Charles University in Prague Faculty of Social Sciences. The Rise of Shadow Banking

Transcription

1 Charles University in Prague Faculty of Social Sciences Institute of Economic Studies MASTER THESIS The Rise of Shadow Banking Author: Bc. Michaela Dovicová Supervisor: PhDr. Petr Teplý, Ph.D. Academic Year: 2013/2014

2 Declaration of Authorship The author hereby declares that he compiled this thesis independently, using only the listed resources and literature, and the thesis has not been used to obtain a different or the same degree. The author grants to Charles University permission to reproduce and to distribute copies of this thesis document in whole or in part. Prague, January 5, 2014 Signature

3 Acknowledgments I would like to express my gratitute to my supervisor, PhDr. Petr Teplý, PhD., for his excellent guidance, encouragement and valuable comments while I was writing my thesis. I would like to thank PhDr. František Čech for his suggestions. Also, I am thankful to my family for their support. Special thanks belongs to Mgr. Miloš Hrachovec for standing by my side and reviewing the thesis.

4 Abstract Recent financial crisis resulting in global financial instability pointed at the importance of growing shadow banking. Shadow banking activities are generally defined as banking-like activities outside of regulated banking. In this thesis, we study theoretical background of shadow banking, its regulation and supervision. Despite the data availability problem, a qualitative analysis is performed to estimate the volume of the European and the U.S. shadow banking sectors from 2006 until 2013 Q2. European shadow banking system hit its bottom of EUR 8.3 trillion (19% of total European bank assets) in 2008 Q4. Nowadays, it equals to EUR 9.3 trillion (21% of total European bank assets). U.S. shadow banking sector attained its maximum of USD 20.7 trillion (163% of total U.S. bank assets) in 2008 Q1. Nowadays, it equals to USD 15.6 trillion which also equals to total U.S. bank assets. Moreover, we concentrate on Chinese money market funds and French and UK repo markets, since these represent an important part of shadow banking. Quantitative analysis studies relationships among traditional banking, shadow banking and economy itself in France and UK. Results show that if repo transactions, GDP and government debt increase, total bank assets increase. Furthermore, if money market fund assets decrease and GDP increases, repo transactions increase. The impact of total bank assets on repo transactions differs in studied countries. JEL Classification Keywords G01, G21, G23, G28 shadow banking, shadow banking regulation, repurchase agreement, money market fund Author s Supervisor s michaela.dovicova@seznam.cz teply@fsv.cuni.cz

5 Abstrakt Nedávna finančná kríza, ktorá mala za následok globálnu finančnú nestabilitu, poukázala na dôležitosť rastúceho tieňového bankového systému. Aktivity tieňového bankového systému sú všeobecne definované ako aktivity podobné bankovým aktivitám odohrávajúce sa mimo regulovaný bankový systém. V tejto práci sa zaoberáme teoretickou stránkou tieňového bankovníctva, jeho reguláciou a dohľadom nad ním. Napriek problému s nedostupnosťou dát, je vykonaná kvalitatívna analýza s cieľom odhadnutia objemu európskeho tieňového bankového sektoru a tieňového bankového sektoru v Spojených štátoch amerických pokrývajúca obdobie od roku 2006 až do 2013 Q2. Európsky tieňový bankový systém dosiahol svoje minimum 8,3 bilióna EUR (19 % celkových európskych bankových aktív) v 2008 Q4. V súčasnosti je jeho hodnota 9,3 bilióna EUR (21 % celkových európskych bankových aktív). Tieňový bankový sektor v Spojených štátoch amerických dosiahol svoje maximum 20,7 bilióna USD (163 % celkových amerických bankových aktív) v 2008 Q1. V súčasnosti je jeho hodnota 15,6 bilióna USD, čo tiež predstavuje celkové americké bankové aktíva. Ďalej sa zameriavame na čínske peňažné trhové fondy a francúzske a britské trhy dohôd o spätnom odkúpení, keďže predstavujú významnú časť celkového tieňového bankovníctva. Kvantitatívna analýza študuje vzťahy medzi tradičným bankovníctvom, tieňovým bankovníctvom a ekonomikou ako takou vo Francúzsku a vo Veľkej Británii. Výsledky ukazujú, že ak rastú HDP, vládny dlh a objem dohôd o spätnom odkúpení, tak rastú aj celkové bankové aktíva. Okrem toho, keď klesajú aktíva peňažných trhových fondov a rastie HDP, tak rastie aj objem dohôd o spätnom odkúpení. Vplyv celkových bankových aktív na objem dohôd o spätnom odkúpení sa v študovaných krajinách líši. Klasifikace JEL G01, G21, G23, G28 Klíčová slova tieňový bankový systém, regulácia tieňového bankového systému, dohoda o spätnom odkúpení, peňažný trhový fond autora vedoucího práce michaela.dovicova@seznam.cz teply@fsv.cuni.cz

6 Contents List of Tables List of Figures Acronyms Thesis Proposal viii ix xi xii 1 Introduction 1 2 Theoretical Background of Shadow Banking Recent Financial Crisis Definition of Shadow Banking Shadow Banking Activities Shadow Banking Entities Subsequent Risks Regulation and Supervision FSB s Regulation FSB s Workstreams FSB s Actions FSB s Following Steps Basel III U.S. Regulation Obstacles to the Regulation Shadow Banking System Assessment of the Shadow Banking Sector European Shadow Banking Sector Estimation of the European Shadow Banking Sector... 38

7 Contents vii The U.S. Shadow Banking Sector Estimation of the U.S. Shadow Banking Sector Comparison of the European and the U.S. Shadow Banking Sectors Money Market Funds Chinese Money Market Funds Repo Market French Repo Market UK Repo Market Empirical Study of Shadow Banking The Model Variables Estimation Expectations and the Hypotheses Results Further Research Opportunities Conclusion 71 Bibliography 78 A Appendix I

8 List of Tables 2.1 Characteristic features of traditional and shadow banking Shadow banking activitites Securitisation: main features FSB s workstreams Impact of the crisis, Europe Impact of the crisis, the U.S Descriptive statistics Sign expectations, model Sign expectations, model Model 1, OLS estimation for France, full results Model 1, OLS estimation for the UK Coefficient summary for model Model 1, Heteroscedasticity-corrected estimation for France - subset Model 2, Heteroscedasticity-corrected estimation for France Model 2, Heteroscedasticity-corrected estimation for the UK Coefficient summary for model A.1 Model 1, OLS estimation for UK, full results II A.2 Model 1, Heteroscedasticity-corrected estimation for France - subset, full results III A.3 Model 2, Heteroscedasticity-corrected estimation for France, full results IV A.4 Model 2, Heteroscedasticity-corrected estimation for the UK, full results V

9 List of Figures 2.1 On-balance sheet intermediation Off-balance sheet intermediation Simplified credit intermediation chain in shadow banking system Securitisation issuance in the euro area and the U.S., in EUR billion Annual averages of daily financing by U.S. government securities primary dealers Split of total European MMFs by country of domicile MMFs total balance sheets, in EUR millions Risk analysis framework of interconnectedness between traditional and shadow banking entities General FSB principles of regulatory measures Size of the European shadow banking sector European CP market, outstanding amounts European repo market, outstanding amounts European MMF market, total assets European ABS market, outstanding balances European hedge fund market, total assets European shadow banking sector vs. total bank assets European shadow banking as a share of total bank assets Shadow banking liabilities vs. traditional banking liabilities, in USD trillion Size of the U.S. shadow banking sector U.S. CP market, outstanding amounts U.S. repo market, outstanding amounts U.S. MMF market, total assets U.S. ABS market, outstanding balances

10 List of Figures x 4.15 U.S. hedge fund market, total assets U.S. shadow banking sector vs. total bank assets U.S. shadow banking as a share of total bank assets Comparison of the European and the U.S. shadow banking sectors Chinese MMF, total assets French repo market UK repo market Total bank assets, France Total bank assets, UK Number of MFI, France Number of MFI, UK A.1 Banks assets and liabilities to non-bank financial intermediaries I

11 Acronyms ABCP Asset-Backed Commercial Paper ABS Asset-Backed Security AFME Association for Financial Markets in Europe AFT Agence France Trésor BCBS Basel Committee on Banking Supervision CP Commercial Paper EC European Commision ECB European Central Bank FED Federal Reserve Bank FSB Financial Stability Board GDP Gross Domestic Product GSE Government-Sponsored Enterprise ICMA International Capital Market Association IMF International Monetary Fund IOSCO International Organisation of Securities Commissions MBS Mortgage-Backed Security MFI Monetary Financial Institutions MMF Money Market Fund NAV Net Asset Value OLS Ordinary Least Squares SIV Structured Investment Vehicle SPV Special Purpose Vehicle UK United Kingdom U.S. United States of America

12 Master Thesis Proposal Author Supervisor Proposed topic Bc. Michaela Dovicová PhDr. Petr Teplý, Ph.D. The Rise of Shadow Banking Topic characteristics Recent financial crisis in 2008 pointed out various problems regarding regulation and supervision of the banking system. The regulation was not adequate and supervision was rather ineffective. There has been an increase in the credit activities of non-bank institutions by which the Financial Stability Board was attracted. The response to this potential thread did not wait long, since the instability of shadow banking became apparent in recent financial crisis. Financial Stability Board defines shadow banking as the system of credit intermediation that involves entities and activities outside the regular banking system. Since the regulation and supervision was primarily focused on the banking sector and shadow banking existed with minimal regulatory risk constraints, nowadays shadow banking is becoming a subject of strengthen oversight and regulation. Size of the shadow banking system recorded a significant growth from 2002 to 2010 when it grew from USD 27 trillion to USD 60 trillion. Market in the United States of America (U.S.) is very important one with the proportion between 35% and 40%. It is necessary to add that the importance of the European market has significantly increased as well, namely British, Dutch and German markets having the highest shares. Financial Stability Board has commenced not only with the definition of the shadow banking principles for monitoring and regulation, but also initiated the mapping process to identify and assess systemic risks involved. Moreover, it identifies the scope of a possible regulatory measure. Having defined shadow banking as outside the regular banking area, it is

13 Master Thesis Proposal xiii necessary to name the entities involved and study their activities. As soon as any entity engages itself into one of the following activities, it is counted as a part of shadow banking. Let us mention some of these activities: ˆ acceptance of funding with deposit-like characteristics done by Money Market Funds or other investment funds, ˆ performance of the maturity and/or liquidity transformation by finance companies or securities entities like Asset-Backed Commercial Paper (ABCP) conduits, Structured Investment Vehicles (SIVs) or Special Purpose Vehicles (SPVs), ˆ undergoing of the credit risk transfer, ˆ use of the direct or indirect financial leverage in case of investment funds or hedge funds, ˆ securitisation and securities lending, ˆ repurchase transactions (repo). This thesis will study these activities with the emphasis on Money Market Funds and repo market. Since shadow banking is a hot topic in these days, we will do the research in this area with the focus on U.S. and European markets. Moreover, China is becoming more and more important player in the world market. Therefore we will study Chinese Money Market Funds as well to be able to capture the situation there. Hypotheses 1. European shadow market increased in volume during the recent financial crisis while the situation in the U.S. went other direction, i.e. during the crisis, the amount of shadow banking activities declines in volume. 2. Strong relationship between U.S. Money Market Funds and European Banks due to the differences in legal framework among countries. We assume that financial crisis was a breaking point after which the markets started to behave in different directions. 3. Significant rise in the volume of shadow banking activities in the Chinese market, mainly in terms of Money Market Funds. 4. When talking about French shadow market, we think that boom in volume of repo transactions was significant due to the onset of the financial crisis in the U.S.

14 Master Thesis Proposal xiv Methodology The thesis will study several parts of shadow banking, mainly Money Market Funds and repo market. Literature proves that there has been a significant increase in the amount of repo transactions in Europe during the recent financial crisis. Recent literature sources point out that availability of data related to shadow banking is lacking. Hence, the authors try to develop their own datasets based on partial data provided by firms and banks. We will try to develop a model that will allow us to study repo market in Europe what will be very challenging. There exists a significant link between U.S. and European banks in terms of hedge funds. We will try to find the relationship between U.S. Money Market Funds and European banks as well. Our dataset will be prepared using the data published by European Central Bank, FITCH, BIS and International Capital Market Association. Despite the gap in the data, we will try to prepare a dataset using mainly European Central Bank (ECB) statistics, especially Euro Area Accounts MFI and OFI. Furthermore, we will try to focus on French shadow banking, namely repo transactions. We will work with the data published by the Banque de France covering the crisis period. We will analyse available data to prove our hypotheses. We will devote a part of this thesis into the study of the ECB repo rate. We will develop a model what will allow us to explain the repo rate itself. Outline 1. Introduction 2. Theoretical background of shadow banking 3. The repo market during the recent financial crisis 4. U.S. Money Market Funds and European banks 5. Conclusion Core bibliography 1. Adrian, T. & A. B. Ashcraft (2012): Shadow Banking Regulation. Staff Reports No.559, Federal Reserve Bank of New York. 2. Bouveret, A. (2011): An Assessment of the Shadow Banking Sector in Europe. Available at:

15 Master Thesis Proposal xv 3. ECB (2012): Shadow Banking in the Euro Area. An Overview. Occasional Paper Series No.133, European Central Bank. ISSN EC (2012): Shadow Banking. Green Paper, European Commission. 5. FSB (2011): Shadow Banking: Strengthening Oversight and Regulation. Recommendations of the Financial Stability Board. 6. FitchRatings reports 7. Ghosh, S., I. Gonzalez del Mazo & I. Ötker-Robe (2012): Chasing the Shadows: How Significant Is Shadow Banking in Emerging Markets? Number 88, Poverty Reduction and Economic Management Network (PREM). The World Bank. 8. Pozsar, Z. (2011): Institutional Cash Pools and the Triffin Dilemma of the U.S. Banking System. IMF Working Paper 190, International Monetary Fund. 9. Pozsar, Z. & M. Singh (2011): The Nonbank-Bank Nexus and the Shadow Banking System. IMF Working Paper 289, International Monetary Fund. 10. Ricks, M. (2010): Shadow Banking and Financial Regulation. Working Paper 370, Columbia Law and Economics. 11. Singh, M. & J. Aitken (2010): The (sizeable) Role of Rehypothecation in the Shadow Banking System. IMF Working Paper 172, International Monetary Fund. Author Supervisor

16 Chapter 1 Introduction Recent financial crisis pointed to the increasing trend in activities performed outside the regular banking sector, in shadow banking. These types of activities are very attractive thanks to their common feature, i.e. none or weak regulation. However, these activities represent also a threat of systemic risk for the whole financial system. It was financial crisis that made it very clear that shadow banking contributes to financial instability. Therefore, the interest in shadow banking increased immediately. Financial institutions, academics and regulators started to pursue their research. Increasing literature regarding the studies of shadow banking proves growing trend in the research of this area of economy. At first, it was quite difficult to define specifically what shadow banking is. After several attempts by researchers, Financial Stability Board unified and presented its definition. According to the FSB (2011a), this term is officially defined as the system of credit intermediation that involves entities and activities outside the regular banking system. Once the problem is defined, studies may be performed. The call for regulation of shadow banking activities came along with financial stability deterioration. After the onset of recent financial crisis, there have been many improvements in regulation and supervision of the shadow banking system. However, one of the ongoing unsolved issues of shadow banking is the uniform methodology capturing its size. The objective of this thesis is to deal with a concept of shadow banking system. This recent phenomenon caught the interest of many institutions, standard-setting bodies, researchers, academics and ours as well. One of the most important issues regarding shadow banking is to capture its actual size. That is the reason why we perform a challenging assessment of the shadow

17 1. Introduction 2 banking sector in Europe and in the United States of America. We put emphasis on development on these markets mainly during and after the recent financial crisis. We believe that on one hand, shadow banking in the United States of America experienced a decrease during the financial crisis and on the other hand, European shadow banking was growing despite the presence of the crisis. A rising importance of China in global financial market and an attractiveness of money market funds confirmed our intention to study the development in Chinese market of money market funds. Since the repo market plays the most important role in the whole European shadow banking, we perform an analysis of this sector in France and United Kingdom. Furthermore, we study shadow banking more profoundly. Therefore, we develop a new framework that studies the interconnectedness of traditional banking, shadow banking and economy as a whole. The thesis is structured in a following manner. Chapter 2 provides a definition of shadow banking system, its activities and involved entities pointing out the risks that shadow banking system brings to the financial market. Chapter 3 deals with the regulation and supervision of shadow banking with emphasis on the regulation framework introduced by Financial Stability Board. Chapter 4 studies the shadow banking system empirically. Foremost, it measures its size in Europe and in the United States of America. Subsequently, it presents an empirical study of relationships among traditional banking, shadow banking and economy itself. Finally, it outlines further research opportunities. Chapter 5 summarises our findings.

