for capital constrained banks that do not want to raise capital in the markets, or banks that want use their capital in a most efficient way.

2.3.3 Literature Review of Post-crisis Regulation

Albeit there is little evidence in literature about the effectiveness of the newly implemented regulations, some papers have assessed the above-mentioned regulations. For instance, it is argued that the retention of interest in the securitized assets can serve as a signaling device. This means that banks that securitize high quality assets can use the retention rate as a signaling mechanism to signal the quality of the underlying assets. Empirical evidence points out that this signaling mechanism can partially solve the asymmetric information problem. Banks can mitigate the adverse selection problem by retaining the junior tranche of the securitized portfolio or to make a commitment to keep monitoring the borrowers. However, it is further found that first loss provisions increase the bank’s risk of default, and that the retention rate increases banks insolvency risk for ‘large-scale’ issuers, while the opposite holds for ‘small-scale’ issuers. This indicates that regulators should also take into consideration the strategy and size of particular securitizations (Sarkisyan and Casu, 2013).

The paper of Gu and Wu (2014) also studies the impact of the risk retention rate and disclosure requirement. The paper argues that both regulations are effective in reducing the informational loss of the investors with regard to the originators, but that a flat 5% retention requirement is not optimal for all underlying asset classes. The authors further argue that there are also regulatory costs: the retention rate can worsen adverse selection since investors can no longer differentiate for low quality assets whether the 5% retention is the optimal choice for the bank, or because it is the minimum threshold. Hence, information asymmetry on low quality assets can even be aggravated, and can therefore undermine the effective informational revelation of securitization intensity. Also, due to signaling costs, the information disclosure requirement can lead to distortion of the banks’ securitization intensity away from what would be the efficient level of securitization activity. According to Gu and Wu (2014), the major distortion is the asymmetric information of securitization and not the securitization intensity itself that should be reduced. Nevertheless, information disclosure is a useful complement to risk retention

(24)

when investors are risk averse. Finally, they argue that a variable retention rate for different markets and separate asset classes that vary in riskiness would be more effective instead of the 5% rate. In addition, Hatorri and Ohashi (2011) point out that the risk retention might actually increase moral hazard, due to the fact that originating banks have to retain some of the securitized loans, they tend to engage into less securitization, and might therefore be less incentivized to carefully screen the borrowers.

Empirical evidence that tests the effectiveness of the leverage ratio introduced under Basel III estimates that the leverage ratio should lead to more stable banks, but since there are no risk-weights attached, there could be an increased incentive by banks to take risk. This in turn, can be absorbed by higher capital requirements and the leverage ratio prevents banks from operating under excessive leverage both on and off-balance sheet (ECB 2015). Other papers estimate that the Basel III accord could lead to a reduction in systemic risk, but that the capital charges should be substantially larger than the ones specified in the current Basel III framework (Poledna, Bochmann, Thurner 2017).

2.3.4 Conclusion

After the shadow banking system breakdown, regulators started to assess the flaws of the system in which several weaknesses came to light. The securitization process suffered from moral hazard and adverse selection, and the complexity and opaqueness of the securitization transactions caused lack of understanding of the true credit risk transfer in the securitization process. Regulatory reforms such as the risk retention rate, the due diligence and disclosure requirements and the Basel III accord, aim at mitigating these problems. Existent literature points out that albeit the retention and disclosure requirements can reduce informational asymmetry and moral hazard, there are also regulatory costs that need to be taken into consideration. In the following section, the impact of these post-crisis regulations on the securitization activity of banks and their contribution to systemic risk will be empirically assessed for our sample of banks.

(25)

3. Empirical Analysis

This section will develop the three hypotheses and the rationale behind them. In addition, this section explains the methodology that is used to empirically examine our hypotheses. It further elaborates on the data selection and presents several descriptive statistics of the independent and dependent variables used. Thereafter, this section discusses the estimation results and performs several robustness checks.

3.1 Hypothesis development

In this thesis, three hypotheses will be tested. The first hypothesis follows from the above-discussed findings by Acharya et al. (2013), who conclude that capital-constrained banks in the period 2001-2009, set up more conduits than financial institutions not subject to capital requirements, and moreover that these conduits were set up with guarantees that circumvented the Basel capital charges. This relationship was stronger for the economic capital ratio of banks, measured as book equity over assets, than for the regulatory capital requirements. This evidence however, dates from the period before the new Basel III framework that mitigates the incentive to set up securitization transactions to lower regulatory capital. Therefore, in this paper we focus again on regulatory arbitrage, but now with the new regulatory framework in place, which might gives us some insight whether new regulation changed mitigates regulatory arbitrage.

Despite this new regulation, we still expect this negative relationship between economic capital and securitization activity of banks to hold. First of all, while maintaining their regulatory capital ratios, banks might be reluctant to hold higher amounts of capital than necessary. Funding or raising equity is typically more expensive for a bank than debt financing

,

3 and lowering economic capital or increasing effective leverage could increase their return on equity. Additionally, capital constrained banks can apply for SRT, therewith freeing up capital that can be used for other means, such as extending new credit to the economy (BoE and ECB, 2014). At last, securitization grew because of stricter capital requirements in the first place. Therefore, our first hypothesis is:

Hypothesis 1: Capital-constrained banks are more likely to engage in shadow banking activities

The second hypothesis is regarding the impact of the before discussed post-crisis regulation on the securitization activity of originating banks. We witnessed a decrease in securitization activity after the 2007-2009 financial crisis, and the second hypothesis aims to examine if the post-crisis regulation played a significant role in this reduction. Although the regulation does not aim to reduce securitization activity by banks, on the contrary, regulation needs to take away uncertainty about the future of securitization and increase confidence in the securitization markets, there might still be impediments to

The impact of post-crisis financial regulation on the shadow banking system : a panel data analysis

UNIVERSITY OF AMSTERDAM

The Impact of Post-Crisis Financial Regulation on the Shadow

Banking System: A Panel Data Analysis

Master Thesis

July 2017

MSc. Finance: Corporate Finance and Banking

Eveline van ’t Spijker 10400982

The Impact of Post-Crisis Financial Regulation on the

Shadow Banking System

Abstract

Contents

List of figures and tables

1. Introduction

2. Literature Review

2.1 Shadow Banking

2.2 Development of the Shadow Banking System

2.3 Post-Crisis Assessment of the Shadow Banking System

3. Empirical Analysis

3.1 Hypothesis development

,