An Alternative Approach to Liquidity Risk Management of Islamic Banks 9783110579994, 9783110582901, 9783110580150

Citation preview

Mohammed Habib Dolgun, Abbas Mirakhor An Alternative Approach to Liquidity Risk Management of Islamic Banks

De Gruyter Studies in Islamic Economics, Finance and Business

Edited by Abbas Mirakhor and Idris Samawi Hamid

Volume 7

Muhammed Habib Dolgun, Abbas Mirakhor

An Alternative Approach to Liquidity Risk Management of Islamic Banks

ISBN 978-3-11-057999-4 e-ISBN (PDF) 978-3-11-058290-1 e-ISBN (EPUB) 978-3-11-058015-0 ISSN 2567-2533 Library of Congress Control Number: 2020944388 Bibliographic information published by the Deutsche Nationalbibliothek The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data are available on the Internet at http://dnb.dnb.de. © 2021 Walter de Gruyter GmbH, Berlin/Boston Cover image: nnnnae/iStock/Getty Images Plus Typesetting: Integra Software Services Pvt. Ltd. Printing and binding: CPI books GmbH, Leck www.degruyter.com

Preface Liquidity risk can be defined as a shortcoming to cover financial liabilities, with its management related to guaranteeing expected and unexpected cash outflows. Liquidity is the lifeblood of banking as demonstrated by the 2008 global financial crisis. An examination of this important issue in an Islamic banking context is crucial to promoting efficiency, growth and resilience of the Islamic financial industry. While there are several studies on the performance, growth and efficiency of Islamic banks, empirical studies from the regulatory and supervisory perspectives are limited. First of all, this book seeks to fill this gap by examining the liquidity risk management of Islamic banks and discuss many regulatory issues. Secondly this book chooses to concentrate on quantitatively assessment of liquidity risk management of, officially known as, Participation Banks (PBs) in Turkey, which is an increasingly important G20 and emerging market economy. In the following chapters, the short-term and long-term determinants of Turkey’s PBs’ liquidity holdings are examined using monthly data over an eightyear period from 2007–2015, based on a liquidity model that incorporates Sukuk, interbank market rate, required reserves, inflation rate and credit default swap rate (CDS) and these factors are compared with factors affecting the liquidity of conventional banks. Long run analysis and Granger analysis show that the liquidity management of PBs has a causal relation with market liquidity (directly) and funding liquidity (indirectly). Furthermore, the liquidity management of PBs and conventional banks (CBs) of Turkey are assessed using panel data. In this model, it is found that there is a direct relationship between the liquidity of Islamic banks and their capital adequacy ratio, inflation rate, required reserves, total assets, and deposits. Moreover, it is evidenced that there is a negative significant relationship between liquidity and return on assets of PBs. Interbank rate and government bonds show an insignificant impact on the liquidity of PBs. If banks were forced to keep a higher level of highquality liquid instruments, their liquidity would be positively affected. This shows that more high-quality liquid instruments and a risk-sharing regulatory framework may provide the inner adjustment process through which any mismatch regarding maturity, risk, value or linkage with the real economy is corrected systematically. Assessing liquidity risk for Islamic banks with stress testing is a new topic that has, as of now, not been sufficiently researched. Conducting stress testing is subject to several limitations, such as a narrowed focus on the banks’ activities – mainly managing their assets and liabilities. Systemic factors, market-wide stress, and disruption of several key functions are beyond the scope of individual banks. This book attempts to develop a stress test to assess the effects of Basel standards on the liquidity of Islamic banks, by developing appropriate scenarios. By using stress testing, it is shown that PBs would have enough liquidity even under severe market stress. This means there is enough room for revision of the current regulatory https://doi.org/10.1515/9783110582901-202

VI

Preface

framework of Islamic banks to encourage greater participatory behavior amongst Islamic banks. Lastly, an alternative regulatory framework for the liquidity of Islamic banks, based on the findings of the models discussed in the book and capital market standards will be outlined as a policy recommendation. Additionally, the alternative regulatory treatment for Perpetual Sukuk and Esham are proposed. The book will conclude by proposing several regulatory recommendations and policy tools.

Contents Preface 1

V

Introduction, Objectives and Motivation 1 1.1 Background and Context of the Book 1 1.2 Problem Statements 2 1.3 Objectives of the Book 5 1.4 Motivation, Justification and Research Questions 5 1.5 Theoretical Foundation and Measurement of Liquidity Risk 1.6 Methodology and Data 7 1.7 Significance of the Study 9 1.8 Organization of the Book 10

6

2

Importance, Drivers and Implications of Liquidity Management, and Liquidity Standards 11 2.1 Literature Review on Conventional Banks’ Liquidity 11 2.2 Discussions on Liquidity Management in Islamic Banks 16 2.3 Literature Discussion on Liquidity Stress Testing 20 2.4 Basel Liquidity Regulations and the Importance of the LCR 25 2.5 Liquidity Types 26 2.6 Liquidity Regulations and Islamic Banks 27

3

Liquidity Tools for Liquidity Management in Islamic Banks 3.1 Liability (Deposit) Management 31 3.2 Financing Asset Management 32 3.3 Bank Reserves and Cash 32 3.4 Deposit Insurance 33 3.5 Interbank Deposits and Loans 33 3.6 IILM Sukuk 34 3.7 Commodity Murabahah (GCC & Malaysia) 36 3.8 Central Bank Wadiah Acceptance 37 3.9 Mudarabah Interbank Investment 37 3.10 Sale and Buy-back Agreements 38 3.11 Government Investment Issue (GII) 39 3.12 Bank Negara Monetary Notes-i (BNMN-i) 39 3.13 When Issue (WI) 39 3.14 Islamic Accepted Bills (IAB) 39 3.15 Islamic Negotiable Instruments (INI) 40 3.16 Ar-Rahnu Agreement-I (RA-I) 40 3.17 Sukuk Bank Negara Malaysia Ijarah (SBNMI) 41 3.18 Cagamas Mudarabah Bonds (Malaysia) 41

30

VIII

Contents

3.19 3.20 3.21 3.22 3.23 3.24 3.25 3.26

Islamic Private Debt Securities 41 Bai’ Al-Inah 41 Salam Sukuk 42 Certificates of Deposits (CDs) 42 Tawarruq 42 Central Bank Participation Papers 43 Ijarah Certificates (Shihab Certificates) 43 Central Bank Musharakah Certificates 43

4

Basel Regulations, Liquidity Management and Regulations in Turkey’s PBs 45 4.1 Turkish Banking System and Participation Banking 45 4.2 Turkey’s Liquidity Regulations for PBs 49

5

Dynamic Determinants of Liquidity Management in Turkish Participation Banks 55 5.1 Introduction 55 5.2 Empirical Approach and Data 56 5.3 Results and Discussion 61 5.4 Conclusion of Time Series Analysis and Policy Implications 76

6

Panel Data Analysis For Evaluating Effects of Liquidity Standards 6.1 Introduction 79 6.2 Methodology 79 6.3 Data 86 6.4 Discussion of Model and Summary of Variables 86 6.5 Results of Panel Models 88 6.6 Liquidity Regulations and Bank Financing 110 6.7 Explaining the Findings 118

7

Stress Testing For PBs 121 7.1 Introduction 121 7.2 Investigating the Main Vulnerabilities of Islamic Banks 124 7.3 The European Banking Authority (EBA) EU-wide Stress Test 125 7.4 Methodology and Stress Test Methods: The Case of Islamic Banks 126 7.5 Liquidity Stress Testing Model for PBs 128 7.6 Discussion of the Results of Models 135 7.7 Conclusions and Final Remarks 139

79

Contents

8

New Regulatory Framework for PBs’ Liquidity Management 141 8.1 Introduction 141 8.2 Fundamental Challenges that should be addressed for Liquidity Management of PBs 144 8.3 Cash and Profit Relations 145 8.4 Introducing a New Framework for Addressing Liquidity Risks 151 8.5 Central Banks and Monetary Policy 153 8.6 A Proposal for Treatment of Esham 153 8.7 EXIMS – Export-Based Musharakah Securities 155 8.8 The IFSB Supervisory Framework 155 8.9 Designing a New Liquidity Ratio and the Net Stable Funding Ratio 156 8.10 Transparency and Symmetric Information 160 8.11 Differentiation between Local and International Active Islamic Banks 161 8.12 New ALA Approach 162 8.13 Summary, Conclusion and Policy Recommendations 163

References

171

Appendices

179

List of Abbreviations List of Figures

211

List of Tables

213

Index

215

209

IX

1 Introduction, Objectives and Motivation 1.1 Background and Context of the Book The recent global financial crisis has brought to the fore the inadequacy of existing institutional frameworks for financial stability, and subsequently called for its strengthening. Following the crisis, financial and monetary authorities, such as central banks, the Bank of International Settlements (BIS), the Financial Stability Board (FSB) and Islamic Financial Services Board (IFSB), consider financial stability to be more important than the prevailing emphasis on financial liberalization or financial engineering. To this end, they have recommended numerous mechanisms, tools and regulatory frameworks for the effective supervision of financial intermediaries, especially global systemically important banks (G-SIBs). Islamic banks are no exception since they are governed by these new developments. While it is not obvious whether the Islamic banks would be affected positively or negatively by these developments, the present developments may crowd out Islamic financial institutions in the long run if they are not properly calibrated in accordance to their needs and requirements. While the resilience of the Islamic banking system during the recent global financial crisis highlights its potential contributive role to financial stability, particularly in countries with a significant presence of Islamic banks, there remain several concerns. Although there are no G-SIBs in the Islamic banking space, domestic systemically important banks (D-SIBs) that have more than 15% market share could be perceived as “too big to fail” in some countries (IFSB, 2015a). In addition, the Islamic financial markets in these countries are still in their infancy stage. Thus, they may not be able to withstand challenges and risks stemming from adverse systemic and financial shocks. For sustained real sector financing, the Islamic financial industry requires policy settings and a combination of micro and macroprudential tools in order for build-ups of financial imbalances and vulnerabilities to be addressed and systemic risk mitigated. A key risk factor relating to financial stability that deserves urgent attention is liquidity risk management. It is well recognized that sustainable liquidity risk management is central to the continuous financing and protection against systemic risks for banks. (Some of these risks include herding, unstable capital flows, vulnerable financial structures, and liquidity risks of counterparties, asset managers, market liquidity, business cycle and other market anomalies). In short, it is key to financial stability. This book seeks to evaluate the relation between liquidity risk management and financial stability in Islamic banks by (i) examining at short-term and long-term factors with different models and (ii) by conducting stress testing under several scenarios and (iii) developing an alternative regulatory treatment for liquid instruments. https://doi.org/10.1515/9783110582901-001

2

1 Introduction, Objectives and Motivation

1.2 Problem Statements The Islamic banking industry is confronted with several challenges regarding liquidity management. Since an Islamic Financial Institution (IFI) is not allowed to use interest-based financing resources from interbank money markets or other resources (such as using the central bank’s interest-based facilities via open market channels) and is not allowed to transfer its debt, Islamic banks have disadvantages concerning liquidity management compared to conventional banks. It is also a well-known fact that there is a dearth of Shariah-compliant securities or highly liquid Sukuk in many jurisdictions (IFSB, 2012; IFSB, 2015a). In addition, even if there are several Shariah-compliant Sukuk or securities in some jurisdictions, the secondary markets for these assets are thin or underdeveloped. The absence of Shariah-complaint lender of last resort (LoLR) facilities in many countries places further constraints on the ability of Islamic banks to mitigate liquidity risk (Mohammad, 2015). Having the relevant lender of last resort facility may promote moral hazard amongst these banks1 and consequently attenuate the efficiency of liquidity management. However, without such facilities, Islamic banks are not able to protect themselves against sudden liquidity changes or increasing stress in market liquidity. These factors and many others can impact the performance, growth, and portfolio management of Islamic banks as well as the confidence of investors. Moreover, the distinctive behavior of Islamic banks concerning asset-liability management, capital adequacy requirement, loan portfolio risk-taking and interbank demand hinder their capacity to undertake a comparable liquidity transformation to their conventional counterparts. The Basel Committee (BCBS) issued a revised Liquidity Coverage Ratio (LCR) in 2013 to strengthen banks’ liquidity management with the goal of promoting a more resilient banking sector (BIS, 2013). This was followed by the Net Stable Funding Ratio (NSFR) in October 2014.2 These requirements were to be enforced by January 2015 and January 2018 respectively. In 2015, with some changes, the IFSB adopted the LCR and the NSFR for Islamic banks (IFSB, 2015). The implementation of the liquidity coverage ratio (LCR) under the new standards may be challenging for Islamic banks. While the LCR and the NSFR are designed to improve

1 For example, according to the study of Drechsler et al. (2016) on EU banks, in general, weaklycapitalized banks use more LoLR loans to buy risky assets. This is one of the unintended consequences of the LoLR mechanism and using LoLR loans for investing risky assets may hamper the efficiency of pass-through mechanism of monetary policy and result in a reallocation of risky assets among banks. Drechsler et al. (2016)'s results show the importance of bank supervision for preventing moral hazard behavior. Although monetary authorities can control usage of LoLR loans, Islamic banks should be supervised carefully for preventing such risks. 2 See the document of the BIS, “Basel III: the net stable funding ratio” at http://www.bis.org/bcbs/ publ/d295.htm.

1.2 Problem Statements

3

banks’ resilience to short-term liquidity shocks by holding High-Quality Liquid Assets (HQLA) as reserve, these requirements compel Islamic banks to keep cash on their books due to a lack of short-term highly liquid and Shariah-compliant financial assets.3 Consequently, the efficiency, resilience, and profitability of Islamic banks can be adversely affected. Moreover, there are interconnections and interactions between liquidity standards and other standards including capital standards and leverage. The capital adequacy standards, particularly Basel II risk-weight requirements, as well as the capital conservation and countercyclical buffers4 further restrict Islamic banks’ liquidity management. Under the risk-weight mechanism, the partnership based investments are treated as high risk-weight due to the presence of counterparty risks.5 Accordingly, the current regulatory framework applies higher risk-weights than appropriate to Islamic banking assets.6 Like the LCR and NSFR, the mandatory capital conservation and countercyclical buffers force Islamic banks to hold cash. The new regulatory standards will also constrain Islamic banks with high international activity due to its new requirements for total loss absorbing capacity or leverage requirements. Although the Basel III framework allows local authorities to use their discretionary power in granting preferential treatment7 to certain assets, most supervisors of Islamic banks have a tendency to mimic the conventional regulatory framework to

3 The IILM was established to facilitate cross-border liquidity management amongst Islamic financial institutions by making available a variety of instruments of acceptable features and characteristics. This is achieved through the use of IILM instruments by Islamic financial institutions as instruments for liquidity management and as eligible collateral for interbank transactions, and central bank financing, or through trading of IILM instruments amongst IBs in the secondary market. However, the IILM instruments have not reached a sustainable critical mass due to its infancy. 4 BCBS developed the countercyclical capital buffer for ensuring that banking sector capital requirements take account of the macro-financial environment in which banks operate (https://www. bis.org/bcbs/ccyb). It is used as a buffer of capital to achieve the broader macroprudential goal of protecting the banking sector from periods of excess aggregate credit growth and reduce the risks in downturns. That means it has countercyclical nature both with excess credit growth and credit losses. It is calculated as the weighted average of the buffers in effect in the jurisdictions to which banks have a credit exposure. 5 Counterparty risk in Islamic contracts is the risk to another party to fulfill its contractual obligations, especially related with profit reporting, business strategies and providing timely symmetric information to banks. 6 Some regulatory authorities, including EBA, allow banks to use internal ratings-based approach instead of standardised approach. In this approach, banks calculate risk weight for capital requirements for credit risk by themselves according to criteria specified and required by regulators (Montes et al., 2016). Risk weigh calculated for credit risk under the internal approach are very low than those calculated under the standardised approach. However, many Islamic banks, including Turkey’s PBs, are not allowed to use internal approach. 7 Preferential treatments have been introduced since the introduction of Basel I. Some countries use these treatments for special sectors that prefer to protect. For example, GCC countries (Bahrain, Kuwait, Oman, Qatar, UAE and Saudi Arabia) decided mutually to apply preferential treatments for

4

1 Introduction, Objectives and Motivation

avoid non-compliance to Basel standards and possible negative assessment by international organizations and international market players. Mechanical application of the recommended methodology of the BIS and the IFSB standards may not be appropriate especially for risk sharing and partnership-based financial instruments. The risk-sharing based instruments, i.e. Mudarabah and Musharakah contracts are classified as illiquid instruments. The partnership-based business model of these instruments as noted above compel authorities to apply higher risk-weight than mark-up based products under the Basel II capital requirements. As a result, the risk-sharing instruments are not preferred by Islamic banks since they have negative bearings on liquidity management. Although new Islamic financial products have been developed in recent years, e.g. Esham,8 GDP-linked Sukuk and commoditybased Sukuk, there has been no attempt to evaluate the regulatory treatment of these instruments with respect to liquidity and capital requirements and standards. Against these limitations, there seems to be an urgent need to enhance the liquidity management of Islamic banks especially through the adoption of stress testing. Essentially, stress testing is an approximate estimate of how the value of a portfolio changes when there are large changes to some of its risk factors. While stress testing has been a popular tool in assessing the vulnerabilities of a group of institutions or of the financial system, its application to specific risk is limited. Indeed, stressing testing to liquidity risk is even more limited and under-developed, while its application to Islamic banks is virtually absent. For example, the European Banking Authority in its last stress test only assessed liquidity risks related to the cost of funding and not to the size and quality of liquidity buffers. In the US, liquidity stress tests play only a subordinate role to capital stress tests. The BIS working papers (BCBS 2013a; BCBS 2013b) identify gaps in current liquidity stress testing and suggest several areas for developing more robust liquidity stress testing. In light of its importance for liquidity risk management, the liquidity stress testing applicable to Islamic banks needs to be developed as a way to enhance their liquidity risk management. At present, Shariah-conscious investors are at a disadvantage when it comes to participating in financial services provided by Islamic banks. At the financing end, they face financing rates higher than the lending rates of conventional banks. Meanwhile, on the other side of the banks’ balance sheet, they receive a lower profit share compared to interest rates paid by conventional banks. Facing these, the Shariah-conscious investors may eventually opt out from the Islamic financial system or the financial system entirely. The uneven treatment that these investors receive would likely jeopardize financial inclusion, which is purported to be one of the goals of Islamic finance. Accordingly, there is a need to improvise the current

credit risk under standardised approach, which do not specifically cover Islamic banks’ assets (see RCAP report of Saudi Arabia at http://www.bis.org/bcbs/publ/d335.pdf). 8 Esham is a Shariah-compliant long-term instrument. Esham will be discuseed in detail in the seventh chapter.

1.4 Motivation, Justification and Research Questions

5

regulatory framework including the liquidity standards such that the interests of these investors are protected and financial inclusion is enhanced. In essence, the liquidity standards can be effective in protecting these investors and strengthening the financial inclusion of those who are unwilling to be bankable under the conventional banking system due to religious reasons.

1.3 Objectives of the Book This book attempts to demonstrate that Muslim-majority countries may well be able to fortify their financial system against asset price sensitivities and global financial shocks by reforming their regulatory framework of liquidity management such as to facilitate further the development of Islamic banks toward achieving the objectives of Islamic principles. The foregoing discussion highlights the importance of liquidity risk management. Accordingly, the present book has the following six objectives: 1. To critically study the liquidity risk management of Islamic banks and discuss the tools used for liquidity risk management. 2. To examine the short run and long run determinants of liquidity holdings of Islamic banks; 3. To examine the effects of financing on liquidity risk of PBs and compare results with that of CBs and also assess the impact of net cash outflows on the liquidity risk of PBs. 4. To develop a liquidity stress testing model for measuring Islamic banks’ liquidity, cash outflows and the stability of their funding under several assumptions; 5. To recommend the appropriate regulatory treatment (liquidity coverage ratio) for risk-sharing instruments (Esham and Sukuk) as alternative macroprudential tools; and 6. To recommend an alternative mechanism using capital market standards for liquidity requirements of Islamic banks to mitigate certain risks. In this context, a new framework for liquid instruments will be developed. For the first three objectives, the focus will be on the roles of the stock of HQLA introduced by Basel III, financing, net cash outflows and the macroeconomic environment; which are widely viewed to be a key in liquidity risk management. On the bases of the results from (1), (2) and (3), the book provides recommendation for improvements of the liquidity risk management of Islamic banks as stated in (4) and (5).

1.4 Motivation, Justification and Research Questions Liquidity risk management bears important implications for financial stability. It directly benefits stability by encouraging banks to reduce the risks on their balance sheets and by facilitating the liquidation of assets in a crisis. In conventional literature,

6

1 Introduction, Objectives and Motivation

various studies have attempted empirical assessments of the liquidity – stability relations and identification of liquidity risk determinants. An illustrative list of these studies include Aldasoro and Faia (2016), Bonner (2014), Wagner (2007), Banerjee and Mio (2014), Leykun (2016), and Hong et al. (2014). While the evidence to date for the liquidity – stability relations remains mixed, they all highlight the importance of effective liquidity risk management and hence the need to continuously identify factors that account for banks’ liquidity risk. Among the factors considered to be significant are capital adequacy ratio, total loan to total asset ratio, total deposit to total asset ratio, and net cash flows (Leykun, 2016; Hong et al., 2014). Despite their acknowledged importance to liquidity risk of conventional banks, whether they are central to Islamic banks remain largely unknown. The limited understanding of liquidity risk of Islamic banks as well as the lack of well-developed liquidity frameworks for the Islamic banking system serves as motivations and justifications for the present book. More precisely, driven by the implications of liquidity risk management for financial stability and in light of the introduction of new regulations for liquidity management, this book assesses macro and micro prudential instruments for effective liquidity risk management and hence the stability of Islamic banks by taking the Turkish banking system as a case study. It is postulated in this book that liquidity formation of Islamic banks is negatively affected by conventional regulations. To validate this postulation, this book will examine the determinants of liquidity holdings in Turkish Islamic banks by addressing the several issues. In line with the stated research objectives, the book raises the following research questions to be addressed: 1. Can the liquidity requirement ratio of Islamic banks be explained by financial factors such as the stock of Sukuk, financing, deposits and return on equity? 2. Can the liquidity requirement ratio of Islamic banks be explained by macroeconomic factors such as inflation, interbank money market rate or government bond rate? 3. Do these determinants affect the liquidity risk of Islamic banks in the short run or in the long run, or both? 4. Are there any significant differences between the factors that affect the liquidity of the Islamic banks and the liquidity of their conventional counterparts? 5. Is there any shortage of liquidity for PBs under stress conditions? 6. What is the appropriate regulatory treatment for risk-sharing instruments? 7. What is an alternative mechanism using capital market standards for the liquidity requirements of Islamic banks to mitigate certain risks?

1.5 Theoretical Foundation and Measurement of Liquidity Risk Liquidity risk is defined as a shortcoming to cover financial liabilities (Banerjee and Mio, 2014) and it may affect bank runs through systematic and idiosyncratic channels (Deming and Hong, 2012; Roman and Sargu, 2015). Moreover, the maturity

1.6 Methodology and Data

7

transformation of short-term deposits into long-term loans results in liquidity risk (Vodova, 2011; Vodova, 2013; Bonner, 2014). Since the liquidity risk management of banks has serious implications on its overall macroeconomic and financial stability (Roman and Sargu, 2015; Banerjee and Mio, 2014; Bonner, 2014), several studies have evaluated the liquidity risk management of banks in the context of regulations and regulatory framework, reaching different conclusions concerning the effects of regulatory standards. Numerous quantitative findings show some positive evidence for reforming standards and applying tighter regulations (Morgan and Pontines, 2013; Banerjee and Mio, 2014; Fuhrer et al. 2017; Vazquez and Federico, 2015; Hoque et al. 2015; Morgan and Pontines, 2013; Admati et al. 2013). Conversely, the negative effects of these regulatory changes on liquidity risk were reported by many studies (Gorton and Winton; Barth et al. 2013; Calluzzo and Dong, 2015; Fukuyama and Matousek, 2011). Further, it is also suggested that regulations are not effective in influencing, or have limited effects, on the risk appetite of banks (Demirgüç-Kunt and Detragiache, 2011; Cathcart et al. 2015; Calluzzo and Dong, 2015; Hong et al. 2014). In both conventional and Islamic finance literature, there are many studies conducted on liquidity risk and determinants of liquidity risk management (Roman and Sargu, 2015; Ogilo and Mugenyah, 2015; Bonner, 2014; Aldasoro and Faia, 2016; Wagner, 2007; Krasicka and Nowak, 2012; Vodova, 2011; Ahmed et al., 2011; Vodova, 2013; Mongid, 2015 and Ariffin, 2012; Cuccinelli, 2013; Leykun, 2016; Laštůvková, 2016; Distinguin et al., 2013; Hong et al. 2014; Akhtar, 2007; Cucinelli, 2013; Angora and Roulet, 2011; Banerjee and Mio, 2014; Akhtar et al., 2011; Mohamad et al., 2013; Boudt et al., 2013; Ergec and Arslan, 2013; Amin, 2016).

1.6 Methodology and Data Based on these different theoretical foundations, it is assumed that regulations are very effective on banks’ liquidity risk management and therefore models in which liquidity risk is used as a dependent variable and calculated according to BIS standards have been developed. Hence, liquidity risk is used as a dependent variable in many studies on Islamic banks or conventional banks, or on both types of banks (Mohamad et al., 2013; Roman and Sargu, 2015; Ogilo and Mugenyah, 2015; Amin, 2016; Bonner, 2014; Banerjee and Mio, 2014; Aldasoro and Faia, 2016; Wagner, 2007; Krasicka and Nowak, 2012; Vodova, 2011; Ahmed et al., 2011; Vodova, 2013; Mongid, 2015; Ariffin, 2012; Cuccinelli, 2013; Leykun, 2016; Laštůvková, 2016; Distinguin et al., 2013; Hong et al., 2014). However, there are different applications on usage of liquidity risk metrics in banking literature. Prior to the introduction of the LCR framework, most empirical studies employed accounting-based measures of liquidity risk, including loans to total assets ratio (Roman and Sargu, 2015), total deposits to total assets ratio (Mohamad et al. 2013) and loans to total assets ratio (Amin, 2016). While these measures are easily computed

8

1 Introduction, Objectives and Motivation

from banks’ balance sheets, they contain two main weaknesses. First, they do not capture the different implications of liquid assets, degree of liquidity treatment of liquid assets, as well as the impact of cash inflows and cash outflows on liquidity risk. Second, as argued by Poorman and Blake (2005) and Distinguin et al. (2013), using such metrics for liquidity could be inaccurate under several conditions. Distinguin et al. (2013) in particular provides an example of a large regional bank (the Southeast Bank of Miami) that went bankrupt in September 1991 because of its inability to repay some liabilities claimed on demand despite its liquid assets to total assets exceeding 30%. Distinguin et al. (2013) and BIS (2013) suggest using liquidity indicators that embed information about liquid assets as well as cash outflows. Subsequently, new metrics of liquidity risk based on high-quality liquid assets, cash inflows and cash outflows have been developed by the Basel committee (BIS, 2013). Numerous new metrics of liquidity risk are adopted in conventional banking literature. Bonner (2014) uses liquidity ratio as a dependent variable defined by the Dutch Central Bank (DLCR – Dutch Liquidity Coverage Ratio) according to Basel requirements. Moreover, Banerjee and Mio (2014) employ the share of HQLA to total assets as a dependent variable to analyze the behavioral reaction of banks to a tightening of liquidity regulation. Aldasoro and Faia (2016) use liquidity coverage ratio, while Wagner (2007) adopts liquid asset ratio as a measure of liquidity. Cucinelli (2013) uses both liquidity coverage ratio and net stable funding ratio for measuring the liquidity risk. Leykun (2016) estimates liquidity risk by calculating the liquid assets to total deposit ratio. Distinguin et al. (2013) use a liquidity creation indicator computed as weighted sums of assets and liabilities, where weights are assigned according to their liquidity levels, and divided by total assets. Laštůvková (2016) defines liquidity as positive flows (cash inflows), the negative flow (outflows) and the net change, in which four methods were used to calculate this dependent variable. With regard to Islamic banks, Krasicka and Nowak (2012) and Ahmed et al. (2011) use liquid assets to total assets as a measure of liquidity. Liquid assets to total funding (Mongid, 2015) and the total liquid assets of liabilities (Ariffin, 2012) are other proxies used for liquidity risk for Indonesian Islamic rural banks and Malaysian Islamic banks respectively. To the best of the researchers’ knowledge; no study has specifically investigated the effects of liquidity requirement ratio or liquidity coverage ratio in a dual banking system, where Islamic banks and conventional banks coexist. The present book measures liquidity risk using the liquidity requirement ratio. The book adopts two empirical approaches to examine determinants of liquidity risk in the dual banking system – time series econometric modelling and panel data modelling approaches. The time series econometric modelling is to uncover especially short-run relations among the variables under study. For concreteness, panel data modelling is used to complement the time series modelling. In panel modelling, both bank-specific factors and macroeconomic factors are examined in terms of their impact on liquidity requirement ratio. Panel data is also utilized to allow the banks to be heterogeneous, collect more informative data, more variability, less collinearity among

1.7 Significance of the Study

9

the variables, more degrees of freedom and more efficiency (Baltagi, 2005). Therefore, in panel data, justifying and measuring effects on liquidity may become easier to detect. Since monthly data is used in this book following Haan and End (2013), Bonner (2014) and also Vodova (2013), it is assumed that the disadvantages associated with using panel data as stated by Baltagi (2005), such as using a short period while relying on infinitiy individuals are not applicable. The “Fixed-effect models” as well as by “Random-effect GLS regression models” are used in the analysis. After the global financial crisis, stress testing has become a very popular and important instrument for evaluating the resilience of portfolios of banks and assessing their riskiness under certain assumptions. There are many different stress testing models used in the literature (Zangari, 1996; Čihák, 2007; Schimeder et al., 2012; Čihák, 2014; Van den End, 2012). To perform a stress test, it is needed first to define the typical structure of assets and liabilities of an Islamic bank. The stress testing is used to understand macro scenario effects on PBs’ liquidity management. For stress testing, the typical structure of assets and liabilities of an Islamic bank are first defined and then the method of Implied Cash Flow Analysis (ICFA) is used based on Čihák (2007), Schimeder et al. (2012), later Čihák (2014) and IFSB (2015b). These empirical methodologies will be explained in detail in later chapters. Banking data in the fourth and fifth chapter are mainly collected from the Central Bank of the Republic of Turkey (CBRT), Banking Regulation and Supervision Agency (BRSA), PBs Association of Turkey’s databases and checked with the banks’ quarterly financial positions. Macroeconomic variables (CDS, government bond rate and interbank rate) are extracted from the Bloomberg Terminal.9 The inflation rate of Turkey was derived from Turkish Statistical Institute. Several statistical programs are used10 in this book for quantitative analyses.

1.7 Significance of the Study On the whole, this study will help regulatory and supervisory bodies to reform their regulatory framework for Islamic banks to extend their playing field. Moreover, this book will be helpful to practitioners and monetary authorities in conducting new 9 The Bloomberg Professional service (Bloomberg Terminal) is a software solution and flexible platform for financial professionals, academicians and public authorities who need real-time data, news and information. For more information, please see http://www.bloomberg.com. 10 Eviews is used (Which is a statistical, forecasting, and modeling tools with a simple objectoriented interface) for time series analysis and Stata 14 statistical package is employed for panel data analysis (Stata is a general-purpose data analysis and statistical software package used by researchers and professionals. It is a complete, integrated statistics package that provides a broad range of statistical analyses, plus data management, graphics, simulations, and custom programming. For more information, please see http://www.stata.com). For Stress testing, an Excel-based model is applied, and a VAR based results is used for developing assumptions.

10

1 Introduction, Objectives and Motivation

policies for liquidity management of banks. Additionally, researchers studying regulations and supervision areas can further investigate the factors affecting the liquidity position of Islamic banks by collecting more micro-level data from different countries and by using the limitations of this study as a stepping-stone.

1.8 Organization of the Book In addition to the introduction, this book is composed of seven chapters. Chapter 1 highlights the main purpose of the book. In particular, it introduces the research background and explains the research problems, followed by motivations and the research objectives and questions. The chapter concludes with the significance and scope of the study. Chapter 2 focuses on the review of literature regarding the importance of liquidity management and standards, as well as their drivers and implications. Chapter 3 discusses liquidity tools used by central banks for liquidity management of Islamic banks. Chapter 4 gives necessary information about Basel requirements, Turkish Banking systems, liquidity management, their regulations and supervision framework. Chapter 5 discusses the liquidity management of PBs in Turkey. A unique VAR based model for liquidity management of these banks is designed which is also applied to Turkish conventional banks. Chapter 6 covers panel data models for Turkish PBs’ bank-level data. These panel models are developed to see the effects of liquidity regulations on Turkish PBs and conventional banks. Chapter 7 presents the application of the stress testing model and its outcomes to Turkish PBs using bank level data. In this stress testing, the main vulnerabilities of Islamic banks are identified, primary issues are diagnosed, and the outputs of the scenario will be mapped into a form useful for analyses. Chapter 8 develops an alternative approach to liquidity management of Islamic banks and introduces reformed standards for Islamic banks. It is then concluded with policy recommendations for regulators, proposals and shortcomings of the study.