18 Chapter 2 Theoretical Background of Shadow Banking For past several years, an importance of the shadow banking system has increased significantly. Recent financial crisis pointed at the link of shadow banking system with global financial stability. This chapter is devoted to the theory behind shadow banking, mostly how shadow banking is defined. First, we focus on the recent financial crisis with the emphasis on the shadow banking system. Then, the definition of shadow banking is presented. Subsequently, the aim is to explain the activities that shadow banking entities perform. We close this chapter with the study of risks that result from the shadow banking system. 2.1 Recent Financial Crisis It was the recent financial crisis that signalised the instability of global shadow banking system. FSB (2011a) proves that shadow banking activities have an important impact on the global financial stability. Huge unregulated amount of financial flows led to the financial crisis in the U.S. which spread widely to other economies in the world. During this crisis, significant amount of bank-like activities took place outside the regular banking system. These financial flows came rapidly into the attention of all financial bodies because of the common special trait of these funds, i.e. very low degree of regulation and supervision or even no regulation at all. Ricks (2010) proposes a possibility of safety net extensions in order to cover shadow banking activities. Consequences of such safety net extensions would

19 2. Theoretical Background of Shadow Banking 4 mean huge government interventions into the financial system. Furthermore, the author argues that instability of shadow banking should be decreased only by regulatory risk constraints. However, this solution of shadow banking problems would be rather incomplete and very costly. There are many incentives in favour of shadow banking activities. One of them is probably an avoidance of regulation and supervision that is imposed on the regular banking system. However, the consequences are carried by the whole financial system. In 2012, European Commision (EC) prepared a detailed study of shadow banking in order to summarise current development and present reflections on this subject. In order to highlight the consequences of shadow banking, EC (2012) illustrates an example: operations circumventing capital, accounting rules and transferring risks outside the scope of banking supervision played an important role in the build-up to the recent financial crisis. Gorton & Metrick (2012) see the run on repurchase (repo) transactions as the main problem in the recent financial crisis. Depositors were afraid of losing their money by selling collateral in the market due to bank failures and investors replied by increased haircuts. As we show in Subsection 4.1.2, the repo market is one of the most important parts of shadow banking. ECB (2012) further states that runs on other mostly shadow banking entities, such as Money Market Funds (MMFs), enhanced the depth of the financial crisis. Adrian & Ashcraft (2012b) suppose that structured securities were the main reason of the recent financial bubble and collapse. Securitisation is a process involving structured securities, as we show in Section Also, it is one of the main and core activities of the shadow banking system. 2.2 Definition of Shadow Banking The term shadow banking system was firstly introduced by Paul McCulley in September He claims it to be the whole alphabet soup of levered up non-banking investment conduits, vehicles and structures. (McCulley (2007)) From that time on, this term has been used by many authors and many institutions but no exact definition was introduced. However, many partial definitions of the shadow banking system elaborated by different authors, investors or institutions exist. The official definition was set by Financial Stability Board (FSB) in According to the FSB (2011a), shadow banking is officially defined as the

20 2. Theoretical Background of Shadow Banking 5 system of credit intermediation that involves entities and activities outside the regular banking system. These kinds of operations inject extra funding and liquidity into the financial market. Adrian & Ashcraft (2012b) define shadow banking as a web of specialised financial institutions that channel funding from savers to investors through a range of securitisation and secured funding techniques. Also, the same authors define shadow banking activities as banking intermediation without public liquidity and credit guarantees. Other economists prefer to define shadow banking in rather informal way. Gorton & Metrick (2010) say that this system performs the same functions as traditional banking, but the names of the players are different and the regulatory structure is light or nonexistent. Figure 2.1: On-balance sheet intermediation Source: Author based on Gorton & Metrick (2010) Note: $ (USD) here represents the flow of money Since the official definition is fairly broad, FSB (2011a) explains precisely what the shadow banking system is. The entities and activities involved in credit intermediation are either extending the credit or facilitating its intermediation. The extension of credit can be direct or a part of a chain of credit intermediation. However, credit intermediation does not include only on-balance sheet transactions, but also derivatives and other off-balance sheet transactions. Altogether, they form credit intermediation chain. All those transactions and activities are done outside the regular banking system. Therefore, the supervision and regulation are often not present at all or the degree is much smaller. This indicates a serious problem. That is why Chapter 3 studies regulation and supervision issues.

21 2. Theoretical Background of Shadow Banking 6 Shadow banking activities are very often used because they provide extra source of financing or liquidity. This is a highly positive characteristic for the participants of shadow banking. Moreover, FSB (2011b) adds that it is a source of efficient credit in the economy. Credit intermediation inside the regular banking system is described in Figure 2.1. It is a transfer of funds between suppliers and borrowers. Usually, there is one entity, a bank, involved. Figure 2.2: Off-balance sheet intermediation Source: Author based on Gorton & Metrick (2010) Note: MMF - money market fund, SPV - special purpose vehicle, $ (USD) here represents the flow of money In comparison, we add the scheme of credit intermediation outside the regular banking system. It is captured in Figure 2.2. Generally, in the shadow banking system, there are more entities involved in such operations than in the regular banking system. These entities are called shadow banking entities and we will study them in Subsection An excellent overview regarding credit intermediation, its reasons of existence, its functioning and also necessary regulation is elaborated by Adrian & Ashcraft (2012a). As we have already shown, a bank or a bank-owned entity can be easily involved in the shadow banking system, e.g. by accepting a fund from a shadow banking entity. Activities performed by a regular bank are regulated and overseen as there are many regulations imposed on the traditional banking system.

22 2. Theoretical Background of Shadow Banking 7 However, there is a significant gap in the regulation of the shadow banking system. Figure 2.3: Simplified credit intermediation chain in shadow banking system LOAN ORIGINATION commercial banks, mortgage/consumer finance companies LOAN WAREHOUSING ABCP conduits, SPVs SECURITISATION SIVs, SPVs DISTRIBUTION AND WHOLESALE FUNDING MMFs, hedge funds, commercial banks Source: Author based on FSB (2011a) Note: ABCP - asset-backed commercial paper, SPV - special purpose vehicle, SIV - structured investment vehicle, MMF - money market fund Table 2.1: Characteristic features of traditional and shadow banking Feature Traditional banking Shadow banking structure simple complex funding deposits wholesale instruments exposure full term short-term relationship balance sheet securities main risk credit risk market risk type of return spread fee typical intermediary low-roe bank high-roe non-bank Source: Author based on ICMA (2012) Note: ROE - return on equity, ROE=Net Income/Shareholder s Equity FSB (2011a) defines credit intermediation as a chain of operations between entities. Common activities of credit intermediation chain are captured in

23 2. Theoretical Background of Shadow Banking 8 Figure 2.3. Credit intermediation is performed using financial transformation, i.e. a way by which the credit is transmitted within the shadow banking system. These operations may involve higher level of risk. We may also talk about the risk redistribution. ICMA (2012) further studies three types of financial transformation, i.e. credit, maturity and liquidity transformation. After having defined shadow banking, let us present a comprehensive comparison of characteristic features of both traditional and shadow banking systems in order to understand important differences between them. This comparison in captured in Table Shadow Banking Activities Since the definition of the shadow banking system is rather general, we find it crucial to define shadow banking activities properly. As it was shown in Section 2.2, there are many definitions of the shadow banking system. Therefore, there are many activities that are counted as the shadow banking activities, i.e. maturity and liquidity transformation, leverage, credit transformation, credit risk transfers, securitisation, securities lending, repurchase transactions, deposit-like funding, etc. The brief definition of mentioned shadow banking activities can be found in Table 2.2. Once the shadow banking activities are defined, it should be possible to trace them. However, there are many shadow banking entities that do not provide information about their activities. Therefore, the size of shadow banking can hardly be estimated. There is an attempt of a proxy measurement of non-bank credit intermediation in FSB (2011b). It was done in Australia, Canada, Japan, Korea, United Kingdom (UK), U.S. and euro area. Based on this information, in 2002 the shadow banking system was estimated to be USD 27 trillion. In 2007, it boomed to USD 60 trillion and stayed at the same level in In Chapter 4, we focus on the size of the shadow banking system and we perform an estimation of shadow banking on our own. Securitisation As already mentioned in Table 2.2, securitisation is one of the most common activities of shadow banking. Literature proves that many authors have an interest in studying this activity. Gorton & Metrick (2010) name this activity as one of three that should be taken look at when regulating the shadow bank-

24 2. Theoretical Background of Shadow Banking 9 Table 2.2: Shadow banking activitites Activity Maturity transformation Liquidity transformation Leverage Credit transformation Credit risk transfers Securitisation Securities lending Repurchase transactions Deposit-like funding Definition use of short-term assets to finance long-term liabilities creating liquidity for the saver and exposing intermediary to rollover and duration risks use of liquid instruments to finance illiquid assets use of debt financing in order to multiply gains enhancement of the credit quality of debt issued by an intermediary through the use of priority of claims (seniority of deposits or equity) way of credit risk separation from other types of risks by passing it to other investors pooling of loans into a portfolio, selling such portfolio to SPVs that finances them by selling securities in the capital markets lending of securities by institutional investors to banks and broker-dealers against collateral of cash or securities short-term transactions between two parties in which one party borrows cash from the other by pledging a financial security as collateral use of funds with characteristics of deposits but outside the regular banking system Source: Author based on Acharya & Öncü (2010), Adrian & Ashcraft (2012b), FSB (2011a), EC (2012), FSB (2012), Gorton & Metrick (2010) and Radević & Lekpek (2010) Note: SPV - special purpose vehicle ing system. ECB (2012) also focuses on the securitisation as one of the key components of shadow banking. For better understanding of the term securitisation, we present the definition by the ECB (2012). First step necessary for the securitisation is pooling and structur-

25 2. Theoretical Background of Shadow Banking 10 ing of loans into term Asset-Backed Security (ABS). Usually, these loans can be mortgages, auto loans, student loans, etc. This pooling is done by broker-dealers ABS syndicate desks. ABS warehousing often uses repo agreements for funding. Similarly, pooling of ABS into certificate of deposits is done by broker-dealers ABS syndicate desks. Then, limited purpose finance companies, such as SIVs, conduits and credit hedge funds, are in charge of ABS intermediation itself. They raise the funding not only in repo transactions, ABCPs, Medium Term Notes and bonds, but also in wholesale funding that is funded by regulated and unregulated money market intermediaries, e.g. MMFs. Securitisation is defined as mainly short-term source of funding. ECB (2012) further explains that this characteristic makes such activity more likely to be an object of runs and drying-up of liquidity. Table 2.3: Securitisation: main features Activity Funding Entity ABS origination/ ABCP conduits ABS warehousing ABS SPVs repo broaker-dealers ABS issuance/ CP SPVs CDO issuance CDO broaker-dealers ABS intermediation ABCP SIVs MTN conduits capital notes hedge funds repo wholesale funding repo securities lenders ABCP cash funds ABCP MMF Source: Author based on ECB (2012) Note: ABS - asset-backed security, ABCP - asset-backed commercial paper, SPV - special purpose vehicle, CP - commercial paper, CDO - collateralised debt obligation, SIV - special investment vehicle, MTN - medium term note, MMF - money market fund. Underlined features are especially important for banks in European Union. Securitisation is an important part of credit intermediation that facilitates

26 2. Theoretical Background of Shadow Banking 11 maturity and liquidity transformation. In Table 2.3, we capture the U.S. securitisation activities, funding and involved entities. The U.S. financial market is undoubtedly more developed and bigger in size than euro area market. The ECB (2012) proves that securitisation issuance in the euro area was smaller than in the U.S. before the crisis. In the euro area it was 5% of the Gross Domestic Product (GDP) and in the U.S. it was 12% of the GDP. Figure 2.4: Securitisation issuance in the euro area and the U.S., in EUR billion Source: ECB (2012) According to the ECB (2012), in the U.S. market, all securitisation activities named in Table 2.3 take place while in the euro area market, the most important activities involve the use of residential mortgage-backed securities. Furthermore, ECB presents the findings of the Banking Supervision Committee about the securitisation issuance. Despite the crisis, the issuance in the euro area interestingly continued albeit in lower amounts. The situation in the U.S. was very different. In order to show the securitisation issuance in the euro area and in the U.S. graphically, we use the graphs published by ECB. These graphs can be found in Figure 2.4.

27 2. Theoretical Background of Shadow Banking 12 Repurchase Transactions Let us have a detailed look at the example adapted from Acharya & Öncü (2010) in order to fully understand what a repo transaction is. There are two parties, a primary security dealer and its client. A dealer needs money, therefore he calls a client. They exchange a Mortgage-Backed Security (MBS) for USD 100 for a week. The understanding is that after a week, the dealer will return USD 105 and will get back the MBS. There is an interest of USD 5, principal of USD 100 and MBS is collateral that secures this loan. If the dealer fails to repay his debt of USD 105 to the client at the end of the week, the client will keep the MBS. If the client sells the borrowed MBS before the end of the week, then the client will have to buy it back in order to return it back to the dealer. If the dealer accepts, the client can instead buy a substitute (most likely cheaper) MBS. If the client fails to return the MBS or accepted substitute, then the dealer keeps USD 100 without paying extra USD 5. This example shows a repo agreement from the dealer s point of view (sale and purchase) and reverse repo agreement from the client s point of view (purchase and resale). Repo transactions may involve the counterparty risk due to the value of collateral that can be different from the principal of the loan. Therefore, there is a haircut, i.e. if the MBS is of USD 100, the loan should be only of USD 90 with a 10% haircut to the dealer. It is a margin set by the client in order to be protected against the value loss of MBS in case of dealer s failure. If the dealer is not the owner of the MBS, he needs to have USD 10 in order to buy it at the end of the repo. The MBS is dealer s asset, so his leverage is 10 times. There can be an opposite situation regarding the haircut, i.e. if the MBS is worth only USD 90 but the loan is USD 100. However, it is usually the dealer who faces a haircut. If the client has only USD 10, then he has to sell the MBS for USD 90 to someone else in order to provide a USD 100 loan. This makes a client s leverage be 10 times. If none of the parties have money/mbs, there is no haircut, but then they are infinitely leveraged.