2 Importance, Drivers and Implications of Liquidity Management, and Liquidity Standards 2.1 Literature Review on Conventional Banks’ Liquidity The 2008 financial crisis has led to shifts in the global regulatory framework, drawing the attention of many supervisory authorities to the need for minimizing risk in financial institutions. After the crisis, there has been an increase in liquidity requirements, Tier I capital requirements and the quality requirements of capital. Basel III reforms have been developed to improve banks’ ability to absorb shocks, enhance their risk management and governance, and strengthen their transparency and disclosures. At the end of May 2016, 24 FSB jurisdictions had Basel III risk-based capital rules in force. Final rules on liquidity (LCR) have been issued in all but two jurisdictions, and are in force in 15 jurisdictions (representing 61% of the market). All large internationally active banks already meet the fully phased-in, risk-based minimum capital requirements; 80% of these banks meet or exceed the minimum liquidity standards – the LCR and the Net Stable Funding Ratio (NSFR). Most banks have adjusted to higher capital requirements through the accumulation of retained earnings rather than by cutting back sharply on lending or raising margins. Additionally, global systemically important banks (G-SIBs) have been designated. Since the issue of liquidity risk management of conventional banks has gained the attention of various studies, it has been discussed by many quantitative studies which have reached different conclusions regarding the effects of introducing new regulatory standards. Several studies find some positive evidence for reforming standards and applying tighter regulations (Morgan and Pontines, 2013; Banerjee and Mio, 2014; Fuhrer et al. 2017; Vazquez and Federico, 2015; Hoque et al. 2015; Morgan and Pontines, 2013; Admati et al. 2013). Hence, Morgan and Pontines (2013) suggest that increasing liquidity standards can have important effects on financial development. This finding was partly supported by Banerjee and Mio (2014) who found that UK banks increased the share of HQLA and reduced short-term intrafinancial loans and short-term wholesale funding after the introduction of tighter regulations by the UK Financial Services Authority (FSA) in 2010. A recent study of Fuhrer et al. (2017) investigate the added value of high-quality liquid assets (HQLA) defined by the Liquidity Coverage Ratio (LCR) requirements and find evidence for the existence of an HQLA premium in the order of 4 basis points. Vazquez and Federico, (2015) suggest that banks with weaker structural liquidity and higher leverage in the pre-crisis period were more likely to fail after the crisis. Moreover, it is claimed that banks in countries with stricter capital requirements have lower systemic risk contributions (Hoque et al. 2015). Morgan and Pontines (2013) find that liquidity standards and rules for OTC derivatives could have important effects on https://doi.org/10.1515/9783110582901-002

12

2 Importance, Drivers and Implications of Liquidity Management

financial development in the Asia region. Admati et al. (2013) claim that social cost of forcing banks to keep more capital in their financial poisitons is not high and this may increase their financial performance. In contrast, several studies find negative effects of these regulatory changes (Gorton and Winton; Barth et al. 2013; Calluzzo and Dong, 2015; Fukuyama and Matousek, 2011). While it is expected that capital standards would decrease vulnerabilities in terms of market and funding liquidity, Gorton and Winton (2000) show that the increase of capital requirements has an inverse impact on banks’ supply of deposits, hence reducing the liquidity provision role of banks. In this sense, new Basel standards have the potential to decrease efficiency and increase vulnerabilities if not calibrated for specific needs of different banking models. Barth et al. (2013) demonstrated that banking restrictions are negatively related to bank efficiency. Besides, Calluzzo and Dong (2015) discovered that while these institutions have become less risky on a standalone basis after the crisis, the financial market has become more vulnerable to systemic contagion. In this context, Fukuyama and Matousek (2011) find that Turkish banks reacted positively to consolidation and restructuring. Bank efficiency had gradually improved, and the cost-efficiency scores peaked immediately, but they also found a gradual deterioration of bank efficiency from 2004 to 2007, and they expounded this negative trend by the strict regulatory rules imposed by BRSA. Barth et al. (2004) suggest that restricting banking activities reduce banking stability and increase the probability of a banking crisis. Calluzzo and Dong (2015) point out that although new regulations have made improvements, the major risks remained. As mentioned above, while financial institutions have become less risky individually after the crisis, Calluzzo and Dong (2015) claim that the financial market has become more vulnerable to face off contagion effect and even more integrated financial system might experience more synchronized shrinkages in future crises. Wagner (2007) finds evidence for an increased liquidity of bank assets which increase banking instability and make them more prone to externalities associated with banking failures. According to his findings, keeping higher liquid assets is an incentive for banks to reduce the risks on balance sheets because banks can liquidate these assets in times of stress, but it is costly for financial stability because it may trigger asset fire sales during stress and endanger financial stability. Nevertheless, some claim that regulations are not effective in influencing or have limited effects on the risk appetite of banks (Demirgüç-Kunt and Detragiache, 2011; Cathcart et al. 2015; Calluzzo and Dong, 2015; Hong et al. 2014). For example, Demirgüç-Kunt and Detragiache (2011) indicate that bank supervision and regulation have a small effect on bank risk. It is suggested that there was a minimum implication of the Basel II regulation in the financial crunch (Cathcart et al., 2015). DeYoung and Jang (2016) studied US commercial banks’ liquidity management before the implementation of the Basel III liquidity regulations and found evidence that the data is consistent with a liquidity management regime, which means that these banks were

2.1 Literature Review on Conventional Banks’ Liquidity

13

managing their liquidity quite well before the introduction of an LCR framework. They also realized that big banks set lower targets for liquidity ratios, implying that there is a negative impact of the size of banks on their liquidity ratios. Although most banks held more liquid assets against liquid liabilities than strictly required, the fundamental risks remained (Calluzzo and Dong, 2015). Although the results of Cucinelli’s (2013) study show that there is a relationship between the liquidity coverage ratio and credit rating in short-term, there is no correlation between the long-term liquidity metrics and a probability of default. Hong et al. (2014) observe potential links between Basel III liquidity risk measures and bank failures using a model that differentiates between idiosyncratic and systemic liquidity risks and find that while the LCR has limited effects on bank failures. They reached a conclusion that an effective framework for liquidity risk management may target liquidity risk at both the idiosyncratic level and the systemic level. However, different regulations and factors may have much stronger effects on liquidity, such as swap lines, idiosyncratic channels, regulatory arbitrage, macroprudential tools and asymmetric information. In this context, the study of Andries et al. (2017) on the effects of international swap lines regulated by the Swiss National Bank (SNB) shows that stock prices of both local and less well-capitalized banks, and banks with high foreign currency exposures responded more strongly to SNB swap lines. They derive a conclusion that swap lines regulated by central banks enhance market liquidity and reduce risks associated with micro-prudential issues. On the other hand, Leykun (2016) uses the fixed-effect unbalanced panel data estimation technique and shows that capital adequacy ratio, total loan to total asset ratio, and total deposit to total asset ratio impacts the liquidity risk of commercial banks negatively and significantly. Liquidity risk may affect bank runs through systematic and idiosyncratic channels as supposed by Deming and Hong (2012). This finding has important consequences for the new standards of the management of liquidity risk because the systematic channel is related to market liquidity whereas the idiosyncratic channel is related to funding liquidity risk, bank’s lending ability and misconduct risk. In this context, regulations are needed to specify the current risks and emerging new risks. Bonner et al. (2013) reveal that if there is no liquidity regulation, the determinants of banks’ liquidity buffers will be a combination of bank-specific and country-specific variables. Despite the fact that liquidity regulation covers most of the incentives, a bank’s disclosure requirement and size are not regulated enough. Cornett et al. (2011) claim that banks holding more illiquid assets on their balance sheets increased asset liquidity and reduced lending after the crisis. Hoque et al. (2015) find that regulatory restrictions, official supervisory power, capital stringency, along with special monitoring can explain bank risk in both crises. While measuring liquidity risk with the LCR and NSFR, Angora and Roulet (2011) highlight the relationship between liquidity risk and return on assets (ROA), the natural logarithm of total assets, the ratio of loans to customers and total loans, GDP annual

14

2 Importance, Drivers and Implications of Liquidity Management

growth rate, and the spread between the interbank rate and central bank policy rate. In general, their study highlights that the liquidity risk ratio has a negative and significant relationship with size, regulatory capital, and total assets, but has a positive and significant correlation with macroeconomic variables (especially monetary policy rate and GDP). Ogilo and Mugenyah (2015) find that, collectively, capital adequacy, liquid asset ratio, ownership type, size, and leverage were significant determinants of liquidity risk. It is worth mentioning the regulatory arbitrage of new standards that may result in negative externalities. In particular, Kara and Ozsoy (2016) examine the interaction between capital and liquidity standards in a model characterized by fire sale externalities and show that banks respond to stricter capital requirements by decreasing their liquidity ratios. Their results show that capital standards can lead to less severe fire sales by addressing the inefficiency and reducing risky assets compared to liquidity regulations. Since the regulators set high capital ratios, this leads to inefficiently low levels of risky assets and liquidity. Macroprudential tools are used for complementing capital regulations, improving financial stability, and allowing for a higher level of investment in risky assets. However, there is no standard for macroprudential tools in developed countries and emerging economies. Banking regulations may have different negative impacts on credit channels. It is claimed that there is a trade-off between banking regulations that promote adequate credit allocations and aggregate credit supply (Boissay and Collard, 2016). If there is a scarcity of high-quality liquid assets in the market, then capital regulations debilitate liquidity of banks. After the last global crisis, risk-weights were increased, and higher liquidity ratios were introduced. These regulatory changes implicitly accept that the risk-weight system of Basel II increased the risk of collapse. On the other hand, some experimental papers showed opposite findings. Cathcart et al. (2015) note that those changes to risk-weight categories were not among the fundamental problems seen in the subprime crisis. This suggestion implies the minimum implication of the Basel II regulation during the crunch. They demonstrate that these dynamics are governed by a formula linking the two ratios together that derives from the sensitivity of the risk-based capital ratio to a change in its risk-weight (Cathcart et al., 2015). Cornett et al. (2011) discovered some evidence that funding liquidity is more sensitive to banks’ capital resources in times of crisis and suggest that changing capital requirements when the banking industry is under financial stress might notably affect their funding liquidity. Benmelech et al. (2017) discovered that during the global financial crisis, the collapse of the asset-backed commercial paper market reduced the financing capacity of non-bank lenders as captive leasing companies in the automobile industry; implying that illiquidity in the market caused a credit supply shock which resulted in the decline in auto sales. The “pure” lending transmission channel estimates the impact of capital requirements on either lending interest rates or credit growth directly. For example,

2.1 Literature Review on Conventional Banks’ Liquidity

15

de-Ramon et al. (2012) sought to estimate the relationship between loan spreads and aggregate bank capital ratios using UK data from 1992 to 2012. They found that, in the long run, UK spreads are increased directly by 9.4 basis points for a one percentage point increase in total capital requirements. Moreover, Sutorova and Teply (2013) estimate that lending rates increased by 19 basis points for a one percentage point increase in capital resources using a sample of 594 European banks for the period 2006 to 2011. Fraisse et al. (2015) study French banks and find that a one-percentage point increase in capital requirements leads to approximately 10-percentage reduction in total lending. Asymmetric information and transparency also influence liquidity in the market. Covitz et al. (2013) found that asymmetric information has important implications for the risk tolerance degree of asset-backed commercial paper investors. Increasing transparency and decreasing asymmetric information may positively reduce liquidity run and contagion effects in different markets. Since liquidity is volatile, banks are subject to inherent liquidity mismatches (Praet and Herzberg, 2008). Therefore, the use of a composite indicator of market liquidity is preferred (Kerry, 2008), based on a series of measures developed in the academic literature; which shows that market liquidity can fall rapidly in times of stress and highlighting the importance of managing liquidity risk in the financial system. It is claimed that the current conventional regulatory framework concentrates on the needs of advanced countries. In this sense, BIS (2014) admit that several regulatory reforms could have potential adverse implications for some emerging markets (EMDEs). Several studies considering the lending channel look at how other factors may have differential impacts on credit in different economic environments. Carlson et al. (2013) find that commercial real estate, business, and industrial lending appear more sensitive to the size of the capital surplus compared to any other lending. The impact on funding liquidity was extensively researched in the period before the financial crisis. The results from that time are consistent with more recent papers, although the pre-crisis discussion focuses more on the various ways in which banks respond to changes in their capital resources. Berrospide and Edge (2009) estimate a dynamic model where capital ratios follow a partial adjustment process. They found that banks with a higher capital surplus will experience higher loan growth post-adjustment. According to the bank liquidity and financial stability research by Valla and Saes-Escorbiac (2006) on UK banks, the liquidity ratio as a measure of the liquidity is dependent on several factors; the probability of obtaining support from the lender of last resort, interest margin as a measure of opportunity costs of holding liquid assets, bank profitability, loan growth, size of the bank, GDP growth as an indicator of a business cycle, and short-term interest rate capturing the monetary policy effect. Along different lines, Almeida et al. (2017) find that sovereign debt deterioration can have a significant impact on financial markets and non-bank corporates through a

16

2 Importance, Drivers and Implications of Liquidity Management

credit rating channel. In their model, firms reduce their investment and reliance on credit markets due to a rising cost of capital following a sovereign rating downgrade. According to the Modigliani-Miller theorem, if there is a cost of financing and tax burden, firms prefer to issue debt-like instruments because these tools are tax deductible. On the other hand, the segmentation of the deposit and equity markets also explains why high leverage is attractive for a bank. Allen and Carletti (2013) add financial friction in the form of intermediation costs in their model. In equilibrium, the cost of equity financing is equal to that of deposit funding plus intermediation costs. Higher leverage can, therefore, be justified from a bank perspective. Higher risk weights and capital requirements reduce the return on equity (ROE). In order to keep their return stable, banks raise credit rates with the risk that higher lending rates may result in lower lending, and thus, reduce economic activity. Slovik and Cornède (2011) show that one percentage point increase in the ratio of capital to risk-weighted assets will push up bank funding liquidity spreads by 14.4 basis points on average. It is claimed that adverse effects on credit supply can be offset by accommodative monetary policy in the short run (Slovik and Cornède, 2011) although, in the long run, monetary policy that is too accommodative might lead to excessive risk-taking by banks.

2.2 Discussions on Liquidity Management in Islamic Banks Many issues and proposals regarding financial reform agenda are in the initial stages of implementation, and the full impact of these reforms on Islamic banks have not yet been studied deeply. Nonetheless, there are some concerns about the effects of Basel standards, including the Basel III liquidity standards, on the portfolio selection of Islamic banks. The new Basel III liquidity framework is expected to lead to implementation challenges for Islamic banks in the coming years due to the limited availability of high-quality liquid assets and difficulties in calibrating the structure to suit the practices of Islamic banks. Islamic banks have some structural differences from conventional banks in the context of contracts and the liability side of their balance sheet (Verhoef et al., 2008). Deposits of Islamic banks are composed mainly of three classes of accounts: Current account deposits, saving deposits, and investment deposits. Investment accounts are divided into restricted investment accounts and unrestricted investment accounts. Even in the case of the latter, account holders have the option to withdraw their investments before maturity. This creates a possibility of premature withdrawal by account holders when there is a mismatch between investors’ expectation of return and the actual return. In a perfect liquidity management system, it may be theoretically possible to estimate the likely demand for withdrawal. If this estimation is nearly perfect, these banks then need to keep only that estimated portion as liquid cash or nearcash items. In practice however, most Islamic banks hold high levels of excess cash.

2.2 Discussions on Liquidity Management in Islamic Banks

17

Unfortunately, the effects of these new standards on Islamic banks have not been analyzed enough in research, and at present there are only a few studies on these regulatory issues. Mongid (2015), for example, explores the liquidity risk management of small Indonesian Islamic banks and claims that capital adequacy as well as asset management and leverage determine liquidity risk of Islamic banks. However, the shortcoming of his study is that the effects of macroeconomic factors or economic environment variables were not included in the model; such as inflation rate, central bank fund rate or interbank rate. Ashraf et al. (2016) studied the impact of the NSFR on the financial stability of Islamic banks and found that the modified NSFR will have a positive influence on the financial stability of Islamic banks and it was found that the marginal impact of the NSFR on stability weakens as the size of the bank increases. This means that if Islamic banks continue to grow, the positive implications of the NSFR on this sector will diminish. One of the recent studies on empirical analysis of the bank-specific, financial, and macroeconomic determinants of the performance of Islamic and conventional banks in Pakistan has been done by Rashid and Jabeen (2016). By using an unbalanced annual panel data and the GLS regression, they showed that the performance of Islamic banks could be explained by operating efficiency, deposits, and market concentration. Rashid and Jabeen (2016) showed that the impact of the lending interest rate on performance is negative for Islamic banks. Islamic banks in many countries show greater potential for growth, intensifyingthe specific challenges for these banks that need to be evaluated differently from conventional banks. For this reason, many international organizations, supervisory and regulatory authorities, and policy makers have examined various aspects of Islamic financial intermediation, each from their own perspective. Most of this research provides evidence of greater resilience (Bourkhis and Mahmud, 2013) and better financial performance of Islamic banks during crises thanks to their higher asset quality despite the lack of high-quality liquid instruments. Moreover, this greater degree of resilience may be related to small bank effects (Hasan and Dridi, 2010). Even though Islamic banks have to keep significant funds as cash to meet regulators’ requirements, Ismal (2010) suggests that Indonesian Islamic banks have historically managed liquidity well, though the industry is too fragile. He shows a trade-off between self-insurance against liquidity risks and opportunity costs of holding liquid assets. A proper understanding of the macroeconomic situation and regulatory framework is thereforeimportant for addressing this trade-off. Basel regulations were developed for internationally active and large banks, while most Islamic banks are small and are subject to these liquidity standards. The majority of these small banks are “scale inefficient” in the context of profitability and capitalization – the primary determinants of Islamic banking efficiency (Rosman et al. 2014). In this context, Saeed and Izzeldin (2014) suggest that a decrease in default risk is associated with lower efficiency levels. However, Alam (2013) finds that regulations increase the technical efficiency for Islamic banks and

18

2 Importance, Drivers and Implications of Liquidity Management

suggest that Islamic banks appear to be technically efficient in stricter regulatory conditions. However, tighter regulatory conditions improve risk-transferability of banks, which result in avoiding risk-sharing contracts. Since Alam’s (2013) study covers data for the year 2006–2010, new studies are needed to evaluate the effects of Basel reforms on the efficiency of Islamic banks based on more granular data. Macroeconomic control variables can also influence the behaviour of Islamic banks in managing liquidity according to Mohamad et al. (2013) based on evidence of Malaysian banks. The same result was found by Krasicka and Nowak (2012) who suggest that Malaysian Islamic banks have responded to economic and financial shocks in the same way as Malaysian conventional banks. Using parametric and non-parametric classification models, Khediri et al. (2015) find that Islamic banks are, on average, more liquid and profitable than conventional banks. Soylu and Durmaz (2013) show that PBs in Turkey have effective and reasonably strong rates of profitability, despite a lower level of profitability amongst its conventional banks. The PBs in Turkey are noticeably influenced by interest rates (Ergec and Arslan, 2013), and being influenced by interest rates may relate to being governed by the same banking law and not having any specific money market specifically for PBs. On the other hand, specific studies on determinants of liquidity management of Islamic banks have shown that the probability of occurrence of irregular liquidity withdrawals and a liquidity run is very moderate in these banks. This may be the direct result of the satisfactory performance of Islamic banks and the confidence of their depositors (Ismal, 2010). It is accepted that short-term instruments are more liquid than long-term assets. Short-term excess liquidity is managed through the money markets. Several countries have developed money markets for Islamic banks, where Islamic banks can convert their assets into liquid assets when needed. These mechanisms allow Islamic banks to manage liquidity while staying profitable. Islamic banks may have excess liquidity because of the regulatory framework that requires that LCR standards be met. This excess liquidity can decrease Islamic banks’ profits and limit their market share. Variably, Akhtar (2007) suggests that regulators should adopt different approaches for the Islamic financial system. Asutay (2013) claims that the secular regulatory framework has an effect on Participation banking in Turkey, while Akkizidis and Khandelwal (2008) suggest that any significant losses on Permanent Musharakah contracts may debilitate any further continuation of the business. For this reason, it is right to give a higher risk-weight for these partnership-based contracts. Haan and End (2013) find that most banks hold more liquid assets against liquid liabilities than strictly required, whereas more solvent banks hold less liquid assets against their stock of liquid liabilities. This suggests an interaction between capital and liquidity buffers. Hardy (2012) says that in Turkey, the state bureaucracy and military routinely resisted Islamic influence in business and finance throughout the Republican period following the reforms done in the 1920s. However, specific needs of Islamic financial

2.2 Discussions on Liquidity Management in Islamic Banks

19

institutions should be taken into consideration when a regulatory framework is developed (Hesse et al., 2008). It is suggested that the implementation of reforms in emerging markets needs to be sequenced, taking into account several factors. Adam and Thomas (2005) claim that the current regulatory framework should protect the confidence of Muslims in the system as well as their investments. Since this structure is draconian towards Islamic contracts by imposing the current risk-weight system, as a result, Islamic banks restrict their activities and are reluctant to introduce new products based on risk sharing. It is claimed that the risk-sharing feature of Islamic banks may ease to absorb external shocks and restrain against cash outflows because under a risk-sharing system, there would be better transformation between liabilities and assets (Chapra, 1992; Mirakhor, 2012; Maghrebi and Mirakhor, 2015). However, Chong and Liu (2009) suggest that most of the deposits are not invested in risk-sharing financing. Instead of questioning the reasons behind this behavior of banks, they conclude that Islamic banks should be regulated and supervised under the same regulations as conventional banks because they are offering similar products (Chong and Liu, 2009). In this context, the researchers have to answer this question: “If Islamic banks are offering the same products and they are being regulated in a similar manner as conventional banks, why do we need Islamic banks?” On the other hand, Maghrebi and Mirakhor (2015) try to show that the procyclicality of the financial system can be reduced by using the risk-sharing modes of financing real investment in the public and private sector. It is claimed that the economic incentives for credit risk transfers and speculative activities can be mitigated by the risk-sharing principle underlying Islamic finance (Maghrebi and Mirakhor, 2015), because forward looking expectations can be governed by the optimal resource allocation, which is maintained by the risk-sharing mechanism. The essence of the problems in Islamic banking is well addressed in the following paragraph: The spectrum of Islamic finance instruments runs the gamut between short-term liquid, lowrisk financing of trade contracts to long-term financing of real sector investment. The essence of the spectrum is risk sharing. At one end, the spectrum provides financing for purchase and sale of what has already been produced to allow further production. At the other end, it provides financing for what is intended or planned to be produced. In this spectrum, there does not seem to be room provided for making money out of pure finance where instruments are developed that use real sector activity only as a virtual license to accommodate what amount to pure financial transactions. (Mirakhor, 2012, p. 32)

Mirakhor et al. (2012) explore the characteristics, operations, and benefits of a comprehensive risk-sharing financial system for long-term economic and social prosperity to present the development of Islamic finance as a complete financial system. In this mechanism, a free flow of information in the market may increase efficiency (Mirakhor, 2010) and decrease vulnerabilities. To reinforce the efficiency of market operations, trust needs to be established among participants, transaction

20

2 Importance, Drivers and Implications of Liquidity Management

costs minimized, and rules set up to internalize externalities of two-party transactions. Trust can be reinforced by decreasing asymmetric information and providing the regulatory base for different expectations and beliefs of investors. Siddiqi (2001) claims that in an Islamic framework a real alternative to interest on loans is to finance companies on a profit-loss sharing basis. He recommends a shift from debt-based transactions to investment-based funding. Çizakça (2014b) gives two historical examples for stating the importance of risk sharing in the success of partnership-based contracts. The first example relates to the success of risksharing Mudarabah contracts in 13th century Venice, in which these contracts were used to finance new merchant class in foreign trade. The second is observed in late 19th century Germany where loan markets were replaced with capital markets to finance industry by using Mudarabah-like products and recorded huge success. Çizakça (2014b) derives that risk sharing had replaced risk shifting as the most important method of finance in Germany during this time. One of the earlier studies on efficiency of products of Islamic banks was conducted by Rosly and Bakar (2003), who found that Islamic banks that provide markup-based financing products are less likely to outshine mainstream banks regarding efficiency. They claim that these products lack moral content and providing these contracts does not increase the efficiency of Islamic banks. According to a survey by the IMF (2015a) sent to regulators in many countries, Islamic banking practices are relatively accepted in regulation, although this is not apparent to Turkey. Regarding clear recognition of Islamic banking, 21 of the 29 respondents (72%) indicated that the legal and regulatory framework explicitly recognizes Islamic banking practices, products or institutions. Concurrently, 76% of survey respondents reported that Islamic banking was being conducted by a standalone Islamic bank in their jurisdictions. It is revealed that 16 of the 29 respondents (55%) indicated that Islamic banking was being carried out by a conventional bank in their jurisdictions which means these countries give license for Islamic windows.11

2.3 Literature Discussion on Liquidity Stress Testing Stress testing is an important instrument for evaluating the resilience of bank portfolios and assessing their riskiness under certain assumptions. There are many different stress testing models used in the literature. One of the early adopters of stress tests at the beginning of the 1990s was J. P. Morgan, who used Value at Risk (VaR) methodology to quantity market risk (Zangari, 1996). In this simple method,

11 Islamic windows are special licences given to conventional banks to make Shariah-compliant banking.

2.3 Literature Discussion on Liquidity Stress Testing

21

first, certain assets in a portfolio were selected and then the forward rate was calculated. The next step was to estimate variances and covariances of each asset. Then Value at Risk was calculated by competing weights on assets. Using this method, Zangari (1996) estimated the potential loss over a specified period of movements in asset prices with a certain probability. Stress tests were originally designed for the context of Pillar 2 (BCBS, 2005). It was expected that these tests would help a bank to evaluate its performance and its capacity to mitigate risks and absorb losses. In the beginning, a micro prudential approach was dominant in formulating these tests. Indeed, stress testing was formulated and then tried; to test the idiosyncratic risks of banks under adverse scenarios. Since these tests were not designed with a macroprudential approach, banks, by implementing these tests could not see the contagion effect of their portfolio choices and therefore undermined the results. Banks used these tests to internalize the unexpected losses and protect the deposit insurance funds. Later, stress tests become more popular, and these tests were formulated as a more macroprudential tool by BIS (BIS, 2015). After the global financial crisis, first generation stress tests failed to adequately forecast interlinkages and the nexus between solvency risk and liquidity risk within and across banks (BIS, 2015). This low-quality stress testing could not show the increased vulnerabilities in many economies, and new models were developed to cover all interlinkages that may lead to contagion effects. For this reason, after the global financial crisis, supervisory authorities have been designing more sophisticated supervisory approaches for stress testing models and include liquidity to capture contagion risks in markets. Demekas (2015) suggests that there are essential differences between the traditional microprudential and the “new generation” macroprudential stress tests. Macroprudential stress tests introduce general equilibrium dimensions. In this sense, the results of the stress tests depend on idiosyncratic factors and systemic factors, as well as interactions with other financial institutions. For this reason, some researchers have developed macroprudential stress tests based on general principles hingeing on macro factors (Greenlaw et al., 2012; Borio et al., 2012). Macroprudential stress testing for the financial system is a functional instrument of central banks and supervisory authorities to assess the impact of marketwide scenarios. Such tests on liquidity risk can enhance insight into the systemic dimensions of liquidity risk. These exercises can also contribute to market participants’ awareness of systemic risks. However, at the beginning, liquidity risk was not included in second generation macro stress-testing models due to the quantitatively challenging multiple dimensions of liquidity risk (IMF, 2008) but later, a new stress-testing model for liquidity risk was developed by IMF staff (Čihák, 2014) and used by many banks and researchers (IFSB, 2015). Since internationally active financial institutions can be systemically important in their relationship with other banks, ensuring the soundness of each institution is necessary for the financial system as well as for the real sector which

22

2 Importance, Drivers and Implications of Liquidity Management

needs banking funds for investments. The occasional failure of individual institutions can be a problem for non-financial corporations too. Mitigating risk of these banks does not mean to mitigate entirely all risks in the financial transactions. Rather, it aims to prepare the parties to bear risks inherent in the model of financing. In this sense, a financial transaction is expected to be based on risk sharing between parties that have equal rights to access information. The important point is to maintain an environment in which financial institutions continue to be capable of providing intermediation services. Abdymomunov and Gerlach (2014) developed a new method for conducting stress testing in which they generate yield curve scenarios for banks’ exposure to the internal rate of return (IRR). They claim that their method gives yield-curve scenarios with a wider variety of slopes and shapes than other scenarios developed with the historical and hypothetical methods proposed in the literature by extracting data from only a large US bank (most probably a G-SIB bank). On the other hand, the last global financial crisis highlighted the importance of systemic risk intensely. The solvency of individual banks was not a good proxy for systemic risk. Instead, it was proposed as starting from an aggregate metric of systemic risk and measuring the value added by individual banks to this metric. Adrian and Brunnermeier (2011) use the conditional VaR (CoVaR) test in which the value at risk (VaR) of the financial system is a dependent on institutions under distress. They defined an institution’s contribution to systemic risk as the difference between CoVaR conditional on the system being under distress and the CoVaR in the median state of the system (Adrian and Brunnermeier, 2011). Moreover, liquidity stress testing should cover both market liquidity and funding liquidity. Goodhart et al. (2006) found that liquidity indirectly plays a role through the credit supply of banks to other banks and consumers while the default is endogenous within the system. Liquidity shortages may come from the bank’s liability side, due to depositor runs (liquidity issue) or withdrawals of interbank deposits (creditability issue). The last global financial crisis showed that market liquidity could evaporate very quickly, affecting the market value of banks’ assets which may trigger fire sales (selling assets at heavily discounted prices) and herding in an asset sale. Therefore, stable funding and the amount of cash outflow should be added to the stress testing. Pagratis et al. (2016) used a stress test methodology for bank liquidity risk and estimated the aggregate liquidity shortfall in the US commercial banking system at the height of the 2007–09 crisis. According to their test, although large banks experienced the greatest liquidity shortfall in the first quarter of 2008, small banks faced liquidity shortfalls before the crisis (Q4 2007). During this time, large time deposits constituted the fundamental funding vulnerability to the system, and government securities largely dominated other classes of liquid assets to be used as a liquidity backstop. The first effects of liquidity shortfall have been seen on securities and other liquid assets,

2.3 Literature Discussion on Liquidity Stress Testing

23

however banks do not systematically consider second round effects12 triggered after first round effects that can amplify losses (Van den End, 2012). These can be caused by reputational risks and collective responses of market participants, in which banks are not able to protect themselves against such risks. Using stress tests may decrease information asymmetries as shown by Quijano (2014) for US banks. Petrella and Resti (2013) examined the 2011 European stress test exercise and found that these tests produce important information for market participants. Banks with a large amount of liquidity have an incentive to under-provide liquidity to benefit from fire sales of assets (Acharya et al., 2008) if there is imperfect competition among banks. The interaction between funding liquidity and market liquidity is researched in the study by Brunnermeier and Pedersen (2008), who showed the relation between funding liquidity and market liquidity with modeling liquidity spirals. In this sense, market illiquidity increases funding constraints as a result of higher margins, enlarges shocks that hit traders, and induces them to reduce trading positions. Cornett et al., (2011) show that banks with more stable sources of funding are better able to continue lending. However, Covas and Driscol (2014) find negative and significant impacts of LCR requirement on loan growth by using the Dynamic Stochastic General Equilibrium (DSGE) model. It is claimed that behavioral responses are related to funding liquidity risks of banks (Aikman et al., 2009). Funding strains increase the default risk of banks which may occur at a certain stress level resorting to fire sales of assets. The funding liquidity risk is explained as an endogenous outcome of the interaction between market liquidity risk, solvency risk and the funding structure of banks (Gauthier et al., 2010). Later it was realised that the systemic risk in the financial system could be explained only by taking into account the network effects13 and liquidity risk (Gauthier and Souissi, 2012). Spill-over effects14 occur due to the network effects among banks. In the liquidity stress tester model of Van den End (2012) studies second round feedback effects and shows that the second round effects both have an impact on the LCR through additional haircuts on assets and net outflows. In one of the studies conducted by Diamond and Rajan (2005), it is found that a withdrawal of liquidity deteriorates aggregate liquidity shortages. Boss et al. (2006) have developed a system in which models for market and credit risk are brought together and connected to an interbank network module. This is similar to the framework developed by Alessandri et al. (2009) which also takes into account the

12 Second round effects are the effects that are seen after the LCR has lost its functions as a buffer. After the LCR becomes ineffective to mitigate negative externalities that arise because of adverse interactions of bank behaviour, the second round effects become more severe. 13 Multiple financial institutions are interconnected through an interbank network. The network effect is related with the effect coming from the interbank markets and interconnectedness (Gauthier and Souissi, 2012). 14 Spill-over effects are a secondary effect following a primary effect.

24

2 Importance, Drivers and Implications of Liquidity Management

effects of network spillover risks and asset-price feedback. Considering the impact of network spillover risks and liquidity risks, such effects may result in failure for some banks. Feedback effects arising from market and funding liquidity risk are missing in most macro stress-testing models of central banks. Such results are featured in models with margin-constrained traders, as in Brunnermeier and Pedersen (2008). They model two “liquidity spirals”, where, in the first, market illiquidity increases funding constraints through higher margins and then a shock to traders funding decreases market liquidity because traders decrease their positions because of margin requirements. Cerutti and Schmieder (2014) suggest that current stress tests suffer from a weakness in assessing potential risks embedded in banking groups’ geographical structures since these banks have consolidated balance sheets. Working on the Hong Kong banking sector, Wong et al. (2010) used a panel probit model to identify the leading indicators of banking distress and find macroeconomic fundamentals, currency crisis vulnerability, credit risks of banks and non-financial companies, asset price gaps and credit growth as important leading indicators (Wong et al., 2010). Financial system stress tests provide information on the behaviour of the system under shocks. These tests are helping practitioners to evaluate the significance of the system’s vulnerabilities. They also help to improve a forward-looking perspective and analyze the financial system as a whole while mitigating risks more effectively. In this context, the EU-wide stress test assesses the resilience of financial institutions in the EU to adverse market developments and evaluates the potential for systemic risk to increase in situations of stress. Schmieder et al. (2012) argue that a natural counter-balancing role is played by central bank funding. In the case of a severe crisis, central banks can act as a lender of last resort. For instance, some central banks entered swap agreements with the Fed (United States Federal Reserve Bank) during the global financial crisis in which they could supply their domestic banks with much-needed dollar funding.15 In fact, the Fed became the global US$ lender of last resort during the crisis, providing liquidity to large international banks such as Barclays and UBS, as well as to domestic US financial institutions. For Islamic banks, this issue is rather complicated. Theoretically, if Islamic banks were to have a liquidity problem, big central banks could not provide liquidity because of non-availability of Shariah-compliant products to trigger a last resort mechanism in many countries – including US, EU, UK and many others.

15 See more details in https://www.federalreserve.gov/monetarypolicy/bst_liquidityswaps.htm.

2.4 Basel Liquidity Regulations and the Importance of the LCR

25

2.4 Basel Liquidity Regulations and the Importance of the LCR The financial crisis of 2008 showed that liquidity is essential for the proper functioning of financial markets and the banking sector (Bonner, 2014; BIS, 2013c). Before the crisis, thanks to a low level of monetary policy rate in developed markets, funding rates were very low, and funding liquidity was very high. After the crisis, market conditions changed very rapidly, and market liquidity and funding liquidity evaporated within days and hours. According to the BIS (2013c), the primary objective of the LCR is to promote the short-term resilience of the liquidity risk profile of banks. The LCR is expected to improve the banking sector’s ability to absorb shocks arising from financial and economic stress. The LCR is designed to act as a buffer against liquidity run. The LCR was developed as a minimum level of liquidity for internationally active banks (BIS, 2013c). It is proposed that the LCR will contribute to banking liquidity as well as market liquidity. Banks are expected to meet this standard, but national regulatory authorities may set higher minimum levels of liquidity. Many countries including Turkey have already started to implement the LCR since 1 January 2015 with the implementation of a minimum requirement at 60%, which will rise in equal annual steps to reach 100% on 1 January 2019. This progressive approach was recommended by the Basel Committee though many countries are already applying a higher level of LCR. A bank should have an adequate stock of HQLA for surviving 30 days of the stress scenario, which consists of cash or assets that can be easily converted into cash with acceptable loss. The LCR has been defined by the Basel committee with this formula: LCR =

High Quality Liquid Assets Net Cash Outflows

HQLA is divided as Level 1 and Level 2 assets. Level 1 assets consist of cash, bank reserves, central bank reserves and securities issued or guaranteed by governments, central banks or multilateral development banks (MDBs). There are two types of Level 2 assets: Level 2A and Level 2B. These assets consist of public sector entity bonds (PSE) that have high quality, covered bonds and securities issued or guaranteed by MDBs. Net cash outflows are calculated as the difference between cash outflow and cash inflows. Net cash outflows are related to funding liquidity. There is a misunderstanding regarding the LCR’s implementation which was developed as a minimum level of liquidity for internationally active banks, and not for all local banks (BIS, 2013c). Haan and End (2013) claim that most banks hold more liquid assets than strictly required by standards. Table 2.1: shows the LCR’s implementation deadline. The primary objective of LCR is to keep funding liquidity and market liquidity stable. The LCR is equal to HQLA/Net Cash Ouflows (NCO). Since the target level of LCR is 100% (minimum level will be 100% after 1 January 2019), it is assumed that LCR would be equal to 1. In this context, it is possible to accept that the LCR = 1.

26

2 Importance, Drivers and Implications of Liquidity Management

Table 2.1: The LCR Implementation Deadline.

Minimum LCR

 January 

 January 

 January 

 January 

 January 

%

%

%

%

%

Source: BIS.

And then, the following equation is written: LCRðtÞ =

HQLA =1 NCO ðCOF ðtÞ − CIF ðtÞÞ

(1)

If LCR = 1 and HQLA = LA, after making some adjestments, such equation is derived: COF ðtÞ − CIft ≤ LAt

(2)

Then, this equation is rewritten as such: COF ðtÞ ≤ CIft + LAt

(3)

Equation 3 shows that there is a relation between three variables. COFðtÞ means cash outflows, whereas CIf t is cash inflows and LAt means liquid assets. This equation indicates that the LCR needs to be higher than 100% of net cash outflows. For Islamic banks, the right side of the equation should be naturally larger than the left aspect of the equation since Islamic banks take deposits from customers for both investment and saving purposes. Savings accounts are more prone to cash outflows while investment accounts, theoretically, should be stable. For this reason, the 100% level of the LCR requirement may not be appropriate for Islamic banks.

2.5 Liquidity Types Liquidity can be defined as the ease of converting assets to cash or the equivalent. Cash, central bank reserves, and bank reserves are considered as liquid assets. IMF (2015b) defines market liquidity as the ability to rapidly execute sizeable securities transactions at a low cost and with a limited price impact. The Basel Banking Committee contemplates instruments as liquid if they can be converted into cash with maximum 10% loss. There are three kinds of liquidity: funding, monetary and market liquidity (Yildirim, 2009; Kerry, 2008). Funding liquidity refers to the bank’s ability to mobilize from its interbank or deposit base easily. Monetary liquidity is related to monetary policy and shows the stance of monetary conditions. Market liquidity is related

2.6 Liquidity Regulations and Islamic Banks

27

to market conditions – the illiquidity of the market reduces the efficiency in the intermediation costs and can potentially inhibit economic growth (Kerry, 2008). Market liquidity can be fragile, that is, prone to evaporation in response to shocks. It is expected that if there are efficient and transparent market infrastructure, market liquidity will be high, which will result in low transactions costs. The asset side of the balance sheet is related to the market liquidity, and the liability side of the balance sheet is related to the funding liquidity risk (Brunnermeier and Pedersen, 2008). Theory suggests that market liquidity can affect funding liquidity in stabilising and destabilising manners (Boudt et al., 2013). Liquidity contagion across markets can occur, and regulations may decrease market liquidity. Tightening of monetary conditions might affect market liquidity as well as funding liquidity. In general, liquidity risk is considered as a determinant of other risks, such as credit risk (Bissoondoyal and Treepongkaruna, 2011) or a determinant of bank performance (Arif and Anees, 2012).16

2.6 Liquidity Regulations and Islamic Banks According to IFSB (2015a), all supervisory authorities should put into effect the LCR requirement for Islamic banks starting from 2015 with the minimum level of 60%. The IFSB regulation requires an Islamic bank to find enough stock of unencumbered high-quality liquid assets that can be liquefied easily with no loss, or slightly acceptable loss. In this sense, when an asset is sold in in the market, 10% of loss is accepted as a threshold for treating these assets as liquid assets. According to IFSB (2015a), the LCR would be composed of Stock of Shariahcompliant HQLA and total net cash outflows calculated for the next 30 calendar days. Total net cash outflows are calculated as total gross expected cash outflows minus lesser of total expected cash inflows or 75% of total expected cash outflows. IFSB (2015b) sets the LCR based on a scenario in which there are both idiosyncratic and market-wide shocks. According to the IFSB (2015b), the HQLA are divided into two broad categories, or levels, almost the same as the Basel LCR standards: Level 1 and Level 2. Level 1 assets can constitute an unlimited share of the pool and are not typically subject to a haircut under the LCR. These assets are coins, banknotes, central bank reserves, Sukuk and other Shariah-compliant marketable securities issued or guaranteed by sovereigns, central banks, public sector entities, multilateral development banks (MDBs) or relevant international organizations such as the International Islamic Liquidity Management Corporation (IILM) which are assigned a 0% risk-weight under IFSB-15. In this sense, coins and banknotes are historically

16 Drehmann and Nikalaou (2009) show that funding liquidity risk can be measured by using publicly available data.

28

2 Importance, Drivers and Implications of Liquidity Management

liquid assets if they are local-currency denominated. Central bank reserves include required reserves as well as reserves kept by banks to the extent that the central bank policies allow them to be drawn down in times of stress. Moreover, the IFSB accepts the IILM Sukuk as Level 1 liquid asset, although many central banks treat it as Level 2A or Level 2B liquid asset. Moreover, the IFSB receives securities issued by public sector entities in the same category with securities issued by sovereigns. Furthermore, Sukuk and other Shariah-compliant negotiable instruments issued by sovereign or central banks that have a non-zero risk-weight, but are issued in domestic currencies by the sovereign or central bank in the country in which the liquidity risk is being taken, or in the IIFS’s home jurisdiction, are accepted as Level 1. These should be traded in a market characterized by a low level of concentration and regarded as “a reliable source of liquidity at all times. ” IFSB (2015b) has not given a reason for insisting on “a reliable source of liquidity at all times” for Islamic banks. Moreover, the IFSB advises banks to issue Ijarah-based Sukuk because the IFSB requires that securities have secondary markets and, at present, only Ijarah-based Sukuk can be traded in the secondary market. Level 2 assets compromise Level 2A and Level 2B assets as permitted by the supervisory authorities. Level 2A assets are subject to a 15 percent haircut applied to the current market value of each asset and limited to the Shariah-compliant marketable securities/Sukuk issued or guaranteed by sovereigns, central banks, PSEs, MDBs or relevant international organizations, which are assigned a 20 percent riskweight under IFSB (2015b); i.e. Shariah-compliant securities and Sukuk that satisfy conditions stated by the IFSB (2015b). In this provision, the IFSB mentions the internal rating which can mean that Islamic banks can use internal approaches instead of “standard approaches” if their regulatory bodies approve. These assets should be traded in a deep secondary market and be accepted as liquid assets, which mean these assets’ price should not decrease by ten percent over a 30-day period. The Level 2B assets are Sukuk and other Shariah-compliant securities backed by commodities and other real assets that satisfy conditions stated by IFSB (2015b), subject to a 25% haircut. However, Sukuk and other Shariah-compliant securities that meet all of the requirements stated by the IFSB (2015b) may be included in Level 2B, subject to a 50% haircut. Also, Shariah-compliant equity shares that satisfy all conditions may be included in Level 2B, subject to a 50% haircut if they meet all conditions; such as, not issued by a financial institution or any of its affiliated entities; exchange-traded and centrally cleared; a constituent of the major stock index in the home jurisdiction or where the liquidity risk is taken; denominated in the domestic currency of an IIFS’s home jurisdiction or in the currency of the jurisdiction where its liquidity risk is taken and being traded in a capital market characterized by a low level of concentration and being regarded as a reliable source of liquidity at all times. Sukuk and other Shariah-compliant marketable securities issued by sovereign or central banks rated BBB+ to BBB- that are not included in Level 1 assets may be included in Level 2B assets with a 50% haircut.

2.6 Liquidity Regulations and Islamic Banks

29

As another option, Islamic banks may use Alternative Liquidity Approach (ALA) treatments only when there is evidence of a genuine shortfall in the HQLA in the domestic currency after being decided by the regulatory authority. In this case, banks have three options. The first choice is to use contractually-committed liquidity facilities from the relevant central bank with a fee. This facility can be used under Wakalah, Mudarabah or Commodity Murabahah contracts. There are certain Shariahrelated issues in this matter that should be handled very carefully. The second option is to use foreign currency denominated HQLA to cover domestic currency liquidity needs. In this option, the Islamic Development Bank (IDB) Sukuk or IILM Sukuk can be used. The last option is to use additional Level 2 assets with a higher haircut decided by regulatory bodies. Fuhrer et al. (2017) share that Swiss banks are permitted to use either option two or three to fulfil their requirements even if the approach was not expected to be used by advanced economies. Unfortunately, many Islamic banks are not permitted to use such options.

3 Liquidity Tools for Liquidity Management in Islamic Banks There are several tools for liquidity management in Islamic banks such as Sukuk, Tawarruq, participation papers, Commodity Murabahah (Dusuki, 2007), Cagamas Mudarabah bonds, interbank deposits, interbank financing, interbank Murabahah, central bank Wadiah acceptance, Islamic repo and interbank Wakalah (Verhoef et al., 2008). Malaysian Islamic banks are very fortunate that they have some liquidity management tools. However, PBs in Turkey do not have such a wide choice of instruments. They have to invest in Sukuk, Commodity Murabahah or stay in cash. Obviously, these tools are not adequate for ensuring an efficient risk management of Turkish PBs because of the supply and demand gap of these instruments. According to IFSB research (2014), high-quality liquid assets of seven countries mainly consist of Level 1 assets (Table 3.1.).

Table 3.1: Islamic Banks HQLA Stocks (World). Assets Class

Share

Level  Assets

.

Cash and Cash Equivalents Total Central Bank Reserves % Zero Risk-weighted Assets Other Sovereign Assets which don’t have % Zero Risk-weight

. . . .

Level A Assets

.

Guaranteed by Public Authorities Corporate Bonds (AA and higher) Covered Bonds (AA and higher)

. . 

Level B Assets

.

Total B Assets Other B Assets

. .

Source: IFSB, 2014.

This research shows that Islamic banks store 30% of their liquid instruments in central banks account or cash account. One of the reasons may be to protect themselves against market liquidity and constraints related to regulatory standards. Many metrics are used for evaluating the liquidity of assets including bid-ask spread (the difference between demand price and supply price of an asset), turnover ratio (the percentage of an asset that has been replaced at a given time) and Amihud illiquidity https://doi.org/10.1515/9783110582901-003

3.1 Liability (Deposit) Management

31

defined and formulated by Amihud et al. (2005). By applying these metrics, the liquidity level of assets is determined. Liquidity management requires proper management of liabilities and assets, and there are various ways of matching liabilities with assets. The following section will discuss how these are managed and what tools and instruments are used to do this crucial function, including regulatory requirements and deposit insurance.

3.1 Liability (Deposit) Management Deposits of Islamic banks, which are on the liability side of the balance sheet, are mainly of three classes of accounts: demand, saving and investment deposits. First class is current account deposits, similar to demand deposits, are guaranteed in capital value. In the current account deposits, the bank provides safe custody, Amanah (or safekeeping), checks and other services such as drawing money on demand. Demand deposits are not entitled to any bank’s profit, and the bank uses demand deposits at its own risk, but they keep a small portion of these deposits as legal reserve at the central bank. Saving deposits, on the other hand, can be withdrawn on demand. Some saving deposits may share profits on the basis of a minimum balance maintained within a specific period of time from time to time, and the provisions of maintaining legal reserves are sometimes applied to saving deposits. Lastly, investment deposits are based on the unrestricted Mudarabah contract between the depositor and the bank, in which the bank is authorized to use it for any investment project not prohibited by the Islamic principles. Investment deposits are not guaranteed in capital value, and do not yield a fixed rate of return. Instead, profit or losses from the bank’s operations are distributed according to negotiated proportions. Profits are distributed either at maturity or sometimes advances are paid to depositors at regular intervals and adjustments are made at maturity. Legal reserves are not kept against investment deposits since the bank cannot guarantee them. Sometimes, special purpose investment deposits, which operate on restricted Mudarabah (on specific investment operation), are managed by the bank and profit and losses are distributed according to the agreed ratio. Thus, when the liquidity management on the liability or deposit side of the Islamic bank’s balance sheet is discussed, the matter is to meet the demand for cash outflows. In general, all deposits by their nature are short term with a maximum period of one year or less. Even in the case of investment accounts, the depositors have an option to withdraw the money before its maturity. It may be noted here that deposits are made not with an objective of getting regular income but for capital appreciation with profit/loss sharing arrangement. There is a possibility of pre-mature withdrawal by account holders when there is a mismatch between investor’s expectation of return and the actual return. Thus the Islamic banks are required to keep

32

3 Liquidity Tools for Liquidity Management in Islamic Banks

adequate cash or cash equivalents to meet the demand. But in practice, majority of these depositors generally renew their deposits after the expiry period. For all practical purposes, these deposits are, thus, in effect, medium or long term. In a perfect liquidity management system, it may be theoretically possible to estimate the likely demand for withdrawal. If this estimation is nearly perfect, these banks need to keep only that estimated portion as liquid cash or near cash items. However, most of the Islamic banks do not follow this principle and keep a lot of excess money in liquid assets.

3.2 Financing Asset Management On the asset side, the evaluation of liquidity risk management takes into account Islamic banks’ efforts to monitor financing, arrange proper financing allocation, tackle financing default under unpleasant economic conditions and handle liquidity shortages (Ismal, 2010). One of the vital issues is to choose a preferred method of financing. Long-term financing or short-term financing can be very effective in liquidity management. By choosing a rational method of financing, Islamic Financial Institutions can mitigate their liquidity risk. Murabahah is chosen overwhelmingly by banks even though it is not re-tradable in the secondary markets before maturity – it lends itself to any fixing of maturity at the time of contract. This is a feature which is not possible in Mudarabah or Musharakah contracts where accrual of profit to the bank is tied with the timing of the project cycle.

3.3 Bank Reserves and Cash The reserve (statutory) requirement is a central bank policy that sets the minimum fraction of customer deposits and notes that commercial banks must hold as reserve. These required reserves are usually in the form of cash stored physically with the bank or deposits made with the central bank. The bank reserve requirement is a tool for monetary policy operations. By changing the required reserve ratio, the central bank can change the value of money multiplier as well as the monetary base which is used to control total credit in an economy. When the central bank wants to tighten the money supply (contractionary monetary policy) in the economy, it raises the reserve ratio, which results in the rate in the market to go up and thus the country’s borrowing reduces. In case of expansionary monetary policy, the reserve ratio is lowered to allow the money supply to expand. In this context, Islamic banks should have different required reserve scheme than the mechanism used for conventional banks. If a central bank gives interest for required reserve to the banks, it should also provide alternative tools for Islamic banks.

3.5 Interbank Deposits and Loans

33

3.4 Deposit Insurance Deposit insurance is a measure implemented in many countries to protect bank depositors from losses caused by a bank’s inability to pay its obligations when it is due. Deposit insurance is an integral component of an effective financial safety net that promotes financial stability and provides depositors with protection. It is required for any country to have some form of deposit insurance as due to the current fractional-reserve banking system. It is claimed that deposit insurance promotes confidence in the financial system by protecting depositors against loss of their deposit and also strengthens the existing regulatory and supervisory framework by providing incentive for sound risk management in the financial system. Moreover, deposit insurance contributes to the stability of the financial system by dealing with bank failure expeditiously and reimbursing depositors. A deposit insurance scheme is likely to increase deposits in the country’s banks both from local and foreign depositors, which can increase the rate of foreign investment in the country. Moreover, deposit insurance enables banks to increase its money supply. On the other hand, it is suggested that the deposit insurance system can raise a moral hazard and agency problem in the economy. Existence of such a scheme is likely to encourage excessive risk taking on the part of depositors as well as the banks accepting the deposits (Schich, 2008). The concept of deposit insurance in the Islamic financial industry is relatively new. Currently, only Malaysia and Sudan have implemented an Islamic deposit insurance system for Islamic banks. In Turkey, an Islamic deposit insurance system was developed in 2001, however it was absorbed into the conventional system in 2005 (Arshad, 2011). Malaysia has applied the Deposit Insurance System since September 2005 and all commercial and Islamic banks, including foreign banks operating in Malaysia, are compulsory member of this system. The Shariah-compliant design is based on an arrangement of guarantee with fee or Kafalah bil ujr. In 2010, this system was enlarged to include Takaful and insurance companies in order to protect investors in the event of an insurer member failure.

3.5 Interbank Deposits and Loans Conventional banks can easily convert their short-term surplus funds into liquid assets or cash. These monetary assets are short-term instruments which can be easily liquidated. On the other hand, Islamic financial institutions cannot deal with conventional money markets, and therefore, cannot use these interest-based money market instruments. IBs need financial instruments which are, on the one hand consistent with Shariah legal opinion, and on the other, allow Islamic banks to achieve the two divergent objectives of liquidity and profitability, especially in the

34

3 Liquidity Tools for Liquidity Management in Islamic Banks

short-term. Since Islamic banks do not have access to an Islamic money market, they need alternative mechanisms for adjusting their portfolios over the short-term, and secondly, have a channel for the transmission of monetary policy. Islamic banks have the legal right to establish mutual inter-bank reciprocal loan agreements in several jurisdictions in which there is interbank market for these banks. Although there are many problems in terms of the Islamicty of these schemes, the participants of the system benefit from short-term Mudarabah funds and all remaining volume is sterilized. At the international institutional level, there has been some progress in the form of the Liquidity Management Centre in Bahrain that have mechanism for its participants to use reciprocal funds for their liquidity risk management. As is well known, central banks have lending and borrowing facilities for the banks under their purview and can do outright sell and buy of a defined group of instruments for monetary policy purposes if it requires. In order for a bank to borrow from a central bank, it first needs to have a certain amount of collateral. The Bank Negara Malaysia has the first Islamic Interbank Market in the world which tries to facilitate Islamic liquidity management among its Islamic banks in its jurisdictions. Many short-term and long-term instruments have been used in this market for liquidity risk management of Islamic financial institutions. This system tries to manage mismatch between liability and assets of Islamic banks. In this context, financial instruments and interbank investment are used for surplus banks to channel funds to deficit banks, thereby maintaining the funding and liquidity mechanism necessary to promote stability in the system. However, there are many issues regarding the operational mechanism of this system and in building an Islamic money market. First of all, in a truly Shariah-compliant banking system there would be no need to have a money market since the capital market would have functions to manage liquidity surplus or deficit. Secondly, there is no real transaction based on assets in the money market, which is hypothetically based on demand and supply of money. Lastly, if Islamic banks would have efficient and effective liquidity management, they would manage their liquidity matches in a perfect way and would not need any money market.

3.6 IILM Sukuk The nine central banks and one multi-development bank (Indonesia, Kuwait, Luxembourg, Malaysia, Mauritius, Nigeria, Qatar, Turkey and the United Arab Emirates as well as the Islamic Development Bank) have become founding members of the International Islamic Liquidity Management Corporation (IILM) in 2010. The IILM focuses on issuing adequate, high-rated and liquid supply of Shariahcompliant instruments to facilitate efficient and effective liquidity management solutions for the Islamic financial services (Aysan et al., 2013). The main goal of the IILM is to enhance cross-border investment flows and improve the financial

3.6 IILM Sukuk

35

stability of its member countries by issuing short-term Sukuk which can be used by Islamic banks for liquidity management. Moreover, the IILM acquires sovereign assets from its member countries and finance-member countries. Furthermore, Sukuk issued by the IILM might be accepted and used efficiently by central banks in their monetary policy applications. Table 3.2 compares IILM Sukuk with selected liquidity tools. In principle, central banks accept the IILM Sukuk as eligible collateral and provide the relevant regulatory treatment for liquidity coverage ratio and capital adequacy ratio. It can also be used as reverse-repo instrument in tandem with local Islamic securities for the purpose of sterilization of ample liquidity in both the Islamic and conventional banking systems. The IILM Sukuk can be traded on secondary markets efficiently through a strongly established chain of primary dealers. Table 3.2: Comparing Selected Liquidity Tools. Instruments

Tenor

Currency

Tradability

Rating

Central bank Islamic Instruments Short-term Sukuk Commodity Murabahah Interbank Mudarabah Interbank Wakalah Islamic repo IILM Sukuk

 months- year  months- year  week- months Overnight –  month – months Overnight- month  months- year

Local Local Any Any Any Local US$

Limited Limited No No No No yes

Unrated Unrated Counterparty Counterparty Counterparty Unrated A

Hence, through the establishment of a liquid secondary market over time, the IILM Sukuk can help facilitate the short-term flow of funds. Being issued in reserve currencies like US$, the IILM Sukuk can also represent a solid alternative as a foreign currency market instrument. It is expected that the IILM Sukuk will be used as an Islamic benchmark yield curve in the future. Table 3.3 shows a comparison between sovereign Sukuk and the IILM Sukuk. Although sovereign Sukuk has low risk weight for capital adequacy calculations Table 3.3: Comparison between Sovereign Sukuk and the IILM Sukuk. Sovereign Sukuk

The IILM Sukuk

Liquid % Risk Weight Deep Secondary Market High Volume Long-term Local Currency Local Market Eligible for Collateral

Liquid % Risk Weight New secondary market High Volume Short-term US Dollar International Market Eligible for Collateral

36

3 Liquidity Tools for Liquidity Management in Islamic Banks

and is accepted as liquid instruments for open market operations, the IILM Sukuk has been treated with 20% risk weight and several central banks accept the IILM Sukuk as an eligible instrument for open market operations.

3.7 Commodity Murabahah (GCC & Malaysia) Commodity Murabahah is a popular mode of investment used by Islamic banks to address their liquidity problems, though it is criticized in many aspects. It is a form of a short-term finance based on the Murabahah contract and generally used for buying and selling commodities in the international market. Commodity Murabahah is a widely-used technique in managing short-term liquidity by Islamic banks in GCC countries – especially Saudi Arabia, Bahrain and the United Arab Emirates. In the GCC, the instrument is based on commodities traded in the London Metal Exchange (LME). The Central Bank of Malaysia (Bank Negara Malaysia) introduced its Commodity Murabahah Program (CMP) as part of its initiative to support the development of Islamic Finance in the country. The CMP was officially endorsed as a permissible instrument to be used in the financial market and Islamic Interbank Money Market (IIMM) by the Shariah Advisory Council of the Central Bank of Malaysia on 28 July 2005, and was subsequently introduced in February 2007. The main purpose of this program was to offer Islamic financial institutions a new instrument in managing liquidity in the IIMM (BNM, IIMM, Bank Negara Malaysia, 2014). The CMP is based on the Tawarruq contract to enable liquidity management in Islamic Banks, using Crude Plam Oil (CPO) as the underlying asset (Dusuki, 2007). The CMP is the first ever commodity-based transaction that uses CPO as its underlying asset. In CMP, the commodity is traded on a spot basis with 100% payment of the purchase price. The purchased commodity is then sold to a third party on a Murabahah (cost-plus sale) basis for a deferred payment with an assigned maturity period, with spot delivery of the sold commodity. Dusuki (2007) suggests that the CMP allows avenues for Islamic banks and the Central Bank of Malaysia to effectively manage bank liquidity and at the same time give vibrancy to the Interbank Money Market (IIMM). As a result, Islamic banks that have excess liquidity can manage their funds productively, while any banks facing temporary liquidity problems due to mismatch between assets and liabilities can also benefit from this CMP instrument. Commodity Murabahah is accepted among a few jurists. Table 3.4 compares commodity Murabahah with the IILM Sukuk. According to a reviewed ruling by the Islamic Fiqh Academy in December 2013, the Academy clarified its stand on Tawarruq by distinguishing and classifying between “real Tawarruq” (Tawarruq haqiqi) and “organized Tawarruq” (Tawarruq munazzam). Organized Tawarruq as practiced by most banks are deemed to be synthetic and fictitious as Bai’ al-Inah and hence disallowed (Dusuki, 2007).

3.9 Mudarabah Interbank Investment

37

Table 3.4: Comparison between Commodity Murabahah and the IILM Sukuk. Commodity Murabahah

The IILM Sukuk

Illiquid % or higher Risk Weight No Secondary Market Up to the issuance Short-term or long-term US, EURO or local currency Local or international Market Not eligible for Collateral Not eligible for Reserve Management

Liquid % Risk Weight New secondary market High Volume Short-term US Dollar International Market Eligible for Collateral Eligible for Reserve Management

3.8 Central Bank Wadiah Acceptance This product basically is used in Malaysia by BNM to place Islamic banks’ surplus in the central bank on the concept of Al-Wadiah, which means to place an institution’s money in a safekeeper. In this mechanism, BNM acts as the custodian for the banks’ funds under several conditions specified and accepted by both parties. BNM does not have any legal requirements to pay any return to the bank. However in practice BNM pays return which is equal to market return. The return paid by the BNM is perceived as a gift, which is called Hibbah. Central bank Wadiah acceptance is one of the open market operation tools used by BNM to absorb excess liquidity from Islamic banking sector. Moreover, it is useful for excess liquidity holder to place its excess liquidity instead of staying in cash without any return. However, this product is very controversial because it is very similar to conventional open market operation tools. In contrast, Indonesia also uses the Wadiah acceptance and the rate of return is tied to market rates, which are in turn tied to recent realized profits. The rate of bonus of SWBI is the lower than the rate of return of the Islamic interbank money market and the rate of return of a Mudarabah time deposit. The rate of return is decided by the market instead of central bank policy setters in Indonesia.

3.9 Mudarabah Interbank Investment Mudarabah Interbank Investment is used as an interbank short-term liquidity management instrument by Islamic banks in Malaysia. It is claimed that Mudarabah Interbank Investment is an alternative to the conventional money market, with the activities based on a profit-sharing principle, as opposed to interest on borrowing/

38

3 Liquidity Tools for Liquidity Management in Islamic Banks

lending activities in the conventional money market. It refers to a mechanism whereby a deficit Islamic banking institution can obtain investment from a surplus Islamic banking institution, based on Mudarabah (profit-sharing) to source, invest and square their short-term fund. Much of the money that is lent in the interbank market is for very short periods of time, ranging from overnight to 12 months. In this mechanism, banks with surplus funds are the “rabbul-maal” and the central bank is the entrepreneur. The rate of return is based on the rate of gross profit before distribution for investment of one-year of the investee bank. The profit-sharing ratio is negotiable among both parties. It is not possible to decide the rate of return at the time of negotiation. It should be realized at the end of the period. The principal invested shall be repaid at the end of the period, together with a share of the profit arising from the use of the fund by the investee bank (BNM, 2014).17 Unfortunately, the LIBOR benchmark is used to determine the profit rate. BNM has introduced this due to banks’ tendency to under-report profits and thus pay less to the bank that supplied the funds.

3.10 Sale and Buy-back Agreements From the Shariah view point, repos are no more than short-term loans being provided by an investor to the dealer (issuer) and guaranteed by certificates under consideration. In addition to this, the ownership of the certificate is not transferred in all cases to an investor; it remains with the dealer. Hence, the transaction is merely a provision of money for a short term at interest. For this reason, an alternative contract is needed, which is the Sale and Buy Back Agreement (SBBA). In this contract there are two parties. First one is seller of the instruments and second one is the buyer of instruments. Both parties agree on price at the spot and then both parties sign a different agreement in which the buyer promises to sell the product back to the seller and the seller promises to buy the poduct from the buyer. If there is deep and developed secondary market, both parties can manage their liquidity risk without facing big losses. Although this is not the best solution of the challenge and includes many problems, Islamic banks can use this mechanism if they do not have any better mechanism as an alternative to conventional REPO money market instrument. However, supervisiory authorities should supervise and review these transactions very carefully. Otherwise, in practice, these products can be very similar to conventional products.

17 http://iimm.bnm.gov.my/

3.14 Islamic Accepted Bills (IAB)

39

3.11 Government Investment Issue (GII) This is one of the first instruments used by central banks to facilitate the liquidity risk management of Islamic banks in Malaysia. The Malaysian Parliament passed the Government Investment Act in 1983 to enable the Government of Malaysia to issue non-interest bearing certificates known as Government Investment Certificates (GIC) which have now been replaced with Government Investment Issues (GII) which was introduced under the concept of Qard al-Hasan. This Qard al-Hasan mechanism does not satisfy the GII as tradable instruments in the secondary market. The price at which the papers are sold or purchased by the players is determined by BNM, which maintains a system to record any movement in the GII. On 15 June 2001, the Government of Malaysia with advice from Bank Negara Malaysia, issued a 3-year GII of RM2billion (US$514.49 million) under a new concept of of Bai’ Al-Inah. This instruments is tradable and has very deep secondart market under the concept of debt trading (Bai’ ad-Dayn), which is seen very controversial by many scholars.

3.12 Bank Negara Monetary Notes-i (BNMN-i) These notes are securities issued by Bank Negara Malaysia to replace the Bank Negara Negotiable Notes (BNNN) to facilitate the liquidity management of Islamic banks but all Islamic financial institutions can invest in these instruments. The instruments have maturity from one year to three years.

3.13 When Issue (WI) When Issue is a transaction of sale and purchase of debt securities before the securities are issued. In such cases, the securities are traded when they are announced. However, the transaction is settled only after the issuance of the securities.

3.14 Islamic Accepted Bills (IAB) Islamic Accepted Bills is an alternative to conventional bills of exchange, which was invented to encourage both domestic and foreign trade through an Islamic financing mechanism. Islamic Accepted Bills is formulated on the Islamic principles of Murabahah and Bai’ al-Dayn. These products are used to finance imports and exports as well as local sales and purchases. Imports and local purchases: Under this category, financing against import bills is provided under the Murabahah mode of financing, where the customer as

40

3 Liquidity Tools for Liquidity Management in Islamic Banks

an agent of the bank purchases the required goods. At the time of acceptance of documents, the bank sells the goods to its customer on a deferred payment basis. The bank pays the seller at maturity of acceptance. Upon maturity of Murabahah financing, the customer pays back the bank the cost of goods plus profit margin. Exports and local sales: The sale of goods by the bank to the customer on a deferred payment term constitutes the formation of debt. This is securitized in the form of a bill of exchange drawn by the bank on, and accepted by, the customer for the full amount of the bank’s selling price payable at maturity. If the bank decides to sell the Islamic Accepted Bill to a third party, then the concept of Bai’ alDayn applies, whereby the bank will sell the instrument at the agreed price. An exporter who had been approved for an IAB facility will prepare the export documentation as required under the sale contract or letter of credit. The export documents shall be sent to the importer’s bank. The exporter shall draw on the commercial bank a bill of exchange and this will be the IAB. The bank shall purchase the IAB at a mutually agreed price using the concept of Bai’ al-Dayn and the proceeds will be credited to the exporter’s account. This concept of Bai’ alDayn is being used in Malaysia only and has little acceptance with majority of Shariah scholars. Mostly, scholars allow sale of debt at par value only, whereas the IABs in Malaysia are being sold at a discount.

3.15 Islamic Negotiable Instruments (INI) “Islamic version” of Negotiable Instruments of Deposits (NID) was introduced in 1993 in Malaysia by the BNM based on two different structures. First, one is negotiable Islamic Debt Certificate, which is structures according to deferred payment sale. These contracts are permissible only in Malaysia. Second contract is called Islamic Negotiable Instruments of Deposits, in which the “concept” of Mudarabah is applied. In this mechanism, to a sum of money deposited with the Islamic banking institutions and repayable to the bearer on a specified future date at the nominal value of Islamic Negotiable Instruments of Deposits plus declared dividend. This contract is very controversial and accepted only in Malaysia.

3.16 Ar-Rahnu Agreement-I (RA-I) Under this product, the lender provides a loan to the borrower without expecting any return based on the concept of Qard al-Hasan. The borrower pledges its securities as collateral for the loan granted. The RA-I is an instrument available to the Central Bank of Malaysia (BNM) as a liquidity management tool for its money market operations approved in 2003. If the borrower fails to repay the loan, the

3.20 Bai’ Al-Inah

41

lender has the right to sell the pledged securities and use the proceeds from the sale of the securities to settle the loan. In practice, the borrower pays an amount of return as a gift to the lender based on on the average interbank money market rate.

3.17 Sukuk Bank Negara Malaysia Ijarah (SBNMI) One of the widely used instruments is the BNM Sukuk Ijarah, which is based on the “sale and lease back” concept, a structure that is widely used in the Middle East and in many other countries. In this structure there is a special purpose vehicle, BNM Sukuk Berhad established by BNM to issue the Sukuk Ijarah based on BNM’s own assets. The assets are first sold to the SPV and then leased back to BNM for rental payment consideration, which is distributed to investors as a return on a semiannual basis. This instruments is almost similar to Sukuk. Upon maturity of the Sukuk Ijarah, which will coincide with end of the lease tenure, BNM Sukuk Berhad will then sell the assets back to Bank Negara Malaysia at a predetermined price. Bank Negara Malaysia issues this instrument on a regular basis with subsequent issues.

3.18 Cagamas Mudarabah Bonds (Malaysia) The Cagamas Mudarabah Bond was introduced on 1 March 1994 by Cagamas Berhad to finance the purchase of Islamic housing debt from financial institutions which provide Islamic house financing to the public. The Cagamas Mudarabah Bond is structured using the concept of Mudarabah where the bondholders and Cagamas will share the profits according to the agreed profit-sharing ratios. These products are used by Islamic banks for their liquidity mismatches.

3.19 Islamic Private Debt Securities Islamic Private Debt Securities (IPDS) was first introduced in Malaysia in 1990. The IPDS which are currently outstanding in the market were issued based on the Shariah-compliant concept of Bai’ Bithaman Ajil, Murabahah and Mudarabah.

3.20 Bai’ Al-Inah Bai’ al-Inah refers to the selling of an asset by an Islamic bank to the customer through deferred payments. At a later date, the Islamic bank will repurchase the asset and pay the customer in cash terms. This is not a widely practiced product and

42

3 Liquidity Tools for Liquidity Management in Islamic Banks

has controversies attached to Bai’ al-Inah mode of financing. Majority of the scholars do not approve this product but it is being used in the Malaysian market only. This controversial prıduct was introduced by BNM in 1999, as a last resort funding facility to cover an Islamic banking institution’s deficit position. In the year 2000, BNM introduced the “Bank Negara Negotiable Notes” (BNNN) which was based on the concept of Bai’ al-Inah. The Government Investment Issue (GII) which was issued in 2001 by the BNM on behalf of the Government of Malaysia was also based on Bai’ al-Inah contract. Due to Shariah unacceptability of Bai’ al-Inah contract in other regions, there has been many steps taken by regulatory authorities in Malaysia to gradually shift away from Bai’ al-Inah contracts.

3.21 Salam Sukuk The Central Bank of Bahrain has issued a series of new Islamic financial instruments designed to broaden the depth and liquidity of the market. One of the most popular instruments is the short-term government bills structured under Salam, having generally a maturity of 91 days and not tradable in the secondary markets. The underlying asset in this transaction is aluminum. The Government of Bahrain will sell aluminum to the buyer. In exchange for the advance payment that will be paid by the Islamic banks, the Government of Bahrain will undertake to supply a specified amount of aluminum at a future date. At the same time the Islamic banks will appoint the Government of Bahrain as an agent to market the quantity of aluminum at the time of delivery through its channels of distribution.

3.22 Certificates of Deposits (CDs) The GCC Central Banks issue Certificates of Deposits (CDs) for IFIs and this instrument is accepted as collateral by the central banks for Collateralized Commodity Murabahah (CMM) agreements which are used for the purpose of repo-like lending by central banks.

3.23 Tawarruq Reverse Murabahah-type contracts (Tawarruq) are routinely used by the Central Bank of Kuwait as a means to absorb structural longer-term liquidity from Islamic banks. This product is also used by Malaysian Islamic banks via the buying and selling of palm oil. There are at least three parties in this contract. This contract is used to manage excess liquidity in the Islamic banking sector.

3.26 Central Bank Musharakah Certificates

43

3.24 Central Bank Participation Papers One of the most interesting monetary policy instruments for Islamic banks is the Central Bank Participation Papers (CBPP), which were introduced and widely used by the Central Bank of Iran or CBI. CBPP were first issued in March 2001. The bearer securities are issued in 6 or 12 months maturities with quarterly coupon payments, exempted from tax. The CBPP are issued based on a portfolio of completed infrastructure projects previously financed by the central bank credit to the government, and yield a predetermined rate of return presumed to approximate the returns on underlying assets. The initial thinking was to create a marketable money market instrument which would empower the central bank to regulate liquidity and provide a viable instrument for liquidity management by commercial banks. The ultimate objective was to foster the development of an interbank market (Haque and Mirakhor, 1999). However, the final design turned out to be somewhat different, which limited their effectiveness. These products have only been issued to nonbanks in the primary market at pre-announced rates of return. Commercial banks are obliged to rediscount them in the secondary market at par and guarantee the initial yield to maturity.

3.25 Ijarah Certificates (Shihab Certificates) Shihab Certificates represent investment in an open-ended fund operated in favor of the Central Bank of Sudan (CBOS). The fund, being a liquidity management instrument of the CBOS, invests in property acquired from, and leased back to, the CBOS. Certificates of the fund are available only to banks, government and quasigovernment funds and financial institutions through the CBOS. Rental income is distributed monthly and is expected to yield between 10% to 12% per annum. The Sudan Financial Services Company issues these certificates. The relationship between Sudan Financial Services Company and the investors is based on an agency agreement.

3.26 Central Bank Musharakah Certificates Sudanese Central Bank Musharakah Certificates (CMCs) and Government Musharakah Certificates (GMCs) are equity or Musharakah-based instruments, which were launched in 1998 and 1999 respectively. CMCs are used against Bank of Sudan ownership in commercial banks and are used as an indirect instrument to regulate and manage liquidity within the banking system. GMCs are used against the Ministry of Finance’s ownership in some profitable public and joint venture enterprises. The certificates are meant to regulate and manage liquidity within the economy as a whole.

44

3 Liquidity Tools for Liquidity Management in Islamic Banks

The instruments enable the government to raise funds through the issuance of securities that promise the investors a negotiable return that is linked to the development in government revenue (a share in government revenue, for example) in return for their investment in the provision of general government services. The fund user is the government, which builds the infrastructure, and the fund suppliers are people who have savings to invest. The intermediaries are Islamic banks and other financial institutions. Once the government issues a particular sharing certificate, the Islamic banks and other financial institutions buy it out of the funds accumulated in their investment accounts or as insurance, as the case may be. The government’s obligation to pay yearly dividends could be met through the bank, to which it may transfer its share of the declared profits. If the certificates have a maturity date when the government is offering to pay back the capital (with the final year’s profit or minus the losses), this obligation too, can be met through the banks. The government does not deal with the public directly. It transacts only with the Islamic banks and other financial institutions, which transact with the public. Those who do not wish to keep the certificate till maturity can sell, and those who missed the issue when it launched can buy. Thus, it will function like any other market (Hunt 2013).

4 Basel Regulations, Liquidity Management and Regulations in Turkey’s PBs 4.1 Turkish Banking System and Participation Banking The Turkish banking sector, including Participation Banking, has shown a relatively robust performance in recent years. The average annual growth rate is around 18% in the last six years. The banking sector’s asset size increased more than the growth of the country’s GDP and reached US$813 billion at the end of 2015. Although the Turkish banking system is growing rapidly, it still has a high-quality capital base with its conventional banks especially having well-developed liquidity management. Asset quality of the banking sector remains unchanged after the global financial crisis, and non-performing loan ratio stays flat. There are 52 banks (34 deposit banks, 13 development and investment banks and 5 PBs). Banks continue to meet liquidity coverage ratios while external financing demand and conditions on accessibility to external financing are still necessary elements of liquidity risk management. The weighted average maturity of the banking sector’s external liabilities began to be extended in February 2015 and reached 51 months by the end of 2015. In Turkey, there is no separate banking legislation that regulates PBs (Turkey, 2016). As at the end of November 2016 there were five PBs in the country: Albaraka, Kuveytturk, Turkey Finance, Ziraat Participation Bank and Vakif Participation Bank (established in 2015). The Licence of Bank Asya was terminated at the end of July 2016 by the Banking Regulation and Supervision Agency (BRSA).18 Aysan et al. (2013) claimed that although PBs have only around 5% of market share, they still can play a pivotal role in channelling idle capital into more productive sectors. PBs need liquid instruments not only for investment but also for counter-balancing cash outflows. These banks, in the “liability” side of their balance sheets, use a profit-and-loss-sharing methodology. On the “assets” side, almost all of their financing facilities are used for tangible projects in the real sector (household or companies), which have higher risk-weights. For this reason, they need to stay in cash or invest in liquid instruments. Table 4.1 shows an overview of PB assets between 2001 and 2015. According to the table, the ratio of total PB assets to the banking sector’s assets increased from

18 Within the scope of Turkish Banking Law No. 5411 (Law), Bank Asya (Asya Katılım Bankası A.Ş.) was audited by the BRSA in 2015 and it was found that the Bank had made many illegal transactions and given credits to a group of customers without obeying regulatory rules. After that, it was decided to transfer the Bank to the Savings Deposit Insurance Fund under the provisions of paragraph (b) of the first paragraph of Article 71 of the Law. Then the bank was tendered for sale in 2016, but there were no offers. After that, the bank’s license was canceled in July 2016. https://doi.org/10.1515/9783110582901-004

46

4 Basel Regulations, Liquidity Management and Regulations in Turkey’s PBs

Table 4.1: Total Asset (Thousand TRY), Asset Growth Rate and Share. Year

Banking Sector

PBs

              

,, ,, ,, ,, ,, ,, ,, ,, ,, ,,, ,,, ,,, ,,, ,,, ,,,

,, ,, ,, ,, ,, ,, ,, ,, ,, ,, ,, ,, ,, ,, ,,

Growth Rate (%)

Share (%)

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

Source: Banking Regulation and Supervision Agency (BRSA).

1.08% to 5.1%. However, shareholder’s equity remains at the same level at the end of 2015. Non-performing loans to gross loans ratio worsened due to the problems related with Bank Asya which was nationalized in 2016. Liquid assets of PBs to total assets ratio increased from 17.4% to 23.5% between 2010 and 2015, whereas both return on assets (ROA) and return on equity (ROE) ratio decreased significantly, from 2% to 0.4% and 16.9% to 4.1% respectively (Table 4.2). This illustrates that while PBs become more liquid, they also become less profitable, implying there is a trade-off between staying liquid and making a profit. It is interesting that this decreasing profit cycle occurred concurrently with the implementation of Basel III liquidity standards, particularly after 2013. The implication that there is a trade-off between profitability and liquidity rule compliance can be gleaned from Appendix A which shows the assets, cash, securities, financing given, deposits, equity and net profit of PBs. Moreover, PBs keep a large proportion of cash in their balance sheets for the sake of safeguarding themselves against internal liquidity mismatches (Figure 4.1). Holding more cash for precautionary needs is the direct result of regulatory restrictions on these banks, which may have significant adverse effects on financial development (Morgan and Pontines, 2013). It is assumed that these banks need to protect themselves against risks related to market liquidity in times of stress (Kerry, 2008). For this reason, these banks stay in cash (Figure 4.1), resulting in decreases in their profitability, evidenced by the declining return on equity ratio over the years (Appendix B). On the capital side, PBs’ capital adequacy ratio (CAR) is around 14.8% (Appendix C), indicating that

47

4.1 Turkish Banking System and Participation Banking

Table 4.2: Overview of Turkish Participation Banking Sector (%).

Total Assets to GDP (%) Financing to GDP (%) Liabilities to GDP (%) Deposits to GDP (%) NPLs / Total Financing (%) Credit Growth (YoY) (%) Cost / Income (Efficiency) ROA ROE Financing to Deposit ratio Leverage Ratio Liquid Assets / Assets CAR













. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

Sources: CBRT.

these banks are well-capitalized although they have many regulatory disadvantages regarding risk-weight of their assets.

PBs’ Cash (Thousand, TL) 60,00,000 50,00,000 40,00,000 30,00,000 20,00,000 10,00,000 08/15

10/14

03/15

12/13

05/14

07/13

02/13

09/12

11/11

04/12

01/11

06/11

08/10

03/10

10/09

12/08

05/09

07/08

09/07

02/08

04/07

0

Total PBs Figure 4.1: PBs’ Cash Assets. Source: BRSA of Turkey, CBRT EVDS.

After Special Finance Houses were identified as banks in the new Banking Law 2005, their financial figures began to grow rapidly (annual average 28.7%) and the growth rate did not slow down between 2010 and 2015. As Figure 4.2 indicates, their total assets haveincreased by US$13 billion (46%) just in the last five years.

48

4 Basel Regulations, Liquidity Management and Regulations in Turkey’s PBs

Total Assets, Deposits, Credits and Equity of PBs (2005–2015) (USD Mn.)

50,000

5,000

40,000

4,000

30,000

3,000

20,000

2,000

10,000

1,000 0

0 2005 2006 2007 2008 2009 Credits

2010

Total Assets

2011

2012

Deposits

2013

2014

2015

Equity (right)

Figure 4.2: Total Assets, Deposits, Credits and Equity of PBs (2005–2015, US$ Mio). Source: BRSA of Turkey, CBRT EVDS.

PBs offer mainly Murabahah, Ijarah and Musharakah financing products. Murabaḥah continues to dominate the participation banking sector contracts with a share of more than 95% of total financing, indicating that there is a need to develop new products as well as increase the effectiveness of existing products. The utilization of Ijarah (leasing) contracts has started to gain pace for the last five years and has reached 5.2% of total financing by the end of 2015 (Figure 4.3). This is attributed to the removal of tax disadvantages on Ijarah transactions.

100

0.27

99

1.35

0.41 0.77

0.23

0.2

1.3

1.66

0.16

0.34

0.3

2.93

98

4.47

97

5.15

96 95

98.37

98.8

98.45

94

98.13 96.9 95.17

93

94.54

92 91 2009

2010

2011

Murabaha

2012 Ijarah

Figure 4.3: PBs Products Share. Source: BRSA of Turkey and CBRT EVDS.