28 2. Theoretical Background of Shadow Banking 13 Figure 2.5: Annual averages of daily financing by U.S. government securities primary dealers Source: Acharya & Öncü (2010). Half of all repo transactions takes place at night, we call them overnight repos. According to Acharya & Öncü (2010), commercial banks, investment banks, hedge funds, mutual funds, pension funds, MMFs, municipalities, corporations and other owners of large amounts of extra cash are the participants of repo transactions in the U.S.. In Figure 2.5, we can see the growing trend of the amount of repo transactions from Gorton & Metrick (2010) explain this growth in the amount of repo transactions by the rapid increase of money handled by institutional investors, pension funds, mutual funds, states, municipalities and non-financial firms. What they have in common is safety of the investments, interest and cash flexibility, in other words characteristics of the demand deposit-like products. These objectives can be easily attained using repo transactions. Repo transactions can also be used in other ways than deposit-like. They are often used to: ˆ hedge derivative positions and primary security issuance, ˆ keep no arbitrage relationships between cash and synthetic instruments, ˆ take short positions in security market, ˆ obtain leverage. (Gorton & Metrick (2010))

29 2. Theoretical Background of Shadow Banking 14 A study performed by Mancini et al. (2013) proves that central banks also participate in repo transactions. Moreover, they proved that the main driver of repo rates is central bank s liquidity provision. Direct intervention of a central bank implies change in the repo rate, thus the volume of repo transactions. In case of ECB, an excess liquidity causes a decrease in the volume of repo transactions. The incentives behind the repo transactions are very obvious. The interest rates on repo transactions are better in comparison to deposits at commercial banks. Also, there is an insurance incentive because large deposits are not insured while there is collateral as a mean of insurance in the repo transaction. (Acharya & Öncü (2010)) Gorton & Metrick (2010) emphasise that rehypothecation played also an important role in the recent financial crisis. As we already know, there is collateral used in each repo transaction. Singh and Aitken (2010) elaborated a paper on this topic, covering mostly the situation in the UK. They define rehypothecation in a following way. If collateral posted by a client to the prime broker is used as collateral in prime broker s other transactions, then the rehypothecation takes place. The prime broker can raise its funding in this way again and again. If we add the rehypothecation to the total estimated amount of shadow banking, the whole system gets larger than documented according to the International Monetary Fund (IMF) calculations presented in the paper by Singh & Aitken (2010). In the most extreme case by the end of 2007, the difference was more than USD 4 trillion. Figure 2.5 presents only partial data on repo transactions, since there is no official (or complete) data on the U.S. repo transactions. Depicted data show only primary dealer banks in the U.S. which is only a fraction of the U.S. market. Therefore, we can only estimate the actual size of the world repo market. ECB (2012) emphasises that repo transactions are the key source of funding in both U.S. and euro area. Thus, repo market is an important part of shadow banking. Further research on repo activities was performed by Adrian et al. (2012). This paper defines and studies repo market in the U.S. It also provides an explanation what data is needed for the repo market overview. Moreover, Krishnamurthy et al. (2012) study U.S. repo market and its importance before, during and right after the recent financial crisis with emphasis on the sources of repo funding. Knowing the importance of repo market in terms of shadow banking, we turn our attention to the study of this market in Section 4.3 in more details.

30 2. Theoretical Background of Shadow Banking Shadow Banking Entities After defining the shadow banking activities, Ricks (2010) summarises the entities that perform these activities: repo-financed dealer firms, securities lenders, SIVs, SPVs, ABCP conduits, credit-oriented hedge funds and most importantly MMFs. MMFs transform short-term credit into true demand obligations. These entities acquire short-term financing using the money markets and afterwards invest them in long-term financial assets. Moreover, Schwarcz (2012) adds to this list also government-sponsored companies and many investment banks. Gorton & Metrick (2010) did not forget about mortgage brokers who are also an important part of the shadow banking system. From all above mentioned information, it is obvious that shadow banking activities are privately financed. Adrian & Ashcraft (2012b) explain that credit intermediation operations are done without public funding and liquidity backstops. They are purely privately financed, often financed by credit put options. Adrian & Ashcraft (2012b) focus on the entities involved in the shadow banking system. Their aim is to present the frictions of shadow banking in order to understand it better. Adrian & Ashcraft (2012b) study ABCP conduits, tri-party repo market, MMFs and securitisation activities. Money Market Funds Money market funds have always been used as cash management services with high liquidity, stability and competitiveness of their market-based yields. They have been believed as immune to runs and financially stable. However, the recent financial crisis caused a significant change in all above mentioned qualities and advantages of MMFs. MMFs influenced negatively traditional banking sector and contributed to financial instability during and after the crisis. (Bengtsson (2013)) Kacperczyk & Schnabl (2010) study whether organisational form affects risk taking on the example of MMFs. They also present the definition of the MMFs and study their activities. MMFs invest in various types of short-term financial instruments, e.g. commercial papers, certificates of deposits, repo agreements or T-bills. Usually, the MMFs are owned by a stand-alone investment company or a financial conglomerate. Gorton & Metrick (2010) add that there is a strict regulation imposed on the U.S. MMFs. Usually, in Europe, MMF is set up as an Undertaking for Collective In-

31 2. Theoretical Background of Shadow Banking 16 vestment in Transferable Securities which makes it subject to the regulation imposed in the fund s domicile. It led to big differences among European countries. (Bengtsson (2013)) Figure 2.6: Split of total European MMFs by country of domicile 37% European MMF market 25% 23% 4% 4% 7% France Luxembourg Ireland Italy Nordic countries Other Source: Author based on Bengtsson (2013) For better illustration, let us present the European MMF market showing countries shares depicted in Figure 2.6. We can directly see that France is the MMF market leader with 37% share, followed by Luxembourg with 25% share and Ireland with 23% share. MMFs have been very popular and very often used in the U.S. since 1970s. In 2006, MMFs managed almost one fourth of the U.S. business short-term assets. Their assets increased over time and reached USD 3.8 trillion in Thanks to this evolution, they are one of the most significant financial product innovations over last 50 years. (Gorton & Metrick (2010)) As written by Gorton & Metrick (2010), the main characteristic of the MMF is that they try to keep the Net Asset Value (NAV) at USD 1 per share. Hence, they can be direct competitors to demand deposits. ECB (2012) informs that on 16 September 2008, the NAV dropped below USD 1 after Lehman bankruptcy and it led to the run on MMFs. Thanks to the U.S. MMF structures, there is a support of fund sponsors that helped to keep NAV at USD 1. Suddenly, after the Lehman Brothers bankruptcy, regulators and politicians realised that the whole financial stability is in question, so they had to come up with some safety measures. Firstly, it was Germany which announced that their Central bank will secure the liquidity of MMFs via temporary special

32 2. Theoretical Background of Shadow Banking 17 liquidity provision against collateral. Immediately, Germany was followed by Luxembourg and other countries supported by the EC. ECB was pushed to grant the liquidity. (Bengtsson (2013)) Gorton & Metrick (2010) also inform about the impact of the runs on MMFs. Due to these runs caused by Lehman bankruptcy, the MMFs had to sell their assets at fire sale prices. As there was a massive outflow of assets, the U.S. government introduced its Temporary Guarantee Program for MMFs. Figure 2.7: MMFs total balance sheets, in EUR millions Source: ECB (2012) Based on the data collected by the ECB, we can conclude that from the second quarter of 2011, the amount of MMF assets in the U.S. was EUR 1.83 trillion and in the euro area it was EUR 1.1 trillion. 1 Figure 2.7 plots the data for better comparison. When looking at Figure 2.7, we can spot an obvious difference in amounts because of different importance of MMFs in each market. Due to this fact, the U.S. market experienced bigger fall in the MMFs assets under management. Moreover, we will study the development on European and U.S. MMF markets in more details in Section and Section respectively. 1 MMFs in the euro area are slightly of a heterogeneous group.

33 2. Theoretical Background of Shadow Banking Subsequent Risks Shadow banking activities were very attractive thanks to the lack of regulation imposed. It was a way how to avoid the financial regulation completely. And probably that was also the reason why many investors decided to be a part of it. However, the shadow banking system brought important risks into the system that affected financial stability of the whole financial market. This became obvious right after the onset of the recent financial crisis. Since we defined the shadow banking activities as activities outside the regular banking system, Meeks et al. (2012) add that these activities have very similar risks as regular banking activities but less capital is involved. Moreover, maturity and liquidity transformation is done in very similar way as in the regular banking system. Furthermore, as Adrian & Ashcraft (2012b) claim, there are no direct public sources of funding and tail risk insurance in the shadow banking system. Schwarcz (2012) states that if shadow banking stays unregulated, it will lead to systemic risks in the financial system. Many authors publish papers on systemic risk as a result of the shadow banking activities. Let us mention Adrian & Ashcraft (2012b), EC (2012), ECB (2012), The Squam Lake Group: Baily et al. (2011), etc. High leverage is another source of the systemic risk. It may also have an important impact on the fragility of the financial market. EC (2012) confirms that shadow banking system has become highly leveraged but in practice it stayed unregulated. ECB (2012) points out another consequence of the high leverage. This leverage with asset price bubbles cause pro-cyclicality in the financial system via interlinkages between traditional and shadow banking or via investment of traditional banking in financial products issued by shadow banking. As we already stated, shadow banking provides short-term funding which is used to cover long-term needs. Therefore, a risk of liquidity discontinuities arises. Schwarcz (2012) further defines that if a traditional bank uses shortterm deposits to fund itself with causing liquidity discontinuities, it is called a bank run. Moreover, EC (2012) supports that this short-term financing may face the risk of sudden and huge withdrawals by clients. Bank runs have an important impact on the whole economy. They cause system risk. Let us recall that shadow banking system and regular banking system are very closely connected. Therefore, the systemic risk is transferred

34 2. Theoretical Background of Shadow Banking 19 Figure 2.8: Risk analysis framework of interconnectedness between traditional and shadow banking entities Source: FSB (2013a) Note: OFI - other financial intermediaries via interlinkages of traditional banking and shadow banking and via contagion effects. ECB (2012) emphasises that systemic risk is driven by those interlinkages. Let us see the graphical view of the risks caused by interconnectedness between shadow and traditional banking explained by FSB in Figure 2.8. FSB (2013a) explains that Figure 2.8 depicts how different types of shadow banking entities are connected to regular banks. Different types of shadow banking entities may introduce different types of risks to traditional banking system via this connection. Let us remind the factors of these risks. It is credit intermediation, maturity and liquidity transformation and leverage. They might lead to credit and funding risk shown in Figure 2.8. Based on this framework, the FSB can focus on risk arising from the interconnectedness of traditional and shadow banking. Moreover, FSB can improve the analysis, measurement and regulation policy implications. The result of the third, most recent FSB analysis (end of 2012 data) regarding the interconnectedness can be found in Appendix A in Figure A.1. Up to now, we have been describing that systemic risk is an important consequence of unregulated shadow banking but let us mention what the systemic risk causes. Apart from contagion, ECB (2012) states that it leads to important catalyst effects for liquidity and solvency risks. On one hand, we have to admit that shadow banking activities led to significant economic efficiencies, but on the other hand the crisis proved that shadow

35 2. Theoretical Background of Shadow Banking 20 banking generates the risk of contagion. (Adrian & Ashcraft (2012b)) Due to the interconnections between regular and shadow banking, the spill over effect poses an important risk. FSB (2011b) warns that the longer and the less transparent credit intermediation chain gets, the bigger risks may appear. Contagion effect resulting from the interconnections is also an important part of the Occasional Paper by ECB (2012). Therefore, authors emphasise the importance of these interlinkages when preparing the shadow banking regulation. Since the MMFs are an important part of shadow banking, The Squam Lake Group: Baily et al. (2011) study MMFs, its risks and point out the need for regulation. They believe that MMFs pose significant systemic risk. Also, they claim that fund investors and systematically important borrowers are exposed to runs. Transparency of actual NAV and fire sales of funds assets are an important issue for such runs. Its impact on the whole market is indisputable. They estimate that total investments into U.S. and similar European MMFs are over USD 2 trillion. Any such run on MMFs would lead to a disaster that we experienced after the fall of Lehman Brothers when the Commercial Paper (CP) markets failed. Due to the interconnections, the whole financial market might consequently fail. Furthermore, run on repo is the main concern of various authors. Let us mention some of them. The insight into the repo runs is presented by Copeland et al. (2011) with the emphasis on the U.S. tri-party repo market. Gorton & Metrick (2012) dedicate their paper to the study of run on repo and they believe that together with securitisation, it was the nexus of the recent financial crisis. Moreover, Martin et al. (2010) develop a model in order to study whether these runs can be prevented. They define two constraints, a solvency and a liquidity constraint. The solvency constraint ensures the existence of sufficient current and future profits to repay all investors with no renewed loan. The liquidity constraint ensures the existence of sufficient resources at the date of the run. The authors prove that if none of these constraints is violated, a run can be prevented. Besides the runs on repo, repo markets are also subject to other risks such as counterparty credit risk, market risk and operational risk. Counterparty credit risk is a risk of default of one party of a repo transaction. Market risk is the main risk on repo market. It is given by the price volatility or by the collateral value realised in a sale. Operational risk is connected with the transfer and management of collateral. There are different tools that are created in order to minimise those risks, e.g. use of collateral and initial margins, position limits

36 2. Theoretical Background of Shadow Banking 21 with counterparties, daily marking to market of collateral or concentration limits for specific securities. All these risks are reflected in the repo rate. Central bank utilises several tools for the regulation of the repo market, e.g. it sets the repo rate, it can expand the collateral list for repo operations or it can change the supply of government securities in order to stimulate the volume of repo transactions. (BIS (2008)) Furthermore, the literature review of repo risks is presented in Mancini et al. (2013). Other risks arising from repo activities and securities lending are defined by FSB (2013c). Nowadays, shadow banking is increasing also in emerging markets and developing countries. The question whether increasing shadow banking poses a significant risk for the financial system in emerging markets and developing countries or it is only the case of developed countries is studied by Gosh et al. (2012). They show the evolution of the shadow banking sector in some of the emerging markets. For example, the shadow banking sector is rising significantly in Philippines, Indonesia, Thailand, China, Bulgaria or Croatia. Despite the relatively small presence of shadow banking in emerging and developing countries, it is already a significant part of their economies.

37 Chapter 3 Regulation and Supervision In Chapter 2, we defined shadow banking and we also pointed out potential risks that come with shadow banking system. As we mentioned, there was a call for regulation due to revealed evidence after the recent financial crisis. Therefore, FSB and other international standard-setting bodies reacted to this call and started a long process of preparation of regulation measures and recommendations for the oversight. In this chapter, we focus on the reasons why the regulation of shadow banking is needed. Then, we present the FSB regulation, Basel III with its link to shadow banking and the U.S. shadow banking regulation. This chapter is concluded with emphasis on the obstacles faced by the shadow banking regulation. Moreover, ICMA (2012) summarises other reasons why the regulators should be concerned about shadow banking sector. We picked some of them in order to point out the importance of shadow banking regulation. First, it is the size of shadow banking that is getting bigger. Second, the fact that shadow banking entities are unregulated or weakly regulated when compared to traditional banking. Third, the regulatory arbitrage that is taking place when traditional banking participates in shadow banking. Fourth, it is the interconnection of the whole shadow banking chain that may be a channel of systemic risk. Fifth, the lack of transparency can hide important risks and subsequently end up in the panic. Significant tightening of banks capital and liquidity requirements was the first impulse after the crisis aiming at the shadow banking regulation. Since the activity is always moving towards less regulated sector, regulators started to focus on the regulation of securitisation, MMF, hedge funds, etc. The risk

38 3. Regulation and Supervision 23 shifted from the regulated sector to less regulated sector. (Eurofi (2011)) 1 The FSB prepared a monitoring framework to asses shadow banking risks and during the summer of 2010, they started to test this framework. Throughout this period, they made some important adjustments to improve the assessment of monitoring risk factors and regulatory arbitrage. It was in November 2010 at the Seoul Summit, when the G20 leaders pointed at the importance of strengthening regulation and supervision of the shadow banking system. They turned to the FSB and other international standardsetting bodies in order to prepare a set of recommendations to strengthen the oversight and regulation of the shadow banking system by mid (FSB (2011a)) The FSB promptly reacted to the G20 leaders call and organised a workshop in London already in December In order to formulate the recommendations, they formed a Task Force with following objectives: ˆ clarification of the shadow banking system, ˆ set out of potential approaches for overseeing the shadow banking system, ˆ examination of possible regulatory measures that would apply to the systemic risk and regulatory arbitrage concerns posed by the shadow banking system. (FSB (2011a)) 3.1 FSB s Regulation In 2011, during the Summit in Cannes, the G20 agreed on policy creation regarding the shadow banking system. The FSB strategy on doing so was defined in two concepts, i.e. monitoring of the shadow banking system and strengthening the oversight and regulation of the shadow banking system. (FSB (2013c)) FSB further explains detailed design of general principles for regulatory measures. The principles are depicted in Figure 3.1. Let us briefly comment Figure 3.1. First, when making a policy, it is necessary to focus on the target of the regulation. Second, a measure should be proportionate to the shadow banking risks. Third, a measure should be forward-looking and adaptable to possible risks. Fourth, the effectiveness is a must when looking for a balance. 1 Eurofi is a European think-tank, chaired by J. de Larosière, dedicated to financial regulation and supervision and a platform for exchanges between the industry and public authorities.