2013 Musharakah

2014

2015

4.2 Turkey’s Liquidity Regulations for PBs

49

Asset-liability transformation is more challenging for PBs due to the nature of deposits in these banks. Historically PBs have a liquidity deficit at one month, and increases especially after 2013. Although all the banks in the Turkish banking sector have a liquidity deficit, the problem is more severe for PBs due to a shortage of limited opportunities for liquid assets and deposit-type products to attract investors for long-term maturities (Turkey, 2016). There is no return for special current accounts, and liability of PBs covers only the principal. Special currency accounts represent deposits in Turkish Lira and foreign currencies, mainly the US dollar and the Euro. Special current accounts within the Participation Banking sector is convenient, and easy to access via various channels like call centres, online banking, automated teller machines; 24 hours a day, 365 days a year, for branches all over the country. Participation accounts are defined in the Article 3 of Turkish Banking Law (published on 1/11/2005 on Official Gazette with number of 5411) as: Accounts constituted by funds collected by PBs that yield the result of participation in the loss or profit to arise from their use by these institutions, which do not require the payment of a pre-determined return to their owners and that do not guarantee the payment of the principal sum.

PBs are maintaining reasonable levels of capital as required by the Banking Regulations and Supervision Agency (BRSA). Currently, capital requirements are determined mostly by applying Basel rules developed for conventional banks with some exceptions. For example, BRSA follows alpha factor of less than 1 to recognize the profit-and-loss sharing nature of deposits. Given this approach, and considering that PBs are operating on a profit-and-loss sharing basis, these levels of capital are sufficient. There is an increase of leverage ratio which is caused by the issuance of Tier-II Sukuk by PBs. Liquidity ratio of PBs (21%) is within prevailing industry norms. The liquidity ratio of PBs in Turkey is very close to Malaysian Islamic banks (around 22%). The average financing rate of these banks is around 69%.

4.2 Turkey’s Liquidity Regulations for PBs In Turkey, the regulatory framework is based on the Banking Act 5411 promulgated in November 2005. There are regulations on supervision, measurement, and evaluation of liquidity coverage ratio requirements of conventional banks which also cover PBs. However, there is currently no special regulatory framework for Participation banking in Turkey, and there are very limited provisions, guidelines, and procedures dealing with this sector. The law distinguishes between deposit and Participation banking, and regulations governing fund collection and fund utilization are distinguished between the two types of banks. The law also takes into account the nature

50

4 Basel Regulations, Liquidity Management and Regulations in Turkey’s PBs

of the profit and loss participation accounts, and allows for a different calculation method for Capital Adequacy Ratio (CAR) for PBs. PBs, a component of the banking system in Turkey, have brought idle funds into the system. These banks have provided alternative financial opportunities to manufacturers, business people, and those who previously shunned interest ratebased financing. However, these banks’ growth has particular challenges, notably, the absence of a well-formulated regulatory framework for their specific needs, a high risk-weight mechanism for participatory contracts, cash-based capital base, challenging liquidity management, and fragile trust of investors. Currently, the BRSA does not require PBs to follow any of the IFSB or AAOIFI standards. Instead, standards for conventional banks issued by Basel are applied to PBs (Turkey, 2016). It is apparent that there is a need to strengthen the regulatory and supervisory frameworks of these banks. In this context, the IFSB standards can be used, which recommends the application of different regulations for calculation of capital adequacy ratio as well as the introduction of the Profit Equalization Reserve and Investment Reserves for mitigating displaced commercial risk and withdrawal risk. Currently, BRSA allows PBs to allocate reserves as 5% for future distribution, but it is only a non-binding recommendation. The policy of financing in instalments and monthly instalment repayments has regulated the cash flow and liquidity needs of PBs and strengthened financing security. Most of the assets of PBs are Murabahah financing (Table 4.3). Partnership-based financing products (Musharakah and Mudarabah) have only 2% share. Table 4.3: Asset of Participation Banks. Assets Murabahah Financing Ijarah Istisna Salam Musharaka (PLS Partnership) Mudarabah (PLS Partnership) Total

Liabilities       

Capital Profit Loss Sharing Accounts Special Current Accounts

  

Total



PBs keep a high proportion of requirement ratio up to one month for safeguarding themselves against inherent liquidity mismatches (Figure 4.4). Keeping more liquid instruments may help PBs to protect themselves against the risk related to market liquidity in times of stress. However, the liquidity requirement ratio of these banks is more than the 60% threshold suggested by the Basel Committee for 2015. The banking system in Turkey has a liquidity deficit, and this deficit is funded by the Central Bank of the Republic of Turkey (CBRT). CBRT has six types of instruments

4.2 Turkey’s Liquidity Regulations for PBs

51

Liquidity Requirement Ratio

160 150 140 130 120 110 100

04–16

10–15

04–15

10–14

04–14

10–13

04–13

10–12

04–12

10–11

04–11

10–10

04–10

10–09

04–09

10–08

04–08

10–07

80

04–07

90

Figure 4.4: Liquidity Requirement Ratio (million TL) – Up to 1 Month. Source: BRSA of Turkey and CBRT EVDS.

for the implementation of Open Market Operations (OMO). Due to its monetary policy, the Repurchase Agreement (Repo) facility appears to have been the main instrument for liquidity injection before 2016. These instruments include outright purchases of government bonds/Sukuk, repo, reverse repo, liquidity bills, TL deposit buying auctions and outright sales of government bonds/Sukuk. CBRT is not only injecting liquidity with conventional tools into the banking system but is also trying to offer liquidity to the PBs. Especially after 2013 CBRT was very active in attempts to change its legislation and introduced several mechanisms for this specific purpose. As the Participation banking industry has grown and has taken a bigger share in the banking system in Turkey, the CBRT has started to adopt several amendments in its operational framework, contracts, and tools to level the playing field for PBs concerning access to liquidity facilities of the CBRT (Turkey, 2016). The CBRT can inject/sterilize liquidity permanently into/out of the banking system in case of liquidity deficit/surplus via outright purchases/sales of domestic (Turkish Lira-denominated) sovereign debt securities/Sukuk issued by the Undersecretariat of Treasury and its Special Purpose Vehicle (SPV, Asset Leasing Company of the Treasury). Moreover, in the case of liquidity deficit or surplus, the CBRT can inject/sterilize liquidity temporarily into/out of the banking system by using repo/reverse repo of domestic sovereign debt securities issued by the Undersecretariat of Treasury and its SPV (Turkey, 2016). In recent years, while repo maturities have varied from overnight (O/N) to 3-month maturities depending on liquidity conditions, a reverse repo is typically O/N. The CBRT has 1-week repo funding at its 1-week repo rate. Reverse repo is conducted at a borrowing rate of the CBRT. In 1-week repo auctions, only Turkish Lira Treasury securities are repurchased, and all other eligible collaterals can be used

52

4 Basel Regulations, Liquidity Management and Regulations in Turkey’s PBs

only for the purpose of margin management. If PBs do not have enough Turkish Treasury lease certificates, they may not use this facility. PBs, after the amendment made in OMO contracts in May 2013 (Turkey, 2016), started to use sale-based repo facilities and outright-purchased facilities comfortably. Outright purchases/sales and repo/reverse repo transactions are based on sale contract and contains a bilateral trade between the CBRT and its counterparties. Repo is the main OMO tool since mid-2008 due to the liquidity deficit within the Turkish banking system. Since PBs do not have access to Borsa Istanbul Organized Repo-Reverse Repo Market instruments, the CBRT allows these banks to do O/N repo against TRY Treasury Lease Certificates under its own Turkish Lira Market (Turkey, 2016). If a PB demands to use deposit facilities, for every one unit of borrowing from the CBRT, it should provide at least 30% of Turkish Lira government securitiesbased collateral (Turkey, 2016). The remaining part can be made up of other eligible collaterals such as Sukuk issued in FX denomination issued by Turkish Treasury, IILM Sukuk, FX deposits and banknotes (only US$ and EUR). Banking and insurance transactions tax applies, which is a kind of expense tax. The CBRT is offering liquidity instruments for PBs in several ways. A Wa’ad (promise)-based instrument to buy back/sell back Sukuk is commonly used by PBs. Although in this mechanism, the CBRT’s promise is not binding, and that of PBs as the counterparty is binding according to the open market operations contract signed between the CBRT and a participation bank to ensure Shariah-compliance as amended in 2013. For excess liquidity problems, the reverse repo mechanism was introduced and can be used in the same manner. The CBRT buys and sells back only TRY-denominated domestic sovereign bonds and Sukuk as the underlying asset (collateral). Other collaterals can be used only for haircut or margin management purposes. The CBRT has allowed PBs to access its repo facilities in 2011 and facilitated their access by amending its open market operations framework agreement (Open Market Operations Agreement) in 2013. The CBRT has also expanded its eligible collateral base for its liquidity facilities to ease PBs’ access to liquidity facilities, simplified and reduced haircut rates for its eligible collateral base within the context of the “Road Map during the Normalization of Global Monetary Policies” published on 18 August 2015. The CBRT has taken many actions at various times to enhance its liquidity facilities to improve the efficiency of its operational framework and to facilitate liquidity management of PBs. For instance, CBRT has become a shareholder of the International Islamic Liquidity Management Corporation (IILM) – a supranational institution established with this specific purpose on 25 October 2010 by central banks. The IILM is a multilateral organization to develop and issue high-quality, short-term, Shariah-compliant financial instruments to facilitate effective crossborder liquidity management for institutions that offer Islamic financial services.

4.2 Turkey’s Liquidity Regulations for PBs

53

Recently, CBRT has amended the guidelines for Open Market Operations, Interbank Money Market Transactions and Foreign Exchange and Banknotes Transactions to include IILM Instruments/Sukuk denominated in US$ and EUR as eligible collateral. Moreover, the IILM Sukuk has been accepted as a qualified instrument for reserve management as of 2013. On the other hand, PBs do not have access to exchange-traded securities issued under the Borsa Istanbul repo-reverse repo market due to its non-conformity with Islamic Law. Due to the non-availability of exchange-traded facilities, PBs are unable to benefit from less costly funding opportunities (Turkey, 2016). Likewise, PBs have limited opportunities for placing very short-term excess liquidity. Due to the pure debt nature of standing deposit facilities offered by CBRT for liquidity absorption, PBs are unable to access typical instruments offered to conventional banks. Since standing deposit facilities are based on loan contracts (whereas as the repo is a sale contract), they cannot be used by PBs. In other jurisdictions, central banks have designed an instrument based on collateralized commodity Murabaḥah (trade finance), however the same cannot be offered in Turkey due to legal restrictions of CBRT on trading of commodities according to the current Central Bank Code. Table 4.4 shows the facilities available to PBs.

Table 4.4: The Utilization of CBRT Facilities by PBs. The Utilization of CBRT Facilities by PBs Liquidity Facilities and Turnover Amounts (Million TRY) Facility Repo Auctions Repo Quotation Outright Purchases Deposit Lending Intraday Liquidity

 ,

 ,

 , 

 , ,

 , , .

 ,.

Total , , . . ,.

Source: CBRT.

PBs can benefit only from an outright purchase facility, repo, and late overnight liquidity facility repo. They do not have the opportunity to access other facilities offered by CBRT such as deposit lending, intraday liquidity, late overnight liquidity facility depo19 and open market operations funding through Borsa Istanbul (BIST) repo market.

19 Depo facility is used by CBRT to let banks place their excess liquidity or find liquid sources for their shortage in a mechanism, in which CBRT plays “blind broker” role (CBRT organizes transactions but banks borrow Money from each other without knowing their counterparties).

54

4 Basel Regulations, Liquidity Management and Regulations in Turkey’s PBs

Last but not least, the availability of a lender of last resort facility is vital for PBs. These are emergency facilities that can be used by the banking system in case of a liquidity deficit or surplus, as well as the absence of an alternative resource of funds/credible counterparty to place excess funds. Punitive rates are applied in these ‘lender of last resort’ facilities. These facilities used to be only deposit facilities based on loan contracts and typically O/N facilities. As of May 2016, the CBRT has started to allow PBs to use late liquidity facilities based on repo/reverse repo contract if deemed necessary. These services are available only in exceptional cases.

5 Dynamic Determinants of Liquidity Management in Turkish Participation Banks 5.1 Introduction Since the 2008 global financial crisis, central banks and global financial regulatory authorities have developed a new regulatory framework. Given the importance of instituting more robust and effective supervision of banks, it was expected that the newly-developed standards would reduce volatility and protect banks against different risks in the markets. This new regulatory framework has positive and negative implications for Islamic financial institutions. Firstly, there is a concern that it will crowd out Islamic financial institutions from the playing field in the long run because the regulatory framework does not consider the specific characteristics of Islamic banks. In principle, the resilience of Islamic banks has potential implications for financial stability in countries where these banks have systemic importance. However, there are many issues that affect the soundness of the financial markets and market liquidity in Muslim countries. This can range from liquidity shortages, herding, fire sales, capital flows and financial structure, to the treatment of counterparties and market liquidity. A robust Islamic financial services industry therefore requires the strengthening of policy infrastructures for the mitigation of systemic risk, and the use of an appropriate combination of micro and macro prudential tools to address the build-up of financial imbalances that may affect financial stability. In this sense, having sustainable liquidity management is essential to enhancing financial stability. The resilience of Islamic banks has potential implications for financial stability particularly in jurisdictions where Islamic banks are systemically important. Although there is no G-SIFIs in the Islamic banking space, domestic systemically important banks (D-SIBs) could be perceived as “too big to fail” in some countries. For example, Malaysia-based Maybank or Saudi Arabia-based Al-Rajhi Group are important local banks in their countries as their share in their banking system and can be accepted as D-SIBs. However, the Islamic banking industry faces several challenges in the area of liquidity management. The attitude of Islamic banks towards Shariah-compliant contracts, capitalization, portfolio management, risk-taking, and interbank demand impedes them from undertaking a comparable and competitive liquidity conversion on par with conventional banks. Besides, implementation of liquidity coverage ratio (LCR) under the new Basel standards is challenging for Islamic banks. Islamic banks also have disadvantages in terms of benefiting from central bank funding for their liquidity management due to insufficient Shariah-compliant open market mechanisms in many countries. In Turkey, PBs can enter open market operations with the CBRT for selling lease certificates to the CBRT for receiving cash for managing their https://doi.org/10.1515/9783110582901-005

56

5 Dynamic Determinants of Liquidity Management in Turkish Participation Banks

liquidity. When PBs do not have enough lease certificates, they cannot find liquidity for its operations, which means their funding liquidity is affected by monetary liquidity to some extent. Furthermore, regulatory requirements may impose stress on funding liquidity and market liquidity. For example, if a regulator changes the liquidity treatment of an asset, the demand for the asset will suddenly align to the regulatory change, because the cost of this asset changes. Since PBs do not have extensive access to the interbank money market in Turkey, they have to price the variations in the market and follow their conventional counterparts to eliminate the potential arbitrage between these two markets. Research on financial stability in emerging economies has been gaining traction in recent years owing to the rapid growth and internationalization of these markets. Turkey is a member of the Basel Committee on Banking Supervision and has recently developed a medium and long-term plan for growing the Islamic banking sector in the country. The Turkish Government aims to expand the Participation banking market share to 15% by 2023 from 5.1% in 2012 (IFSB, 2015a; Thomson Reuters, IRTI, CIBAFI, 2014). In this context, Turkey offers an interesting case for analysis since the Turkish economy has become more integrated with the global economy over the last three decades. In fact, the extent of trade integration is indicated by the fact that Turkish business cycles are closely linked with those of European Custom Union members (Erden and Ozkan, 2014). In this chapter, the time series based liquidity model and Granger causality relations are employed to analyze the short run and long run relations among selected variables related to Basel III standards. Sub-section 5.2 discusses the empirical approach and data used in this chapter. The discussion of results as well as the conclusion and policy implications are provided in sub-section 5.3 and 5.4, respectively.

5.2 Empirical Approach and Data This chapter adopts a time series liquidity model to analyse short-term and long-term factors that affect liquidity management of Islamic banks in Turkey. Regulations can be accepted as country-specific effects, which intensify benefits of time series analysis. For the sake of enhancing the liquidity management of PBs, time series analysis is better suited initially. In the upcoming chapter, a panel data analysis will be conducted for a comparison. One of the underlying assumptions of classical regression is to accept that all variables are stationary. Engle and Granger (1987) establish that a co-integrated system can be represented by an error correction structure that integrates both changes and levels of variables to ensure that all the elements are stationary. Vector Auto Regression (VAR) is used to analyse the relationships among different factors. VAR provides a convenient representation for both estimation and forecasting of systems

5.2 Empirical Approach and Data

57

of economic time series (Sims, 1980; Litterman, 1986). The stationarity of the time series is vital for the validity of the VAR analysis. The relationships between the variables employed based on Granger Causality analysis and Impulse-Response Function (IRF) analyses within the VAR method are investigated. It is expected that the time series should not include trends or fixed seasonal patterns. Hence, economic time series have to be transformed before the properties hold. Two-unit root tests on the individual stochastic structure which are the Augmented Dickey-Fuller (ADF) and Phillip-Perron (PP) tests are used in the examination. The lag is determined based on Schwarz Criterion (SC) which is commonly used for financial data. If the variables are identified as stationary in the level, standard VAR analysis applies to the level data; however, if the variables are found out to be I(1), a cointegrating relationship between these variables is searched. The Johansen and Juselies (1990) test is applied to examine the presence of cointegration. The ARDL approach is then used for PBs to see short run and long run relations and lessen the challenges of unit root analysis. To investigate the causality and the direction of influence of one variable on another variable, the bivariate Granger causality test based on Granger (1969) is adopted to the PBs data and CBs data. The Impulse Response Function (IRF) for Islamic banks is utilized to investigate the response of each variable over a particular period of a given shock. The model used in this chapter is derived from previous literature (Cucinelli, 2013; Angora and Roulet, 2011; Bonner, 2014; Ahmed et al., 2011; Banerjee and Mio, 2014; Akhtar et al., 2011; Mohamad et al., 2013; Boudt et al., 2013; Ergec and Arslan, 2013; Amin, 2016). Angora and Roulet (2011) define liquidity risk with LCR and NSFR and add several independent variables to the model, which are ROA, the natural logarithm of total assets, the ratio of loans to customers and total loans, GDP annual growth rate, the spread between the interbank rate and central bank policy rate, etc. Cucinelli (2013) uses both the LCR and NSFR as dependent variables and add credit rating as well as bank size, bank specialization, and net interest margin. For Islamic banks, Ahmed et al. (2011) calculate liquidity ratio as liquid assets to total assets. Moreover, many studies used new metrics for conventional banks, such as Bonner (2014), who uses liquidity ratio defined by Dutch Central Bank (DLCR- Dutch Liquidity Coverage Ratio) according to Basel requirements as a dependent variable. Moreover, Banerjee and Mio (2014) used the share of HQLA to total assets as a dependent variable to analyze the behavioral reaction of banks to a tightening of liquidity regulation. To the best of the research knowledge, no study has specifically investigated the effects of liquidity requirement ratio or liquidity coverage ratio in the dual banking system for Islamic banks and conventional banks in Turkey. In this book, liquidity requirement ratio is used as a dependent variable for both time series. Time series model is employed to help understand especially short-run Granger causality relations among variables. Ergec and Arslan (2013) use interbank rate as well as credit and deposits in their model. Ogilo and Mugenyah (2015) use capital adequacy, a

58

5 Dynamic Determinants of Liquidity Management in Turkish Participation Banks

liquid asset to total asset ratio, total asset size, and leverage for evaluating the determinants of liquidity risk. Table 5.1. shows list of variables used for the model. The model is formulated as follows: LRR = α + β1 RR + β2 ROE + β3 CAR + β4 INF + β5 CDS + β6 INTER + β7 BOND + β8 CRE + β9 SUKUK + β10 DEP + ε

(4)

Table 5.1: List of Variables. Variable (short form)

Variable (Long form)

Description

Source

LRR (Dependent)

Liquidity Ratio High-Quality Liquid Assets/ Net Cash outflows

BRSA Monthly Bulletin

RR

Required Reserves

Required reserves of bank kept as short term assets

BRSA Monthly Bulletin and checked with CBRT EVDS

ROE

Return on Equity

Net Profit / Equity

BRSA Monthly Bulletin

CAR

Capital Adequacy Ratio

Tier  Capital + Tier  Capital / Risk-Weighted Assets

BRSA Monthly Bulletin

INF

Inflation Ratio

Consumer Price Index

TUIK – Turkish Statistical Institute

CDS

Credit Default Swap

CDS rate of the Country

Bloomberg Terminal

INTER

Interbank

Interbank Money Market Rate

Bloomberg Terminal

BOND

Treasury Bonds

 Months Treasury Bond Return

CBRT EVDS

CRE

Total Credits

Total Financing

BRSA Monthly Bulletin

SUKUK

Securities

Lease Certificates and Sukuk

BRSA Monthly Bulletin and TKBB Database

DEP

Deposits

Total Deposits

BRSA Monthly Bulletin

Ε

ε

Error Term

A

α

Constant

Dependent Variable: Liquidity, the dependent variable, is measured as the ratio of liquid assets to net cash outflows based on the previous studies (Bonner, 2014; Van den End, 2012; Cuccinelli, 2013; Angora and Roulet; Leykun, 2016). Bank liquidity is expected to be dependent on the individual behavior of banks, market, macroeconomic

5.2 Empirical Approach and Data

59

environment and structural factors (exchange rate regime, regulations, etc.). Assets that are accepted as high-quality liquid assets are cash, central bank reserves, public bonds or treasury bills (highly rated), financial assets held by multinational investment banks and sovereign bonds in foreign currency with high ratings. Independent variables are required reserves, Sukuk/securities stocks (mainly lease certificates of Turkish Treasury and Sukuk issued by other banks and MDBs), the logarithm of total credits (CRE), the logarithm of total deposits (DEP), and return on equity (ROE), capital adequacy ratio (CAR), inflation ratio, CDS of Turkey, three-month Treasury bond return. Required Reserve: The required reserve (RR) is used as a tool in monetary policy for controlling market liquidity and banks’ liquidity. By changing the required reserve ratio, the central bank can change the value of the money multiplier as well as the monetary base. When the central bank wants to tighten the money supply in the economy, it raises the reserve ratio for increasing the interest rate in the market. In the case of expansionary monetary policy, the reserve ratio is lowered, to allow for the expansion of money supply. Since, some components of required reserves are not treated as high-quality liquid assets (such as gold, keeping Euro instead of TL), it is accepted that there is no perfect collinearity between liquidity and required reserves and Appendix J supports this initial intuition. Increasing required reserves might decrease the profit of banks, if the central bank would decide to not give any profit/interest rate for keeping required reserves. Decreasing profit may affect net cash outflows, which may impair to the liquidity ratio. Therefore there is no endogenity problem for including required reserves to the model. Sukuk: Sukuk or securities are included to the model as an independent variable. Sukuk show all Sukuk (short run or long run) purchased by PBs for investment. Securities are included as a variable for conventional banks. It is expected that if Sukuk stocks increased, the liquid assets ratio would increase. However, all Sukuk invested by PBs are not accepted as high-quality liquid assets. For example, corporate Sukuk are not treated as HQLA. Prior to 2016, the IILM Sukuk was not accepted as HQLA by the BRSA. This shows that there is no perfect correlation between Sukuk and LRR. Inflation: The inflation rate (one-year ex-post inflation) and CDS rate of Turkey are derived from the Bloomberg database. The inflation rate is calculated with the following formula: INF = 100*(log (CPIt) – log (CPI t-12). The inflation rate is the general price changes calculated based on a basket of goods and services. It is assumed that inflation impact banks’ cost of funding, and high inflation forces banks to keep higher liquid assets for managing the liquidity and decreasing the cost of funding. For this reason, it is used as an independent variable by many studies including Amin (2016), Azam et al. (2013) and Adrian et al. (2017) and Ergec and Arslan (2013), who provided evidence showing the negative impact of inflation on liquidity risk. The PBs in Turkey are noticeably influenced by interest rates (Ergec and Arslan, 2013).

60

5 Dynamic Determinants of Liquidity Management in Turkish Participation Banks

CDS: The CDS rate is obtained through the Bloomberg database and is calculated as the theoretical CDS spread implied by the bond yield (Adrian et al. 2017). Two components of TED spread are used following Deming and Hong (2012) who calculated TED spread as the spread between three months’ treasury bills rate and three months LIBOR rate. The BOND and INTER variables are formulated to see the effects of market liquidity on banking liquidity (Adrian et al., 2017; Praet and Herzberg, 2008). These variables have been used by many researchers as control variables (Akhtar et al. 2011; Demirgüç-Kunt et al. 2003; Deming and Hong, 2012; Ergec and Arslan, 2013). CAR: Capital adequacy ratio (CAR) is an important variable in the interaction between liquidity and capital of the bank. Bonner (2014) shows that there is no perfect correlation between risk-weighted assets and liquid assets. He shows that the correlation coefficients of liquid assets and risk-weighted assets differ noticeably across countries. Leykun (2016) employs capital adequacy ratio as an independent variable to examine the determinants of liquidity risk. Moreover, Akhtar et al. (2011) use this variable with four other variables in their model (asset size, capital adequacy ratio, credit to assets ratio, cash to total assets ratio) to show the impact of the bank-specific factors of profitability on the performance of Islamic banks. ROE: Return on Equity (ROE) is calculated by dividing net profit to the shareholders’ equity. If net profit increases, the ROE will increase since the shareholders’ equity has stable position in short run. Therefore it is expected that there would be positive relation between the ROE and the liquidity ratio. Inter and Bond: Interbank rate and Treasury Bonds rate are included to the model to assess the impact of changes of interest rate in the market on the activities of PBs. Credit: Total credit show financing provided by the PBs or credits given by the CBs. Deposits: Total deposits cover saving accounts, restricted investment accounts, unrestricted investment accounts for PBs and include all deposit accounts for CBs. Several variables of the model have also been used in Mohamad et al. (2013) model who studied Malaysian Islamic banks. They used total assets, CAR, KLIBOR, total credit to total assets ratio, profit ratio and inflation.20 Bonner (2014) used liquidity with capital adequacy ratio, interest rates and the inflation rate in his model. Arif and Anees (2012) compared the profitability of banks with their deposits, cash reserve, liquidity gap and non-performing loans (NPL). Data in this chapter are mainly provided by the Central Bank of the Republic of Turkey (CBRT), Banking Regulation and Supervision Agency (BRSA)'s monthly bulletin, PBs Association of Turkey’s databases and Bloomberg terminal.

20 Total assets were tested in the first model but showed high correlation with liquidity ratio and other variables. Hence it was omitted from this first model.

5.3 Results and Discussion

61

5.2.1 Research Questions This chapter attempts to compile evidence for the following issues: 1. Can the liquidity requirement ratio of PBs be explained by financial factors such as the stock of Sukuk, financing (credit), deposits and return on equity? 2. Can the liquidity requirement ratio of Islamic banks be explained by macroeconomic factors such as inflation, interbank money market rate or government bond rate? 3. Do these determinants affect liquidity risk of PBs in the short-run or the longrun, or both? 4. Are there any significant differences between the factors affecting the liquidity of PBs and the liquidity of their conventional counterparts?

5.3 Results and Discussion 5.3.1 Descriptive Statistics There are four PBs and 20 CBs (both foreign and local deposit banks) covered by this analysis. Investment banks and several particular public project banks (Eximbank, Ilbank, and Turkish Development Bank) are omitted from the analysis because these banks do not have liquidity requirement ratio data. The study covers the period between April 2007 and August 2015 due to data availability. The data is on a monthly basis as the new liquidity framework concentrates on monthly data and cash inflows and outflows are calculated based on one month. Before Basel III, liquidity ratio and analysis were done with quarterly data. However, after the introduction of LCR, quarterly data is no longer adequate to cover changes in cash flows. The empirical investigation is started by presenting descriptive statistics of the model. PBs and conventional banks descriptive statistics are presented in Appendix G. The information regarding variable distribution are derived from these statistics, in which mean value refers to a measure of the average of the underlying variables over the examined period, whereas minimum values indicate lowest value and maximum values show the highest value in the sample. Additionally, standard deviation defines how much a variable shows variation or diversification from the average value. The Jarque-Bera test is done to see a goodness-of-fit test if the sample data has the skewness and kurtosis matching a normal distribution. The skewness value shows the asymmetry of the probability distribution of a real-valued random variable about its mean and kurtosis in statistics is defined as an amount of the “taildness” of the probability distribution. After taking the log of liquidity ratio, financing (credit) and deposits, the standard deviations are reduced. Skewness of some variables is near zero, but skewness of some variables is around 1. Kurtosis of the variables is between

62

5 Dynamic Determinants of Liquidity Management in Turkish Participation Banks

1.72 and 7.39. All variables for PBs have positive means and probability shows statistically significant results for all variables. Similarly, descriptive statistics for conventional banks show that all variables in the model have statistically significant relations and taking the log of liquidity ratio, credit and deposits reduced the standard deviations. Skewness of all variables is around zero. Kurtoses of the variables are generally between 2 and 3. The number of observations of both models is 101, enough for time series analysis. Preliminary analyses show that there are no perfect collinearity relations among variables because there are no linear relations among variables (Appendix D and Appendix E). Homoscedasticity holds for the equation. Therefore, the model in this chapter can be accepted as a best linear unbiased estimator.

5.3.2 Unit Root Tests The stationarity of variables is required for having unbiased and accurate results of time series analysis. This can be determined via unit root tests such as ADF and PP tests (Table 5.2). These tests show that some variables have unit root at the level and some at first level I(1). For this reason, it is accepted that all variables are I(1).

5.3.3 Cointegration Test Results After the unit root test, dynamic relationship among variables has to be tested. If there is cointegration, this means that there is a short-term or long-term relation. It is possible for a variable to deviate from long-term relations. The Johansen-Julius test is used in which all variables have relations with their own values as well as past values of other variables. Each variable is treated as an endogenous variable and depends on its own lags and the lags of other variables. First, the hypothesis that there is no cointegration is tested and then Trace and Maximal Eigenvalue test results are used. The null hypothesis of zero cointegration vectors against at most one is tested. If this is rejected, the null of one against at most two is researched and so on. The optimal lag specified for PBs is two whereas the optimal lag found for conventional banks is three. According to the Trace cointegration test results (Table 5.3), there is more than one cointegration for PBs and conventional banks. However, according to the Max Eigenvalue cointegration test results, one cointegration result has been found for PBs, but there is more than one cointegration for conventional banks. In this case, if the system is bi-variate, trace statistic is used. Since Max Eigenvalue has greater power in the event of a multivariate frame, it is better to follow Max Eigenvalue Statistic.

−.** −.

−.***

−.***

−.***

−.***

−.***

−.**

−.

−.

−.

−.

−.**

−.

−.

−.

DEP

RR

BOND

INTER

CAR

CDS

ROE

SUKUK

−.

−.***

−.

−.

−.

−.***

−.***

−.***

−.***

−.***

−.***

−.***

. Difference

−.

−.*

−.

−.

−.

−.

−.

−.

−.**

−.***

−.***

T-Stats

−.***

−.***

−.***

−.***

−.***

−.***

−.***

. Difference

−.

−.***

−.

−.

−.

−.

−.

−.

−.

−.*

−.***

T-Stats

−.***

−.***

−.***

−.***

−.***

−.***

−.***

−.***

. Difference

Phillips-Perron

CONVENTIONAL BANKS Augmented Dickey-Fuller

Notes: 1)*, ** and *** indicates significance levels respectively at 0.10 ( 10%), 0.05%5 and 0.01 (1%). 2) Sukuk is included for PBs whereas securities have been included for conventional banks.

−.***

−.

−.***

−.

CRE −.

−.**

−.***

T-Stats

Phillips-Perron

−.***

.***

. Difference

PBS

LINF

LRR

T-Stats

Augmented Dickey-Fuller

Table 5.2: Unit Root Test Results.

5.3 Results and Discussion

63

64

5 Dynamic Determinants of Liquidity Management in Turkish Participation Banks

Table 5.3: Cointegration Test Results. PBs Null Test Statistics Hypothesis Trace None At Most  At Most  At Most  At Most  At Most  At Most 

.* .* .* . . . .

Max Eigen .* . . . . . .

CBs

Critical Values (%)

Test Statistics

Critical Values (%)

Trace

Max Eigen

Trace

Max Eigen

Trace

Max Eigen

. . . . . . .

. . . . . . .

.* .* .* .* .* .* .*

.* .* .* .* . . .

. . . . . . .

. . . . . . .

Note:* denotes significant at 5% significance levels.

Cointegration test results show that there is one cointegration for PBs at 5% significance level according to the Max Eigenvalue Statistic. Having cointegration means that there are long-term relations among variables, but this is not enough to know which variable is endogenous or exogenous. Since it is not known which variable is leading others, Vector Error Correction (VECM) model with Johansen-Juselius test is applied. Then, the weak exogeneity of error term with Chi-square test and p-values is tested. For conventional banks, there is more than one cointegration at 5% significance level according to Trace statistics and Maximum Eigenvalue. If there is a cointegration, it means that non-causality between the variables is rejected (Engle and Granger, 1987). Indeed, there should be causation in at least one direction. It seems that it would be better to use unrestricted the VAR model for conventional banks because there is no specific theoretical foundation for conventional banks liquidity risk analysis. VAR model assumes that all variables are endogenous. VAR expresses each variable in the system to be a function of its own lags and lags of the remaining variables in the system. The advantage of VAR model is that it captures empirical regularities in the data with minimal theoretical restrictions. In this sense, first, ARDL approach is applied and liquidity of PBs for the long-term is checked. Then, the Granger causality test is applied to both conventional banks and PBs to see the causality relations among variables. Next the impulse response analysis applied to PBs is discussed.

5.3.4 ARDL Approach for Liquidity Time Series Model of PBs The application of ARDL approach to cointegration contains estimating the conditional error correction version of the ARDL model (Pesaran et al., 2001) for liquidity

5.3 Results and Discussion

65

ratio and independent variables. Bounds test was used for the presence of a longrun relationship between liquidity and independent variables using two separate statistics. The first involves an F-test on the joint null hypothesis that the coefficients on the level variables are jointly equal to zero (Pesaran et al. 2001). The second is a T-test on the lagged level dependent variable. The statistics have a non-standard distribution and depend on whether the variables are individually I(0) or I(1). In this test, maximum dependent lags are specified as 4. The Akaike information criterion (AIC) is used for model selection. Among the dynamic regressors, there is one dependent and ten independent variables. All variables’ lag forms were used. 98 observations are included after making certain adjustments. The sample size covers data for the period between August 2007 and August 2015. Instead of the conventional critical values, this test involves two asymptotic critical value bounds, depending on whether the variables are I(0) or I(1) or a mixture of both. If the test statistic exceeds their own upper critical values, then there is evidence of a long-run relationship; if it breaks below critical values, the null hypothesis of no cointegration cannot be rejected and if it lies between the bounds, it is inconclusive to make an inference. If the test statistic exceeds its upper bound, then the null hypothesis of no cointegration can be rejected regardless of the order of integration of the variables. The numbers of models evaluated are 195.312.500. According to the test results, F-statistics was found as 4.280632, which is above 1% critical value. This means that the model has one cointegration and all variables can be accepted like I(0). Selected ARDL model is ARDL (3,2,0,0,2,0,2,3,2,0). ARDL cointegration test shows that there are significant and positive relationships between liquidity and deposits, return on equity, government bond rate, and capital adequacy ratio (Table 5.4). Coeffient of deposits is the highest one among positive and significant relationships and it is around 0.74, which means 1 percentage change increase in deposits positively impacts liquidity with 0.74%. Coefficents of return on equity, capital adequacy ratio and government bond are very low. However, there are negative and significant relationships between financing (credit), required reserves, Sukuk stocks, and interbank rate. Among these relationships, coefficient of financing (credit) is bigger than coefficient of other significant factors. According to ARDL results, there is no significant relationship between liquidity, CDS rate and inflation rate. There is a negative relationship between liquidity and interbank rate. These results show evidence for responding to research questions. First of all, financing, deposits, return on equity, required reserves, Sukuk, interbank rate, government bond rate, and capital adequacy ratio are determinants of liquidity risk. Secondly, some macroeconomic factors (government bond rate and interbank rate) are significant for impacting liquidity whereas inflation rate is found insignificant over liquidity requirement rate of PBs according to ARDL approach. The same result was found by Aldasoro and Faia (2016) that LCR will reduce systemic risk at the

66

5 Dynamic Determinants of Liquidity Management in Turkish Participation Banks

Table 5.4: ARDL Cointegration Tests. Variable

Lag

Coefficient

DCDS



−.

−.

DCRE



−.

−.*

DDEP



.

. ***

DLINF



.

.

ROE



.

.**

DRR



−.

−.***

DSUK



−.

−.***

INTER



−.

−.***

BOND



.

.**

CAR



.

.***

.

.***

C

t-statistic

Notes: 1) The upper limit of the critical value for the F-test (all I(0) variables) is 1.98 (5%) and 2.41 (1%) and for the t-test, 4.280632 (1%). Critical values obtained from Pesaran et al. (2001). 2) *, **, *** indicate respectively 10%, 5% and 1% significance level.

beginning phase and then increase the risk in the final phase. They claim that high LCR reduce the insurance function of interbank markets, and LCR impose unnecessary liquidity shortages on banks, which are mildly leveraged, and market makers21 when applied equally to all banks. For this reason, Aldasoro and Faia (2016) find evidence for differentiating the LCR requirements for different banks. These results between liquidity and its determinants exist in short-run. For long-run relations, the determinants in the following sub-section are studied.

5.3.5 Long Run Relations of PBs The conditional long run model can then be produced from the reduced form solution of Eq. (4), when the first-differenced variables jointly equal zero. The long run coefficients and error correction model are estimated by the ARDL approach to cointegration, where the conditional ECM in Eq. (4) is estimated using OLS and then the Schwarz–Bayesian criteria is used to select the optimal lag structure for the ARDL specification of the short-run dynamics. The ECM is then used to conduct the

21 Market makers are acting as primary dealers to provide liquidity in interbank markets.

5.3 Results and Discussion

67

conventional Granger non-causality tests, as noted in Granger et al. (2000). Cointegration (CointEq(-1)) is statistically significant and the coefficient is −0.36385. T-statistic is −7.786565 and significant at 1% level. This shows that there is a meaningful long run cointegration. Long run cointegration is found in the following equation: LRR = 0.0011*CDS − 0.0642*CRE + 0.4353*DEP − 0.000029*INF − 0.0139*ROE + 0.166*RR − 0.0101*SUKUK + + 0.0403*INTER − 0.0564*BOND + 0.0335*CAR + 20.9793

(5)

Long run Coefficients can be seen in Table 5.5. Alternatively, heteroscedasticity and autocorrelation is controlled. For heteroscedasticity, Breusch-Pagan-Godfrey test was used. Results of this check indicate that heteroscedasticity, auto, and serial correlation do not exist here. Evidently, heteroscedasticity hypothesis was rejected. Then Breusch-Godfrey serial correlation LM test is used, in which lag was specified as 4. The results show that there is no serial correlation.

Table 5.5: Long Run Coefficients. Variable CDS LCRE LDEP

Coefficient . −. .

t-statistic .* −. .

INF

−.

−.**

ROE

−.

−.**

RR

.

.

−.

−.

INTER

.

.

BOND

−.

SUK

CAR C

. .

−.* . .***

Note: *, **, *** indicate respectively 10%, 5% and 1% significance level.

Long run relations show that there is a negative relation between inflation and liquidity ratio of PBs and it is statistically significant at 10% level. However, coefficients of these variables are very small, which means these variables will significantly but weakly impact the liquidity risk management in the long run. Since liquidity risk is affecting banking sector mainly in the short run, the cofficients of the variables

68

5 Dynamic Determinants of Liquidity Management in Turkish Participation Banks

impacting liquidity in the long run are expected to be small comparing to factors influencing in the short-run. Although there was no significant relation between inflation and liquidity of PBs in short-run, this long run relationship, found above, shows that there is a lag for the inflation rate to impact liquidity of PBs. When inflation increases, the cost of liquid assets increases, therefore, banks want to keep less liquid assets in the long run, as expected. This relationship evidences that macroeconomic factors impact liquidity in the short run, except inflation rate, which is affecting the liquidity of these banks in the long run. On the other hand, there are no significant relations found between liquidity, financing (credit), and deposits in the long run. Return on equity is also found to be negatively correlated with liquidity at 5% significance level. Accordingly, there is a clear trade-off between keeping more liquid assets and making a profit. Other variables do not have statistically significant relations with the liquidity. An increase in the CDS rate (market liquidity shortage) corresponds to an increase in banking liquidity. The results are consistent with other findings in the literature. PBs’ ROE has adverse effects on the size of its liquidity as evidenced by Bonner et al. (2013). Although Gehde-Trapp et al. (2015) find that illiquidity strongly affects CDS premiums; it is established that when CDS rate increases, banks increase their liquidity by using their liquid assets stocks as a countercyclical buffer. Akhtar et al. (2011) found that the relationship of networking capital to net assets and the size of banks with liquidity risk were positive but insignificant for banks studied in Pakistan. It was also found that the capital adequacy ratio in Islamic banks and return on asset in conventional banks have a positive but insignificant relationship with liquidity risk. Bonner (2014) finds that an increase in profit and deposits correlates to an increase in liquidity holdings. An increase of an institution’s capital ratio increases its liquidity holdings. While an institution’s profit and capital ratios have moderate effects on the size of its liquidity buffer, its deposits from clients have a large impact. The positive effect of deposits is also found by Haan and End (2013) and is likely attributable to a lack of funding diversification. Banks with large amounts of deposits are highly concentrated in a single source of financing and, therefore, are especially prone to liquidity risks. Adrian et al. (2017) study an inactivity of dealer financial positions after the financial crisis of 2007–2009, and use high-frequency trade and quote data for US Treasuries and corporate bonds, but they do not find clear evidence of deterioration in market liquidity.

5.3.6 Granger Causality Test Results for Turkish Banks Some tests were conducted to determine the patterns of Granger causality between each variable pair. Correlation does not necessarily imply causation. Granger (1969) approaches the question of whether X causes Y equivalent by asking how much of

5.3 Results and Discussion

69

the current value of Y can be explained by past values of X, and then to see whether adding lagged values of X can improve the explanation. Y is said to be Grangercaused by X if X helps in the prediction of Y, or equivalently, if the coefficients on the lagged X’s are statistically significant. Granger causality measures precedence and information content. According to the results of the VEC Granger Causality or Block Exogeneity Wald Tests, PBs’ liquidity are related to financing (credit) at 1% significance level, to Sukuk stocks at 5% significance level, and to interbank interest rate at 10% significance level (Table 5.6). This means that financing volume is directly affecting liquidity of the PBs because financing provided to household and real sector are not highly liquid instruments. If banks are forced to keep a higher level of liquidity, they have to decrease financing provided to the real sector because it is found in ARDL approach that there is a negative and significant relationship between liquidity and financing of PBs. Accordingly, both Sukuk and the interbank money market interest rates affect the liquidity of PBs, not vice versa. Sukuk is related to the highly-liquid instrument stocks whereas interbank money market rate shows the level of market liquidity. This means that PBs need Sukuk for their liquidity management and cannot remain independent from the interbank money market rates when managing their liquidity. This result shows that PBs liquidity management is causally related to market liquidity, not vice versa. Bonner (2014) found that liquidity regulation is positively associated with the growth of bank claims, short-term interest rates, as well as lending rates to the private sector but during crises, all three previously significant factors change their sign and are therefore negatively associated with liquidity regulation. This result can be interpreted as liquidity regulation having adverse effects on bank lending and short-term interest rate in stress times. Additionally, there are Granger causality relations between interbank interest rate and inflation rate, required reserves, and CDS rate of the country. Inflation rate shows that interbank rate is affected by the inflation rate. Indeed, the relation between required reserve, CDS rate and interbank rate is in accord with expectations, since CDS rate exemplifies tensions in the market, and required reserve has direct influence on funding liquidity of banks. It can be argued that there is indirect relation between liquidity ratio and required reserves, inflation rate, and CDS rate, implying that funding liquidity has indirect causal relation with market liquidity. Andries et al. (2017) found that international swap lines that directly influence interbank interest rate and CDS rate of the country not only enhanced market liquidity, but also reduced risks associated with micro-prudential issues. One of the reasons for this outcome is that both PBs and conventional banks have been regulated by the same conventional regulatory framework. PBs have to meet conventional liquidity standards when managing their assets. This framework requires arranging liquidity management with the aim of avoiding possible arbitrage between the two markets. For this reason, PBs have to look at market liquidity. Since

. (.)

. (.)

. (.)

. (.)

. *** (.)

. * (.)

. (.)

. (.)

. (.)

. (.)

. (.)

. (.)

. (.)

. (.)

CRE

DEP**

RR

BOND

INTER ***

CAR

CDS

. (.)

INF

. (.)

LRR

INF

LRR**

Dependent Variable

Table 5.6: VECM Based Granger Results (PBs).

. ** (.)

. * (.)

. (.)

. (.)

. (.)

. ** (.)

. ** (.)

.* (.)

CRE

. (.)

. (.)

. (.)

. (.)

. (.)

. (.)

. (.)

. (.)

DEP

. (.)

. (.)

. ** (.)

. (.)

. ** (.)

. (.)

. (.)

. (.)

RR

. ** (.)

. (.)

. (.)

. (.)

. (.)

. (.)

. (.)

. (.)

BOND

. (.)

. (.)

. (.)

. (.)

. *** (.)

. (.)

. (.)

.* (.)

INTER

CAR

. (.)

. (.)

. (.)

. (.)

. (.)

. (.)

.* (.)

. (.)

statistics of lagged st differenced term [p-value]

Independent Variable

. (.)

. *** (.)

. (.)

. (.)

. (.)

. (.)

. (.)

. (.)

CDS

. (.)

. (.)

. (.)

. (.)

. (.)

. (.)

. (.)

. (.)

. (.)

ROE

. (.)

. (.)

. (.)

. (.)

. (.)

. (.)

. (.)

. (.)

.** (.)

SUKUK

70 5 Dynamic Determinants of Liquidity Management in Turkish Participation Banks

. (.)

SUKUK

. (.)

. * (.) . (.)

. *** (.) . (.)

.*** (.) . (.)

. (.)

Notes: 1) *, **, *** indicate respectively 10%, 5% and 1% significance level. 2) The figures in parenthesis are the standard errors.

. (.)

ROE *** . (.)

. (.) . (.)

. (.) . (.)

. (.) . (.)

. ** (.) . (.)

. (.)

5.3 Results and Discussion

71

72

5 Dynamic Determinants of Liquidity Management in Turkish Participation Banks

they are under the same regulatory framework and the same economic system, PBs, which have only 5 to 6% of the market share, have to follow their conventional counterparts. This means that PBs are forced to follow conventional banks to catch up margins. Therefore, the availability of low-risk investments and short-term Sukuk would be a factor that significantly reduces the opportunity cost of PBs. Granger analysis for conventional banks shows that there are significant causality relations between liquidity ratio and credit and deposits (Table 5.7). Coefficents of both deposits and credit are bigger than 1%, which means a one percentage change in these factors causally impacts to change in liquidity of banks. This means that when credit and deposits change, liquidity ratio will change. Furthermore, liquidity ratio has causal relations with inflation rate, interbank rate, and capital adequacy ratio. It can be concluded that there are bilateral relations between credit and liquidity ratio. This means that credit volume indirectly affects liquidity of the banks since credits given to household and real sector are not highly-liquid instruments. If banks are forced to keep a higher level of liquidity, they have to decrease credit given to the real sector. However, liquidity of conventional banks has statistically insignificant relationships with the interbank rate and securities, though these variables show significant relationships with the liquidity ratio of PBs. On the other hand, interbank rate has causal relations with the liquidity of conventional banks. This indicates that conventional banks’ liquidity is driving the market liquidity, and interbank rate is dependent on the liquidity of conventional banks. As it is known from the Granger causality test of PBs, their liquidity is dependent on the interbank rate, and interbank rate is dependent on the conventional banks’ liquidity. This means that PBs’ liquidity has indirect causal relations to the conventional banks’ liquidity. Since the Granger causality tests only provide information on the direction of the variables’ causal interactions (Mansor and Shah, 2012), impulse-response functions are simulated to further assess the dynamic interactions of the variables.

5.3.7 Impulse Response Test Results for PBs Impulse response function tests are conducted to analyze the effects of a liquidity shock. Vector Decomposition (VDC) test results correspond to impulse response analysis obtained for PBs. The impulse-response function analysis is used to complement the Granger analysis to get the magnitudes and signs of a variable’s responses to impulses by other variables. The impulse-response functions can capture both direct influences of a variable on another variable, as well as indirect influences that are propagated through other variables. Granger-causality may not explain all relations and the interactions between the variables. Therefore, it is interesting to know the response of one variable to an impulse in another variable, and to investigate these relationships in a higher dimensional system. If there is a

. (.)

. (.)

. (.)

. (.)

. (.)

. (.)

. (.)

. (.)

CRE***

DEP

RR*

BOND***

. (.)

INF

.* (.)

LRR

INF**

LRR***

Dependent Variable

. (.)

. (.)

. (.)

.*** (.)

.** (.)

CRE

Table 5.7: VECM Based Granger Results (CBs).

. (.)

. (.)

.* (.)

. (.)

.** (.)

DEP

. (.)

. (.)

. (.)

.** (.)

. (.)

RR

. (.)

. (.)

. (.)

. (.)

. (.)

BOND

.*** (.)

. (.)

. (.)

. (.)

. (.)

. (.)

INTER

CAR

.** (.)

. (.)

. (.)

. (.)

.** (.)

. (.)

statistics of lagged st differenced term [p-value]

Independent Variable

. (.)

. (.)

. (.)

.** (.)

.** (.)

. (.)

CDS

. (.)

. (.)

. (.)

. (.)

. (.)

. (.)

ROE

(continued )

. (.)

. (.)

. (.)

.** (.)

. (.)

. (.)

SEC

5.3 Results and Discussion

73

.* (.)

. (.)

. (.)

. (.)

.*** (.)

. (.)

. (.)

. (.)

CAR***

CDS***

ROE

SEC**

.** (.)

. (.)

. (.)

. (.)

. (.)

CRE

.** (.)

. (.)

. (.)

. (.)

. (.)

DEP

. (.)

. (.)

.** (.)

. (.)

.*** (.)

RR

. (.)

. (.)

.** (.)

.** (.)

. (.)

BOND

. (.)

. (.)

. (.)

. (.)

INTER

CAR

. (.)

. (.)

.*** (.)

.* (.)

statistics of lagged st differenced term [p-value]

Independent Variable

Notes: 1) *, **, *** indicate respectively 10%, 5% and 1% significance level. 2) The figures in parenthesis are the standard errors.

. (.)

INF

.** (.)

LRR

INTER ***

Dependent Variable

Table 5.7 (continued )

. (.)

. (.)

.** (.)

. (.)

CDS

. (.)

. (.)

. (.)

. (.)

ROE

. (.)

.* (.)

.** (.)

.*** (.)

SEC

74 5 Dynamic Determinants of Liquidity Management in Turkish Participation Banks

75

5.3 Results and Discussion

reaction of one variable to an impulse in another variable, it is called the latter causal for the former. The Y-axis of Figure 4.1 shows the shock given to the LCR and the X-axis shows the response of a variable within ten periods. When a shock is given to the liquidity for PBs, Figure 5.1 depicts that the bond, interbank rate, capital level, CDS rate, return on equity and Sukuk stocks impulse on liquidity, meaning that these six variables have causal relations with liquidity. This result has important implications for regulators, monetary authorities, and banks because it shows that liquidity shocks have an effect on interbank rate which may increase market stress. Moreover, liquidity shocks have the potential to melt down the capital level of banks. The relationship between liquidity and CDS indicates that low

Response to Cholesky One S.D.Innovations ± 2 S.E. Response of LLCR to LLCR

.06 .04

.008

.02

.004

.00

.000

–.02

–.004

–.04

1

2

3

4

5

6

7

8

9

10

Response of LDEP to LLCR

.02

–.008

.02 .01 .00 –.01 1

2

3

4

5

6

7

8

9

10

1

2

3

4

5

6

7

8

9

10

Response of BOND to LLCR 1.0

40

0.5

20 0.0

0

–.01

–20 1

2

3

4

5

6

7

8

9

10

Response of INTER to LLCR

.8

–40

–0.5 1

2

3

4

5

6

7

8

9

10

Response of CAR to LLCR

.6

2

3

4

5

6

7

8

9

10

Response of ROE to LLCR 4

–.4

1

2

3

4

5

6

7

8

9

10

Response of SUKUK to LLCR

400 300

2

200 100

0

0 –2

–100 1

2

3

4

5

6

7

8

9

10

–200

4

5

6

7

8

9

10

Response of CDS to LLCR

–20

–.2 1

3

0

.0

–4

2

20

.2 .0

1

40

.4

.4

–4

–.02

Response of RR to LLCR

60

.00

–8

Response of LCRE to LLCR

.04 .03

80

.01

–.02

Response of LINF to LLCR

.012

1

2

3

Figure 5.1: Impulse Response Results for PBs.

4

5

6

7

8

9

10

–40

1

2

3

4

5

6

7

8

9

10

76

5 Dynamic Determinants of Liquidity Management in Turkish Participation Banks

level of liquidity in banks decreases demand for government bonds, which will then increase risk premium. Increasing risk premium will affect the CDS rate of the country. The relationship between Sukuk stocks and liquidity is evident – if there is liquidity in banks, banks demand more Sukuk. The impulse of return on equity in liquidity is also evident in the findings of long-run relations of ARDL model, where there is negative and significant influence of return on equity on liquidity. On the other hand, liquidity responds when other variables, i.e., return on equity, financing (credit), deposits, and capital adequacy ratios are subject to shocks. The overall impression is that the responses of these variables are close to equilibrium and return to that state within ten periods (ten months). In actuality, ten months is longer period for liquidity risk management of PBs. However, it is possible for PBs to stabilize their cash flows in this period. It can be interpreted that both market liquidity and profitability of PBs have impulse response on liquidity management of these banks.

5.4 Conclusion of Time Series Analysis and Policy Implications The consequences of the last global financial crisis have compelled policymakers, regulators and market players to increase their focus on mitigating liquidity risks. Since Islamic banks are more real-sector oriented than their conventional counterparts, management of liquidity risk of Islamic banks may have more significant effects on financial stability and the economy. However, financing provided to the real sector has cyclical effects on growth. The adoption of appropriate macro-prudential measures and liquidity management mechanisms that are conducive to the development of Islamic banks is therefore essential. In this chapter, a time series based liquidity model is used to assess the liquidity management of banks in Turkey based on unit root analysis, cointegration test, ARDL test, and long-run and short-run analyses. Long-run and shortrun relations are present among the variables. The summarizing results of the model are compared with results of models used in the relevant literature at Table 5.8. Pieces of evidence were found that inform the research questions. According to the ARDL approach, it is found that financing (credit), deposits, return on equity, required reserves,, interbank rate, government bond rate and capital adequacy ratio are determinants of liquidity risk. Secondly, some macroeconomic factors (government bond rate and interbank rate) are significant for impacting liquidity. Thirdly, whereas inflation rate is found insignificant over liquidity in the short run, it is considered negative and meaningful for impacting liquidity of PBs in the long run. As a response to the fourth question, there is direct relationship between liquidity and financing volume of PBs, but indirect relation of conventional banks. Lastly, Granger analysis shows that Islamic banks’ liquidity management has causal relation with market liquidity. Additionally, both Sukuk and the interbank money market interest rates affect the liquidity of PBs but not vice versa. If the Turkish Treasury increases issuance of lease certificates, liquidity risk of PBs will decrease. PBs

5.4 Conclusion of Time Series Analysis and Policy Implications

77

Table 5.8: Summarizing Results. Dependent Variable Liquidity Granger

ARDL Short-Run

ARDL Long-Run

Other Studies

PBs

CBs

PBs

PBs

Required reserves

Ins.

Ins.

(−)

Ins.

Return on Equity

Ins.

Ins.

(+)

(−)

(−): Ariffin (); Bonner () (+): Trad et al. (); Vodova (); Roman and Sargu ()

Capital Adequacy Ratio

Ins.

Ins.

(+)

Ins.

(+): Roman and Sargu (); Chen et al. (); Amin (): Admati et al. (); Ogilo and Mugenyah (); Mongid (); Vodova ()

Inflation

Ins.

Ins.

Ins.

(−)

(−): Amin (); Trad et al. (); Munteanu ()

CDS

Ins.

Ins.

Ins.

(+)

(+): Gehde-Trapp et al. ()

Interbank Rate

Yes

Ins.

(−)

Ins.

(+): Angora and Roulet (); Hong et al. (); Vodova () (−): Vodova (),

Government Ins. Bond

Ins.

(+)

Ins.

(+): Hong et al. ()

Credit (Financing)

Yes

Yes

(−)

Ins.

(−): Leykum (); Hong et al. (); Covas and Driscol () (+): Vodova ()

Securities/ Sukuk

Yes

Ins.

(−)

Ins.

(+): Haan and End ();

Deposits

Ins.

Yes

(+)

Ins.

(−): Leykum () (+): Haan and End (); Bonner ()

Notes: 1) (+): positive and significant, 2) ins: insignificant

cannot remain independent from the interbank money market rates when managing their liquidity. This result shows that PBs liquidity management has causal relations with market liquidity in the short run but not vice versa. Furthermore, liquidity ratio is indirectly related to required reserves, inflation rate and CDS rate. This implies that funding liquidity has an indirect causal relationship with market liquidity.

78

5 Dynamic Determinants of Liquidity Management in Turkish Participation Banks

Moreover, the impulse response analysis strengthened the Granger analysis by showing that bond, interbank rate, capital level, CDS rate and return on equity impulse impact liquidity. This means that these variables have causal relations with liquidity. When other variables are subject to shocks, liquidity responds to the return on equity, financing, deposits and capital adequacy ratio. It can be argued that market liquidity impacts banks’ liquidity, and banks’ liquidity responds to profitability of PBs. The summarized results of both ARDL and Granger tests are reflective of the historical experiences of PBs. The importance of proper liquidity management is put to the test in the case of Ikhlas Finans (Turkey) which faced liquidity problems as well as other problems, including misconduct risk, resulting in its liquidation process in 2001. The lack of Shariah-compliant instruments might force these banks to invest funds into fixed return and long-term financing projects. Thus, the availability of Shariah-compliant instruments in the system is of utmost importance for proper management of liquidity and also for the development of the scheme itself. This is also indicated by Obiyathulla (2008) who states that the outstanding growth performance of Islamic banks in Malaysia owes much to the initiative of establishing the Islamic Interbank Money Market, and issuing highly-liquid Shariah-compliant tradeable instruments. Islamic banks were formed long after the conventional banks were. To provide services for Muslims and other investors, they should be allowed advantages and incentives. Even if these banks have to compete under the same conditions as other banks, the development of these banks will be positively affected provided they are on a level playing field. Liquidity management is at the forefront of the areas most needed to support these banks. Issuing high-quality liquid assets, maintaining Shariah-compliant lender of last resort mechanism, and establishing an Islamic interbank money market, are the essential elements that should be made available to Islamic banks with some urgency.

6 Panel Data Analysis For Evaluating Effects of Liquidity Standards 6.1 Introduction The results of the time series based model shows remarkable consequences for liquidity management of PBs. These outcomes would have crucial implications for evaluating a new regulatory framework for Islamic banks in coming chapters. However, it may be possible for researchers to critique these results regarding data quality because the time series analysis is based on cumulative data from banks (both PBs and CBs). For this reason, in this chapter, the analysis is improved by using micro-level data, and adopting panel data models (Vodova, 2011; Bonner, 2014; Vodova, 2013; Haan and End, 2013) for analyzing factors that affect liquidity management of PBs and CBs of Turkey. For this purpose, several models are developed by including different control variables one by one. Following Kassimatis and Clark (2015), Bonner (2014) and Hilscher and Nosbusch (2010), adding new variables increases the adjusted R-squared of panel models, which is further evidence of the incremental information generated by the new variables. Panel data has many advantages compared to time series analysis (Gujurati, 2011) and involves a cross-sectional dimension and a time series dimension. Although the collection of data from panel analysis is more costly than the collection of time series data, panel data provides more accurate inference of model parameters. There is more degree of freedom in panel data. For complex relations or behavior of markets, panel data captures more information. Furthermore, omitted variables problem can be addressed by using panel data. Finally, dynamic relationships among variables can be covered with the panel data methods.

6.2 Methodology While the horizontal cross-sectional data provides information for only one period for many units, the time series data only gives information for a unit according to periods. If information is requested pertaining to both periods and units, panel data should be used. Moreover, the exclusion of exploratory variables leads to deviation estimates. A similar situation applies to the exclusion of time series variables, which always affect the behavior of units in the same way but differently at each period. Hence, the panel data resolves this problem. It is claimed that panel data is superior to cross-section or time-series, offering a great number of advantages as stated in Baltagi (2005). In particular, panel data allows for the banks to be heterogeneous, offers more informative data, more variability, less collinearity among the variables, more degrees of freedom, and more https://doi.org/10.1515/9783110582901-006

80

6 Panel Data Analysis For Evaluating Effects of Liquidity Standards

efficiency (Baltagi, 2005). Therefore, in panel data, the effects on liquidity may be easier to justify, measure and detect. Indeed, it is expected that panel data would lead to more efficient estimates and reduce the endogeneity problem. Hence, panel data models permit the investigation of problems that cannot be dealt with solely by cross-sections or by time series. Panel variables are strongly balanced, which infers that all banks have data for all months. Since monthly data is used in this book, following Haan and End (2013), Bonner (2014) and also Vodova (2013), it is assumed that the disadvantages of using panel data stated by Baltagi (2005) will not be encountered, such as using a short period while relying on infinitiy individuals. Moreover, there is a possibility in panel data that the individual units are inter-dependent, which is called crosssectional dependence. The panel data also assumes that both slopes and intercepts are constant across banks. The fixed-effect and random-effect models may contribute to controlling the intercept bias, but still, there may be present slope biases that could occur if the effects of a given independent variable are different for each bank or sector. In line with the objectives of this study, two main estimators in this section of the book are used, which are fixed-effect and random-effect for the static panel model. Derivation of panel data estimator: It is difficult to make a real distinction between fixed-effect and random-effect models because each one has their own technical and conceptual advantages, drawbacks, and the assumptions made by both. In panel data analysis, the individual effect contains a constant term and a set of individual variables which may be observed or unobserved. If this individual effect contains unobserved variables and, at the same time is correlated with the vector of explanatory variables, the fixed-effect should be utilized. The random-effect assumption is that the individual specific effects are uncorrelated with the independent variables. The fixed-effect model assumes that there is a correlation between the individual specific effect and the independent variables. If the random-effect models assumption holds, the random-effect model is more efficient than the fixed-effect models. However, if this assumption does not hold, the random-effect model is inconsistent and not efficient (Gujurati, 2011; Baltagi, 2005). The problem of choosing between fixed-effect and random-effect would be resolved after conducting the Hausman test. The within R-squared22 is used for fixed-effect but it is not relied on any R-squared for random-effect estimators but

22 The Stata output gives three different values of R-squared; these are within, between and overall. The between R-squared is giving the variance between separate panel units for the model account, whereas the within R-squared is showing the variance within the panel units. And the overall R-squared is a weighted average of these two.

6.2 Methodology

81

report the within R-squared based on below theoretical framework: It is considered that the model is fitting to the form: yit = α + xit β + Wi + εit

(6)

In this equation, xit is a matrix of explanatory variables with β being the vector of the estimated coefficients and Wi+εit is the error term. The unit-specific residual is Wi, differing among units and its value is constant for any specific individual while εit is the standard residual with the usual properties; mean equal to zero, uncorrelated with itself and others, and homoscedastic in nature i.e. εit ≈ i.i.d.N(0,σ2). Before making necessary assumptions for estimation, it is required to provide more information on reasons for using R-squared by performing some useful algebra on equation (6). Whatever the properties of Wi and εit, if (4) is true, it must also be true that; y
i = α +
xi β + Wi +
εi

(7)

X X X Where
yi = y =Ti ,
xi = x =Ti and ε
i = ε =Ti . Besides, after subtracting t it t it t it (7) from (6), the following equation is developed: ðyit −
yi Þ = βðxit −
xi Þ + ðεit − ε
i Þ

(8)

After solving these three equations, now there is a basis for calculating β. Moreover, when fixed-effect “within estimators” are employed (Stata command xtreg, fe), OLS is used to estimate equation (8). The “between estimators” (Stata command xtreg, be) are found by performing OLS on equation (6). Lastly, random-effect estimators are nothing more than a weighted average matrix of the estimates produced by the “within and between estimators”. When the Stata command xtreg, fe performs OLS on (8) and hence the informed the within R-squared is the ordinary R-squared. The Stata command xtreg, re applies OLS on (8) and consequently none of the R-squared directly applies to it. In summary, R-squared overall corresponds to (6), R-squared between to (7) and Rsquared within to (8). Based on these correspondences, for fixed-effect estimators, the true R-squared is the R-squared within, while in the case of random-effect, none of the R-squared reported has the properties of the R-squared. Therefore, it is not relied on any R-squared within for random-effect, but only within R-squared is shared in tables. List of Banks: The model used in this chapter covers several panel data analyses. First of all, the panel data models are developed to cover both Turkish Participation and conventional banks (24 banks) based on the same variables in the previous timeseries model. Then panel analyses for Turkish PBs (four banks) are studied based on bank level data and these analyses are compared with the Turkish conventional banks (20 banks). Although there are 52 banks in Turkey, many of them are investment banks or state banks, and do not have liquidity ratio data. Since the model is a balanced panel, the banks that do not have data for relevant periods are omitted. The list of banks can be seen in Table 6.1.

82

6 Panel Data Analysis For Evaluating Effects of Liquidity Standards

Table 6.1: List of Banks included for Panel Data Analysis. Panel A: PBs

Panel B: Conventional banks

Albaraka

Ziraat

Turkish Bank

Kuveytturk

Halkbank

ING Bank (Oyakbank)

Türkiye Finans

Vakıfbank

Fibabank

Bank Asya

Tebbank

ICBC(Tekstil Bank)

Akbank

Finansbank

Şekerbank

HSBC

Garanti

AlternatifBank

İşbank

Burganbank

Yapıkredi

Denizbank

Arap Turk Bank

Anadolubank

Dependent Variable: In these models, the dependent variable was liquidity requirement ratio which was announced by the Banking Regulation and Supervision Agency. For bank-level data, the data provided by Central Bank of the Republic of Turkey are used. The Liquidity Requirement Ratio is calculated similarly as the Liquidity Coverage Ratio. In this data, there are some missing values, but these data are calculated based on the formula developed by the BIS (2013c) and IFSB (2015) used by regulatory authorities. On the other hand, a dummy variable is used for the LCR announcement date to see the effects of new regulations on banks’ portfolios. Independent Variables: In randomly effective static panel data models, it is very common to use the least squares or generalized least squares (GLS) if estimated parameters are known. In these base balanced panel analyses, the same variables used in the previous model are employed. These are; total assets (AS), required reserves (RR), credit (CRE), deposits (DEP), capital adequacy ratio (CAR), securities (SAFS), return on equity (ROE), interbank rate (INTER), government bond rate (BOND), and inflation (INF), to alleviate the endogeneity problem to the base model. The inflation, credit, and deposits are used in lag form. Hence, the same relevant variables used in the previous model, and by many researchers were chosen for this base model. Variables added for Robustness Check: However, several variables are added to the base model for robustness check, one by one. These are short-term assets (Bonner, 2014), high quality liquid assets (Banerjee and Mio, 2014), cash, total assets (Bonner, 2014; Cucinelli, 2013; Ahmed et al., 2011; Angora and Roulet, 2011),

6.2 Methodology

83

capital (Bonner, 2014; Akhtar et al., 2011; Vodova, 2011; Vodova, 2013) and net profit (Bonner, 2014; Ahmed et al. 2011), return on assets (Bonner, 2014; Amin, 2016; Roman and Sargu, 2015; Chen et al., 2015; Vodova, 2013) and short-term assets to total assets ratio-LATA, (Banerjee and Mio, 2014), are added to the previous model as new variables. One of the contributions of the book is to test the net cash outflows (NCO) to assess the relationship between liquidity risk and this variable. Short-term assets: Short-term assets are the assets that have a maturity of less than one year. Since all short-term assets are not accepted as high-quality liquid assets, it is expected that there would not be any high correlation between shortterm assets and long-term assets. High-quality liquid assets: This variable is added for robustness check. Assets accepted as high-quality liquid assets are cash, central bank reserves, high-quality public bonds or treasury bills, financial assets held by multinational investment banks, and sovereign bonds in foreign currency with high ratings. Correlation matrix shows that there is no perfect collinearity between this variable and liquidity ratio, because an increase in this variable does not mechanically increase the LCR. Since cash is also accepted as HQLA, increasing cash may decrease profitability which may trigger net cash outflows. At the end, if an increase of net cash outflows exceeds an increase of HQLA, the LCR decreases. Cash: Cash is added for robustness check. Increasing cash may also increase the LCR but it harms the profit, which may trigger the cash outflows. Therefore there is no perfect collinearity between the dependent variable and cash. Total assets: Total assets as an independent variable is used in many studies to measure the effect of changes in size on liquidity. The initial expectation is that if banks increase their asset size they will have a high liquidity ratio. Capital: Regulatory capital (CAP) is included for robustness check of capital adequacy ratio. Net profit: Net profit is important for understanding the relationship between liquidity and efficiency of these banks. Return on assets: ROA is measured by net income to total assets ratio (Roman and Sargu, 2015; Amin, 2016; Angora and Roulet, 2011; Ahmed et al., 2011; Akhtar et al. 2011). Amin (2016) found that a 1% increase in return on asset will reduce liquidity risk by 1.3%. Short-term assets to total assets ratio: Since all short-term assets are not accepted as high-quality liquid assets, it is wanted to see the effects of increasing short-term assets’ share in total assets on liquidity of banks. If increasing this ratio increases the liquidity, this can be interpreted as the banks investing in liquid short-term assets. Net cash outflows: Net cash outflows are calculated as the difference between cash outflow and cash inflows. Net cash outflows are related to funding liquidity. One of the significances of the study is to include net cash outflows for evaluating the determinants of liquidity. Leykun (2016) and Hong et al. (2014) also used net

84

6 Panel Data Analysis For Evaluating Effects of Liquidity Standards

cash flows as an independent variable but their dependent variable is not liquidity coverage ratio. The base model is as follows: LCRY = αi + β1it *DUMMY1 + β2it *RR + β3it *ROE + β4it *CAR + β5it *LINF + β6it *CDS + β7it *INTER + β8it *BOND + β9it *LCRE + β10it *LDEP + β11it *SAFS + εit

(9)

i = 1, 2, 3, 4, 5 and t = 1, 2, 3 . . . 101 Where, L is the natural logarithm of a variable, εit is the common error term, and α is the bank-specific effect. Despite the fact that including various bank-specific variables may raise a multicollinearity issue, there are no perfect collinearity relations among variables according to the initial analyses (except between deposits and credit; between bond and interbank rate). Appendix I and Appendix J show that observations for all variables (except among deposits and credit and also among bond and interbank rate) in the correlation matrix show that all of the correlation coefficients are below 80%. However, government bond and interbank rate are expected to be correlated. The same correlation is expected between deposits and credit (financing). Equation (9) can be rewritten in the following format; in which some variables are grouped: LCRY = α + β1it *DUMMY1 + β2it *Size + β3it *Market Liquidity + β4it *Macro Economic Variable + β5it *Banking Liquidity + β6it *Profitability + β7it *Capital + ε

(10)

Where Size consists of credit (CRE) and deposists (DEP), banking liquidity refers to required reserves (RR), securities available for sale (SAFS). Cash and high-quality liquid assets (HQLA) are added for robustness check. Macro economic variable includes inflation (INF) and cds of the country (CDS). Market liquidity covers interbank interest rate (INTER), and government bond rate (BOND). Profitability includes return on equity (ROE). Net profit (NP) and return on assets (ROA) are included for robustness check. Liquidity, the dependent variable, is measured as the liquidity coverage ratio as defined and formulated by BIS (2013c). This ratio is found by dividing high-quality liquid assets stocks by the net cash outflows. The LCR standards were announced at the end of 2013, and many banks started to apply these standards before its implementation day. The BRSA has been announcing this ratio since January 2014. However, liquidity requirement ratio is calculated similarly as the LCR. Banks’ liquidity is expected to be dependent on individual behaviour of banks, market, macroeconomic environment, and structural factors (exchange rate regime, regulations, etc.). Short-term assets, high-quality liquid assets, cash, total assets, capital, net profit and net cash outflow are added to the model as control variables. Table 6.2. shows the list of variables and gives information about the source of data.

6.2 Methodology

85

Table 6.2: List of Variables. Variables

Description

Source

LRR Liquidity Ratio: High-Quality Liquid Assets / CBRT (Dependent) Net Cash outflows RR

Required reserves of bank kept as shortterm assets

CBRT

ROE

Return on Equity: Net Profit / Equity

CBRT

Capital Adequacy Ratio: Tier  Capital + Tier  Capital/RiskWeighted Assets

CBRT

CAR

LINF

Inflation Ratio (log): Consumer Price Index

TUIK – Turkish Statistical Institute

CDS

Credit Default Swap rate of the Country

Bloomberg Terminal

INTER

Interbank Money Market Rate

Bloomberg Terminal

BOND

Treasury Bonds:  Months Return

CBRT EVDS

LCRE

Total Credits (Financing) (log)

CBRT

SAFS

Securities available for sales: Lease Certificates and Sukuk

CBRT and checked with TKBB Database

LDEP

Total Deposits (log)

CBRT

LHQLA

High-quality liquid assets (log)

CBRT

LSTA

Logarithm of Short-term assets (log)

CBRT

CASH

Cash

CBRT

LAS

Total assets (log)

CBRT

LCAPITAL

Logarithm of Shareholders equity (log)

CBRT

NCO

Net Cash Outflow

CBRT

ROA

Return on Equity: Net Profit/Total Shareholders’ Equity

CBRT and checked with banks’ quarterly balance sheets

LATA

Short-term assets to Total Assets Ratio

CBRT and checked with banks’ quarterly balance sheets

Dummy

Regulations announced for the LCR (Dummy)

ε

Error Term

α

Constant

86

6 Panel Data Analysis For Evaluating Effects of Liquidity Standards

Several independent variables are the same as in the first time series based model too. These are Sukuk stocks (mainly lease certificates of Turkish Treasury and Sukuk issued by other banks and MDBs), logarithm of total credits (CRE), logarithm of total deposits (DEP) and return on equity (ROE), the required reserve (RR), logarithm of inflation rate, CDS rate of Turkey, three months treasury bills rate (BOND) and three months LIBOR rate (INTER) and Capital adequacy ratio (CAR) (Akhtar et al. 2011; Demirgüç-Kunt et al. 2003; Deming and Hong, 2012; Ergec and Arslan, 2013; Praet and Herzberg, 2008; Bonner, 2014; Mohamad et al. 2013; Arif and Anees, 2012).

6.3 Data Monthly bank-level data over the period of April 2007 to August 2015 is used (Bonner, 2014; Haan and End, 2013). Banking data in this chapter are mainly extracted from the Central Bank of the Republic of Turkey (CBRT), Banking Regulation and Supervision Agency (BRSA), PBs Association of Turkey’s databases and checked with Banks’ quarterly financial positions. Macro-economic variables (such as CDS, government bond rate and interbank rate) are extracted via Bloomberg Terminal. The inflation rate of Turkey was derived from Turkish Stats. In this section, Stata statistical software is utilized because of having more choices than Eviews, and is one of the most reliable statistical software programs used by researchers.

6.4 Discussion of Model and Summary of Variables In the studies using panel data, it is necessary to include the variation in the differences between the units, or the differences arising from differences occurring over time. One way of doing this is to assume that the current variation leads to changes in some, or all, of the coefficients of the regression model. Models where the coefficients are assumed to change with units, or units with respect to time, are called “fixed-effect models”. Since lagged liquidity buffers may be correlated with the panel-level effects and the time dimension is relatively limited, there is a risk that the estimator will be inconsistent. The Turkish banks’ (all banks) descriptive statistics are presented in Table 6.3. For PBs and CBs, a summary of variables is provided in Appendix L and Appendix M, respectively. Presenting descriptive statistics of models (mean, standard deviation, minimum and maximum) is the starting point for the empiricial assessment. In this context, mean values refer to a measure of the average of the underlying variables over the examined period, whereas, minimum values indicate lowest value and maximum values show highest value in the sample. Also, standard deviation defines how

6.4 Discussion of Model and Summary of Variables

87

much a variable shows variation or diversification from the average value. Mean and standard deviation have positive values for all samples, but minimum values have negative values for net profit, return on equity, return on assets and logarithm of liquidity risk of all banks, PBs and CBs. Table 6.3: Summary of Variables. Variable

Obs

Mean

Std.Dev.

Min

Max

STA

,

..

..

.

..

HQLA

,

..

..

.

..

LCR

,

.

.

.

.

CAR

,

.

.