39 3. Regulation and Supervision 24 Figure 3.1: General FSB principles of regulatory measures focus assessment proportionality effectiveness forwardlooking and adaptable Source: Author based on FSB (2013d) Fifth, regular assessment of the effectiveness of a regulatory measure is a key to the improvement FSB s Workstreams FSB defines their workstreams in order to decide on the necessary regulatory actions. These workstreams are presented in Table 3.1. FSB (2011b) further explains that tasks regarding the workstreams were directly assigned to institutions in charge. Namely, the Basel Committee on Banking Supervision (BCBS) is in charge of the first workstream, the International Organisation of Securities Commissions (IOSCO) is in charge of MMF, IOSCO together with BCBS are responsible for securitisation and the Task Force is in charge of the rest. They all report directly to FSB. Thanks to their reporting, the FSB can adjust the regulation. To prevent the risk of contagion and spill over effects, one of the bases of the FSB is sharing experience internationally in monitoring and regulating shadow banking.

40 3. Regulation and Supervision 25 Table 3.1: FSB s workstreams Workstream The regulation of banks interactions with shadow banking The regulatory reform of money market funds The regulation of other shadow banking entities Definition examination of enhanced consolidation for prudential regulatory purposes, concentration limits, risk weights for banks exposures to shadow banking entities examination of regulatory action related to money market funds new workstream related to examination of other shadow banking entities The regulation of securitisation examination of retention requirements and transparency The regulation of securities lending and repos new workstream regarding the examination of securities lending and repo transactions including margins and haircuts Source: Author based on FSB (2011b) FSB s Actions In the recent FSB s publication form August 2013, an overview of the policy recommendations regarding shadow banking is presented. Let us summarise the FSB s actions in order to strengthen oversight and regulation of shadow banking: ˆ mitigation of the spill over effect between the traditional banking system and shadow banking system, ˆ reduction of the vulnerability of MMFs to runs, ˆ assessment and alignment of the incentives regarding the securitisation, ˆ damping of the financial stability risks and pro-cyclicality regarding repo activities and securities lending, ˆ assessment and mitigation of systemic risks posed by other shadow banking institutions and activities. (FSB (2013b))

41 3. Regulation and Supervision 26 In order to monitor the global shadow banking sector, FSB began annual monitoring exercises starting in FSB (2013b) proudly announces that they succeeded to cover 90% of global financial assets. The improvements are expected thanks to the data gathering via this annual monitoring exercise. FSB also observed positive results, as the whole size of the shadow banking system decreased FSB s Following Steps The following path of the FSB shadow banking regulation comes from the Saint Petersburg Summit (September 2013). Let us briefly summarise these steps: ˆ adjustment of recommendations on traditional banks interactions with shadow banking institutions, ˆ improvement of suggested minimum standards on haircuts in securities financing transactions, ˆ review of securitisation markets, ˆ improvement of data collection and aggregation, ˆ enhancement of information-sharing process on high-level, ˆ timing of policy recommendations. (FSB (2013b)) In the most recent monitoring report, FSB (2013a) admits that data obtained from countries in the sample is very often not granular enough. In order to specify the size of shadow banking and narrow down the data collected, more granular data is highly preferred. According to the measurement, in 2012, the non-bank financial intermediation was USD 71 trillion. The FSB uses this indicator as a proxy of global shadow banking system. In this measurement, 25 countries were covered. 2 If more granular data is used, the estimate is narrowed down. However, only 20 countries were covered in the narrowed estimation. 3 In this case, the shadow banking estimate decreases from USD 55 trillion to USD 35 trillion. From our point of view, it is necessary to emphasise that estimation of shadow banking by FSB is rather broad and not precise. This is due to 2 Argentina, Australia, Brazil, Canada, Chile, China, Germany, France, Hong Kong, Indonesia, India, Italy, Japan, Korea, Mexico, Netherlands, Russia, Saudi Arabia, Singapore, Spain, Switzerland, Turkey, United Kingdom, United States and South Africa. 3 Germany, France, Italy, Netherlands and Spain were omitted.

42 3. Regulation and Supervision 27 the definition of shadow banking estimate they apply. FSB s estimate includes activities performed by other financial intermediaries, i.e. all financial institutions that are not classified as central banks, banks, insurance companies, pension funds, public financial institutions or financial auxiliaries. However, shadow banking estimate should also capture shadow banking activities performed by institutions that are not necessarily shadow banking institutions. We have already shown in Section 2.2 that some of the activities performed by traditional banking entities are counted as shadow banking activities. During the crisis, many traditional banks used shadow banking sector to raise their funds. Moreover, the FSB s more precise estimation excludes countries that represent an important share of the global shadow banking system. 3.2 Basel III Basel III is an international framework prepared by BCBS in order to strengthen the capital and liquidity regulation, supervision and risk management of the banking sector. Basel III aims to improve the sector s ability to absorb shocks caused by financial and economic stress, to ameliorate the risk management and governance and to strengthen banks transparency and disclosures. (BCBS (2010c)) As for the capital requirements, the BCBS adjusts limits and sets minima at all times as follows: Common Equity T ier 1 4.5% RW A T ier 1 Capital 6.0% RW A T otal Capital = T ier 1 Capital + T ier 2 Capital 8.0% RW A where RW A stands for risk-weighted assets. (BCBS (2010a)) BCBS (2010b) further defines the liquidity coverage ratio so that banks maintain an adequate level of unencumbered high-quality liquid asset that can be turned into cash. It facilitates banks liquidity needs for a 30 calendar day time horizon under liquidity stress scenario specified by supervisors. This liquidity coverage ratio is specified as follows:

43 3. Regulation and Supervision 28 Stock of high -quality liquid assets T otal net cash outflows over the next 30 calendar days 100% ECB (2012) is persuaded that thanks to Basel III, the banks may try to avoid higher risk weights and capital requirements using securitisation or to avoid limitations to leverage using investments into the shadow banking entities. equity. These entities have higher leverage in order to get higher return on However, more stringent capital and liquidity requirements for credit institutions may lead to an increase in the shadow banking activities. The Basel III can leave traditional financial institutions no other choice than to look for their funding in the shadow banking sector. (ECB (2012)) Eurofi (2012) also states the opinion that with Basel III, the shadow banking sector is expected to increase. Since the regulation imposed by Basel III is very strict, banks cannot borrow from the institutions under this regulation. Shadow banking sector is the best alternative. Therefore, Basel III indirectly pushes traditional financial institutions outside the regulated banking sector. (ECB (2012)) Pozsar & Singh (2011) also hold the view that riskier activities are performed outside the traditional banking system. Therefore, these efforts end up in increasing shadow banking after all. Moreover, Shen (2013) points out that Basel III regarding capital adequacy has not adopted appropriate measures to deal with potential risks arising from shadow banking. Regulation imposed on the traditional banking in the form of Basel III has the opposite effect on shadow banking than shadow banking regulators would prefer. This suggests that separate regulation of the shadow banking system was needed. 3.3 U.S. Regulation Question of the shadow banking system regulation is relevant not only for standard-setting institutions but also for other economists. Gorton & Metrick (2010) published their proposition of regulation. They focus on three main parts of the U.S. shadow banking, i.e. money market mutual funds, securitisation and repo transactions. Government-guaranteed insurance and strict guidelines on collateral are successful ways of the regulation of privately created

44 3. Regulation and Supervision 29 money based on the historical experience. 4 Therefore, they propose a combination of these two methods as an efficient regulation of the shadow banking system. Schwarcz (2012) points out that the attention of regulators should be turned to maximisation of economic efficiency and minimisation of shadow banking potential to increase systemic risk instead of increasing efficiency and limiting shadow banking per se. In 2010, as a response to the recent financial crisis, U.S. government prepared a Dodd-Frank Act in which the regulation was turned to the shadow banking system as well. 5 As described by Adrian & Ashcraft (2012b), Dodd-Frank Act is targeted to change the risk retention of securitisation transactions, mainly ABS and ABCP. Furthermore, the Volcker rule and other restrictions embodied in the Dodd-Frank Act prohibit banks from sponsoring private funds, including ABCP conduits, and impose strict restraints regarding the connections between banks and ABCP conduits. Additionally, this regulation obliges these ABS sponsors to disclose their activities. U.S. Securities and Exchange Commission also prepared a regulation AB that aims to regulate, monitor and measure the ABS sector. Similarly, the Dodd-Frank Act is targeted at repo activities and MMFs. The support of the public sector aimed at the broker-dealers of repo activities is harshly restricted. When regulating MMFs, the main goal is to decrease MMFs fragility and their inclination to runs which may lead to financial instability. Gorton & Metrick (2010) express their disappointment that this reform did not encompass the system as a whole. Moreover, Adrian et al. (2013b) not only present an overview of the U.S. regulation framework but also introduce a program for monitoring financial stability with emphasis on shadow banking. 3.4 Obstacles to the Regulation Lack of available data regarding shadow banking activities is the biggest and the most important problem that shadow banking regulators, financial institutions and researchers face. The authors are forced to prepare their own data in order to study this sector of economy. The same problem arises in case of standardsetting institutions, other financial institutions or central banks, e.g. FSB, IMF, ECB, etc. Lately, there has been a significant evolution in the precise definition 4 Great Depression and Panic of Dodd-Frank Wall Street Reform and Consumer Protection Act

45 3. Regulation and Supervision 30 of the shadow banking entities and activities. Therefore, we hope that data will be easier to collect and publish respectively. ECB (2012) claims that more than 60% of shadow banking activities is linked to the entities that do not provide the high-frequency data. One may think that available data on shadow banking is accurate and can be used for monitoring purposes. However, Bouveret et al. (2013) point out that use of such data for market monitoring is very limited. Further data aggregation is very difficult. Moreover, the importance of available data is explained by FSB. The data is necessary for financial stability monitoring and policy-making. FSB outlines the lesson from the recent financial crisis. Clear and timely data regarding securities financing markets, their evolution and related risks of activities on such markets are essential components for standard-setting authorities. Otherwise, it is only cyclic dealing with relatively late-stage market development that brings systemic risk transmission. (FSB (2013c))

46 Chapter 4 Shadow Banking System Chapter 2 and Chapter 3 were dedicated to theory behind shadow banking and the regulation and supervision of this area of economy. In this chapter, the organisation is as follows. We start with our assessment of the European and the U.S. shadow banking systems and we compare these sectors and comment on obtained data. We study mostly the impact of the recent financial crisis on the consequent development of shadow banking activities. Given the rise in the regulation and supervision of shadow banking, we expect the whole shadow banking sector to be diminished gradually after the recent financial crisis until nowadays. MMFs and repo operations represent the core of the shadow banking system, also pointed out by Adrian et al. (2013a). Therefore, we study Chinese MMFs and French and UK repo markets. Moreover, we introduce our model in order to study the relations among shadow banking, traditional banking and economy itself. At the end of this chapter, we outline further research opportunities. 4.1 Assessment of the Shadow Banking Sector As we have already mentioned, the availability of data regarding shadow banking is rather scarce. Adrian et al. (2013a) point out two reasons why the estimation of the shadow banking system is very difficult. The first is given by the shadow character of the entities involved. Therefore, the data may be completely missing. The second reason is explained by complicated cross country aggregation of data due to the definition of certain shadow banking institutions that may vary across countries because of different legal and regulatory meaning.

47 4. Shadow Banking System 32 Let us first stress out the importance of the measurement of the shadow banking system. According to the IMF study presented by Claessens et al. (2012), it is necessary to measure this sector in order to improve regulation policies and supervision. They also mention that there are big differences in the methodology, since the methodology is not uniform. Apart from big bodies such as FSB, Federal Reserve Bank (FED) and IMF, there is only an attempt of Bouveret (2011) who tried to measure both U.S. and European shadow banking system. However, his research was performed on the data only until Therefore, we will focus on measurement of the shadow banking system and we will extend and improve Bouveret s research. Practice shows that authors often prepare their own data based on several public or private databases. Since the specifically defined methodology of shadow banking measurement is missing, we will follow existing attempts. Using published approaches and authors ideas for future research, we will perform a qualitative analysis in which we assess the size of the shadow banking sector in Europe and in the U.S. in order to test the first hypothesis. Hypothesis 1: European shadow banking market increased in volume during the financial crisis while the situation in the U.S. went other direction, i.e. during the crisis, the amount of shadow banking activities declines in volume. We will first estimate the volume of shadow banking for the period from 2006 until This period of time covers the impacts of recent financial crisis as well as the situation afterwards and actual state. Based on the available data for this research, we may subsequently decide whether the hypothesis will be rejected or not. One of the first attempts of an assessment of European shadow banking sector was performed by Bouveret (2011). The author tried to estimate both U.S. and European shadow banking sectors and compare them. As he mentions, the data for U.S. sector is provided in the FED s flow of funds reports. Unfortunately, there is no such database for European sector. We will follow the definition of shadow banking used in Bouveret (2011). The author computes shadow banking as the sum of maturity transformation data and credit transformation data, i.e. the sum of total outstanding open market paper, total repo liabilities, total shares outstanding of MMF and total liabilities of ABS issuers. In our estimation, we will include data on hedge funds and we will study the development until 2013.