.

.

CASH

,

..

.



..

SAFS

,

..

..



..

RR

,

..

..



..

LATA

,

.

.

.

.

CRE

,

..

..

.

..

AS

,

..

..

.

..

DEP

,

..

..

.

..

CAP

,

..

..

.

..

NP

,

.

.

–.

..

ROE

,

.

.

–.

.

ROA

,

.

.

–.

.

CDS

,

.

.

.

.

INF

,

.

.

.

.

INTER

,

.

.

.

.

BOND

,

.

.

.

.

NCO

,

..

..

.

..

LLCR

,

.

.

–.

.

LAS

,

.

.

.

.

LINF

,

.

.

.

.

LCRE

,

.

.

.

.

LDEP

,

.

.

.

.

88

6 Panel Data Analysis For Evaluating Effects of Liquidity Standards

Descriptive statistics for these banks show that all variables in the model have statistically significant relations. There are 2424 bank level observations of all banks (404 observations for PBs and 2020 observations for CBs) for 20 variables which are generally accepted as sufficient for such panel data analysis. Subsequently, additional analysis is conducted to test the differences between the mean values and variances of all banks, PBs and CBs respectively. According to the Z-test shown in Appendix N, the results suggest that all independent variables have significantly different effects on all banks, PBs and CBs. Z-values are bigger than 1.96, which shows that the variable represents significant differences. Research Questions: This section aims to deliver answers for the following questions: 1. Is there any relationship between liquidity of Islamic banks and market liquidity based on bank-level data analysis? 2. Are there any relationships between investing in liquid assets and short-term assets, high-quality liquid assets, cash, short-term assets to total assets ratio, total assets, capital, profit and net cash outflows of PBs? 3. Are there any significant differences between the factors that affect the liquidity of the Islamic banks and the liquidity of their conventional counterparts?

6.5 Results of Panel Models Panel data can be analyzed by using “fixed-effect models” as well as by using “random-effect GLS regression models,” which is the result of differences in units, or differences in units and time. In random-effect models, the changes that occur according to units, or units and time, are included as a component of the error term in the model. The main reason for this inclusion is to avoid the loss of the freedom level encountered in fixed-effect models. Furthermore, in the random-effect model, the effects observed in the cross-section data, units and time, and the effects outside the sample are all taken into account. Linear random-effect models are estimated via Generalized Least Squares. If there are no omitted variables, the randomeffect model is accepted as more preferable to fixed-effect model because the effects of time-invariant variables can be estimated, and standard errors of estimates tend to be smaller (Williams, 2015; Gujurati, 2011). In the fixed-effect model, each horizontal section has its own fixed intercept value, whereas in the random-effect model, the variable constant reflects an average value of all horizontal section constants. For this reason, the random-effect model is checked and then the fixedeffect model is tested. Random-effect and fixed-effect models are tested for the Turkish banking system, PBs and CBs separately. Each panel covers ten models. Results of all the selected models after conducting Hausman tests (fixed-effect for the Turkish banking

6.5 Results of Panel Models

89

system and PBs; random-effect for CBs) are presented in Table 5.4, 5.5 and 5.6 respectively. Random-effect results for all banks and PBs are shared at the Appendix O and Appendix P respectively. Results of fixed-effect for CBs are available in Appendix R. The p-values are provided in parentheses and the ***, **, *, indicate the significance level at 1%, 5%, and 10% respectively. The within R-squared value is presented at the bottom of each model in the table.

6.5.1 Results for Turkish Banking System Model results for all Turkish banks, PBs and CBs are presented in Tables 6.4–6.6 respectively. The p-values are provided in parentheses and the ***, **, *, indicate the significance level at 1%, 5% and 10% correspondingly. The within R-squared value is presented at the bottom of each model in the table. Coefficients in fixedeffect models are smaller than coefficients in random-effect for all banks. The findings of the fixed-effect models for the Turkish banking sector show required reserves, capital adequacy ratio, interbank rate, credit (financing) and securities available for sale have statistically meaningful and positive relations with the liquidity requirement ratio at 1% significance level (Table 6.4). The signs of these variables are positive as expected, which means the increase in these variables can affect dependent variables positively. Coefficients of required reserves and securities available for sale are bigger than 1, which means a one percentagechange in these factors affects more than one percentage change in liquidity. However, other factors’ coefficients are less than one, implying that a change of these factors impacts liquidity less than the mentioned two factors. These results are consistent with several studies, but different from others. For example, although Admati et al. (2013) and Vodova (2013) suggest positive relations between liquidity and holding more equity, Gorton and Winton (2000) finds that the increase in capital requirements has a negative impact on bank liquidity. However, their study is very dated compared to the latter. For credit (financing) relations, Bonner’s (2014) results suggest that a liquidity requirement increases banks’ demand for long-term loans, making them more expensive, which means there are direct relations between liquidity requirement and liabilities (deposits and other loans, especially long-term loans). Amin’s (2016) results show that capital has a positive and significant impact on liquidity risk. On the other hand, CDS rate and deposits show negative and significant relationships with the liquidity at 1% and 5% significance level respectively. Geode-Trapp et al. (2015) show that CDS trades have a significant price impact on assets and that the price impact functions are upwards sloping. Vodova (2013) finds a negative relationship between interbank rate and liquidity, opposite to the positive impact found in the study. Yet, the findings related to interbank rate is consistent with that of Vodova (2011), who studied factors affecting liquidity risk for

90

6 Panel Data Analysis For Evaluating Effects of Liquidity Standards

commercial banks in the Czech Republic. Although it is found that inflation has insignificant influence on liquidity risk for all banks in the sample, Amin’s (2016) evidence shows that inflation poses significant and negative effects on liquidity risk. The constant term has negative and statistically significant relations. However, within R-squared is low (0.2819), which means all these factors jointly account for 28.19% of the variability in liquidity. In Model 2, total assets as a new variable is added to the previous model, and total assets show positive and significant relationships with liquidity. This means that if total assets size of the banking sector increases, the liquidity of the banking sector will also increase. The coefficient of total assets is 1.4921, which implies that one percentage change in total assets results in 1.4921% change in liquidity of banks. However, after total assets are added as a new variable, credit becomes insignificant. After adding one more variable, within R-squared increases to 32.32%, which means the new variable increases the power of all factors, jointly explaining the impact on liquidity. This result is inconsistent with the findings of Angora and Roulet (2011), and Cucinelli (2013), both studies find that liquidity ratio is negatively related to the asset size. In Model 3, cash is added to the previous model, and cash shows negative and significant relations. The coefficient of cash is very big (5.14) compared to other variables’ coefficients. The magnitude of the coefficient of cash shows that liquidity of banks is very sensitive to the change of cash stock of banks. Moreover, after adding cash as a new variable, the credit starts to show negative and significant relations, which means credit growth negatively affects liquidity risk management of Turkish banks. All other variables however also show the same relations. In Model 4, net profit is added, which shows insignificant relations and within R-squared increases. In Model 5, the short-term asset size is added to the previous model, and it shows negative and significant relations with the liquidity. The coefficient of shortterm assets is bigger than 1% (2.3208). This finding is consistent with the analysis of Bonner (2014). Within R-squared also continues to decrease. In Model 6, the high-quality liquid asset is added to Model 5, and it shows a positive and significant relationship with the liquidity and it has big coefficient (1.5508). All other variables show the same relations. In Model 7, return on equity is added to the previous model, and it shows an insignificant relationship with the liquidity and all other variables keep their positions. In Model 8, the capital size of banks is added, and it shows a negative and significant relationship with the liquidity. Magnitude of the coefficient of capital shows that the liquidity of banks is more sensitive to the change in capital stock than the change in capital adequacy ratio. However, return on equity shows negative and significant relations with the liquidity. This finding is consistent with studies which claimed that profitability has a negative influence on the liquidity of banks (Chen et al., 2015; Akhtar et al., 2011; Roman and Sargu, 2015). Nevertheless,

6.5 Results of Panel Models

91

some studies suggest that profitability positively affects liquidity of banks (Amin, 2016; Ahmed et al., 2011; Arif and Anees, 2012). In Model 9, short-term assets to total assets ratio is added to the model, showing positive and significant relations – but the constant term becomes insignificant, whereas capital adequacy ratio, CDS, high-quality liquid assets and interbank rate become insignificant, but net profit shows positive and significant relations. In the last model (Model 10), net cash outflows are added to the model; it shows significant negative relations. Constant term in the last model becomes insignificant, but within R-squared increases (0.4898).

6.5.2 Results of Models for PBs The core results of the fixed-effect for Model 1 of PBs indicate that required reserves, capital adequacy ratio, inflation rate and financing (credit) have statistically meaningful positive relations with the liquidity requirement ratio at 1% significance level (Table 6.5). The coefficients of required reserves, inflation and financing have more than one percentage, implying that changes in these variables have more than one percentage change in liquidity. The coefficient of securities available for sale in model for PBs is smaller than the coefficient of securities available for sale in the model for all banks. The change in inflation and financing is however more effective on PBs compared to the results for all banks concerning the magnitude of their coefficients. Previous papers also support a positive relationship between capital and liquidity risk (Roman and Sargu, 2015; Chen et al., 2015; Amin, 2016). However, securities available for sale and deposits show negative significant relationships with liquidity contrary to the initial expectation, since customer deposits are the most stable source of funding; the use of deposits to provide most of the financing is expected to lower liquidity risk of banks (Bonfim and Kim, 2011). Additionally, although it is expected that there would be positive relations between securities available for sale and liquidity, such expected relations in the fixed-effect model are not found. On the other hand, CDS, interbank rate, government bond rate and return on equity do not show any significant relations with the PBs’ liquidity. The constant term has negative but statistically meaningful relations. Within R-squared is much larger than in Turkish banking sector model (0.7730) implying that all factors in the model mutually account for 77.30% of the variability in liquidity. In Model 2, total assets as a new variable is added to the previous model, and total assets show positive significant relations with the liquidity and its coefficient is smaller than 1%, although it was given an insignificant result in the randomeffect model (Appendix P). This means that PBs’ liquidity can be increased directly by increasing its asset size. However, this result is inconsistent with the result of Ahmet et al. (2011), who show that there is an insignificant relationship between liquidity risk of Pakistan’s Islamic bank and asset size of these banks. However, it is

()

.*** (.)

–. (–.)

. (.)

. (.)

–.*** (–.)

.*** (.)

. (.)

–. (–.)

.*** (.)

()

.*** (.)

–. (–.)

.*** (.)

. (.)

–.*** (–.)

.*** (.)

–. (–.)

.*** (.)

.*** (.)

RR

ROE

CAR

LINF

CDS

INTER

BOND

LCRE

SAFS

. (.)

.** (.)

. (.)

.*** (.)

()

.*** (.)

–.** (–.)

–. (–.)

.*** (.)

.*** (.)

–.** (–.)

–. (–.)

.*** (.)

–.*** –.*** (–.) (–.)

. (.)

.** (.)

–. (–.)

.*** (.)

()

.*** (.)

–.*** (–.)

–. (–.)

.*** (.)

–.*** (–.)

. (.)

.* (.)

. (.)

.*** (.)

()

.*** (.)

–.** (–.)

–. (–.)

.*** (.)

–.*** (–.)

. (.)

.** (.)

. (.)

.*** (.)

()

All banks: Dependent variable LLR

Table 6.4: Panel Data Regressions Results for All Turkish Banks.

.*** (.)

–.** (–.)

–. (–.)

.*** (.)

–.*** (–.)

–. (–.)

.** (.)

. (.)

.*** (.)

()

.*** (.)

–.* (–.)

–. (–.)

.*** (.)

–.*** (–.)

. (.)

.*** (.)

–.* (–.)

.*** (.)

()

–.*** (–.)

.*** (.)

–. (–.)

. (.)

–. (–.)

–. (–.)

–. (–.)

–.*** (–.)

.*** (.)

()

–.*** (–.)

.*** (.)

–. (–.)

. (.)

–. (–.)

–. (–.)

–. (–.)

–.*** (–.)

.*** (.)

()

92 6 Panel Data Analysis For Evaluating Effects of Liquidity Standards

CONS

NCO

ROA

NP

CAP

LAS

LATA

CASH

HQLA

STA

LDEP

–.*** (–.)

–.** (–.)

–.*** (–.)

.*** (.)

–.*** (–.)

–. (–.)

.*** (.)

–.*** –.*** (–.) (–.)

.*** (.)

–.*** –.*** (–.) (–.)

–.*** –.*** (–.) (–.)

–.*** (–.)

–. (–.)

.*** (.)

–.*** (–.)

–. (–.)

.*** (.)

–.*** (–.)

–.*** (–.)

. (.)

–. (–.)

–.*** (–.)

. (.)

–. (–.)

. (.)

.*** (.)

–.*** (–.)

–.*** (–.)

(continued)

–. (–.)

. (.)

. (.)

.*** (.)

–.*** (–.)

.*** (.)

.*** (.)

.*** (.) .*** (.)

–.*** (–.)

. (.)

.*** (.)

–.*** (–.)

–.*** (–.)

. (.)

.*** (.)

–.*** (–.)

.*** (.)

–.*** (–.)

.*** (.)

–.*** (–.)

–.*** (–.)

–. (–.)

.*** (.)

–.*** (–.)

.*** (.)

.*** (.) –.*** (–.)

–.*** (–.)

–.*** (–.)

–.*** (–.)

–.*** (–.)

–.*** (–.)

–.*** (–.)

6.5 Results of Panel Models

93

.



.

.



.

Prob>F

Ob.s

Within R–sq

.



.

.

FE

()

.



.

.

FE

()

.



.

.

FE

()

.



.

.

FE

()

All banks: Dependent variable LLR

.



.

.

FE

()

.



.

.

FE

()

.



.

.

FE

()

.



.

.

FE

()

Notes: 1) The dependent variable is LLR which represents the banks’ liquidity requirement ratio calculated according to BIS (2013c). 2) *, **, *** indicate respectively 10%, 5% and 1% significance level. The figures in parenthesis are the standard errors. 3) Specification (2) adds an indicator variable LAS (logarithm of total assets). Specification (3) adds an indicator variable CASH (total cash). Specification (4) adds an indicator variable NP (net profit). Specification (5) adds a variable STA (short-term assets). Specification (6) adds a variable HQLA (high-quality liquid assets). Specification (7) adds a variable ROA (return on assets). Specification (8) adds a variable CAP (total shareholders’ equity). Specification (9) adds a variable LATA (short-term assets to total assets ratio). Specification (10) adds a variable NCO (net cash outflows).

.

FE

.

FE

()

F(n,)

ESTIMATOR

()

Table 6.4 (continued)

94 6 Panel Data Analysis For Evaluating Effects of Liquidity Standards

6.5 Results of Panel Models

95

consistent with findings of many others, including, DeYoung and Jang (2016), Mohamad et al. (2013) and Giannotti et al. (2010). In Model 3, cash is included to the previous model, which shows positive and significant relations, and all other variables keep the same significance level with the previous model, but the within R-square increases. In Model 4, net profit is added to the previous model, which shows insignificant relations and a decrease in within R-squared. In Model 5, short-term asset size is included to the previous model, and it shows negative and significant relations with the liquidity which is not expected. The coefficient of short-term assets is 3.64, which means one percentage change in the variable impacts more than 3.6% in liquidity of these banks. Inflation however becomes negative, and financing (credit) shows insignificant relations. In Model 6, the high-quality liquid asset is added to Model 5, and it shows a positive significant relationship with the liquidity which is expected. Its coefficient is around 3%. In this model, financing gives positive and significant results at a level of 10%. All other variables show the same relations except securities which show insignificant relations. In Model 7, return on assets as a new variable is added to the previous model, and it shows a negative and significant relationship with liquidity as expected. Hence, return on assets have a very high coefficient (20.28) in this model, whereas the coefficient of return on equity has less than 1%. These results show an evidence of high negative relationship between return on assets and liquidity of PBs. Besides, net profit and return on equity show a positive significant relationship with the liquidity, but financing and inflation become insignificant. Furthermore, all other variables keep their positions. In Model 8, the capital size of banks is included to the previous model, showing a negative and significant relationship with the liquidity, and return on equity becomes insignificant. In Model 9, short-term assets to total assets ratio is added to the model and shows positive and significant relations, but the constant term continues to show significant results. Whilst cash, total assets, net profit and short-term give insignificant results, return on equity and financing become positively significant after this variable is included. In the last model (Model 10), net cash outflows as a new variable is included to the previous model; showing negative significant relations as expected, which means if net cash outflows increase, liquidity levels of PBs decreas. However, there is not such a relation for the entire banking sector, because if net cash outflows of a bank increase, these cash outflows become inflows for another bank, implying that these cash flows will stay within the banking sector’s boundaries. However, there is a risk to PBs’ cash flows. In the last model, short-term assets and CDS become significant. Also, the constant term is significant and within R-squared reaches 0.9254 in the last model, larger than the Turkish banking sector models’ R-squared in the fixed-effect model. Within R-squared systemically increases and shows that all factors jointly explain 93.69% of the variability in liquidity, which is a very satisfactory level for the model.

.*** (.)

. (.)

.*** (.)

.* (.)

. (.)

–. (–.)

. (–.)

.*** (.)

–.** (–.)

.*** (.)

. (.)

.*** (.)

.*** (.)

. (.)

–. (–.)

–. (–.)

.*** (.)

–.*** (–.)

ROE

CAR

LINF

CDS

INTER

BOND

LCRE

SAFS

()

RR

()

–.** (–.)

.*** (.)

–. (–.)

–. (–.)

. (.)

.** (.)

.*** (.)

. (.)

.*** (.)

()

Table 6.5: Panel Data Regressions Results for PBs.

–.** (–.)

.*** (.)

–. (–.)

–. (–.)

. (.)

.* (.)

.*** (.)

. (.)

.*** (.)

()

.*** (.)

–. (–.)

–. (–.)

. (.)

–. (–.)

– (–.)

.*** (.)

. (.)

.*** (.)

()

. (.)

.* (.)

–. (–.)

. (.)

–. (–.)

.** (.)

.*** (.)

. (.)

. (.)

()

PBs: Dependent variable LRR

. (.)

. (.)

–. (–.)

. (.)

–. (–.)

. (.)

.*** (.)

.*** (.)

. (.)

()

. (.)

. (.)

–. (–.)

. (.)

–. (–.)

. (.)

.*** (.)

. (.)

.** (.)

()

–.*** (–.)

.*** (.)

–. (–.)

. (.)

. (.)

. (.)

.*** (.)

.** (.)

–.*** (–.)

()

–.*** (–.)

.*** (.)

–. (–.)

. (.)

.** (.)

. (.)

.*** (.)

.*** (.)

–.*** (–.)

()

96 6 Panel Data Analysis For Evaluating Effects of Liquidity Standards

–.*** (–.)

–.*** (–.)

ESTIMATOR FE

CONS

NCO

ROA

NP

CAP

LAS

LATA

CASH

HQLA

STA

LDEP

FE

–.*** (–.)

.** (.)

–.*** (–.)

FE

–.*** (–.)

.** (.)

.** (.)

–.*** (–.)

FE

–.*** (–.)

–. (–.)

.** (.)

.** (.)

–.*** (–.)

FE

–.*** (–.)

–. (–.)

.*** (.)

FE

–.*** (–.)

. (.)

.** (.)

.*** (.)

FE

FE

–.*** (–.)

–.* (–.)

–.*** (–.)

–.*** (–.)

.** (.)

FE

–.*** (–.)

–.** (–.)

. (.)

. (.)

–.*** (–.)

FE (continued)

–.*** (–.)

–.*** (–.)

–.*** (–.)

. (.)

.* (.)

. (.)

.*** (.)

.*** (.) . (.)

–.*** (–.)

.*** (.)

.** (.)

–.*** (–.)

–.*** (–.)

.*** (.)

. (.)

–.*** (–.)

.*** (.)

.*** (.)

.*** (.)

–.*** (–.)

–.*** (–.)

.** (.)

.*** (.)

.*** (.)

.*** (.)

.*** (.) .*** (.)

–.*** (–.)

–.*** (–.)

–.*** (–.)

–.*** (–.)

–.*** (–.)

–.*** (–.)

6.5 Results of Panel Models

97

.

.



.

.

.



.

F(n)

Prob>F

Ob.s

Within R–s

.



.

.

()

.



.

.

()

.



.

.

()

.



.

.

()

PBs: Dependent variable LRR

.



.

.

()

.



.

.

()

.



.

.

()

.



.

.

()

Notes: 1) The dependent variable is LLR which represents the banks’ liquidity requirement ratio calculated according to BIS (2013c). 2) *, **, *** indicate respectively 10%, 5% and 1% significance level. The figures in parenthesis are the standard errors. 3) Specification (2) adds an indicator variable LAS (logarithm of total assets). Specification (3) adds an indicator variable CASH (total cash). Specification (4) adds an indicator variable NP (net profit). Specification (5) adds a variable STA (short-term assets). Specification (6) adds a variable HQLA (high-quality liquid assets). Specification (7) adds a variable ROA (return on assets). Specification (8) adds a variable CAP (total shareholders’ equity). Specification (9) adds a variable LATA (short-term assets to total assets ratio). Specification (10) adds a variable NCO (net cash outflows).

()

()

Table 6.5 (continued)

98 6 Panel Data Analysis For Evaluating Effects of Liquidity Standards

6.5 Results of Panel Models

99

In response to the research questions for fixed-effect results of PBs, there is no evidence of any significant impact of market liquidity on the liquidity of PBs. Secondly, all variables included in the base model show significant relationships with liquidity, but the coefficients and signs differ. In this context, short-term assets, capital, return on asset and net cash outflows have a negative sign, but others have positive impacts on liquidity.

6.5.3 Results of Models for CBs The findings of the random-effect GLS regression estimation for Model 1 of conventional banks show very similar results with the panel results for the Turkish banking system, including PBs and conventional banks. This implies that the Turkish banking system is dominated by conventional banks. In the first model, required reserves, capital adequacy ratio, interbank rate, credit and securities available for sale have statistically meaningful positive relations with the liquidity requirement ratio at 1% significance level (Table 5.6). Among these variables, only required reserves and securities have more than 1% coefficients. Others have low level of coeffients. These results imply that banks having a higher capital base, higher credit ratio and higher securities are holding more liquid assets. Moreover, there are positive relations between securities and liquidity. Bonner (2014) finds pieces of evidence that banks’ liquidity position significantly affects their demand for government securities, which means banks increase their demand to hold more long-term government securities for liquidity requirement. The signs of these variables are positive as expected, which means the increase in this variable can affect the dependent variable positively. However, CDS rate and deposits show negative significant relationships with the liquidity at 1% and 5% significance level respectively. Gehde-Trapp et al. (2015) show that CDS trades have a significant price impact on assets and that the price impact functions are upward-sloping. The constant term has negative, but statistically insignificant relations. However, the within R-squared is low (0.2568). In Model 2, a total asset as a new variable is added to the previous model and required reserves, capital adequacy and credit become insignificant, whereas total assets show positive and significant relations with the liquidity (its coefficient is 1.1466%). This means that if the total asset size of the banking sector increases, the liquidity of the banking sector will also increase. This finding is consistent with the findings of DeYoung and Jang (2016), Mohamad et al. (2013) and Giannotti et al. (2010) who find a positive significant relationship between liquidity exposure and size of banks. Nevertheless, contrary to findings in this book, Angora and Roulet (2011) and Cucinelli (2013) find evidence of a negative relationship between liquidity and asset size. This means that if the total asset size of the banking sector increases, the liquidity of the banking sector will also increase.

100

6 Panel Data Analysis For Evaluating Effects of Liquidity Standards

In Model 3, cash is added to the previous model, and cash shows positive and significant relations but the coefficient of cash is very high, meaning that liquidity is very sensitive to the change in cash level. Moreover, required reserves and credit start to show significant relations. The former shows a positive relation, and the latter shows negative relations; with all others showing the same relations. In Model 4, net profit is added, which shows insignificant relations and within R-squared decreases. In Model 5, short-term asset size is included to the previous model, and it shows insignificant relations with the liquidity. Within R-squared also continues to increase. In Model 6, the high-quality liquid asset is added to Model 5, and it shows an insignificant relationship with the liquidity which showed significant results for all banks. All other variables show the same relations except credit, in which significant relations become insignificant with adding a new variable. Fuhrer et al. (2017) find that Swiss banks HQLA increases premium, and a structural shortage of HQLA may increase premiums in the market. In Model 7, return on assets are included to the previous model, and it shows a positive and significant relationship with liquidity, and all other variables keep their positions; except high-quality liquid assets which show a positive significant result, with small changes in coefficients of other variables. The coefficient of ROA is very high compared to coefficients of other variables in the model but it is lower than cash and required reserves. The variables that have higher coefficients affect liquidity more than other variables. In Model 8, the capital size of banks is added, and it shows a negative significant relationship with liquidity. However, return on equity shows negative significant relations with liquidity. This finding is consistent with Ariffin (2012) who suggests that as liquidity risk increases, ROE will decrease. In Model 9, short-term assets to total asset ratio is added to the model, showing positive significant relations, but the constant term becomes insignificant, whereas capital adequacy ratio, net profit, and short-term assets show positive significant relations. In the last model (Model 10), net cash outflows are included to the model, and showing negative and significant relations. Constant term in the last model becomes significant, but within R-squared increases (0.4347). The random-effect model shows that asset size is insignificant for PBs in seven models, but it is positive and statistically significant for PBs in three models (Models 5, 8 and 9) and all the last nine models for conventional banks. A similar result was reached by Ahmed et al. (2011); ie. asset size is insignificant for Islamic banks’ liquidity risk in Pakistan. On the other hand, high-quality liquid assets are found to be positive and significant for PBs’ five models, but it is found to be significant for the last four models of conventional banks. Banerjee and Mio (2014) find that Banks increased the share of HQLA and funding from more stable UK nonfinancial deposits after the introduction of liquidity requirements by the UK financial regulator, while reducing the share of short-term intra-financial loans and short-term wholesale funding.

()

.*** (.)

–. (–.)

–. (–.)

.*** (.)

–.*** (.)

.*** (.)

. (.)

–. (–.)

.*** (.)

()

.*** (.)

–. (–.)

.* (.)

.*** (.)

–.*** (–.)

.*** (.)

–. (–.)

.*** (.)

.*** (.)

RR

ROE

CAR

LINF

CDS

INTER

BOND

LCRE

SAFS

.*** (.)

–.** (–.)

–. (–.)

.*** (.)

–.** (–.)

.*** (.)

. (.)

. (.)

.*** (.)

()

Table 6.6: Panel Data Regressions Results for CBs.

.*** (.)

–.** (–.)

–. (–.)

.*** (.)

–.** (–.)

.*** (.)

. (.)

. (.)

.*** (.)

()

.*** (.)

–.* (–.)

–. (–.)

.*** (.)

–.** (–.)

.*** (.)

. (.)

. (.)

.*** (.)

()

.** (.)

–. (–.)

–. (–.)

.*** (.)

–.** (–.)

.*** (.)

–. (–.)

–. (–.)

.*** (.)

()

CBs: Dependent variable LRR

.** (.)

–. (–.)

–. (–.)

.*** (.)

–.** (–.)

.*** (.)

–. (–.)

–. (–.)

.*** (.)

()

.** (.)

–. (–.)

–. (–.)

.*** (.)

–.* (–.)

.*** (.)

. (.)

–.*** (–.)

.*** (.)

()

–.*** (–.)

.*** (.)

–. (–.)

. (.)

–. (–.)

.*** (.)

–.* (–.)

–.*** (–.)

.*** (.)

()

(continued)

–.*** (–.)

.*** (.)

–. (–.)

. (.)

–. (.)

.*** (.)

–.* (–.)

–.*** (–.)

.*** (.)

()

6.5 Results of Panel Models

101

ROA

NP

CAP

LAS

LATA

CASH

HQLA

STA

LDEP

–.*** (–.)

–.*** (–.)

.*** (.)

()

()

Table 6.6 (continued)

.*** (.)

–.*** (–.)

–.*** (–.)

()

–. (–.)

.*** (.)

()

–. (–.)

–. (–.)

.*** (.)

–. (–.)

.*** (.)

.** (.)

.* (.)

. (.)

.*** (.)

–.** (–.)

–.*** (–.)

.*** (.) .*** (.)

. (.)

.* (.)

.*** (.)

–.*** (–.)

()

. (.)

.*** (.)

. (.)

.* (.)

.*** (.)

–.*** –.*** (–.) (–.)

.* (.)

.*** (.)

–.*** (–.)

()

.*** (.)

–. (–.)

.*** (.)

. (.)

–.*** (–.)

()

–. (–.)

.*** (.)

–.*** (–.)

.*** (.)

. (.) –.*** (–.)

–. (–.)

–.*** (–.)

()

–. (–.)

–.*** –.*** (–.) (–.)

()

–.*** –.*** (–.) (–.)

–.*** (–.)

()

CBs: Dependent variable LRR

102 6 Panel Data Analysis For Evaluating Effects of Liquidity Standards

.

.



.

.

.

.



.

.

Wald chi(n)

Prob>chi

Observations

Theta

Within Rsq

.

.



.

.

RE

–.*** (–.)

.

.



.

.

RE

.

.



.

.

RE

–.*** –.*** (–.) (–.)

.

.



.

.

RE

–.*** (–.)

.

.



.

.

RE

–.*** (–.)

.

.



.

.

RE

–.*** (–.)

.

.



.

.

RE

. (.)

.

.



.

.

RE

. *** (.)

–.*** (–.)

Notes: 1) The dependent variable is LLR which represents the banks’ liquidity requirement ratio calculated according to BIS (2013c). 2) *, **, *** indicate respectively 10%, 5% and 1% significance level. The figures in parenthesis are the standard errors. Specification (2) adds an indicator variable LAS (logarithm of total assets). Specification (3) adds an indicator variable CASH (total cash). Specification (4) adds an indicator variable NP (net profit). Specification (5) adds a variable STA (short-term assets). Specification (6) adds a variable HQLA (high-quality liquid assets). Specification (7) adds a variable ROA (return on assets). Specification (8) adds a variable CAP (total shareholders’ equity). Specification (9) adds a variable LATA (short-term assets to total assets ratio). Specification (10) adds a variable NCO (net cash outflows).

RE

–.*** (–.)

RE

–. (–.)

ESTIMATOR

CONS

NCO

6.5 Results of Panel Models

103

104

6 Panel Data Analysis For Evaluating Effects of Liquidity Standards

In the fixed-effect, within R-squared is 0.4589 for the last panel model of conventional banks whilst it is higher for the last panel model of PBs (0.9369), which is still sufficient for explaining models in both PBs and CBs. Probability and F-statistics show that the fixed-effect model is meaningful for both PBs and the Turkish Banking system. Within R-squared decreased in the robustness check for PBs (0.9266) and decreased for CBs (0.5923). The fixed-effect model for both PBs and CBs show that there is a statistically significant relationship between liquidity and high-quality liquid assets. As a summary of findings in fixed-effect models for all panel models, strong evidence is documented in support of the fact that most bank-specific variables (i.e. credit, deposits, cash, asset size, and profitability) are directly correlated to the changes in liquidity risk of banks in Turkey. More specifically, profitability and deposits have a negative impact on liquidity risk of CBs and PBs; while credit has positive effects on the liquidity of PBs but negative effects on the liquidity of CBs. Capital, required reserves, high-quality liquid assets, cash and total assets have a positive effect on liquidity risk of PBs and CBs – except cash, which is negatively related to the liquidity of CBs. However, for macro factors, the only mild evidence is found in the negative impact of CDS on liquidity risk. Also, inflation and government bonds are insignificant for both types of banks, while the interbank rate has positive impacts on the liquidity of CBs. Additionally, there is evidence of positive effects of net cash outflows on the liquidity of CBs, but there is a negative and significant relationship between net cash outflows and liquidity of PBs. The results of the fixed-effect models indicate that liquidity risk in PBs can be differentiated from liquidity risk in CBs concerning the impact of net cash outflows, credit, cash, CDS and interbank rate. Regressions for PBs present within R-squared between 77% and 93%, while it is between 25% and 43% for CBs, suggesting that these models are well specified and have enough ability to explain the variations in liquidity risk. The consistent increasing value of within R-squared and the estimations of all coefficient variables from Model 1 through 8 indicate that the baseline model is robust for both PBs and CBs. The estimated values of F-statistics and probability indicate that all estimated models for PBs and CBs are a good fit to the data. Fixed-effect models produced, in average, higher standard errors than random-effect models. Since fixed-effect models do not concentrate on information about differences between individuals, variables having little variation over time for each dataset cause the fixed-effect to produce larger standard errors. If there is a choice between omitted variable bias and efficiency of models, more weight is preferred to be given to efficiency. Furthermore, if the number of cross-sectional units in the panel is greater and the period (T) is shorter than the number of cross-sectional samples (N), the random-effect model provides more effective estimates than the fixed-effect model. In contrast, if the number of time periods (T) is large and the number of samples

6.5 Results of Panel Models

105

per section data (N) is small, a small difference is expected between the two estimation results and the fixed-effect model is preferred. The Hausman specification test is conducted to compare the fixed-effect estimates with the random-effect estimates to predict their consistency and efficiency. The Hausman test examines whether the estimated coefficients from the fixedeffect estimation and the random-effect estimation is statistically significant; H0: βRE = βFE. The rejection (p-value < 0.05) of the test is commonly interpreted as a rejection of the random-effect model estimation. The results reveal that for liquidity risk across all models of PBs and banks, fixed-effect seems to be the appropriate estimator, whereas random-effect is appropriate for CBs (Appendix K). Ho is tested, which is the difference in coefficients that are not systematic. The Hausman test results show that fixed-effect model is consistent under Ho, but the random-effect is inconsistent but efficient under this hypothesis. Since the presence of homoscedasticity or serial auto-correlation lead to incorrect standard errors, a modified Wald test is used for the presence of homoscedasticity. A simple test of the first-order autocorrelation consists of regressing the residuals from the model estimation onto their lagged counterparts. It tests the null hypothesis that the variances are equal across all firms. The results from the Wald test are presented in the last row of Appendix N. The result of the test evidences that the test for “homoscedasticity” is rejected in all models, which means that heteroscedasticity is present. Consequently, the “robust” option has been applied in all models (Appendix N).

6.5.4 Robustness Check In this sub-section, the robustness of empirical results in this chapter is explored, and some tests are undertaken to assess empirical findings. First, two types of controls are included: the change in NCO is used to show that findings are not caused by variations in the cash outflows. Next, lagged net profit and lagged cash are included as explanatory variables to adjust for autocorrelation in the LRRR. The pvalues are provided in parentheses and the ***, **, *, indicate the significance level at 1%, 5%, and 10% respectively. The within R-squared value is presented at the bottom of each model in the table. For the robustness check, the overall results of all banks are first examined. These results show that capital adequacy ratio, interbank rate, credit and securities available for sale have statistically meaningful positive relations with the liquidity requirement ratio at 1% significance level (Table 6.7). However, CDS rate and deposits show negative and significant relationships with liquidity. The constant term has negative but statistically insignificant relations. However, the R-squared is 0.2818. In Model 2, total assets are added to the previous model, and capital adequacy and financing (credit) become insignificant whereas total assets show

106

6 Panel Data Analysis For Evaluating Effects of Liquidity Standards

positive significant relations with liquidity. Also, the constant gives significant relations. In Model 3, cash is added to the previous model, and cash shows insignificant relations. In Model 4, net profit is included to the model, which shows insignificant relations, and R-squared increases. In Model 5, short-term asset size is added to the previous model, and it shows insignificant relations with the liquidity. Required reserves then become significant. In Model 6, the high-quality liquid asset is added to Model 5, and it shows an insignificant relationship with the liquidity. All other variables show the same relations, except required reserves, which show an insignificant relationship with the addition of a new variable. In Model 7, return on assets is included to the previous model, and it shows an insignificant relationship with the liquidity. In Model 8, after adding the capital size of banks, the new variable shows a negative and significant relationship with the liquidity. However, return on equity shows negative significant relations with the liquidity. In Model 9, short-term assets to total assets ratio is added to the model – it shows positive significant relations, but constant term becomes insignificant, whereas cash shows negative significant relations and capital becomes insignificant, but net profit shows positive significant relations. In the last model (Model 10), net cash outflows are included to the model; it shows insignificant relations. The constant term in the last model shows insignificant relations, but the within R-squared increases (0.4732). As a last robustness check, with regards to liquidity risk, the different balance sheet proxies proposed by literature, such as the ratio of high-quality liquid assets to total assets or liquid assets to deposits were considered. However, the liquidity requirement ratio proposed by the Basel Committee for enhanced measurement of liquidity risk was instead chosen. Additionally, for robustness check, the overall results of PBs and CBs (Tables 6.8 and 6.9. respectively) are observed. Results of PBs show that required reserves, capital adequacy, and deposits have positive and significant relations but government bond and securities have negative significant relationships. Other variables are insignificant (Table 6.8). For CBs, interbank rate, credit, and securities available for sale, have statistically meaningful positive relations with the liquidity requirement ratio but CDS and government bonds have negative significant relations. Other variables are insignificant (Table 6.9). However, CDS rate and deposits show negative significant relationships with the liquidity. The constant term is significant for PBs but insignificant for the all panel model of CBs. R-squared is 0.6131 for the PBs model, and it is 0.2568 for CBs. In Model 2, total assets are added to the previous model and it is insignificant for PBs, but it is positive and significant for CBs. In Model 3, cash is included to the previous model, which shows positive and significant relations for PBs, but negative and significant relations for CBs. This can be interpreted as CBs having a trade-off between staying in cash and investing in government bonds, whereas PBs do not have such choices because there is a

. (.)

–. (–.)

. (.)

. (.)

–.*** (–.)

.** (.)

. (.)

–. (–.)

. (.)

–. (–.)

.* (.)

. (.)

–.** (–.)

.*** (.)

–. (–.)

.* (.)

ROE

CAR

LINF

CDS

INTER

BOND

LCRE

()

RR

()

–. (–.)

–. (–.)

.*** (.)

–.*** (–.)

. (.)

. (.)

–. (–.)

. (.)

()

–. (–.)

–. (–.)

.** (.)

–.*** (–.)

. (.)

. (.)

. (.)

–. (–.)

–. (–.)

.*** (.)

–.*** (–.)

. (.)

. (.)

. (.)

.* (.)

()

–. (–.)

–. (–.)

.*** (.)

–.*** (–.)

. (.)

. (.)

. (.)

. (.)

()

All Banks: Dependent variable LRR

. (.)

()

Table 6.7: Robustness Check (All banks – PBs and CBs).

–. (–.)

–. (–.)

.*** (.)

–.*** (–.)

. (.)

. (.)

. (.)

. (.)

()

. (.)

. (.)

.** (.)

–.*** (–.)

. (.)

. (.)

–. (–.)

. (.)

()

.** (.)

. (.)

–. (–.)

–. (–.)

–. (–.)

. (.)

–. (–.)

. (.)

()

(continued)

.** (.)

. (.)

–. (–.)

–. (–.)

–. (–.)

. (.)

–. (–.)

. (.)

()

6.5 Results of Panel Models

107

ROA

LNP

CAP

LAS

LATA

LCASH

HQLA

.*** (.)

–.** (–.)

–. (–.)

LDEP

STA

.* (.)

.* (.)

()

SAFS

()

Table 6.7 (continued)

.*** (.)

–. (–.)

–.*** (–.)

.** (.)

()

–. (–.)

.*** (.)

–. (–.)

–.*** (–.)

.** (.)

()

–. (–.)

.*** (.)

–. (–.)

.*** (.)

–. (–.)

.* (.) –. (–.)

–. (–.)

–. (–.)

.* (.)

–. (–.)

–.** (–.)

. (.)

.* (.)

–. (–.)

.* (.)

.*** (.)

.*** (.) .* (.)

–. (–.)

–. (–.)

. (.)

–.*** (–.)

–. (–.)

()

–. (–.)

–. (–.)

. (.)

–.*** (–.)

–. (–.)

()

.*** (.)

. (.)

. (.)

–.*** (–.)

–.** (–.)

. (.)

()

–. (–.)

.*** (.)

–. (–.)

. (.)

. (.) –. (–.)

–. (–.)

–. (–.)

–. (–.)

–.** (–.)

. (.)

()

–.** (–.)

. (.)

()

–.** (–.)

. (.)

()

All Banks: Dependent variable LRR

108 6 Panel Data Analysis For Evaluating Effects of Liquidity Standards

.

.

.

.



–.*** (–.)

.

.



–.*** (–.)

.

.



–.*** (–.)

.

.



–.*** (–.)

.

.



–.*** (–.)

.

.



–.*** (–.)

.

.



–. (–.)

.

.



–. (–.)

–.*** (.–.)

Notes: 1) The dependent variable is LLR which represents the banks’ liquidity requirement ratio calculated according to BIS (2013c). 2) *, **, *** indicate respectively 10%, 5% and 1% significance level. The figures in parenthesis are the standard errors. 3) Specification (2) adds an indicator variable LAS (logarithm of total assets). Specification (3) adds an indicator variable LCASH (logarithm of cash) for robustness check. Specification (4) adds an indicator variable LNP (logarithm of net profit). Specification (5) adds a variable STA (short-term assets). Specification (6) adds a variable HQLA (high-quality liquid assets). Specification (7) adds a variable ROA (return on assets). Specification (8) adds a variable CAP (total shareholders’ equity). Specification (9) adds a variable LATA (short-term assets to total assets ratio). Specification (10) adds a variable DNCO (differenced net cash outflows).

Within R–sq

.





Ob.s

Sigma_e .

–.** (–.)

–. (–.)

CONS

DNCO

6.5 Results of Panel Models

109

110

6 Panel Data Analysis For Evaluating Effects of Liquidity Standards

shortage of liquid assets. In Model 4, net profit is added, which shows insignificant relations for both PBs and CBs but R-squared increases for PBs, and decreases for CBs. In Model 5, short-term asset size is added to the previous model, and it shows insignificant relations with the liquidity for CBs, while showing negative but significant relations for PBs. Moreover, the interbank rate becomes positive and significant for PBs, but securities available for sale become insignificant for both PBs and CBs. Also, required reserves show insignificant relations for CBs’ liquidity. In Model 6, the high-quality liquid asset is added to Model 5, and it shows positive significant relations with the liquidity of PBs, but an insignificant relationship with the liquidity of CBs. Deposits become positive and significant, whereas securities and CDS show negative but significant relations with the liquidity of PBs. Additionally, required reserves become insignificant for PBs. Inflation shows positive and significant relations with the liquidity of CBs. In Model 7, return on assets is included to the previous model, and it shows an insignificant relationship with the liquidity of both CBs and PBs. In Model 8, after adding the capital size of banks, capital and return on equity show negative and significant relationships with the liquidity of both CBs and PBs. Securities available for sale and deposits become insignificant, but net profit becomes positive and significant for PBs. This elucidates that keeping more liquid assets decreases the need to increase capital, and since risk-weight of liquid assets is zero, or very low, banks tend to stand less on capital. In Model 9, short-term assets to total asset ratio is added to the model, showing positive significant relations with the liquidity of CBs but insignificant relations for PBs. CDS, short-term assets, high-quality liquid assets, and cash show insignificant results after the new variable is included for liquidity of PBs. Required reserves, inflation, CDS, interbank rate, total assets, and capital show insignificant results for CBs liquidity. In the last model (Model 10), net cash outflows are added to the model; it shows insignificant relations with the liquidity of CBs but shows negative and significant relations with the liquidity of PBs. Almost all regressions present consistent results for almost all variables indicating that these specification models are robust.

6.6 Liquidity Regulations and Bank Financing Banerjee and Mio (2014) discovered that after the introduction of the LCR which brings together tougher liquidity regulations, banks replaced claims on other financial institutions with cash, central bank reserves and government bonds – and so reduced the interconnectedness of the banking sector without affecting lending to the real economy. It is worth testing whether the PBs follow the same pattern or instead reduce financing to the real economy. Bonner (2014) tests the effects of

.*** (.)

–. (–.)

.* (.)

.*** (.)

. (.)

. (.)

–.* (–.)

–. (–.)

–.*** (–.)

.** (.)

–. (–.)

.* (.)

.*** (.)

. (.)

. (.)

–.** (–.)

–. (–.)

–.*** (–.)

ROE

CAR

LINF

CDS

INTER

BOND

LCRE

SAFS

()

RR

()

Table 6.8: Robustness Checks (PBs).

–. (–.)

.* (.)

. (.)

–. (–.)

. (.)

. (.)

. (.)

.* (.)

. (.)

()

–. (–.)

. (.)

–. (–.)

–. (–.)

. (.)

. (.)

. (.)

. (.)

. (.)

()

. (.)

–. (–.)

–. (–.)

.* (.)

–. (–.)

–. (–.)

. (.)

. (.)

.* (.)

()

–. (–.)

.** (.)

–. (–.)

. (.)

. (.)

. (.)

. (.)

. (.)

. (.)

()

PBs: Dependent variable LRR

–. (–.)

.** (.)

–. (–.)

. (.)

–.** (–.)

. (.)

. (.)

.* (.)

. (.)

()

. (.)

. (.)

–. (–.)

. (.)

–.** (–.)

–. (–.)

. (.)

. (.)

. (.)

()

–. (–.)

. (.)

–. (–.)

. (.)

–. (–.)

–. (–.)

.* (.)

. (.)

–. (–.)

()

(continued)

–. (–.)

.* (.)

. (.)

. (.)

–. (–.)

–. (–.)

. (.)

. (.)

–. (–.)

()

6.6 Liquidity Regulations and Bank Financing

111

ROA

LNP

CAP

LAS

LATA

LCASH

HQLA

STA

LDEP

. (.)

()

Table 6.8 (continued)

. (.)

. (.)

()

. (.)

. (.)

–.* (–.)

()

–. (–.)

. (.)

. (.)

–.* (–.)

()

–. (–.)

.** (.)

–. (–.)

.* (.)

. (.)

.* (.) –.* (–.)

–.** (–.)

–.** (–.)

.* (.)

–. (–.)

–. (–.)

–.** (–.)

. (.)

. (.)

.*** (.)

.** (.)

.** (.) .*** (.)

–. (–.)

.* (.)

–.* (–.)

–.** (–.)

()

–. (–.)

.* (.)

–.* (–.)

–.** (–.)

()

.*** (.)

. (.)

.*** (.)

–.* (–.)

–.** (–.)

()

.* (.)

.*** (.)

. (.)

.*** (.)

.*** (.) . (.)

–.* (–.)

–. (–.)

–.** (–.)

–.** (–.)

()

–.** (–.)

()

–.* (–.)

()

PBs: Dependent variable LRR

112 6 Panel Data Analysis For Evaluating Effects of Liquidity Standards

–.* (–.)

 .

.

 .

.

Ob.s Sigma_e

Within R–sq

.

 .

–. (–.)

.

 .

–. (–.)

.

 .

–.* (–.)

.

 .

–.* (–.)

.

 .

–.* (–.)

.

 .

–.** (–.)

.

 .

–.** (–.)

.

 .

–.* (–.)

–.* (–.)

Notes: 1) The dependent variable is LLR which represents the banks’ liquidity requirement ratio calculated according to BIS (2013c). 2) *, **, *** indicate respectively 10%, 5% and 1% significance level. The figures in parenthesis are the standard errors. 3) Specification (2) adds an indicator variable LAS (logarithm of total assets). Specification (3) adds an indicator variable LCASH (logarithm of cash) for robustness check. Specification (4) adds an indicator variable LNP (logarithm of net profit). Specification (5) adds a variable STA (short-term assets). Specification (6) adds a variable HQLA (high-quality liquid assets). Specification (7) adds a variable ROA (return on assets). Specification (8) adds a variable CAP (total shareholders’ equity). Specification (9) adds a variable LATA (short-term assets to total assets ratio). Specification (10) adds a variable DNCO (differenced net cash outflows).

CONS

–.** (–.)

DNCO

6.6 Liquidity Regulations and Bank Financing

113

()

. (.)

–. (–.)

–. (–.)

. (.)

–.*** (–.)

.** (.)

. (.)

–. (–.)

.* (.)

()

. (.)

–. (–.)

. (.)

. (.)

–.** (–.)

.*** (.)

–. (–.)

. (.)

.* (.)

RR

ROE

CAR

LINF

CDS

INTER

BOND

LCRE

SAFS

Table 6.9: Robustness Checks (CBs).

.* (.)

–. (–.)

. (.)

.*** (.)

–.*** (–.)

. (.)

–. (–.)

–. (–.)

. (.)

()

.* (.)

. (.)

. (.)

.** (.)

–.*** (–.)

. (.)

–. (–.)

–. (–.)

. (.)

()

. (.)

–. (–.)

. (.)

.*** (.)

–.* (–.)

. (.)

–. (–.)

–. (–.)

. (.)

()

. (.)

–. (–.)

. (.)

.** (.)

–.** (–.)

. (.)

–. (–.)

–. (–.)

. (.)

()

CBs: Dependent variable LRR

. (.)

. (.)

. (.)

.** (.)

–.** (–.)

. (.)

. (.)

–. (–.)

. (.)

()

. (.)

. (.)

. (.)

.** (.)

–.* (–.)

. (.)

. (.)

–. (–.)

. (.)

()

–. (–.)

.** (.)

. (.)

–. (–.)

–. (–.)

. (.)

–. (–.)

–. (–.)

. (.)

()

–.*** (–.)

.* (.)

.* (.)

–.* (–.)

–. (–.)

. (.)

–. (–.)

–.* (–.)

–. (–.)

()

114 6 Panel Data Analysis For Evaluating Effects of Liquidity Standards

–. (–.)





Ob.s

.*** (.)

–.** (–.)

–. (–.)

–. (–.)

CONS

DNCO

ROA

LNP

CAP

LAS

LATA

LCASH

HQLA

STA

LDEP



–. (–.)

.*** (.)

. (.)

–.** (–.)



–. (–.)

. (.)

.** (.)

. (.)

–.** (–.)



–. (–.)

. (.)

.** (.)



–. (–.)

. (.)

.** (.)

. (.)





–. (–.)

–. (–.)

–. (–.)

–. (–.)

. (.)



. (.)

–. (–.)

. (.)

–. (–.)

–.** (–.)

(continued)



.*** (.)

–. (–.)

. (.)

. (.)

.*** (.)

–.** (–.)

.*** (.)

.*** (.) . (.)

. (.)

–. (–.)

.*** (.)

–.*** (–.)

. (.)

–. (–.)

. (.)

–.*** (–.)

. (.)

.* (.)

. (.)

–. (–.)

–.** (–.)

. (.)

.*** (.)

. (.)

. (.)

. (.) . (.)

–.** (–.)

–.** (–.)

–. (–.)

–.** (–.)

–.** (–.)

–.** (–.)

6.6 Liquidity Regulations and Bank Financing

115

.

.

.

.

Sigma_e

Within R–sq

.

.

()

.

.

()

.

.

()

.

.

()

CBs: Dependent variable LRR

.

.

()

.

.

()

.

.

()

.

.

()

Notes: 1) The dependent variable is LLR which represents the banks’ liquidity requirement ratio calculated according to BIS (2013c). 2) *, **, *** indicate respectively 10%, 5% and 1% significance level. The figures in parenthesis are the standard errors. 3) Specification (2) adds an indicator variable LAS (logarithm of total assets). Specification (3) adds an indicator variable LCASH (logarithm of cash) for robustness check. Specification (4) adds an indicator variable LNP (logarithm of net profit). Specification (5) adds a variable STA (short-term assets). Specification (6) adds a variable HQLA (high-quality liquid assets). Specification (7) adds a variable ROA (return on assets). Specification (8) adds a variable CAP (total shareholders’ equity). Specification (9) adds a variable LATA (short-term assets to total assets ratio). Specification (10) adds a variable DNCO (differenced net cash outflows).

()

()

Table 6.9 (continued)

116 6 Panel Data Analysis For Evaluating Effects of Liquidity Standards

6.6 Liquidity Regulations and Bank Financing

117

liquidity regulations on financing. Following Bonner (2014) and Banerjee and Mio (2014), dummy variables of the Tobit model are used which is expressed as: Yj = 1 if Yj > 0 Yj = 0 if Yj ≤ 0

(11)

The following Tobit regression is estimated for the Turkish PBs in the sample. A dummy variable is included in Equation 9 for the years 2007–2015 to identify the effects of the LCR standards: θi = β 13 + εi

(12)

where _i = efficiency scores 13 = 2013 LCR launch day Furthermore, to investigate possible determinants of bank efficiency, the following hypothesis at the α = 0.05 significance level has also been tested: H0:βx = 0

(13)

Hi:βx = 0 β is the coefficient of the variable, whereas x represents the intercept of the country-specific banking systems under study. Table 6.10: Liquidity Regulation and Bank’s Financing. Cash

Financing

Regulations

.*** (.)

–.***(–.)

Observations



R



,

Table 6.10 shows that liquidity regulation increased the cash of banks, leading to a reduction in banks’ financing support to their customers. This is consistent with the results of Bonner (2014). The fundamental reason is that liquidity regulation requires banks to permanently stay in cash, high-quality liquid assets, and apparently, increase institutions’ marginal costs of funds, which affect their financing. In this context, raising cash is a buffer for banks against liquidity run. The incentives for banks to hold cash are different from the incentives to keep other liquid assets. It is very likely that banks with high cash holdings do so because of a lack of other investment opportunities, as opposed to incentives related to liquidity risk. When measuring the impact of liquidity regulation, binding and non-binding liquidity requirements cannot be accurately distinguished (Bonner, 2014).

118

6 Panel Data Analysis For Evaluating Effects of Liquidity Standards

6.7 Explaining the Findings Key findings of models are discussed in this sub-section of the book. The fixedeffect models that includes the variation in the differences between the units or the differences arising from differences occurring over time is preferred to the randomeffect models after conducting Hausman tests for PBs and all banks. The results of the fixed-effect models for PBs support the results of the time-series based model performed for PBs and CBs in the previous section. As a summary, the fixed-effect model’ results show pieces of evidence that liquidity risk of banks in Turkey is affected by most bank-specific variables (i.e. financing, deposits, cash, asset size, and profitability) because there is a statistically significant relationship between liquidity and these bank specific variables. Table 6.11: Comparing Findings with Other Studies. Dependent Variable Liquidity All Banks PBs CBs Other Studies Required reserves

(+)

(+)

(+)

Return on Equity

Ins.

Ins. Ins. (–): Ariffin (); Bonner () (+): Trad et al. (); Vodova (); Roman and Sargu ()

Capital Adequacy Ratio

(+)

(+)

Ins. (+): Roman and Sargu (); Chen et al. (); Amin (): Admati et al. (); Ogilo and Mugenyah (); Mongid (); Vodova () (-): Leykum (); Slovik and Cornède (); Angora and Roulet ()

Inflation

Ins.

(+)

Ins. (–): Amin (); Trad et al. (); Munteanu ()

CDS

(–)

Ins. (–)

(+): Gehde-Trapp et al. ()

Interbank Rate

(+)

Ins. (+)

(+): Angora and Roulet (); Hong et al. (); Vodova () (–): Vodova (),

Government Bond

Ins.

Ins. Ins. (+): Hong et al. ()

Credit (Financing)

(+)

(+)

(–)

Securities/Sukuk

(+)

(–)

(+)

Deposits

(–)

(–)

(–)

(–): Leykum () (+): Haan and End (); Bonner ()

Short-term assets

(–)

(–)

(+)

(–): Bonner ()

(–): Leykum (); Hong et al. (); Covas and Driscol () (+): Vodova () (+): Haan and End ();

6.7 Explaining the Findings

119

Table 6.11 (continued) Dependent Variable Liquidity High quality liquid asset

(+)

(+)

(+)

(+): Bonner ()

Cash

(–)

(+)

(–)

Short-term assets to total assets ratio

(+)

(+)

(+)

Ins: Ogilo and Mugenyah ()

Asset size

(+)

(+)

(+)

(+): DeYoung and Jang (); Mohamad et al. (); Giannotti et al. () Ins: Ahmed et al. (); Ogilo and Mugenyah (); Mongid (); Cucinelli () (-): Angora and Roulet (); Cucinelli (); Hong et al. (); Vodova ()

Capital

(–)

(–)

(–)

(–): Gorton and Winton (); Hong et al. ()

Net Profit

Ins.

Ins. Ins. (–): Chen et al. (): Akhtar et al. ()

Return on Assets

Ins.

(–)

Ins. (–): Chen et al. (): Akhtar et al. (): Roman and Sargu (); Hong et al. (); (+): Ahmed et al. (): Arif and Anees (); Trad et al. () Ins: Angora and Roulet ()

Net cash outflows

Ins.

(–)

(+)

(+) Haan and End ()

Notes: 1) (+): positive and significant, 2) (–): negative and significant 3) ins: insignificant.

Table 6.11 shows the results of the fixed-effect model for all banks, only for PBs and for CBs. The table also compares these results with the findings of other studies. Results show that liquidity management of PBs is negatively influenced by profitability and deposits. However, the analyses in this section reveal mixed results for effects of financing on banks’ liquidity management, which is inconsistent with the initial intuition. On the other hand, keeping more cash positively affects the liquidity of PBs because they do not have enough choices for differentiating their alternatives. For CBs, there is a negative correlation between cash and liquidity, which means that CBs have enough alternatives for a cash investment. It is seen that capital, required reserves, high-quality liquid assets and total assets have a positive effect on liquidity risk of both types of banks. If banks want to increase their level of liquidity, they may increase their capital, required reserves and high-quality liquid assets (Table 5.14). However, for macro factors, the only mild evidence is found in the negative impact of CDS on liquidity risk. Also, inflation and government bonds are insignificant for both types of banks, while the interbank rate has positive impacts on the

120

6 Panel Data Analysis For Evaluating Effects of Liquidity Standards

liquidity of CBs. Additionally, there is evidence of positive effects of net cash outflows on the liquidity of CBs, but there is a negative and significant relationship between net cash outflows and liquidity of PBs. The results of the fixed-effect model indicate that liquidity risk in PBs can be differentiated from liquidity risk in CBs regarding the impact of net cash outflows, credit, cash, CDS and interbank rate. The dummy variable shows that banks began to keep more liquid assets after the launch of the new liquidity coverage ratio requirements. Increasing cash is a buffer for banks contrary to liquidity run. The incentives for banks to hold more cash could be related to precautionary reasons, or may be due to a lack of other investment opportunities. When measuring the impact of liquidity regulation, it can be seen that liquidity leads to increased cash in banks, and as a consequence, banks began to decrease financing to its customers. This clearly indicates that there is a trade-off between keeping more liquid assets and financing customers. These results show evidence for generating a buffer for PBs against systemic factors that are useful in the liquidity of PBs. Otherwise, PBs will have to develop an ineffective balance sheet composition to protect themselves against changes in the market liquidity. This will lead to an ineffective Participation Banking sector and amplified financial stability risks.

7 Stress Testing For PBs 7.1 Introduction The current conventional banking system has certain functions that require banks to be dependent on liquidity, primarily market liquidity and funding liquidity, to secure their funding. Liquidity risk is not only a source of banks’ funding risks but also has a strong link to market liquidity. Central banks are conducting open market operations to assist in the liquidity management of banks in order to keep the current conventional system stable enough for serving their primary objectives – price stability and strengthening the financial system’s stability. PBs have to operate under this system and, as a result, are under the pressure of market liquidity and funding liquidity as shown in the first time series based model of this book. In other words, these banks are also vulnerable to the macroeconomic and financial shocks that may increase their liquidity risk. However, since it is not known whether all Islamic banks in the world have enough tools to protect their liquidity, it is better to seek the answer empirically using liquidity stress testing. For this reason, a particular stress test designed for Islamic banks may measure the risks of these banks and test their liquidity management capability. Conducting stress testing necessitates generating several coherent assumptions and scenarios. Banks perform their activities primarily by managing their assets and liabilities. If any mismatch between assets and liabilities occur, this mismatch can result in several internal risks and trigger many other risks. Hence, the first category of risks is related to idiosyncratic factors, mainly factors associated with banks’ internal systems, corporate governance, asset liability management and stakeholder relations. Systemic factors, market-wide stress, and disruption of several essential functions, are not in the sphere of control by individual banks. Systemic risk damages the intermediation capacity of the banking system and can temporarily freeze financial markets. Systemic risks also impair the intermediation function of the banking system by harming financing capacity of banks to the real sector. Banks must give attention to these external factors when managing liquidity, but these risks cannot be eliminated from such a complex financial world by banks alone. For this reason, all stakeholders of the financial system should add value to maintain stable systemic liquidity across markets, instruments, and counterparties. Indeed, the last crisis showed that during times of systemic stress concerning liquidity in the market, correlations between the components of systemic liquidity diverge. In a sense, market participants become color blind, and do not differentiate between colors. This means that investors in times of stress are not differentiating between banks. In this situation, financial instruments issued in advanced countries are regarded as “safe havens” and all other financial instruments in developing and emerging countries would face liquidity evaporation and fire sales, as shown after https://doi.org/10.1515/9783110582901-007

122

7 Stress Testing For PBs

the taper tantrum and the US 2016 presidential election.23 Although it is expected that Islamic banks would be recognized as safe banks because of the absence of high leverage and their inherent risk sharing characteristics, in times of stress, investors show herding features and do not differentiate banks and their modus operandi. Furthermore, investment strategies of internationally active banks require selling risky assets in fire sales during liquidity stress times. As a result of these factors, systemic market liquidity can evaporate quickly for Islamic banks too. In this scenario, Islamic banks would face a challenge in managing their funding liquidity. As a result of funding liquidity, Islamic banks differentiate between their customers, ration financing and, therefore, act pro-cyclically. Stress testing highlights this process, and it is very useful in understanding its workings. As a definition, stress testing is as a way of revaluing a portfolio using a different set of assumptions, or it is an estimation of value changes of a portfolio under massive changes to risk factors. It attempts to measure the sensitivity of a portfolio to a set of extreme and reasonable shocks (Jones et al., 2004). Several shocks are given to test the effect of selected variables on other variables. The test provides an approximate estimate of how the value of a portfolio changes when there are large variations in some of its risk factors. Stress tests were initially developed for use at the portfolio level, to understand the latent risks to a trading book in stress times. Later, they became widely used as a risk management tool by supervision authorities and financial institutions. Stress tests can be used to measure the risks of the whole banking sector, or a group of financial institutions. There are two approaches here: the first is the bottom-up approach used with bank-level data; the second is a top-down approach, which is used for an entire financial system with macro-level data. It is also possible to employ stress testing to examine the impact of changes in the operating environment, macroprudential policies, regulations, and rules. The Stress testing technique is a precise tool for estimating the effects of some shocks with scientific accuracy (Demekas, 2015). The most important part of this method is to develop objective assumptions that reflect basic characteristics of financial institutions and the financial system. Since every assumption can lead to important implications from the results of the analyses, it is vital to pay attention to the model building estimation and process. Therefore, it is important to go beyond mathematical formulas and technical calculations, and to develop precise assumptions for banks undergoing stress testing. Stress tests can be used for assessing an individual financial institution’s soundness or can be utilized for macroprudential analysis of the financial system.

23 Lee, S. (2016). REITs and the Taper Tantrum. Journal of Property Investment & Finance, Vol. 34 Iss: 5, pp. 457–464. And also see: Rosenthal, R. (2016). Emerging Market Outflows Raise Specter of Taper Tantrum Redux. The Wall Street Journal. December 1, 2016.

7.1 Introduction

123

Understanding the sensitivity of the portfolio to changes in various risk factors will help both financial institutions and regulators. Although there are several models that try to capture liquidity, capital risk, and some behavioral responses and feedback effects, it is challenging to connect these models to the regulatory framework, systemic risk and individual behavior of banks. For this reason, macroprudential stress tests should be used as a complement to microprudential stress tests that are conducted at the bank level. Different risks could be analyzed by using stress testing, which are market risk (price changes), credit risk (non-performing loans, counterparty risk or default of borrower), and liquidity risk (inability to convert to cash due to the illiquidity of the asset or product). More complex stress tests including multiple risk scenarios or changes in the macroeconomic environment still serve a purpose: revaluing a portfolio under a different set of assumptions. Stress tests can involve changes in a portfolio, as well as the duration, liquidity, default rates, and recovery rates assumed for this portfolio. Stress tests generally produce a numerical estimate of the change in the value of the portfolio (Demekas, 2015). If the main issue is to understand the sensitivity of the Islamic banks’ portfolios to changes in various risk factors, the stress testing designed for PBs should have a macroprudential perspective because macroprudential policies and financial stability risks have clear effects on these banks. Moreover, incorporating general equilibrium dimensions would help to assess the resilience of the financial system. In this context, the procedure is promising in measuring PBs’ behavioral responses, their interactions with each other and with other economic agents. Since this book focuses on the liquidity management of PBs, the stress testing developed in this book only focuses on liquidity. Stress testing may facilitate an understanding of market players’ behaviors and their potential impact on market liquidity. By designing and developing this test, it would benefit supervisory authorities by informing them of potential policy interventions. Furthermore, it would likely support meaningful understanding of systemic risks, develop new policies, and take measures to mitigate these risks. This chapter is formulated to investigate the answer to the following question:

Research Question: Is there any shortage of liquidity for PBs under stress conditions? The remainder of this chapter focuses on stress testing frameworks used by financial supervisors, macroprudential authorities, and international organizations. The essential aspect of this liquidity stress testing is that it focuses on the macro perspective of stress testing and is based on a top-down approach. In this sense, the chapter attempts to develop stress testing solely for PBs, which are found to be contingent on market liquidity. Additionally, many features of stress testing such as scenario design, calibration of shocks, selection of variables, adverse scenario are

124

7 Stress Testing For PBs

discussed in this chapter but not covered comprehensively in this book. First, the main vulnerabilities of Islamic banks are discussed, and then sources of liquidity stress and funding stress are evaluated. In this stage, information is gathered from macroeconomic and structural indicators. The broad macroeconomic context and fundamental indicators are used to narrow the focus and to understand the Islamic financial system’s vulnerability to shocks and its capacity to absorb the resulting losses. Second, a diagnostic analysis of the main issues is done, followed by the mapping of the scenario into a form, which is useful for analysis of Islamic banks’ financial positions. Subsequently, assumptions are specified, the analysis carried out, and primary results of stress testing are discussed. Third, stress testing is performed based on these assumptions, and the health of the Islamic banking sector will be analyzed by looking at liquidity and exposure to market risks. Finally, the results of the tests are interpreted.

7.2 Investigating the Main Vulnerabilities of Islamic Banks Investigating the main vulnerabilities of Islamic banks necessitates looking at vulnerabilities inherited in the system, many shocks originating from outside the banking sector, as well as idiosyncratic risks within these banks. Macroeconomic factors may affect the resilience of Islamic banks. For example, the inflation rate of the country, CDS rate of the country, interbank rate, market liquidity, and regulations are strongly influencing Islamic banks’ portfolio selections as demonstrated in the previous chapters. Since there is strong interaction between these banks and financial markets, fire sales of illiquid assets and securities during market stress can especially affect Islamic banking sector in a very short period of time. It is a well-known fact that Islamic banks rely on the wholesale funding markets, especially in Turkey. After the global financial crisis, several banks had challenges concerning non-performing loans to other markets and these banks increased their reliance on wholesale funding markets. Over-reliance on wholesale funding can be a serious problem for PBs during stress in the market. Banks cannot depend too heavily on inflows and, therefore, when designing liquidity stress tests, inflow rates should be calibrated conservatively. Turning to outflows, uninsured deposit outflows were one of the drivers of liquidity stress. Additionally, off-balance items can add to stress in some circumstances. For example, PBs have certain commitments to non-financial corporations of special purpose vehicles. During the recent financial crisis, it was seen that commitments to asset-backed commercial paper conduits and other capital market instruments could significantly affect banks’ liquidity positions. Since Sukuk has a similar structure as other asset-backed commercial paper conduits, banks that invest in the Sukuk market may face the same problem in stress times. Since Islamic banks do not rely on margin calls and pre-funding of FX swaps, these banks are on the safe side regarding the effects of these factors on the bank’s

7.3 The European Banking Authority (EBA) EU-wide Stress Test

125

liquidity position. Therefore, these factors are not causally significant in creating liquidity gaps for Islamic banks. On the other hand, liquidity risks stemming from off-balance items and derivative transactions should be handled carefully. As explained earlier, to cope with large net cash outflows, banks hold liquidity buffers. The first buffer is to stay in cash. The second one is used as a well-defined operational mechanism and as infrastructure for their assets. Using three-party’s sale and buy-back agreements is another buffer for Islamic banks. Conventional banks may be able to raise funds through repo with central counterparties (CCPs) if they have such facilities, but Islamic banks do not. Research suggests that the market risk is higher for Islamic banks than for their conventional counterparts (Farooqi and O’Brien, 2015). Moreover, it is suggested that Islamic banks’ assets are exposed to a higher risk than assets of conventional banks (Ariss, 2010). Liquidity risk management requires maintaining an appropriate level of liquid resources for liquidity management purposes. For this reason, many regulatory institutions in Turkey are taking relevant measures to solve PBs liquidity problems. The Turkish Treasury has been issuing lease certificates since 2012, and outstanding lease certificates are about TRY8 billion (US$2.11 billion) as of August 2016. However, there are important vulnerabilities. For example, Islamic banking’s reliance on cash reserves and use of commodities for collateral makes them comparatively more vulnerable to high inflation and change in real economic activities. From a regulatory perspective, addressing vulnerabilities in these banks requires having both macroprudential instruments and microprudential tools and applications. As Pagratis et al. (2016) suggest, all classes of liquid assets are dominated by government securities as a liquidity backstop, and time deposits dominates funding vulnerabilities. Since many governments in advanced economies and emerging economies increased their issuance of government bonds, Sukuk, lease certificates and other securities in the last ten years, there are concerns related to the sustainability of a massive amount of public debt. Stress testing consists of exploring how bank liquidity changes in response to certain external events (shocks). An event, or the aggregate of linked events, has an influence on the values of individual items of the bank’s asset and liability structure, which results in the change of immediate, current, and long-term liquidity. The new liquidity value is compared with the threshold values, and then the conclusion is drawn on the stress tolerance of the bank.

7.3 The European Banking Authority (EBA) EU-wide Stress Test EBA designed the European Union (EU)-wide stress test to assess banks’ resilience to hypothetical external shocks, and to identify vulnerabilities in the EU banking sector. The EBA’s standard methodology is used by all EU supervisory authorities. The macroeconomic scenarios were developed by the European Systemic Risk

126

7 Stress Testing For PBs

Board (ESRB) to evaluate the effect of changes in the economic environment on EU banks. These scenarios cover credit and market risks, exposures towards securitization, sovereign and funding risks. The ESRB developed the adverse scenario which reflects the systemic risks that are considered a threat to the financial stability of the EU banking sector. They assume there is an increase in global bond yields amplified by an abrupt reversal in risk assessment, especially towards emerging market economies. For this reason, the financial authority adds more aspects such as non-performing loans and public debt level. It is also assumed in the stress testing of EBA that there is lack of necessary bank balance sheet repair to maintain affordable market funding. The ESRB conducted this stress test on a sample of 124 EU banks which covers at least 50% of each national banking sector. This test was executed on the assumption of a static balance sheet, which means if any assets matured or impaired will be replaced with similar assets. This static balance sheet assumptation simplify the process of conducting the stress testing. The impact on interest income was also assessed, including the increase in the cost of funding over the stress test time horizon. The market risk section of the test covers all positions exposed to risks stemming from the changes in market prices, including counterparty credit risk. There are important differences between the IMF approach and the European authorities’ approach. The hurdle rates in the IMF stress test were more stringent: there are differences in the calculation of risk-weighted assets and capital ratios, such as the inclusion of operational RWAs in the IMF stress tests. The European authorities use granular data. Some losses are not covered in the IMF stress test due to the lack of historical data.

7.4 Methodology and Stress Test Methods: The Case of Islamic Banks To perform a stress test, it is needed first to define the typical structure of assets and liabilities of an Islamic bank. On the asset side, IBs have cash, central banking accounts, Sukuk, Commodity Murabahah, investments to the authorized capital of other companies, Murabahah financing, profit-loss sharing financing (Musharakah, Mudarabah), Ijarah, and required reserves. On the liability side, IBs have demand deposits, investment deposits, capital, undistributed profit, and saving deposits. At the beginning, Implied Cash Flow Analysis (ICFA) is developed. The rationale behind this analysis is that this approach has higher granularity and uses precise haircuts defined by BIS (2013a) and IFSB (2015b). This method is based on Čihák (2007), Schimeder et al. (2012) and a later version developed by Čihák (2014). Haircut for liquidity inflows, outflows and wholesale funding are derived from Schimeder et al. (2012) and IFSB (2015b). The unique characteristics of Islamic banks with higher

7.4 Methodology and Stress Test Methods: The Case of Islamic Banks

127

granularity are applied, particularly on the assets side. On the liability side, the sudden substantial outflow of funding was assumed. If there is any difference between the approach in this study and the approach developed by Schimeder et al. (2012) and IFSB (2015b), the difference will be explained and the reason behind such change will be discussed. For Liquidity Coverage Ratio analysis, net cash outflows are calculated based on bank level data provided by banks’ annual and quarter-based financial reports. Different haircuts for different assets are applied based on their level of liquid assets. In this sense, certain shocks are applied for increasing of outflows and decreasing of inflows. These shocks were formulated by the IFSB (2015b). However, these shocks are calibrated according to five period impulse responses of the first time-series model used in this book. For the asset side, applications of haircuts for liquid assets are used to counterbalance funding gaps. For the Net Stable Funding Ratio, the tests are adopted according to the final decision on the NSFR by the Basel Committee. First, stable funding is calculated. The same parameters used for the LCR are employed for the NSFR analysis. The first stage in the stress-testing process is the identification stage, where the main vulnerabilities are identified. Besides, in general stress testing, some studies add information regarding broader macroeconomic factors in order to assess the effect of macro factors on the performance of the financial system and discloses potential sources of shocks (Jones et al., 2004). Comprehending the bigger picture assists in the discerning of what is “normal” for an economy and in comparison with other countries. Balance sheet structures derived from aggregate financial statements can indicate significant exposures to particular classes of assets and liabilities, or income sources. This information can be analyzed according to type of financial institution as well as over time to indicate areas of concentration or accumulation of risks. Such aggregated information can be used to analyze growth rates of credit by various types of institutions, and by different sectors, in conjunction with a variety of financial soundness indicators. The health of the financial sector can be evaluated by looking at levels and trends in financial institutions, mainly of liquidity, capital adequacy, asset quality, profitability, and exposure to market risks. In addition to using the broad macroeconomic context and structural indicators, a range of financial soundness indicators (FSIs) developed by the IMF (2013) can be used to narrow the focus and understand the financial system’s vulnerability to shocks and its capacity to absorb the resulting losses. The analysis of FSIs can be enhanced by the information gathered from the macroeconomic and structural indicators discussed above. These indicators are more “micro-level” because they are typically derived from data about individual institutions or sectors. FSIs cover the banking sector, reflecting the central role of the banking sector in many financial systems. The data needed to compile these FSIs are already available. The IMF recommends that the set of FSIs covers additional FSIs for

128

7 Stress Testing For PBs

the banking system as well as FSIs for key nonfinancial sectors, as weaknesses in these sectors are a source of credit risk for banks and, thus, helping to detect banking sector vulnerabilities at an earlier stage (IMF, 2013). This liquidity stress testing does not cover the effects of capital on liquidity or the effects of liquidity on capital, because the stress testing is formulated only for liquidity coverage ratio and this test is not the part of advanced stress testing used for all components of financial positions. The ICFA has higher granularity for only liquidity and uses precise haircuts defined by BIS (2013a) and IFSB (2015b). This method is based on Čihák (2007), Schimeder et al. (2012) and a later version developed by Čihák (2014). The method has more than 150 data for each bank, which are collected from financial positions of banks and gives very detailed information on liquidity. Haircut applied for liquidity inflows, outflows and wholesale funding based Schimeder et al. (2012) and IFSB (2015b) can be calibrated only for liquidity. For assessing the affects of liquidity on capital, more data and new method are needed, which may give less information about the liquidity positions of banks. As Schimeder et al. (2012) summarize the stress-testing tools of this particular stress testing that can be used to run some tests in circumstances where data is very limited to broad asset and liability items. These tests allow linking the liquidity to solvency risks but do not allow to link liquidity to the capital. These tests’ findings provide failure and pass rates and the estimated funding shortfalls for each bank (Schimeder et al. 2012).

7.5 Liquidity Stress Testing Model for PBs For conducting stress testing, satellite models are used to calibrate assumptions and haircuts are applied to variables. The motivation for analyzing the effects of interbank interest rate on PBs funding liquidity is to understand the linkage of market liquidity to PBs’ liquidity. Before conducting stress testing, it is needed to analyze the effects of the interbank interest rate on PBs funding liquidity. For this purpose, a baseline VAR model is developed based on monthly data, and this structural equation is assumed: C0 Yt = Ct Xt + Bεt

(14)

Yt = C*Xt + υt

(15)

With C* = Ct− 1 andut = Ct− 1 Bεt Where Yt is a (nX1) vector of endogenous variables such as certain levels of haircuts applied to liquid assets, C0 is a (nXn) matrix of coefficients of simultaneous relations on endogenous variables, Xt includes lag of endogenous variables, C is the matrix of coefficients on the lagged variables in the model, εt as (nX1) vector of

129

7.5 Liquidity Stress Testing Model for PBs

X  ′ the structural shock is orthogonal, and E εt εt represents variance-covariance εt matrix of the structural shocks. The impulse response functions in Figure 7.1 indicate that one standard deviation increase in interbank rate spread has a significant and persistent negative effect on PBs’ funding liquidity. This finding suggests that shocks on term spread are recognized negatively.