48 4. Shadow Banking System European Shadow Banking Sector The availability of the data is the biggest obstacle for the assessment of the European shadow banking sector. There are many private providers of the majority of shadow banking data but their service is not feasible. Thanks to the definition of shadow banking, we will be using, and several databases, we can compute the European shadow banking system. According to the most recent monitoring exercise capturing year 2012, FSB (2013a) claims that non-bank financial intermediation in euro area was USD 22 trillion. The report further states that in the UK, it was USD 9 trillion. Hence, for covered European part of shadow banking sector, FSB s monitoring exercise results in total size of USD 31 trillion (EUR 23,5 trillion) 1 at the end of However, let us remind that this estimation is rather broad. Figure 4.1: Size of the European shadow banking sector Source: Bouveret (2011) Bouveret (2011) estimated the size of European shadow banking and this study is still quite unique. No such complex European study has been performed ever since. There has also been an attempt to estimate the size of European shadow banking by Tyson & Shabani (2013). The dataset is composed of data taken directly from the financial statements of shadow banking institutions. They use consolidated group financial statements of investment banks and also subsidiary financial statements. However, they performed their 1 Author applied the conversion using ECB exchange rate at the end of 2012: EUR 1 = USD

49 4. Shadow Banking System 34 analysis only on the UK market. This method is time consuming and less accurate. Therefore, we decided to take this study as a good example on which we can build our research. The size of the European shadow banking system computed by Bouveret (2011) can be seen in Figure 4.1. In Figure 4.1, we can see the evolution of shadow banking sector in Europe between 2006 and There has been a decrease in the amount of repo transactions during 2008 but the overall shadow banking activities remain high despite the regulation imposed on shadow banking. According to this research, European shadow banking attained EUR 10.5 trillion in 2010 Q2. Seeing that FSB applied their regulation recommendations on shadow banking only after this study was performed, we find it very interesting to study the shadow banking system from that time on. Such analysis allows observing the impact of shadow banking regulation changes. In order to estimate European shadow banking sector, we will concentrate on CP market, repo market, MMF market, ABS issuance and hedge funds. Unfortunately, the data covering securities lending in Europe is not available. Otherwise, it would be included in our study as well. We also have to mention a study by ECB (2012) in which the authors assess shadow banking in euro area. They confirm the difficulty of such work. Also, they point out the fact that euro area banks use shadow banking institutions as a source of their funding. Now, we present a study of partial shadow banking sectors in Europe in order to analyse them. Afterwards, we put them together to obtain overall shadow banking sector. The same procedure is repeated for the U.S. market. Finally, we compare obtained results and analyse them in more details. Commercial Paper Market To cover the data on CP market, we use the ECB database on Short Term European Paper. Using this data, we cover the issuance of CPs, Euro CPs, Certificates of Deposits and Euro Certificates of Deposits. We obtain the data of outstanding amounts for the period from 2006 Q3 until 2013 Q3 and we plot it in Figure 4.2. What we notice first is no big shock due to the recent financial crisis. From 2006, the market was growing quickly. From 2009, the growth slowed down and stabilised around EUR 400 billion. This way of funding was obviously preferred during and directly after the crisis. Therefore, this market keeps its volume at stable level. The impact of the regulation aimed

50 2005 Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q3 4. Shadow Banking System 35 Figure 4.2: European CP market, outstanding amounts EUR billion European CP market Source: Author based on ECB at downsizing the shadow banking system is not obvious but during last year, the market shows a decreasing trend. Repo Market Figure 4.3: European repo market, outstanding amounts EUR billion 8,000 7,000 6,000 5,000 4,000 3,000 2,000 1,000 0 European repo market Source: Author based on ICMA (2013) In September 2013, the International Capital Market Association (ICMA)

51 2006 Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q2 4. Shadow Banking System 36 published the most recent official European repo market survey with complete data covering all repo business (repo and reverse repo transactions) from 2001 until now. The survey involves 65 institutions. In our analysis, we use the data from 2005 Q4 until 2013 Q2. The development on repo market is shown in Figure 4.3. From the graph, we can see an obvious impact of the recent financial crisis on repo business in Europe. Already in 2008 Q3, the market experienced a loss in the amount of repo activities. During 2008, the total loss in the volume of repo transactions was EUR 1.87 trillion. Despite the presence of the crisis, repo transactions attained its pre-crisis level very quickly. Nowadays, the market is stable around EUR 6 trillion. We cannot spot any significant changes in the volume caused by existing regulation. MMF Market Figure 4.4: European MMF market, total assets EUR billion 1,400 European MMF market 1,200 1, Source: Author based on ECB and Bank of England In order to estimate the MMF market, we use outstanding amounts of total assets of European MMFs. The data is accessible from the ECB database. To improve the measurement, we decided to include the UK market. We include the UK because it is the biggest player on the MMF market in terms of non-euro area countries and this data is public thanks to the Bank of England. The data is in quarterly frequency. Hence, we use data from 2006 Q1 to 2013 Q2 and we plot it in Figure 4.4. MMF market in Europe is very important due to its overall volume. The market peaked at the beginning of 2009 (EUR 1.3 trillion)

52 2007 Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q2 4. Shadow Banking System 37 despite the presence of the crisis. From that time on, MMF activities have a decreasing trend. We believe that this decrease is influenced by the regulation of shadow banking activities and by the supervision applied on MMFs. ABS Market Figure 4.5: European ABS market, outstanding balances EUR billion 2,500 European ABS market 2,000 1,500 1, Source: Author based on Association for Financial Markets in Europe (AFME) AFME prepares a survey on securitisation on the quarterly basis. They monitor the issuance of securities and in their securitisation reports, they publish outstanding balances. Thanks to these reports, we are able to collect the data on ABS from 2007 Q3 to 2013 Q2. Unfortunately, the data for 2006 and the first half of 2007 are not accessible. ABS market in Europe is presented in Figure 4.5. European ABS market reflects interesting changes in this shadow banking activity due to the recent financial crisis. We assume that increase in the volume during 2008 is a result of changes in behaviour of market participants who started to lean towards ABS. They might have changed their ways of funding from other sources to ABS which results in this boom. During the crisis in 2009, the situation is quite stable around EUR 2.2 trillion. However, from 2012, the market is visibly shrinking which may be a result of the regulation recommendations introduced by FSB.

53 2006 Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q2 4. Shadow Banking System 38 Figure 4.6: European hedge fund market, total assets EUR billion European hedge fund market Source: Author based on HedgeFund Intelligence Hedge Fund Market So far, there has been no academic research on hedge funds regarding its volume and its share in the whole shadow banking system. That is the reason why we include hedge funds into our research. We use the data collected by HedgeFund Intelligence, the largest private provider of hedge fund news and data in the world. Figure 4.6 captured hedge fund market for the time period from 2006 Q1 until 2013 Q2. In Figure 4.6, we can directly spot a significant gradual decrease from 2007 until the beginning of 2009 when the market hit its bottom of EUR 237 billion. Overall hedge fund market in Europe has not been recovered to its pre-crisis value, but we may expect it in the following years. We believe that reason of this growing trend is the lack of of regulation Estimation of the European Shadow Banking Sector After having studied all partial shadow banking markets that form the whole shadow banking as we defined it, let us put studied markets together and estimate the whole European shadow banking sector. We prepared a summary of the effect of recent financial crisis on the volume of activities in partial markets. Here, we study the difference in volume over time. We compare the values from 2008 Q2, right before the crisis, and 2010 Q1, after more than a year of the presence of crisis. Final change is presented

54 2006 Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q2 4. Shadow Banking System 39 Table 4.1: Impact of the crisis, Europe Market commercial paper repo money market fund asset-backed securities hedge fund Source: Impact of the crisis Author by arrows where and mean a significant change and and mean a slight change. This summary is presented in Table 4.1. Based on this information, we can conclude the following. It was repo market and ABS market that have experienced significant growth in the volume over chosen time. Hedge fund market decreased heavily. In the CP and MMF markets, the changes were rather small than sharp. Figure 4.7: European shadow banking sector vs. total bank assets EUR billion 12,000 10,000 8,000 6,000 4,000 2,000 0 European shadow banking sector EUR billion 50,000 45,000 40,000 35,000 30,000 25,000 20,000 15,000 10,000 CP Repo MMF ABS Hedge funds Bank assets (rhs) Source: Author based on ECB, ICMA, AFME, Bank of England and Hedgefund Intelligence Let us put all studied partial markets together in order to form European shadow banking estimate. We present our estimate in Figure 4.7. For better orientation, we add total bank assets curve. In the graph of total European shadow banking captured in Figure 4.7, we can see a sharp decrease to EUR 8.3 trillion in 2008 Q4. This decrease is mainly given by the decrease in repo activities which represent the majority of shadow banking activities. However,

55 2006 Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q2 4. Shadow Banking System 40 the whole sector recovered very quickly and already in 2010 Q2, it reached its maximum of EUR trillion. Pre-crisis maximum was EUR 9.9 trillion in 2008 Q2. By the end of 2013 Q2, according to our calculations, the European shadow banking sector equals to EUR 9.3 trillion. Summarily, European shadow banking market decreased before the onset of recent financial crisis, it was rising despite the presence of the crisis, and it has decreasing tendency after the crisis. The estimation of shadow banking in Europe shows that size of this market is still enormous. We believe that regulation aimed at this part of economy has not succeeded yet. Hence, the expectations of downsizing shadow banking system are still not proved. In order to study the effect of shadow banking regulation, we recommend recalculating our estimate in several years. Additionally, as shown in Figure 4.7, we include the European bank assets curve in order to discover how big the shadow banking system is in comparison to the traditional banking system. We use the data available by ECB on total assets of Monetary Financial Institutions (MFI) in euro area and in non-euro area countries and we subtracted the euro system assets and MMF shares/units. There are no big shocks due to the crisis in total bank assets. Thanks to this comparison, we can say that shadow banking ranges from 19% to 25% of the total bank assets. Figure 4.8: European shadow banking as a share of total bank assets 30% 25% 20% 15% 10% 5% 0% Share of shadow banking Source: Author s computations Figure 4.8 shows the share of shadow banking on total bank assets and how

56 4. Shadow Banking System 41 it varies over time. The pre-crisis state shows higher share of shadow banking. Right after the onset of recent financial crisis, the share hits its minimum (19%) and starts rising. From 2010 Q2, when the share of shadow banking is 25% of total bank assets, it decreases. Last observations suggest increasing tendency. The estimation of European shadow banking sector proves that repo market is the biggest and the most important part of it. We also see an importance of MMFs as well. Therefore, we will study these two sectors in more details later in this chapter The U.S. Shadow Banking Sector Figure 4.9: Shadow banking liabilities vs. traditional banking liabilities, in USD trillion Source: Pozsar et al. (2010) Literature proves that many authors tried to estimate the volume of U.S. shadow banking sector. Let us mention the study performed by Pozsar et al. (2010). They provide two measures of the size of the shadow banking system, net and gross. Gross measure counts all liabilities related to securitisation activities (ABS, MBS and Government-Sponsored Enterprises (GSEs) liabilities) and short term money market transactions that are not backstopped by deposit insurance (repo, CP and MMF liabilities). The aim of the net measure is to eliminate double counting. They use the data published in the flow of funds by FED. Their comparison of shadow banking and traditional banking for the

57 4. Shadow Banking System 42 time period from 1990 until 2011 can be seen in Figure 4.9. Note that shadow banking system is much bigger than the traditional banking system already by the end of the 90s. From Figure 4.9, we can spot a significant increase in the shadow banking system from the beginning of new millennium. The impact of recent financial crisis on the size of the shadow banking system is very obvious when this part of economy decreased sharply. The peak of shadow banking was reached in June 2007 and it was estimated at nearly USD 22 trillion. The pre-crisis size of shadow banking was not attained according to this study. Figure 4.10: Size of the U.S. shadow banking sector Source: Bouveret (2011) Note: MMF - money market funds, ABS - asset-backed securities, GSE - government-sponsored enterprises A newer and already mentioned study by Bouveret (2011) also assesses the size of the U.S. shadow banking sector. The result of his estimation is depicted in Figure As compared to the study by Pozsar et al. (2010), the partial markets of shadow banking are separated by different colour in Figure Over time, shadow banking was rising slowly and it peaked in 2007 as already shown in the previous example by Pozsar et al. (2010). This analysis ends in Therefore, it is very interesting how the shadow banking system evolved over past 3 years. Also, its comparison to the traditional banking sector represented by total bank assets is very interesting. This evidence is very different to what we showed in case of Europe. Already mentioned FSB monitoring report also presents its measurement

58 2006Q1 2006Q2 2006Q3 2006Q4 2007Q1 2007Q2 2007Q3 2007Q4 2008Q1 2008Q2 2008Q3 2008Q4 2009Q1 2009Q2 2009Q3 2009Q4 2010Q1 2010Q2 2010Q3 2010Q4 2011Q1 2011Q2 2011Q3 2011Q4 2012Q1 2012Q2 2012Q3 2012Q4 2013Q1 2013Q2 4. Shadow Banking System 43 regarding the U.S. shadow banking sector. FSB (2013a) confirms that U.S. shadow banking is the biggest player among all countries in the sample. The FSB shadow banking estimate is USD 26 trillion by the end of In order to achieve an assessment of the U.S. shadow banking system, we use the same methodology as we did for the calculation of the European shadow banking system. In our analysis, we likewise cover CP market, repo market, MMF, ABS issuance, hedge funds and we include GSE and pool securities extending the credit transformation activities that are present in the U.S. economy. All necessary data is available in the FED s flow of funds database. We cover the same period of time, i.e. from 2006 until Thanks to this period, we are able to study post-crisis situation in the U.S. and observe the evolution in the shadow banking sector until nowadays. Commercial Paper Market Figure 4.11: U.S. CP market, outstanding amounts USD billion 2,500 U.S. CP market 2,000 1,500 1, Source: Author based on FED Figure 4.11 depicts the evolution on the CP market from 2006 until Gradual decrease in volume, visible from the graph, started already in 2007 Q2. From 2010, the market stabilised around the volume of USD 1 trillion. We suppose that after the onset of recent financial crisis, the use of CP as a way of funding, lost its attractiveness to many agents on the market. Moreover, FSB regulation does not show any significant downsizing effect from its introduction during 2011.

59 2006Q1 2006Q2 2006Q3 2006Q4 2007Q1 2007Q2 2007Q3 2007Q4 2008Q1 2008Q2 2008Q3 2008Q4 2009Q1 2009Q2 2009Q3 2009Q4 2010Q1 2010Q2 2010Q3 2010Q4 2011Q1 2011Q2 2011Q3 2011Q4 2012Q1 2012Q2 2012Q3 2012Q4 2013Q1 2013Q2 4. Shadow Banking System 44 Repo Market Figure 4.12: U.S. repo market, outstanding amounts USD billion 3,000 U.S. repo market 2,500 2,000 1,500 1, Source: Author based on FED Since ICMA survey covers only European market, we use FED data in order to capture the size of the U.S. repo market. This data is plotted in Figure Repo market was on a way to achieve USD 2.5 trillion when it started to decrease. With the onset of recent financial crisis the volume of repo transactions decreased even further. In the present days, repo market is not even a half of what it was before the crisis. This may be a proof of the effectiveness of imposed regulation on the repo activities. For last two years, the volume is slightly above USD 1 trillion. MMF Market In order to capture the U.S. MMF market, we plot total assets under MMF management. This data is accessible from the FED database. The development on MMF market is depicted in Figure MMF market was a growing market until the beginning of 2008 when the trend has slowed down but the market was still growing. Figure 4.13, we can see a delay in the response to the recent financial crisis which proves the interest of investors to use MMFs as a way of funding. The MMFs market started to decrease significantly in its volume from Until now, the volume from 2008 was not recovered and market is stabilised around USD 2.5 trillion. MMFs market is bigger in size than repo

60 2006Q1 2006Q2 2006Q3 2006Q4 2007Q1 2007Q2 2007Q3 2007Q4 2008Q1 2008Q2 2008Q3 2008Q4 2009Q1 2009Q2 2009Q3 2009Q4 2010Q1 2010Q2 2010Q3 2010Q4 2011Q1 2011Q2 2011Q3 2011Q4 2012Q1 2012Q2 2012Q3 2012Q4 2013Q1 2013Q2 2006Q1 2006Q2 2006Q3 2006Q4 2007Q1 2007Q2 2007Q3 2007Q4 2008Q1 2008Q2 2008Q3 2008Q4 2009Q1 2009Q2 2009Q3 2009Q4 2010Q1 2010Q2 2010Q3 2010Q4 2011Q1 2011Q2 2011Q3 2011Q4 2012Q1 2012Q2 2012Q3 2012Q4 2013Q1 2013Q2 4. Shadow Banking System 45 Figure 4.13: U.S. MMF market, total assets USD billion 4,000 3,500 3,000 2,500 2,000 1,500 1, U.S. MMF market Source: Author based on FED market, which is opposite to what we have shown in the assessment of European shadow banking. Credit Transformation Market Figure 4.14: U.S. ABS market, outstanding balances USD billion 14,000 12,000 10,000 8,000 6,000 4,000 2,000 0 U.S. market of credit transformation ABS Pool securities GSE Total Source: Author based on FED Due to the variety of securities in the U.S. market, we expand our research to

61 2006 Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q2 4. Shadow Banking System 46 pool securities and GSEs. Such securities cannot be found in Europe. However, the importance of pool securities and GSEs in the U.S. is proven in Figure Credit transformation market is a significant part of the U.S. shadow banking. The volume of ABS is declining slowly over time. On the other hand, pool securities decreased sharply at the beginning of 2010 and now they tend to grow slowly. Only the GSEs activities grew rapidly in 2010 and are quite stable. The delay in the response to the recent financial crisis is significant and the overall decline in credit transformation activities is only slight. The volume of the U.S. market of credit transformation needs to be reduced and regulated due to its total volume. The last observed value amounts to USD 9.4 trillion. Hedge Fund Market Figure 4.15: U.S. hedge fund market, total assets USD billion 1,800 1,600 1,400 1,200 1, U.S. hedge fund market Source: Author based on HedgeFund Intelligence In order to study hedge fund market, we use the same database as in case of Europe, i.e. Hedgefund Intelligence. Total assets under management of hedge funds are shown in Figure Hedge fund market was a growing market until 2008 Q2 when it achieved USD 1.68 trillion. Then, it decreased sharply to USD 1.14 trillion. Therefore, we can definitely say that recent financial crisis had also an impact on the hedge fund activities. From 2009, hedge funds are growing slowly and we can expect them to recover their pre-crisis volume very soon. This implies that regulation of hedge funds is still not very effective. Let

62 2006Q1 2006Q2 2006Q3 2006Q4 2007Q1 2007Q2 2007Q3 2007Q4 2008Q1 2008Q2 2008Q3 2008Q4 2009Q1 2009Q2 2009Q3 2009Q4 2010Q1 2010Q2 2010Q3 2010Q4 2011Q1 2011Q2 2011Q3 2011Q4 2012Q1 2012Q2 2012Q3 2012Q4 2013Q1 2013Q2 4. Shadow Banking System 47 us recall the same finding in case of European hedge fund market where the regulation was not effective as well Estimation of the U.S. Shadow Banking Sector Table 4.2: Impact of the crisis, the U.S. Market commercial paper repo money market fund asset-backed securities pool securities government-sponsored enterprise hedge fund Impact of the crisis Source: Author Similarly as in case of European shadow banking, let us study the impact of the crisis on partial shadow banking markets. Again, we compare the situation on the market before the crisis in 2008 Q2 and after more than a year of crisis in 2010 Q1. Figure 4.16: U.S. shadow banking sector vs. total bank assets USD billion 25,000 U.S. shadow banking sector 20,000 15,000 10,000 5,000 0 CP Repo MMF ABS Pool securities GSE Hedge funds Bank assets Source: Author based on FED and Hedgefund Intelligence.

63 2006Q1 2006Q2 2006Q3 2006Q4 2007Q1 2007Q2 2007Q3 2007Q4 2008Q1 2008Q2 2008Q3 2008Q4 2009Q1 2009Q2 2009Q3 2009Q4 2010Q1 2010Q2 2010Q3 2010Q4 2011Q1 2011Q2 2011Q3 2011Q4 2012Q1 2012Q2 2012Q3 2012Q4 2013Q1 2013Q2 4. Shadow Banking System 48 In Table 4.2, the differences are presented using arrows following the same logic as in case of Europe. In cases of CP market, MMF market, pool securities and hedge funds, the negative impact was harsh. In other cases, the impact was represented by a slight decrease and it was the case of repo market and ABS market. Only GSEs activities grew despite the crisis. We suppose that GSEs activities represented a preferred way of funding, hence its use grew significantly. After detailed partial studies of shadow banking in the U.S., let us put all markets together and obtain our estimate of the U.S. shadow banking system. This estimate is presented in Figure Only because of GSE securities, the whole U.S. shadow banking sector stayed almost as large as it was before This proves our supposition that GSEs were used as an alternative way of funding when the crisis came, while other shadow ways of financing lost their attractiveness. Shadow banking reached its maximum in 2008 Q1 when its volume equalled to USD trillion. Afterwards, it was diminishing and nowadays it reaches about USD 15.5 trillion which means that shadow banking sector in the U.S. shrunk approximately by USD 5 trillion. Let us note that U.S. shadow banking is decreasing very slowly over time. Figure 4.17: U.S. shadow banking as a share of total bank assets 180% 160% 140% 120% 100% 80% 60% 40% 20% 0% Share of shadow banking Source: Author s computations What should be emphasised is the whole size of shadow banking compared to total traditional banking, represented by total bank assets in Figure In Europe, we discovered that shadow banking represents from 19% to 25% of

64 2006 Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q2 4. Shadow Banking System 49 total bank assets. In the U.S., shadow banking ranges from 100% to 163%. Interestingly, during the crisis, this share achieved its maximum and then started to decrease. As shown in Figure 4.17, nowadays, the U.S. shadow banking sector is of the same size as the traditional banking sector, represented by total bank assets. Let us recall that in Europe it is around 20% Comparison of the European and the U.S. Shadow Banking Sectors Figure 4.18: Comparison of the European and the U.S. shadow banking sectors EUR billion 12,000 10,000 8,000 6,000 4,000 2,000 0 Comparison of the shadow banking sectors USD billion 24,000 20,000 16,000 12,000 8,000 4,000 0 Europe (lhs) U.S. (rhs) Source: Author based on ECB, ICMA, AFME, Bank of England, Hedgefund Intelligence and FED We managed to compute the European and the U.S. shadow banking sectors. We have also shown the size of shadow banking compared to traditional banking using total bank assets. U.S. shadow banking is not only of a bigger size but obviously of a bigger importance. In Figure 4.18, we depict our overall estimation of the shadow banking sectors both in Europe and in the U.S. over the time period from 2006 until nowadays. Figure 4.18 can help us in order to conclude our qualitative analysis and test our hypothesis. Let us recall the first hypothesis.

65 4. Shadow Banking System 50 Hypothesis 1: European shadow banking market increased in volume during the financial crisis while the situation in the U.S. went other direction, i.e. during the crisis, the amount of shadow banking activities declines in volume. Based on the calculations we prepared, we may now test the hypothesis. Let us use Figure 4.18 to support our findings. On one hand, the European shadow banking sector decreased to its minimum right after the onset of recent financial crisis. On the other hand, from that time on until 2010 Q2, shadow banking in Europe was increasing. Moreover, it overcame its pre-crisis volume. Decrease in U.S. shadow banking was not that quick but happened in longer run. The whole sector reduced and reached its amount from 2006 Q1. Let us have a look at the situation during the crisis. European shadow banking kept increasing in volume despite the presence of the crisis. The U.S. shadow banking sector decreased. Based on our own measurement of shadow banking sectors in both economies, we cannot reject the hypothesis. 4.2 Money Market Funds It is not only our assessment of the shadow banking sector in Europe and in the U.S. which proves that MMFs are an important part of shadow banking system but also the paper published by Chernenko & Sunderam (2012). Moreover, the paper states that MMFs were one of the channels that spread distress among companies during recent sovereign debt crisis in Europe. Therefore, we look at the Chinese market because the importance of China in the global financial market has been increasing recently Chinese Money Market Funds Chinese shadow banking is nowadays one of the concerns for increasing number of authors. Shen (2013) explains that Chinese shadow banking origins from the rapid growth of underground financing and unregulated off-balance sheet lending. Hsu et al. (2013) compare European and Chinese shadow banking sectors. Their primary goal is to compare its risks rather that to capture its volume. They claim that Chinese shadow banking system is smaller, less developed and less stressed than European. Moreover, Adrian et al. (2013a)) claim that Chinese shadow banking system was growing significantly, since the onset of recent financial crisis in Accordingly, we would expect the MMF

66 4. Shadow Banking System 51 market to grow as well. Furthermore, Gosh et al. (2012) estimate a significant growth in shadow banking with the volume around USD 1.59 trillion in Figure 4.19: Chinese MMF, total assets Source: FitchRatings (2013) Unfortunately, we experience a problem with the availability of data. There are several private providers of total assets under management of Chinese MMFs but we were unable to get an access. Therefore, we use a graph published in the Special Report by FitchRatings (2013) and we perform a qualitative analysis. This graph is presented in Figure It covers the period from 2007 Q3 until 2012 Q4. During first three quarters in 2008, the market was decreasing and fell under CNY 2 trillion (EUR 211 billion) 2 attaining its minimal value. Afterwards, the market was getting bigger and by the end of 2009, it achieved around CNY 2.9 trillion (EUR 295 billion). 3 During 2010 and 2011, it was quite stable. From 2012, the market is increasing and probably getting back to its pre-crisis volume. Hypothesis 2: We expect a significant rise in total assets of MMFs in the Chinese market despite the recent financial crisis. As it is shown in Figure 4.19 published by FitchRatings (2013), there has been a negative tendency in the amount of Chinese MMF assets during 2008 but from 2008 Q3, the total amount was increasing. Based on this study, we cannot reject the hypothesis. 2 Author applied the conversion using ECB exchange rate at the end of 2008: EUR 1 = CNY Author applied the conversion using ECB exchange rate at the end of 2009: EUR 1 = CNY 9.835

67 2004 M M M M M M M M M M M M M M M M M M M M M M M M M M03 4. Shadow Banking System Repo Market Our calculations of the shadow banking sector in Europe showed that repo transactions represent its biggest share. Unfortunately, the data on repo activities is not available for every country. After many attempts to find repo data in particular European countries, we present a of study French and UK repo markets where the data is available French Repo Market Figure 4.20: French repo market EUR billions French repo market % repo volume (lhs) repo rate (rhs) Source: Author based on Agence France Trésor (AFT) and ECB Since the repo transactions became very important in terms of shadow banking activities, we study French market as we believe that this repo market is one of the biggest markets in euro area. This is also proved by the ICMA (2013) survey. To run this analysis, we use the data from the AFT. The AFT publishes the data of outstanding amounts of repo transactions on the monthly basis (from 2004 M11 until 2013 M4). In Figure 4.20, we depict the volumes of repo transactions as well as repo rate. Directly, we can spot a significant decrease in repo activities during 2007 and 2008 from its pre-crisis maximum of EUR billion. This decrease meant a cut in the volume by more than 40%. From the moment of the beginning of recent financial crisis in the U.S., the market started to grow again. Already in 2011, it overreached its pre-crisis

68 2000 Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q3 4. Shadow Banking System 53 volume when it attained EUR billion. Repo activities became an important source of funding for many financial institutions. We believe that from 2012, the regulation on repo transactions came into effect and caused the decrease visible in Figure However, from 2013, the market is growing again. Hence, we may speculate on the effectiveness of regulation of the repo market in France. Moreover, we plot the repo rate given by the ECB in Figure We can see a significant decrease in the repo rate from 4.25% in 2008 M9 to 1.0% in 2009 M5. We believe that this decrease was introduced in order to stabilise the market and prevent it from decreasing even further. Hypothesis 3: When talking about French shadow market, we think that boom in volume of repo transactions was significant in spite of the recent financial crisis that started in the U.S. From Figure 4.20, we may conclude that repo market started to grow despite the presence of the crisis. We should note that repo rate was changed with the onset of the crisis which definitely had an impact on the volume of repo transactions. Nevertheless, based on the qualitative analysis, we cannot reject our hypothesis UK Repo Market Figure 4.21: UK repo market GBP billion 1,200 1, UK repo market % repo volume (lhs) repo rate (rhs) Source: Author based on Bank of England

69 4. Shadow Banking System 54 Since the UK market is one of the biggest in terms of repo transactions in Europe, we decided to study this market. Also, ICMA (2013) confirms its important position on the European repo market. The data is downloaded from the Bank of England. In our analysis, we use this data as a proxy of repo market in the UK. We depict quarterly data from 2000 Q1 until 2013 Q3 in GBP billions in Figure From this graph, we can see an evolution of repo market in the UK. The market was growing gradually until 2009 Q1 when it achieved its maximum of GBP 1,125 billion. We cannot spot any significant decrease as French market experienced but a decrease was present. The volume of repo transactions decreased during 2010 but the market is recovering from that time on. Since the recent financial crisis did not have a significant negative impact on the UK repo market, we may look for a reason why. Therefore, in Figure 4.21, we also show the average repo rate calculated as the average of all repo rates in the UK. 4 We believe that post-crisis increase in the volume of repo transactions is caused by a huge decrease in repo rates from 5.9% in 2007 Q3 to 0.5% in 2009 Q3. We may say that it was a delayed response to the recent financial crisis. However, this decrease is no more present because the market is growing again. Hypothesis 4: We believe that similarly as in the French repo market, the volume of repo transactions in the UK increased after the onset of recent financial crisis in the U.S. Similarly, we use the volume of the repo transactions shown in Figure 4.21 in order to test our hypothesis. We have shown that situation on the UK market was completely different than on the French market. We proved that market in the UK was growing in spite of the recent financial crisis. Moreover, the change in the repo rate helped the market to grow even further. Using this qualitative analysis, we cannot reject our hypothesis. 4.4 Empirical Study of Shadow Banking Empirical research in this area is scarce because of the data availability. Since we have performed an assessment of the shadow banking system and we studied also the developments on the repo markets in France and in the UK, we perform 4 UK repo rates involved in the calculation: overnight, 1 week, 1 month, 3 months and 6 months.

70 4. Shadow Banking System 55 deeper analysis of already mentioned repo markets. Therefore, in this part, we present the empirical study that we have developed on our own. We start with the definition of our model and we describe the variables included in this empirical study. Then, we present our expectations and hypotheses. Finally, we perform the estimation and present obtained results The Model To set a model in this area of economy is rather difficult task. There has been very little effort in developing any econometric model including shadow banking activities in a particular country. Nonetheless, we present two shadow banking models that would explain repo transactions and total bank assets. We believe we do not need to mention that availability of the data regarding country s repo transactions or assets under MMF management is quite a challenge. Therefore, our research will be done on two chosen countries, i.e. France and the UK. We developed a number of models in order to study different relationships among shadow banking, traditional banking and the economy itself. We used our data but there have been many attempts with no significant results. Fortunately, we accomplished to achieve two models with interesting results. Let us define them: total assets t = α + β 1 repo t + β 2 gdp t + β 3 debt t + u t (4.1) repo t = γ + δ 1 mmf t + δ 2 total assets t + δ 3 gdp t + v t (4.2) where t=2004 M11,..., 2013 M4 in case of France and t=2000 Q1,...,2013 Q3 in case of UK. All variables mentioned in Equation 4.1 and Equation 4.2 are described in more details in the next section Variables The choice of variables and the model itself is based on our interest in the relationships among shadow banking, traditional banking and economy itself. Therefore, we came up with variables that represent each sector. Let us emphasise that no such model has been developed and presented in the literature yet. We chose two representative countries, i.e. France and United Kingdom. The studied period and the data frequency differ for chosen countries. This

71 4. Shadow Banking System 56 is caused by different availability of the data. As for French market, we use monthly data from 2004 M11 until 2013 M4. As for UK market, we use quarterly data from 2000 Q1 until 2013 Q3. For the estimation, we apply Ordinary Least Squares (OLS) method and Heteroscedasticity-corrected estimation method. Our choice of countries is not random. We intended to choose a country that is a euro area member and a that is a non-euro area member. First, an important factor of this choice is a share of a country in overall European shadow banking sector. Second, our choice is based on the data availability. Shadow Banking Variables Figure 4.7 capturing the European shadow banking sector helped us with deciding upon the choice of shadow banking variables. In the European shadow banking sector, the repo market plays the most important role. Therefore, we chose the volume of repo transactions as the first variable for our analysis. Moreover, the repo market is a main concern of the study performed by Mancini et al. (2013). This paper focuses on the European repo market, mainly its evolution, activities and risks. In one of their models, they estimate the volume of repo activities by other variables; among others including for example lagged repo volume, average term, composite indicator of systemic stress or haircut. Volume of the repo transactions (repo) represents the size of repo market in the country. We use the data that we have already shown in Figure 4.20 and Figure The choice of repo transactions is given by its share in the whole shadow banking. The second largest part of shadow banking is the ABS market but the data per country is not available. Therefore, we chose the MMF market with only slightly smaller share than ABS market in the overall estimated European shadow banking sector. The study performed by Chernenko & Sunderam (2012) focuses on the MMF market and its link to the instabilities caused by market turmoil. As a variable characterising MMF market, we picked the assets under management of MMFs also thanks to its availability. Assets under management of MMF (mmf) capture the size of the MMF market in the country. We also included refinancing rate (repo rate) that is a policy rate given by the central bank of a country. In case of France, the repo rate is given by the ECB, since the repo rate is valid for each euro area member. However, due to the collinearity problem, we had to drop this variable. In order to study further the shadow banking system, we would extend the

72 2004 M M M M M M M M M M M M M M M M M M M M M M M M M M03 4. Shadow Banking System 57 number of shadow banking variables. Inclusion of outstanding balances of ABS, hedge fund assets, net securities loaned or number of shadow banking institutions in the system would definitely be very interesting to study. Moreover, the extension to volume of repo activities, one might add the data regarding used collateral. However, the data per country is unavailable or unknown. Also, with the current number of observations, we would end up with too many variables and not enough observations in order to perform such analysis. Traditional Banking Variables To characterise the traditional banking system, we chose the volume of total bank assets. We already used this variable when we plotted the whole calculated shadow banking sector in Europe and in the U.S., shown in Figure 4.7 and Figure 4.16 respectively. In this analysis, we use the data per country. Total bank assets (total assets) indicate the size of the traditional banking system in a country. This variable is used in several studies when calculating or estimating the whole shadow banking system. It is a good proxy for traditional banking when compared to shadow banking. For example, this approach was used by Bouveret (2011), Gorton & Metrick (2010) or Pozsar et al. (2010). Figure 4.22: Total bank assets, France EUR billion 9,000 8,000 7,000 6,000 5,000 4,000 3,000 2,000 1,000 0 Total bank assets in France Source: Author based on ECB In order to see the time evolution of total bank assets in France, we plot the data in Figure Similarly, we plot total bank assets in the UK in

73 2000 Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q3 4. Shadow Banking System 58 Figure 4.23: Total bank assets, UK GBP billion 10,000 9,000 8,000 7,000 6,000 5,000 4,000 3,000 2,000 1,000 0 Total bank assets in UK Source: Author based on ECB Figure We use the ECB data on MFI. Total bank assets are calculated in the same way as we calculated total European bank assets. 5 According to the data depicted in Figure 4.22, we can see an increasing trend in total bank assets in France. There are no big ups or downs. The last observed value is EUR 8.3 trillion in 2013 M4. It represents 19% share of European total bank assets. As shown in Figure 4.23, there was an increasing trend before the recent financial crisis. Then, we can spot one obvious increase from 2008 Q3 to 2008 Q4 followed by almost equal decrease from 2008 Q4 to 2009 Q2. Afterwards, total bank assets are quite stable. The last observation equals to GBP 926 billion and it represents 21% share of European total bank assets. Based on this information and similar information from the previous paragraph, we may conclude that by choosing these two countries, we cover approximately 40% of European traditional banking sector on average. Then, in order to encompass traditional banking sector, we decided to include a variable representing the number of MFI in the country. For better illustration, we present this variable in Figure 4.24 for France and in Figure 4.25 for the UK. Figure 4.24 shows that at the beginning of the studied period, there were 5 Total bank assets equal to MFI assets minus euro system assets and minus MMF shares/units.

74 2000 Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q M M M M M M M M M M M M M M M M M M M M M M M M M M03 4. Shadow Banking System 59 Figure 4.24: Number of MFI, France 1,800 1,600 1,400 1,200 1, Number of MFI in France Source: Author based on ECB Figure 4.25: Number of MFI, UK Number of MFI in UK Source: Author based on ECB 1,651 MFI in France. This number was decreasing over time. At the end of the studied period, it reached its minimum of 1,037 MFI. According to Figure 4.25, we can spot a very similar trend in the UK market. The number of MFI is decreasing over time. The maximum number of MFI in the UK was 558 in 2001 Q3. The minimum number of MFI is also at the end of studied period and it is 391 MFI. Interestingly, this number is around one

75 4. Shadow Banking System 60 third of the number of MFI in France. Unfortunately, due to the presence of collinearity, we had to drop this variable from the estimation. There are also some other variables that could be included in the analysis that we decided to perform. ECB (2013) suggests the variables that may represent the traditional banking sector. In order to capture the improvement or aggravation of the banks profitability, the return on equity should be used. ECB further proposes to include the non-performing loans when studying the quality of the assets. We also planned to include the measure of stress in the banking sector. However, these variables are very specific. As our analysis aims, we capture the size of the traditional banking system. On one hand, let us emphasise that because of the limited number of observations, we could not include more variables into our models. On the other hand, studying the relationship between banks profitability or the quality of assets and shadow banking sector would be an interesting idea for further research. Economy Variables Finally, in order to represent the economy itself, there is a vast choice of variables. Let us begin with variables that we include in our analysis. Then, we will mention other possible variables that could represent the economy. The first variable representing the economy itself is gross domestic product (gdp). This variable is used as a proxy of the size of the economy. We use the IMF data for both countries. Then, we add the government debt to GDP which shows how much the country is indebted as a proportion to its GDP. Again, the IMF statistics are used. There are many other macroeconomic variables that might be included in such research. Let us mention inflation, unemployment or consumption. ECB (2013) further proposes equity prices, residential house prices or other different financial indicators. Our target is to study three sectors using the variables that capture their sizes. Also, if we include too many variables, the number of observations is not sufficient for a meaningful analysis. After having introduced all variables involved in the empirical part, let us have a look at the descriptive statistics of the data presented in Table 4.3. Our final dataset is composed of five variables. Four of them, i.e. repo, mmf, total assets and gdp, are represented in national currency billions and debt is described as a percentage. We also tried to estimate our model with

76 4. Shadow Banking System 61 Table 4.3: Descriptive statistics Descriptive statistics of French data repo mmf total assets gdp debt Mean , , Median , , Minimum , , Maximum , , Std. Dev , Skewness Ex. Kurt Descriptive statistics of UK data repo mmf total assets gdp debt Mean , , , Median , , Minimum , , Maximum 1, , , , Std. Dev , Skewness Ex. Kurt Source: Author s computations Note: Std. Dev. - Standard Deviation, Ex. Kurt. - Excess Kurtosis debt variable expressed in national currency billions but the results were not significant. Moreover, we could not even apply the OLS method because of the violation of the assumptions Estimation Expectations and the Hypotheses Table 4.4: Sign expectations, model 1 Variable Sign repo + gdp + debt + Source: Author Model 1, given by the Equation 4.1, explains total bank assets. This model

77 4. Shadow Banking System 62 enables us to study two relationships, i.e. relationship between traditional banking and shadow banking and relationship between traditional banking and economy. We expect positive impact of repo transactions on total bank assets and positive impact of GDP and government debt. Our expectations are based on our assumption that repo activities are aimed to increase the amount of banks assets. Furthermore, we believe that the higher the GDP, the bigger the production capacity, thus the higher the incentives for trade which is boosting total bank assets. The target of this model is to test whether the behaviour in chosen countries is the same or different. Based on the similarities of studied countries, their economic situation and their geographical location, we expect their behaviour to be the same. Therefore, our expectations about the coefficient signs are the same for both countries. Subsequently, we formulate our hypothesis as follows. Hypothesis 5: Total bank assets increase with rising repo transactions, increasing GDP and bigger government debt. In Table 4.4, we present the summary of our sign expectations of coefficients of the first model given by Equation 4.1 with total assets as explained variable. We believe that chosen countries behave in the similar way. signs are expected to be the same. Table 4.5: Sign expectations, model 2 Variable Sign mmf - total assets + gdp + Therefore, the Source: Author Our aim in model 2, given by Equation 4.2, is to explain repo transactions. Firstly, we explain repo transactions by other part of shadow banking, i.e. by MMFs. We expect negative relationship, since we believe that shadow banking activities function as substitutes. Secondly, we explain repo transactions by traditional banking sector represented by total bank assets. In this case, we expect positive influence of total assets on repo transactions. Similarly as for the previous model, we expect that relationship of total bank assets and repo

78 4. Shadow Banking System 63 transactions is mutual, i.e. final positive impact in this model as well. Finally, we add economy variables into the model explaining repo transactions, since we believe there is an impact of GDP on total repo transactions in each country. We expect this impact to be positive. As in the previous model, we expect the same behaviour in both countries. Therefore, expected coefficient signs for France and the UK are the same as well. Subsequently, we define our hypothesis. Hypothesis 6: Repo transactions increase with decreasing assets under MMF management, rising total bank assets and increasing GDP. Similarly as for model 1, Table 4.5 summarises our expected signs of coefficients of the model given by Equation 4.2 where repo is explained variable Results We eliminated plenty of models that either did not fulfilled OLS assumptions or did not give us significant results. In each model, we ran three main tests in order to be sure, whether the OLS method of estimation can be applied, i.e. test for normality of residual, White s test for heteroscedasticity and test for collinearity. As we have already mentioned, because of the collinearity problem, we had to drop two variables. In cases where the non-presence of heteroscedasticity is rejected, we used OLS with Robust Standard Errors and Hetoroscedasticity-corrected method in Gretl. The first method did not correct the heteroscedasticity problem. Therefore, we use Heteroscedasticity-corrected method which eliminates heteroscedasticity problem completely. Model 1 Let us recall Equation 4.1 that represents model 1. Regarding the OLS assumptions, we may present following conclusion. First, at 1% level of significance, we cannot reject the null hypothesis that heteroscedasticity is not present in both cases. Second, at 1% level of significance, we cannot reject the null hypothesis that error is normally distributed and this holds for both countries. Third, according to the test for collinearity, there is no indication of such problem, also for both countries. Based on these tests, we may use the OLS method as the estimation method when explaining total bank assets. In Table 4.6, we present full results for France, including above

79 4. Shadow Banking System 64 mentioned tests. From now on, in the text, we will present only brief results and full results can be found in Appendix A. Table 4.6: Model 1, OLS estimation for France, full results Variable Coefficient Standard Error t-ratio p-value const 29, , repo gdp debt Mean dependent var 6, S.D. dependent var 1, Sum squared resid 15,489,461 S.E. of regression R Adjusted R F (3, 98) P-value(F ) 2.87e 46 Log-likelihood Akaike criterion Schwarz criterion 1, Hannan Quinn 1, White s test for heteroskedasticity Null hypothesis: heteroskedasticity not present Test statistic: LM = with p-value = P (χ 2 (9) > ) = Test for normality of residual Null hypothesis: error is normally distributed Test statistic: χ 2 (2) = with p-value = Variance Inflation Factors Minimum possible value = 1.0 Values > 10.0 may indicate a collinearity problem repo gdp debt VIF(j) = 1/(1 - R(j) 2 ), where R(j) is the multiple correlation coefficient between variable j and the other independent variables Properties of matrix X X: 1-norm = e+008 Determinant = e+016 Reciprocal condition number = e-011 Source: Author s computations

80 4. Shadow Banking System 65 Based on the coefficients we obtained using OLS estimation, we may draw following conclusions for France. The higher the GDP, the higher the total bank assets and the same dependence holds for the government debt, i.e. the higher the debt, the higher the total bank assets. On the other side, the results show that the higher the repo transactions, the lower the total bank assets. This finding disproves our expectations. We expected the rise of total bank assets with rising repo activities. The results show that situation on the French market is opposite regarding the relationship between shadow and traditional banking to what we expected. When looking at the results of the relationship between traditional banking and economy itself, we can say that there is a positive impact of both gdp and debt variables on the amount of total bank assets. This goes along with our expectations. Based on all French results from the OLS estimation, we reject our hypothesis. Table 4.7: Model 1, OLS estimation for the UK Variable Coefficient const 7, ( ) repo (8.6268) gdp (6.0550) debt (8.1215) Source: Author s computations Note: T-ratios(t-ratio=estimated coefficient/standard error) are presented in the parentheses below the computed coefficients. In Table 4.7, we present brief results for the UK, full results can be found in Appendix A in Table A.1. Regarding the behaviour on the UK market, we obtained following results. We proved that there is a positive impact of repo transactions on total bank assets. As we expected, the more repo activities are performed, the higher bank assets there are. Similarly, economy variables, here represented by GDP and government debt, have a positive impact on the total bank assets. This proves our expectations. Therefore, we cannot reject our hypothesis regarding the situation in the UK market. Let us present the summary of coefficient signs, both expected and es-

81 4. Shadow Banking System 66 Variable Table 4.8: Coefficient summary for model 1 Expectation Coefficient sign Estimation France Estimation UK repo gdp debt Source: Author timated for chosen countries in order to recapitulate our findings regarding model 1. This summary is shown in Table 4.8. First, let us recall our hypothesis. Hypothesis 5: Total bank assets increase with rising repo transactions, increasing GDP and bigger government debt. Using the OLS estimation method, we proved that there is a positive impact of GDP and government debt on total bank assets in studied countries. As for the UK, we obtained the same results as we expected regarding the positive impact of repo transactions on total bank assets. However, we are not able to conclude the same behaviour on the French market. Therefore, the results are not conclusive. Based on the outcome of the model we tested, we decided to study the French dataset more deeply. Since our analysis of repo market in France and in the UK showed different behaviour pattern in the trend before the recent financial crisis, we decided to split the dataset. Thus, we study French market from the moment when repo market hit its bottom in 2008 M11 (EUR billion). Studying the period during the crisis allows us to understand the behaviour of French market from different perspective. We dropped the observations before the crisis and we focus only on the period when the crisis was present and the period afterwards. New subset created from original dataset shrunk to 59 observations. This number of observations is still sufficient for an empirical analysis. The method used here is Heteroscedasticity-corrected estimation due to the violation of OLS assumption on heteroscedasticity, i.e. non-presence of heteroscedasticity was rejected. Let us have a look at brief re-

82 4. Shadow Banking System 67 Table 4.9: Model 1, Heteroscedasticity-corrected estimation for France - subset Variable Coefficient const 5, ( ) repo (2.9792) gdp (3.8521) debt (2.5155) Source: Author s computations Note: T-ratios(t-ratio=estimated coefficient/standard error) are presented in the parentheses below the computed coefficients. sults presented in Table 4.9, full results including the tests of OLS assumptions can be found in Appendix A in Table A.2. Thanks to the analysis of French subset, when we dropped the observations before the crisis, we obtain slightly different output from the regression. The biggest change happened in the impact of repo transactions on total bank assets. This impact happens to be positive. For the rest of variables, the impact stayed positive. Therefore, we may conclude that dropping the observations from the precrisis period changed our results significantly. Based on this sub-analysis, we cannot reject fifth hypothesis. Model 2 Let us recall Equation 4.2 that defines model 2. First, we perform the OLS estimation. However, after checking OLS assumptions, we had to reject the null hypothesis that heteroscedasticity is not present. Therefore, we decided to use Heteroscedasticity-corrected estimation method in order to test our hypothesis. First, thanks to this method, the heteroscedasticity problem was solved. Second, for both countries, at 5% level of significance, we cannot reject the null hypothesis that error is normally distributed. Third, we performed the test for collinearity and there is no indication of collinearity problem, also for both countries. Based on these tests, we may use the Heteroscedasticity-corrected estimation method as the estimation method in

83 4. Shadow Banking System 68 Table 4.10: Model 2, Heteroscedasticity-corrected estimation for France Variable Coefficient const 1, ( 4.814) mmf ( 9.287) total assets ( 4.902) gdp (7.075) Source: Author s computations Note: T-ratios(t-ratio=estimated coefficient/standard error) are presented in the parentheses below the computed coefficients. order to explain our variable capturing repo transactions. In Table 4.10, we present brief results of this estimation, full results can be found in Appendix A in Table A.3. Table 4.11: Model 2, Heteroscedasticity-corrected estimation for the UK Variable Coefficient const ( ) mmf ( ) total assets (8.0685) gdp (2.6058) Source: Author s computations Note: T-ratios(t-ratio=estimated coefficient/standard error) are presented in the parentheses below the computed coefficients. Using heretoscedasticity-corrected estimation method, we can conclude following dependencies. Based on the coefficient estimated for the mmf variable, we may say that the higher the assets under MMF management, the lower the repo transactions. This goes along with our expectation that repo activities and

84 4. Shadow Banking System 69 MMF function as substitutes. Here, we proved expected negative relationship. When talking about the relationship between repo transactions and total bank assets, the results show negative relationship. The total bank assets are lowering repo transactions. On one hand, this relationship proves the dependency that we found out previously in model 1 for the whole French dataset. Hence, the relation is mutual. On the other hand, it again contradicts our expectations. Then, according to estimated coefficient of gdp variable, we may say that higher GDP leads to an increase in total bank assets. This relationship proves our expectations. Based on these partial conclusions, we have to reject our hypothesis due to the relationship of total bank assets and repo transactions. Table 4.11 presents brief results of the estimation of model 2 using the UK data, full results can be found in Appendix A in Table A.4. Thanks to the Heteroscedasticity-corrected estimation method, we could perform the analysis in order to test our hypothesis. The UK data draws following conclusions. There is a negative impact of MMF on the repo transactions. This result corresponds to our expectations. Then, as we also expected, the estimation proves that if total bank assets increase, the repo transactions increase as well. Finally, the results show that rising GDP also raises repo transactions. Therefore, our expectation of this relationship was proved. Based on the UK results obtained by the estimation method, we cannot reject our hypothesis. Variable Table 4.12: Coefficient summary for model 2 Expectation Coefficient sign Estimation France Estimation UK mmf total assets gdp Source: Author Let us put the partial results for model 2 from above together and summarise them. Table 4.12 shows the coefficient signs we expected and we estimated for both countries. Let us recall the hypothesis.

85 4. Shadow Banking System 70 Hypothesis 6: Repo transactions increase with decreasing assets under MMF management, rising total bank assets and increasing GDP. Using the Heteroscedasticity-corrected estimation method, we proved that rising assets under MMF management have negative impact on total repo transactions. Also, as we expected the estimation proves that the higher the GDP, the higher the repo transactions. However, we obtained non-conclusive results regarding the impact of total bank assets on repo transactions. We proved our expectations of the behaviour on the UK market but the behaviour on the French market is not as we expected. Therefore, the whole results regarding the explanation of the repo transaction are unfortunately non-conclusive. We tried to perform similar sub-analysis of the French market in order to improve our results. We used the same subset of data as for model 1, i.e. the studied period began in 2008 M11. However, the results did not turn out to be improving or significantly different. 4.5 Further Research Opportunities Since the topic of shadow banking is still quite new, there is a lot of room for further research. Based on the qualitative analysis we accomplished,when we assessed the size of shadow banking, we suggest broadening the definition of shadow banking and adding securities lending if the data is available and reliable. Also, it is worth continuing in the assessment of shadow banking system in order to study the impact of regulation aimed at the downsizing of the whole system. Our quantitative analysis showed that if we included more countries into dataset, we would obtain more information regarding examined relationships. However, in this case, the data availability limits the research. If the availability does not improve over time, we suggest repeat our framework later, e.g. in five years, in order to obtain more observations. Having much more observations in the dataset, it is possible to include other shadow banking, traditional banking and economy variables. We mentioned those variables in Subsection

86 Chapter 5 Conclusion In this thesis, we study the shadow banking system. We define the shadow banking system, its activities and entities that are involved. We point out one of the biggest problems regarding shadow banking, i.e. systemic risk and other types of risks. Also, the data availability regarding shadow banking activities is still an unsolved issue. These problems are faced and being tackled mainly by the Financial Stability Board. Therefore, we present shadow banking regulation and supervision measures introduced and implemented by Financial Stability Board. Let us summarise the findings of both qualitative and quantitative analysis that we performed in this thesis. First, our aim was to estimate the shadow banking sector in Europe and in the United States. A qualitative analysis, we performed, presents these findings. European shadow banking reached almost EUR 10 trillion before the onset of the recent financial crisis in 2008 Q2. Then, shadow banking in Europe shrunk to EUR 8.3 trillion, 19% of total European bank assets, but started to increase already from In spite of the financial crisis, the European shadow banking system was growing. The last observed value from 2013 Q2 indicates that European shadow banking system equals to EUR 9.3 trillion which is 21% of total European bank assets. We proved that development in the United States was quite different. Increasing trend in the shadow banking system was stopped in 2008 Q1 at USD 20.7 trillion, 163% of total U.S. bank assets, and from that time on, it was decreasing gradually. According to the last observation from 2013 Q2, the shadow banking system in the United States equals to USD 15.6 trillion which also equals to the total bank assets in the United States. Although these developed economies have a lot in common, we

87 5. Conclusion 72 showed that European and the U.S. shadow banking systems behaved differently during the crisis. Second, our interest was turned to the study of money market funds in China. The importance of China in the global shadow banking was proved by numerous studies. However, China seems to be standing aside and that is why we believed that financial crisis did not have a negative impact on the performance of money market funds. On one hand, our analysis captures the downturn that this market experienced before the crisis when it fell under CNY 2 trillion (EUR 211 billion). On the other hand, the presence of the crisis did not have a negative impact. Moreover, the money market funds increased their assets to approximately CNY 2.9 trillion (EUR 295 billion) already in 2009 Q4. Third, our study of European shadow banking showed that repo market is the most important part of shadow banking. Financing via repo transactions proved to be very attractive. This finding encouraged us in studying this sector in more details. Based on the share of countries in total European repo market, we picked France and the United Kingdom for our further research. This choice was also limited by the availability of data. During the crisis, French repo market started to increase (also given by the decrease in repo rate from 4.25% to 1.0%) and British market behaved in a very similar way. Interestingly, the behaviour before the crisis was different, i.e. decreasing repo activities in France and gradually increasing trend in the United Kingdom. Already in 2011, French repo market overreached its pre-crisis volume when it attained its maximum of EUR billion in 2011 M10. The market in the United Kingdom was growing gradually until 2009 Q1 when it achieved its maximum of GBP 1,125 billion. Fourth, a quantitative analysis was performed for the purposes of more profound research. It was aimed to study the relationships among traditional banking, shadow banking and economy itself. For this reason, we developed two models. Using Ordinary Least Squares method and Heteroscedasticitycorrected estimation method, we estimated these relationships. Our own dataset covers two European countries, i.e. France and the United Kingdom. The covered time period and data frequency differs, in case of France it is the period from 2004 M11 until 2013 M4 and in case of the United Kingdom it is the period from 2000 Q1 until 2013 Q3. This difference is given by data availability. Using the OLS estimation method for model 1, we proved that there is a positive impact of GDP and government debt on total bank assets in studied countries. As for the UK, we obtained the same results, as we expected, regarding

88 5. Conclusion 73 the positive impact of repo transactions on total bank assets. However, we were not able to conclude the same behaviour on the French market. Therefore, we performed an analysis of French subset, when we dropped the observations prior to 2008 M11. Thanks to this sub-analysis using Heteroscedasticity-corrected estimation method, we obtained the results proving that impact of repo transactions on total bank assets is positive and for the rest of variables, the impact on total bank assets stayed positive. Using the Heteroscedasticity-corrected estimation method, we proved that rising assets under MMF management have negative impact on total repo transactions in both studied countries. Also, as we expected, the estimation proves that the higher the GDP, the higher the repo transactions. Unfortunately, we obtained non-conclusive results regarding the impact of total bank assets on repo transactions. In case of the UK, the estimation proves that if total bank assets increase, the repo transactions increase as well. In case of France, such positive impact of total bank assets on repo trancastions was not proved.

89 Bibliography Acharya, V. V. & S. T. Öncü (2010): The Repurchase Agreement (Repo) Market. Regulating Wall Street: The Dodd-Frank Act and the New Architecture of Global Finance Chapter 11: pp Adrian, T. & A. B. Ashcraft (2012a): Shadow Banking: A Review of the Literature. Staff Report 559, Federal Reserve Bank of New York. Adrian, T. & A. B. Ashcraft (2012b): Shadow Banking Regulation. Staff Report 559, Federal Reserve Bank of New York. Adrian, T., A. B. Ashcraft, & N. Cetorelli (2013a): Shadow Bank Monitoring. Staff Report 580, Federal Reserve Bank of New York. Adrian, T., B. Begalle, A. Copeland, & A. Martin (2012): Repo and Securities Lending. Working Paper 18549, NBER Working Paper Series. Adrian, T., D. Covitz, & N. Liang (2013b): Financial Stability Monitoring. Finance and Economics Discussion Series , Federal Reserve Board, Division of Research & Statistics and Monetary Affairs, Washington, D.C. BCBS (2010a): Basel III: A global regulatory framework for more resilient banks and banking systems. Report, Basel Committee on Banking Supervision. ISBN: BCBS (2010b): Basel III: International framework for liquidity risk measurement, standards and monitoring. Report, Basel Committee on Banking Supervision. ISBN: BCBS (2010c): International regulatory framework for banks (Basel III). Basel Committee on Banking Supervision. Available at:

90 Bibliography 75 Bengtsson, E. (2013): Shadow banking and financial stability: European money market funds in the global financial crisis. Journal of International Money and Finance 32: pp BIS (2008): International banking and financial market developments. Quarterly review, Bank for International Settlements. ISSN X. Bouveret, A. (2011): An Assessment of the Shadow Banking Sector in Europe. Report, Available at: Bouveret, A., J. Jardelot, J. Keller, P. Molitor, J. Theal, & M. Vital (2013): Towards a monitoring framework for securities financing transactions. Occasional Paper Series 2, European Systemic Risk Board. Chernenko, S. & A. Sunderam (2012): The Quiet Run of 2011: Money Market Funds and the European Debt Crisis. Working Paper , Fisher College of Business. Claessens, S., Z. Pozsar, L. Ratnovski, & M. Singh (2012): Shadow Banking: Economics and Policy. Staff Discussion Note 12, International Monetary Fund. Copeland, A., A. Martin, & M. Walker (2011): Repo Runs: Evidence from the Tri-Party Repo Market. Staff Report 506, Federal Reserve Bank of New York. Revised March EC (2012): Shadow Banking. Green paper, European Commission. ECB (2012): Shadow Banking in the Euro Area. An Overview. Occasional Paper Series 133, European Central Bank. ISSN ECB (2013): Financial Stability Review. Review, European Central Bank. ISSN Eurofi (2011): Shadow banking - Improving the consistency of banking and non-banking regulations. In The Eurofi G20 High Level Seminar 2011, Paris. Eurofi (2012): Addressing shadow banking and stability issues in investment fund regulation. In The Eurofi Financial Forum 2012, Brussels. FitchRatings (2013): Asset Management in East Asia. Growth Shifting to Onshore Markets. Special report, Fitch Ratings.

91 Bibliography 76 FSB (2011a): Shadow Banking: Scoping the Issues. A Background Note of the Financial Stability Board. Report, Financial Stability Board. FSB (2011b): Shadow Banking: Strengthening Oversight and Regulation. Recommendations of the Financial Stability Board. Report, Financial Stability Board. FSB (2012): Securities Lendings and Repos: Market Overview and Financial Stability Issues. Report, Financial Stability Board. Interim Report of the FSB Workstream on Securities Lending and Repos. FSB (2013a): Global Shadow Banking Monitoring Report Financial Stability Board. Report, FSB (2013b): Strengthening Oversight and Regulation of Shadow Banking. An overview of Policy Recommendations. Report, Financial Stability Board. FSB (2013c): Strengthening Oversight and Regulation of Shadow Banking. Policy Framework for Addressing Shadow Banking Risks in Securities Lending and Repos. Report, Financial Stability Board. FSB (2013d): Strengthening Oversight and Regulation of Shadow Banking. Policy Framework for Strengthening Oversight and Regulation of Shadow Banking Entities. Report, Financial Stability Board. Gorton, G. & A. Metrick (2010): Regulating the Shadow Banking System. Brookings Papers on Economic Activity. Fall: pp Gorton, G. & A. Metrick (2012): Securitized banking and the run on repo. Journal of Financial Economics 104(3): pp Gosh, S., I. Gonzalez del Mazo, & I. Ötker Robe (2012): Chasing the Shadows: How Significant Is Shadow Banking in Emerging Markets? Number 88, Poverty Reduction and Economic Management Network (PREM). The World Bank. Hsu, S., J. Li, & Y. Qin (2013): Shadow Banking and Systemic Risk in Europe and China. CITYPERC Working Paper Series 02, City Political Economy Research Centre in London. ISSN ICMA (2012): Shadow banking and repo. Report, International Capital Market Association. ICMA European Repo Council.

92 Bibliography 77 ICMA (2013): European repo market survey. Capital Market Association. Number 25, International Kacperczyk, M. & P. Schnabl (2010): Does Organizational Form Affect Risk Taking? Evidence from Money Market Mutual Funds. Available at: Krishnamurthy, A., S. Nagel, & D. Orlov (2012): Sizing Up Repo. Working Paper 17768, NBER Working Paper Series. Mancini, L., A. Ranaldo, & J. Wrampelmeyer (2013): The Euro Interbank Repo Market. Working Paper on Finance 16, Swiss Institute of Banking and Finance, University of St. Gallen. Martin, A., D. Skeie, & E.-L. von Thadden (2010): Repo Runs. Staff Report 444, Federal Reserve Bank of New York. Revised May McCulley, P. A. (2007): Teton Reflections. PIMCO Global Central Bank Focus. Available at: Meeks, R., B. Nelson, & P. Alessandri (2012): Shadow Banks and Macroeconomic Instability. Report, Western Social Science. Pozsar, Z., T. Adrian, A. Ashcraft, & H. Boesky (2010): Banking. Staff Report 458, Federal Reserve Bank of New York. Shadow Pozsar, Z. & M. Singh (2011): The Nonbank-Bank Nexus and the Shadow Banking System. IMF Working Paper 289, International Monetary Fund. Radević, B. & A. Lekpek (2010): Credit Risk Transfer as a Mechanism of Protection against Risks. Series: Economics and Organization 7(4): pp Ricks, M. (2010): Shadow Banking and Financial Regulation. Paper 370, Columbia Law and Economics. Working Schwarcz, S. L. (2012): Regulating Shadow Banking. Review of Banking and Financial Law 31(1): pp Shen, W. (2013): Shadow Banking System in China Origin, Uniqueness and Governmental Responses. Journal of International Banking Law and Regulation 1: pp

93 Bibliography 78 Singh, M. & J. Aitken (2010): The (sizable) Role of Rehypothecation in the Shadow Banking System. IMF Working Paper 172, International Monetary Fund. The Squam Lake Group: Baily, M. N., J. Y. Campbell, J. H. Cochrane, D. W. Diamond, D. Duffie, K. R. French, A. K. Kashyap, F. S. Mishkin, D. S. Scharfstein, R. J. Shiller, M. J. Slaughter, H. S. Shin, J. C. Stein, & R. M. Stulz (2011): Reforming Money Market Funds. A Proposal by the Squam Lake Group. Working Paper 109, Rock Center for Corporate Governance at Stanford University. Tyson, J. & M. Shabani (2013): Sizing the European Shadow Banking System: A New Methodology. CITYPERC Working Paper Series 01, City Political Economy Research Centre in London. ISSN

94 Appendix A Appendix Figure A.1: Banks assets and liabilities to non-bank financial intermediaries Source: FSB (2013a) Note: OFI - other financial intermediaries.