.8

Response of LCR to Cholesky One S.D. Innovations

Response of INTER to Cholesky One S.D. Innovations

6 4

.4

2 .0

0 –2

–.4 2

4 INTER

6

8

10

LCR

2

4

6 INTER

8

10

LCR

Figure 7.1: Response of LCR and Interbank Rate.

The liquidity haircuts, stable funding, cash inflows and outflows scenarios are in line with the country risk assessment of Turkey, and they are used to assess the performance of the PBs under stress scenarios. Quarterly bank-level data that covers 2015 data is used in applying a scenario which is developed uniformly by weighting the banks’ liquid assets and liability items that would be affected by the scenario with stress weights (wi). Then certain haircuts are applied according to their liquidity level. The main objective is to reflect market-to-market losses in the liquidity level of assets. The same logic is applied to the liability haircuts to reflect the effect of fire sales or running to deposits.

7.5.1 Liquidity Risk Stress Testing Process The liquidity risk stress-testing process needs an in-depth knowledge of the LCR components, specific aspects of banks covered for analysis, and well-formulated processes. This test is divided into four parts: investigation; diagnostic (developing and calibrating assumptions), which covers fundamental problems; calculation, which covers several formulas and numerical designs; and interpretation of the results. In the investigation part, the vulnerabilities and main issues are discussed. Then, a scenario fitted to Basel III requirements and IFSB standards for the LCR is developed. Data is obtained from financial institutions’ balance sheets, liquidity reports, and

130

7 Stress Testing For PBs

other sources. The scenario horizon is set at three months. Thereafter, few liquidity buffers are defined. In this test, several thresholds are applied according to BIS and IFSB standards. Lastly, the results are interpreted. The information regarding the weights and haircuts of cash inflows, cash outflows, stable funding and unstable funding will be detailed in following sections.

7.5.2 Assumptions and Specification In this sub-section detailed information about assumptions and specification are given. Mainly suggestions of IFSB (2015b) are followed and several assumptions based on findings in this book are calibrated. More information is given about the calibration, but other assumptions stated and reasoned by the IFSB (2015b) are not explained in details. According to the designed stress testing, systemic risk turns out to be larger if (i) more banks would react since common reactions are more disturbing; and (ii) responses would be more similar. In the model, both the idiosyncratic loss of reputation and the wider systemic effects impact banks’ liquidity buffers through new haircuts on liquid assets and withdrawals of liquid liabilities. PBs have the opportunity to use the mechanism of lender of last resort provided by the Central Bank of the Republic of Turkey according to Shariah-compliant principles. However, using the lender of last resort mechanism requires a sufficient quantity of Shariah-compliant eligible collateral for open market operations. This is also essential for the pass-through mechanism of monetary policy implementation. It is assumed that PBs will react to the expected increase of the structural liquidity deficit at the target rate. Since liquidity stress testing must provide self-insurance, there is no room for LoLR in liquidity stress testing. It is formulated to be a thorough analysis of an institution’s capability to compensate net cash outflows. This liquidity stress testing is employed to help maintaing the liquidity coverage ratio’s spill-over effects on banks’ balance sheets (BIS, 2013). Both BIS (2013) and IFSB (2015b) recommend to calculate the LCR based on stock of Shariah-compliant HQLA and total net cash outflows calculated for the next 30 calendar days. Both standards (BIS, 2013 and IFSB, 2015b) calculate the LCR based on a scenario in which there are both idiosyncratic and market-wide shocks. The shocks specified by IFSB (2015b) are: a partial loss of unsecured wholesale funding capacity, the run-off a proportion of retail funding, partial loss of secured, short-term financing with particular collateral and counterparties, additional contractual outflows, increases in market volatilities impacting the collateral, unscheduled draws of customers and the potential need for the IIFS to buy back obligations. High-quality liquid assets can be seen in Table 7.1. Cash, central bank reserves, and Sukuk are accepted as the Level 1 high-quality liquid assets with 100% treatment according to the IFSB (2015b) but Schimeder et al. (2012) apply a 1% haircut for government securities at the moderate stress scenario and increase this haircut

7.5 Liquidity Stress Testing Model for PBs

131

Table 7.1: High-Quality Liquid Assets for Assumptions. High-Quality Liquid Assets A. Level  assets Coins and banknotes

%

Qualifying central bank reserves (including required reserves).

%

Qualifying Sukūk and other Sharī'ah-compliant marketable securities issued or guaranteed by sovereigns, central banks, PSEs, MDBs or relevant international organisations assigned % a % risk-weight for credit risk under IFSB (b) Qualifying other Sharī'ah-compliant marketable securities issued by sovereign or central banks that have a non-% risk-weight

%

Qualifying foreign currencies’ Sukūk and other Sharī'ah-compliant marketable securities issued by sovereign or central banks that have a non-% risk-weight

%

B. Level  assets (maximum of % of HQLA) Level A assets Sharī'ah-compliant marketable securities issued or guaranteed by sovereigns, central banks, PSEs, multilateral development banks or relevant international organisations, qualifying for a % risk-weighting for credit risk under IFSB (b)

%

Qualifying Sharī'ah-compliant securities (including commercial paper) and Sukūk that satisfy all of the conditions

%

Level B assets (maximum of % of HQLA) Qualifying Sukūk and other Sharī'ah-compliant securities

%

Qualifying Sharī'ah-compliant equity shares

%

Qualifying other Sharī'ah-compliant liquidity instruments that are widely recognised in the jurisdictions of the home country

%

Source: Based on IFSB (2015b).

to 10% for very severe stress scenario. In this sense, it is followed the IFSB (2015b) for treating zero haircut for sovereign Sukuk, because government securities including Sukuk in Turkey historically have very deep secondary markets. While securities that are Shariah-compliant and highly rated are accepted as Level 2A, and other Sukuk and papers are accepted as Level 2B, assets with different treatments (Table 5.6). However, there are certain differences between applications by regulatory bodies. For example, the BRSA of Turkey accepts the IILM Sukuk as Level 2A or Level 2B high quality liquid assets, according to short-term rating of its Sukuk issuance program. If the long-term or equivalent short-term rating of the IILM Sukuk is AAA to AA-, it will be accepted as Level 2A. If the long-term or equivalent short-

132

7 Stress Testing For PBs

term rating of the IILM Sukuk is A+ to BBB-, it will be accepted as a Level 2B asset. In this stress testing, it is treated the IILM Sukuk as Level 2A based on the rationale of its short-term rating (A1). For outflow of funds after the introduction of LCR, all regulatory authorities are looking for demand deposits and term deposits that have less than 30 days’ maturity. Demand deposits are mainly Amanah (safe custody) for PBs and can be withdrawn depending on customers’ need. On the other hand, investment deposits have a long-term maturity, generally more than one year, and operate with the profitloss sharing mechanism. Therefore, these investment deposits are accepted as stable deposits. Retail deposits are defined as deposits placed with an institution by a natural person and are divided into “stable” and “less stable” categories. From the outflows of funds list (Table 7.2:), a 5% haircut is applied to the term deposits stocks and unrestricted profit-loss investment accounts with a residual maturity greater than 30 days. Since the PBs in Turkey do not have restricted profit-loss investment accounts, the unrestricted profit-loss investment account is added to the analysis as the profit-loss sharing accounts (PLSA).

Table 7.2: Outflows of Funds. Outflow of funds A. Retail deposits and PLSA

Haircut

Demand deposits and term deposits (less than  days’ maturity): Stable deposits (Sharī'ah-compliant deposit insurance scheme meets additional criteria)

%

Stable deposits

%

Less stable retail deposits Term deposits and PLSA with residual maturity greater than  days

% %

Source: Based on IFSB (2015b).

Wholesale funding is paramount, especially for the Islamic banking model. For banks, wholesale funding represents a way to expand their portfolio and fulfil funding demand especially when banks have trouble attracting new deposits due to market liquidity conditions or the low-rate environment. Normally, if it is unsecured wholesale funding, it is very profitable for Islamic banks. However, credit market changes may cause a problem for Islamic banks. Unsecured wholesale funding is defined as those liabilities and general obligations that are raised from legal entities, including sole proprietorships and partnerships, and are not collateralized by legal rights. In this context, stable retail deposits are treated with 3% haircut if they are

7.5 Liquidity Stress Testing Model for PBs

133

backed by Shariah-compliant deposit insurance scheme meets additional criteria, other stable deposits are applied a 5% haircut, whereas less stable retail deposits are applied a 15% haircut (Table 7.3:). Among the retail deposits, PLSA with residual maturity greater than 30 days are accepted more stable and low haircut (5%) is applied to them. Table 7.3: Wholesale Funding. B. Unsecured wholesale funding

Haircut

Demand and term deposits (less than  days’ maturity) provided by small business customers: Stable deposits

%

Less stable deposits

%

Operational accounts generated by clearing, custody and cash management activities

%

Portion covered by deposit insurance (deposits that less than , TRY is covered by deposit insurance scheme)

%

Cooperative IIFS in an institutional network (qualifying deposits with the centralised institution)

%

Non-financial corporates, sovereigns, central banks, multilateral development banks and PSEs

%

All amount covered by deposit insurance scheme

%

Other legal entity customers

%

Source: Based on the IFSB (2015b).

For secured funding, transactions with a central bank counterparty or those backed by Level 1 assets with any counterparty, will be treated with zero haircuts. Secured funding can be defined as liabilities and general obligations with maturities of less than 30 days. These obligations are collateralized by legal rights to precisely designated assets owned by the counterparty in the case of bankruptcy, insolvency, liquidation or resolution. In this context, a 15% run-off factor is assigned to secured funding transactions backed by Level 2A assets with any counterparty. Assumptions for additional requirements are stated at Table 7.4. (D). Accordingly, trade finance and secured funding backed by Level 1 assets are applied zero-haircut, because both of them are based on secured collateral or assets and it is easy to liquidize these funding. Generally, trade finance has very short-term maturity and fully secured by a letter of guarantee or other pledged instruments.

134

7 Stress Testing For PBs

Table 7.4: Secured Funding. C. Secured funding Secured funding transactions with central bank counterparty or backed by Level  assets with any counterparty.

Haircut %

Secured funding transactions backed by Level A assets, with any counterparty

%

Secured funding transactions backed by non-Level  or non-Level A assets, with domestic sovereigns, multilateral development banks or domestic PSEs as a counterparty

%

Backed by residential mortgage-backed securities (RMBS) eligible for inclusion in Level B

%

Backed by other Level B assets

%

All other secured funding transactions

%

D. Additional requirements Sharī'ah-compliant hedging (Tahawwut)

%

Undrawn credit and liquidity facilities to retail and small business customers

%

Undrawn credit facilities to non-financial corporate, as well as sovereign, central banks, PSEs and multilateral development banks

%

Other contractual obligations extend to financial institution Trade finance

% %

Any additional contractual outflows

%

Any other contractual cash outflows

%

Secured funding backed by Level  assets

%

Secured funding backed by Level  assets

%

Secured funding backed by other assets close to Level  but not Level 

%

Collateral needed in case of downgrade by three notches

%

Undrawn but committed credit facilities to corporate and retail customers

%

Undrawn but committed liability facilities

%

Portion of assets reinvested

%

Source: Based on the IFSB (2015b).

For required stable funding such as cash, central bank reserves – including all shortterm claims on central banks, will be treated as zero haircut as stated in the Table 7.5. CBRT is applying some reserve option mechanisms for allowing banks to keep more foreign currency in central banks. These options are accepted in the same manner as central bank reserves.

7.6 Discussion of the Results of Models

135

Table 7.5: Required Stable Funding. Required Stable Funding

Haircut

Cash; All central bank reserves; all claims on central banks with residual maturities of less than six months; “Trade Date” receivables arising from sales of financial instruments, foreign currencies, and commodities

%

Unencumbered Level  assets, excluding coins, banknotes, and central bank reserves

%

Unused financings to financial institutions with residual maturities of less than six months, where the funding is secured against Level  assets

%

All other unencumbered financing to financial institutions with residual maturities of less than six months not included in the above categories; Unencumbered Level A assets

%

Unencumbered Level B assets; High-quality liquid assets (HQLA) encumbered for a period (between six months and one year);

%

Unencumbered residential real estate financing with a residual maturity of one year or more and with a risk-weight of less than or equal to % under the standardised approach;

%

Cash, securities or other assets posted as initial margin for Sharī'ah-compliant hedging contracts and cash or other assets provided to contribute to the default fund of a central counterparty;

%

All assets that are encumbered for a period of one year or more; All other assets not included in the above categories, including non-performing financing, financing to financial institutions with a residual maturity of one year or more, non-exchange-traded equities, fixed assets, items deducted from regulatory capital, Takāful assets, and defaulted Sharī'ah -compliant securities

%

Source: Based on the IFSB (2015b).

7.6 Discussion of the Results of Models While developed liquidity stress testing is a top-down model, it is run with banklevel data. In the case of the Turkish PBs, the liquidity positions (both liquid stocks and cash flows) used are those available from the Turkish Banking Association webpage and banks’ financial positions from Public Disclosure Platform (PDP) on a quarterly basis. The data includes on and off-balance sheet items. To examine banks’ responses to funding liquidity shocks, impulse-response functions are used that are derived from the VAR model. The identifying assumption is that variables that come earlier in the ordering affect the variables following contemporaneously, as well as with lags, while the variables that come later affect the previous variables only with lags. In other words, the variables that appear earlier in the ordering are more exogenous than the ones that appear later (or, more formally, in the short run

136

7 Stress Testing For PBs

the former are weakly exogenous for the latter). Robustness checks are performed to test the sensitivity of the outcomes for changes to the ordering of the variables.

Table 7.6: Implied Cash Flow Test (5 Days). Cumulative Withdrawal of Deposits/PLSIA (% of Total)

Cumulative Loss of Shortterm Funding (% of Total)

,%

,%

Period 

,%

Period 

Cumulative Total Loss of Funding (% of Total)

Minimum number of periods of survival

Number of illiquid Banks

Survival -% of PBs

,%





,%

,%

,%





,%

,%

,%

,%





,%

Period 

,%

,%

,%





,%

Period 

,%

,%

,%





,%

Period 

,%

,%

,%





,%

Implied cash flow test shows that at the first period (one month) only one bank becomes illiquid (Table 7.6). However, at the second period, the number of illiquid banks goes up to two, and then after the second period, the two banks stay illiquid, and the two other banks continue to have insufficient liquidity. Since inflow rates are calibrated very conservatively when designing the liquidity stress testing for five-period analyses, it is expected to have more than one illiquid bank in five periods. Cumulative withdrawal of deposits to the profit-loss sharing investment accounts increased to 24% in period 4 and decreased in the last period. Moreover, uninsured deposit outflows that were bigger than 100,000 TRY may be one of the drivers of the cumulative loss of short-term funding, and cumulative total loss of funding reached 75% at the last period.

Table 7.7: Results of LCR based Stress Testing. LCR

Number of Banks

% of Banks

% of Assets

“Not Passed”

< .



,%

,%

“Not Passed”

.–.



,%

,%

“Not Passed”

.–.



,%

,%

“Not Passed”

.–



,%

,%

>



,%

,%

“Passed”

7.6 Discussion of the Results of Models

137

Liquidity Coverage Ratio analysis shows that, on average, PBs have enough liquidity and do not have any problems under the adverse scenario (Table 7.7). Furthermore, Table 7.8. shows the results of LCR based stress testing and stress testing results show that total high-quality assets to total assets ratio are 26.2%, and total cash outflow to total assets ratio is 15.4%. This finding is parallel to the IFSB survey (IFSB, 2014) which found that Islamic banks have 30% ratio of liquid assets to total assets. Table 7.8: Results of LCR based Stress Testing-HQLA Ratios. Cap on Level  assets?

Yes ,%

Total High-Quality Assets (% of Assets) Other Cash Inflow (% of Assets)

,%

Total Cash Inflow (% of Assets)

,%

Total Net Cash Outflow (% of Assets)

,%

According to stress testing, the average liquidity coverage ratio of PBs that pass the tests would be 237% whereas median LCR would be 231%, as shown in Table 7.9. This test result shows that one of the PBs would fail the liquidity stress testing. This means one of the PBs would face liquidity run under the scenarios explained above. Liquidity shortfall would be around TRY2 billion (US$528.31 million). The percentage of liquidity shortfall to the total assets would be around 2%. It can be concluded from this table that PBs liquidity management would not trigger contagion effects and would not cause major financial stability problems under the scenarios of the stress testing because the Turkish banking system asset size is more than TRY2 trillion (US$528.31 billion). Table 7.9: Results of Liquidity Shortfall. Average LCR (weighted-average)



Median LCR



Number of Banks failing the test



Liquidity Shortfall

-,,

Liquidity needed as a % of Total Assets

-,%

However, if one of the PBs were to fail under severe stress times, the vulnerabilities and reputational risks of other banks may increase. As in the first model, it is shown that PBs liquidity management has close relations with market liquidity, and if one of

138

7 Stress Testing For PBs

the PBs default or go under liquidity run pressures, the monetary authority should immediately provide liquidity as a lender of last resort. Most importantly, these PBs should have enough high-quality liquid papers to use the lender of last resort mechanism. Van den End (2012) simulated a liquidity stress testing by using end-2009 data of banks in Netherlands and found that the first round effect erased more that 16 percentage points of the initial LCR of banks and then in the second round this effect increased to 48% of the basic LCR of these banks. He also shared a piece of evidence of strong impact of large banks’ response on financial markets (Van Den End, 2010). Ismal (2010) classifies sources of short-term demand for liquidity for Indonesian banks as first, second and third tiers and finds that based on the liquidity run scenarios, the first tier fails to handle a liquidity run when deposit withdrawals reach 45% of total deposits and the second tier fails to survive in a liquidity run when the withdrawals reach 30% of total deposits. Even during times of net cash outflows, the first and second tier liquid instruments are still able to mitigate the situation. Hence, Van den End’s (2012) liquidity stress tester model reveals that second round feedback effects are determined by the number and size of reacting banks and the similarity of reactions. He shows that the second round effects both have an impact on the LCR through additional haircuts on assets and net outflows. According to the stress test, NSFR ratio of PBs can be seen in the Table 7.10. This liquidity stress testing reveals that PBs need at least 2.6% stable funding compared to their total assets. This table shows that PBs’ stable funding is 10.3% of its total assets. Ultimately, they need more stable funding against further liquidity shortfall under severe market conditions.

Table 7.10: Net Stable Funding Ratio (NSFR). Available Stable Funding (% of Assets)

.%

Required Stable Funding (% of Assets)

.%

Average NSFR



Median NSFR



Number of Banks failing the test



Liquidity Shortfall



% of Total Funding Liabilities

.%

Since the NSFR standards would be effective starting from the 1 January 2019, PBs have enough time to restore their stable funding. Moreover, the results show that all PBs have passed this test and there is no liquidity shortfall.

7.7 Conclusions and Final Remarks

139

7.7 Conclusions and Final Remarks In the fourth chapter it was shown that liquidity risk is not only a source of banks’ funding risk, but is also strongly linked to market liquidity. The current conventional system has made banks increasingly dependent on market liquidity to secure funding by issuing securities on wholesale markets, and by trading credit. As a result, banks have become more vulnerable to macroeconomic and financial shocks that may engender liquidity risk. PBs try to do their best in managing their liquidity under the current liquidity coverage ratio regimes. Compared with their conventional counterparts, PBs still have a substantial shortage of products, high-quality liquid assets, and lack deep and well-developed financial markets. Under these disadvantages, it can be expected that they would expose severe problems when subjected to stress testing. However, their core results show that PBs would not face a large shortage of liquidity. Only one PB had liquidity shortage and failed in this stress testing, though very conservative assumptions were made. These results are very promising in one sense and alarming in other. It is promising because under severe assumptions, only one bank – which is known for having pure liquidity management, has a complicated illiquidity problem. It is alarming because if more liquid assets were not issued by the government, treasury, or central banks, PBs would have substantial cumulative loss of cash outflows under assumptions made in the stress testing of this thesis. Since PBs are small-and medium-sized banks, they are less likely to increase stress in the market. Both idiosyncratic loss of reputation and the wider systemic effects should be taken into consideration. A high level of haircuts on liquid assets applied according to Basel liquidity rules and authorities’ regulations did not worsen the liquidity of these banks. As only one PB may have liquidity run under the scenarios, the spillover effects on freezing credit lines to other banks should be handled very carefully. While the implementation of harmonized liquidity regulation across the global market is a unique and necessary step for the supervision of these institutions, one single metric cannot provide a complete picture of an institution’s liquidity risk profile, especially for Islamic banks. Using various stress tests may reveal some inherent problems in these banks’ current mechanisms related to regulatory framework. Well-tailored specific stress tests are a useful instrument to analyze and understand these banks` vulnerabilities. Although liquidity risk is a small component of macro stress tests, conducting liquidity stress testing for Islamic banks is useful for analyzing their liquidity risk management. For instance, in their recent stress test, the European Banking Authority only accounted for liquidity risks as an assumed increase in funding costs, as opposed to testing the size and quality of institutions’ liquidity buffers. Also, in the US, liquidity stress tests play only a subordinate role compared to capital stress tests. The Bank for International Settlements working

140

7 Stress Testing For PBs

papers (BCBS 2013a and BCBS 2013b) identify gaps in current liquidity stress testing and suggest areas of further improvement regarding liquidity stress testing. Another important factor that needs more attention is the Shariah-compliant deposit insurance system which is very important for Islamic banks as a backstop measure. As well as having such a functional fence, PBs need to have a relevant backstop mechanism that includes operational requirements for the effectiveness of liquidity buffers. Moreover, having enough high-quality liquid papers and having the opportunity to use the lender of last resort mechanism are other backstops that can enhance the safety and security of PBs. Islamic banks, unlike the conventional ones, are not so free to select financial instruments or make financial transactions. Consequently, their assets and liabilities are more exposed to shocks than those of conventional banks. Thus, Islamic banks’ managers must perform stress tests and correct the balance item structure on a regular basis. Furthermore, Islamic banks’ managers should pay special attention to managing the current liquidity, as it is more susceptible to shocks.

8 New Regulatory Framework for PBs’ Liquidity Management 8.1 Introduction The new Basel III liquidity requirements are expected to accelerate standardization of Islamic banking practices and applications, facilitate development of liquidity management framework and enhance risk management practices. Unfortunately, contrary to these expectations, it is claimed that the liquidity requirements have downside effects on the Islamic banking sector in the forms of higher capital charges, excess cash holdings, low-profit assets and rigid regulatory framework which constrains their ability to provide a partnership-based financing. As the current Basel regulations and standards are developed for internationally active banks, their applications to small local banks or banks that do not have systemic importance are unlikely fitting. Most of the Islamic banks are small local banks and, as a result, they have to bear such a high cost of international regulations. In addition, like conventional banks, the critical functions of Islamic banks are hampered by the problem of asymmetric information that generally characterizes financial environments faced by the banking sector. Many stakeholders in the current financial system are excluded from the decision-making process and are not given the appropriate rights of participation. According to a survey conducted by Ziraat PBs, 25% of investors stated that they are excluded from the financial system due to the lack of opportunity to deposit their investment in a Shariah-compliant state bank (Turkey, 2016). These investors have a confidence problem with private Islamic banking practices because of the failure and resolution process of Ikhlas Finans of Turkey in 2001. Introducing a new regulatory framework, or reforming the existing one, will increase Islamic banks’ ability to provide more financing to the real sector and to mitigate liquidity risks. The benefits of these would be reduction of costs associated with financial regulations and supervision, and stability of financial systems based on risk sharing. The unique features of PBs require that these banks’ analytical and policy dimensions should be restructured differently from those of their conventional peers. Analytically, risk identification and surveillance requires an encompassing view of the financial sector. Since these banks have close relations with the real sector, there are important vulnerabilities regarding real-sector financing that should not be ignored, such as counterparty risks or information sharing problems between banks and companies applied for partnership based financing. For this reason, a separate and distinct risk assessment of these banks is needed. Regulating cross-sectoral and crossborder linkages and channels necessitates having certain well-formulated tools to identify underlying risks. Otherwise, overestimation or underestimation of these https://doi.org/10.1515/9783110582901-008

142

8 New Regulatory Framework for PBs’ Liquidity Management

risks is highly possible. Once they are identified, instruments for a policy response could be designed and developed. In this context, transparent communication with all stakeholders is very important. Without having transparent and symmetric information, it will not be possible to have appropriate policies and an effective regulatory and supervisory structure. According to the findings of time-series models, panel data analyses and the stress testing, PBs’ liquidity management should be handled differently from their conventional counterparts’ liquidity management. First of all, these banks have a shortage of liquidity and shortage of stable funds as evidenced by the findings of stress testing. If they are forced to keep the same liquidity coverage ratio with the conventional banks, they have to keep a large amount of their resources as cash, because there is shortage of high quality liquid Sukuk. Secondly, according to the last model of PBs (Table 5.5), return on equity is found positively and statistically significant in affecting liquidity, whereas return on assets is found negatively and statistically significant in affecting liquidity. This means that the capacity of profit in terms of shareholders’ equity to observe the LCR increases as earning increases, whereas the capacity of profit in terms of total assets to observe the LCR decreases as earning increases. Since the ROA ratio is calculated by dividing the net profit to the total assets, increasing ROA ratio means that the bank is investing in assets generating high profit. Since the cash and required reserves are low profit-yielding assets, increasing profit can be interpreted that the bank starts to keep less cash or low-yielding assets in its financial positions. This finding shows that the regulatory framework of PBs should be reformed. Moreover, it is found that there is a positive and significant relation between the CAR ratio and liquidity of PBs (Table 6.5). However, there are many other reforms that should be applied by these banks concurrently. For example leverage ratio24 requires that banks should have at least 3% leverage ratio, in which there is no weightage mechanism and all assets are treated as the same. Under the leverage ratio, banks will face higher alternative cost of keeping cash and low-yielding highquality liquid assets, which necessitates reforming the LCR mechanism for maintaining the current profit level and keeping capital stable. 24 According to the document on “Basel III leverage ratio framework and disclosure requirements”, one of the underlying causes of the global financial crisis was the build-up of excessive onand off balance sheet leverage in the banking system. In many cases, banks built up excessive leverage while apparently maintaining strong risk-based capital ratios. For this reason, the Basel III framework introduced a simple, transparent, non-risk based leverage ratio which is intended to restrict the build-up of leverage, reinforce the risk-based requirements. This leverage ratio is designed to be complementary to the risk-based capital framework. Implementation of the leverage ratio requirements has begun with bank-level reporting to national supervisors of the leverage ratio and its components from 1 January 2013, and will proceed with public disclosure starting 1 January 2015. The final calibration of the ratio will be completed by 2017 and it will be part of Pillar 1 (minimum capital requirement) treatment on 1 January 2018. See http://www.bis.org/publ/bcbs270.pdf.

8.1 Introduction

143

According to Table 6.5, increasing financing enhances the liquidity level, but Table 6.10 gives opposite results that show a negative relation between financing level and liquidity regulation. Moreover, the results of ARDL short run analysis strengthen findings showed in Table 5.10. In other words, financing positively affects liquidity level but liquidity regulations is discouraging banks to give more financing in order to stabilize their LCR level. Furthermore, if deposits decrease the liquidity ratio, banks will have to find stable funds. Otherwise, they would be under the liquidity run during market stress times. As there is a direct relation between market liquidity and liquidity of PBs, the current framework should be reformed to protect these banks against sudden changes in market conditions, because these banks do not have enough Shariahcompliant high-quality liquid papers to use lender of last resort mechanism. Likewise, panel data results show that keeping more cash positively affects the liquidity of PBs because they do not have enough choices for differentiating their alternatives and PBs have to keep a large amount of their resources as cash by sacrificing profit. For CBs, there is a negative correlation between cash and liquidity, which means that CBs have enough alternatives for a cash investment. Moreover, it is found that there is a negative and significant relationship between net cash outflows and liquidity of PBs. If net cash outflows increase, banks’ liquidity decreases. Hence, there is a need to increase the resilience of these banks against net cash outflows by reforming its’ liquidity framework. Table 6.10 shows that liquidity regulations increased cash of banks; consequently, banks began to decrease financing to its customers, which specifies that there is a trade-off between keeping more liquid assets and financing customers. These results show evidence for reforming regulatory framework for PBs’ liquidity management against an ineffecient balance sheet composition. Also, an ineffecient participation banking sector may amplify shocks for financial stability.Additionally, the impulse response analysis strengthened the Granger analysis by showing that when other variables are subject to shocks, liquidity ratio responds to the return on equity, financing, deposits and capital adequacy ratio. These results show that liquidity ratio of PBs is sensitive to the changes in return on equity, financing, deposits and capital adequacy ratio. Moreover, the implications of Sukuk and the interbank money market interest rates have on liquidity of PBs signify the need for better options for resolving the shortage of highly-liquid Shariah-compliant assets. Evidence from the time series analysis shows continuous dependence of the PBs on the interbank money market rates if no revisions made to the current system. For these specific reasons, a more ideal and reasonable regulatory framework should be formulated to extend the playing field of Islamic banks and subsequently to increase their compliance to Shariah rules. It is expected that this reality-based approach will prepare the ground for developing more a “ideal rule-compliant system”. In this book, first of all, the short-run and long-run determinants of liquidity

144

8 New Regulatory Framework for PBs’ Liquidity Management

holdings of Islamic banks are examined and then the effects of financing on liquidity risk of PBs are researched. In this context, results of PBs are compared with results for CBs as well as the impact of net cash outflows on liquidity risk of PBs is assessed. Following that, a liquidity stress-testing model is adopted for measuring Islamic banks’ liquidity under several assumptions. Based on findings of these models, since there is a need for high quality liquid assets that are generating profit, more risksharing instruments will be recommended in this chapter with appropriate regulatory treatment (liquidity coverage ratio) for risk-sharing instruments (Esham and Sukuk) as alternative macroprudential tools; and a new framework for liquid instruments will be developed. In the remainder of this chapter, first it is discussed different aspects of the current system, and then an alternative approach for liquidity risk management of the Islamic banks is developed. Subsequently, liquidity treatment of Sukuk and Esham is analyzed and it is offered a new way of treating thems within the context of liquidity management.

8.2 Fundamental Challenges that should be addressed for Liquidity Management of PBs The results of the analyses suggest fundamental challenges faced by Islamic banks that need to be addressed for their liquidity management. These challenges are: 1. PBs hold more cash than normally needed. 2. High-quality liquid assets are inadequate for their investment, and regulators need to change the regulatory framework for accepting alternative treatments. 3. New Shariah-compliant products or assets must be designed by authorities and banks. Also, these assets should be accepted as collateral by the central banks for open market operations to increase their acceptance in secondary markets. 4. To control the cash outflow and mitigate inherent risk in their assets and liabilities, the Islamic banks should be more transparent and offer new risk-sharing based channels to investors for placing their deposits as long-term investments. 5. The Basel framework should not be applied to small Islamic banks since it is costly for them. Moreover, they are not the source of systemic risk and do not danger the financial stability. 6. Current regulatory framework for liquidity management has a negative effect on PBs credit and a positive effect on PBs` cash stocks. Therefore, it should be reformulated to let these banks to increase their risk sharing financing and keep a reasonable amount of cash. The quantitative analysis of chapters four and five reinforced this conclusion. This means that the current regulatory framework of liquidity based on conventional ontology has direct effects on the portfolio of Islamic banks. PBs are not working in an isolated environment. The macro financial and political environment is affected by conventional ontology, epistemology and methodology. All sectors, mainly the government sector, household sector,

8.3 Cash and Profit Relations

145

foreign trade and corporate sector have many rules that are shaped by conventional thinking. Without reforming some parts of this conventional environment, it would not be easy for Islamic banks to pursue their mission. In this chapter, several recommendations will be made to adress these challenges and direct implications on the portfolio of Islamic banks. In order to implement a new regulatory framework, new models are needed that suit modus operandi of Islamic banks. In addition to new models, the micro-level data and banks-specific information are required to shape and to reform these institutions in a more effective and efficient way. Specific information related to Muslim-majority countries and financial market intelligence are also key components in defining a new model. Human resources and sound judgements are needed for this purpose. The risk-weight of Mudarabah and Musharakah partnership contracts is a discouraging factor to the expansion of partnership-based financing and is profit reducing. Some regulatory authorities impose more than 300% risk-weight for Mudarabah contracts. Moreover, it is claimed that higher capital requirements for partnershipbased contracts reduce their financing. In Turkey, the partnership-based financing share in total assets is as little as two 2% of total assets. Furthermore, higher capital requirements may reduce Islamic banks’ profitability since they would have to increase their capital to keep their capital adequacy ratio stable, or avert from risky investments. An alternative regulatory framework will help authorities to maintain financial stability and support Islamic banks to have a better mechanism for their liquidity management. In the coming sub-sections of this chapter, it is proposed a new approach for liquidity management of Islamic banks. Before that, it is reiterated first the findings on the relationship between cash and profit and the relationship between liquidity and capital of Islamic banks. Then, the new framework is represented and it is discussed for liquidity management of Islamic banks. Subsequently, it is offered a new approach to transparency of liquidity investment of Islamic banks. Thereafter, the regulations of internationally active banks are offered to be differentiated from small local banks’ regulations. Finally, it is proposed a new liquidity treatment of Esham and other risk sharing instruments.

8.3 Cash and Profit Relations Among various aspects, this book evaluates the correlation between cash and profit and between LCR and CAR of PBs. PBs keep high cash and central bank reserves in their balance sheets. As reported in Table 8.1, the correlations between cash and profits are positive for all banks. Except for Bank D, these correlations can be considered high. Meanwhile, except one, the correlations between the LCR and the CAR are negative. The exception is Bank A, whose LCR-CAR correlation is highly positive.

146

8 New Regulatory Framework for PBs’ Liquidity Management

Table 8.1: Correlation of Cash and Profit and Correlation of LCR and CAR. Correlation of Cash and Profit

Correlation of LCR and CAR

BANK A

.

.

BANK B

.

–.

BANK C

.

–.

BANK D

.

–.

Figures 8.1–8.4 provide line plots of cash, net profit, LCR and CAR respectively for the four banks. In Figure 8.1, the plots well reflect the positive cash – net profit and LCR – CAR correlations. Bank A’s trends upward and exhibits two noticeable spikes. Meanwhile, net profit fluctuates widely but with a slight upward trend. As for the LCR and CAR, they witness a decreasing pattern. Bank B’s cash, net profit, liquidity coverage ratio and capital adequacy ratio can be seen in Figure 8.2. Bank B’s cash has a clear upward trend. The drastic increase in its cash holdings suggests that Bank B faces liquidity management problems. Profitability of this bank is compromised with a large amount of cash holdings. Bank C seems satisfactory in managing its cash and net profit as compared to other banks. Its cash level however jumps to a very high amount between 2011 and 2012. In addition, after 2015, its capital adequacy level and LCR diverge with the former moving upward while the latter downward. Recall from the previous chapter that Bank D fails the stress-testing scenario. Figure 8.4 shows that the failure is due to a reduction in cash and huge loss between 2013 and 2015. During this period, its liquidity coverage ratio decreased below 100% level while its capital adequacy ratio stayed remarkably stable. This shows that this bank was under liquidity run and had several problems regarding its liquidity management. These two figures provide important evidence regarding liquidity management of Islamic banks. It can be derived from the figures that a bank with a high level of capital adequacy may have severe problems due to its liquidity management. Capital buffers are not enough to guard against liquidity run. Even liquidity buffers may not protect banks against liquidity run in some circumstances if there is no mitigating mechanism of cash outflows. Figure 8.5 plots cash to asset ratio of the four banks. It can be seen from the figure that the cash holdings as a ratio of total assets of one PB exceeds 10% between 2011 and 2012. Since there is no regulation that places a ceiling on the cash to asset ratio, banks are free to violate their contracts with their fundraisers by staying passive during market stress time. For this reason, it is proposed that there should be a ceiling for the cash account and it should be maintained within 3% of total assets.

2013

2014

2015 I

II

III 2014

IV

I

II 2015

III

Net Profit

Figure 8.1: Bank A’s Cash and Net Profit (left) and CAR and LCR ratio (right).

Cash

LCR_A

CAR_A

100

2012

0 2011

200

100,000

2010

12

300

200,000

2009

13

400

300,000

2008

14

500

15

500,000

400,000

16

600,000

8.3 Cash and Profit Relations

147

2008

2009

Cash

2010

2012

Net Profit

2011

2013

2014

2015

Figure 8.2: Bank B’s Cash and Net Profit (left) and CAR and LCR ratio (right).

0

500,000

1,000,000

1,500,000

2,000,000

2,500,000

3,000,000

3,500,000

100

150

200

250

300

350

400

I

II 2014

III

LCR_B

IV

CAR_B

I

2015

II

III

13.5

14.0

14.5

15.0

15.5

16.0

148 8 New Regulatory Framework for PBs’ Liquidity Management

2008

2009

2011

Cash

2010

2013

Net Profit

2012

2014

2015

Figure 8.3: Bank C’s Cash and Net Profit (left) and CAR and LCR ratio (right).

0

100,000

200,000

300,000

400,000

500,000

600,000

I

II 2014

III

IV

I

II 2015

III

70

80

90

100

LCR_C

CAR_C

12.0

120 110

12.4

130

12.8

13.2

13.6

14.0

8.3 Cash and Profit Relations

149

Cash

Net Profit

Figure 8.4: Bank D’s Cash and Net Profit (left) and CAR and LCR ratio (right).

–1,000,000

–750,000

–500,000

–250,000

2008 2009 2010 2011 2012 2013 2014 2015

80

120

160

200

240

I

II 2014

III LCR_D

IV

CAR_D

I

II 2015

III

16

250,000

14

18

500,000 280

20

750,000

0

22

1,000,000

150 8 New Regulatory Framework for PBs’ Liquidity Management

8.4 Introducing a New Framework for Addressing Liquidity Risks

151

Cash to Asset Ratio 15,00% 10,00%

BANK A

BANK B

BANK C

04/15

10/14

04/14

10/13

04/13

10/12

04/12

10/11

04/11

10/10

04/10

10/09

04/09

10/08

04/08

10/07

0,00%

04/07

5,00%

BANK D

Figure 8.5: Cash to Asset Ratio (PBs in Turkey, Bank Level Data).

From a profit ratio perspective, as plotted in Figure 8.6, a similar pattern is seen across all four PBs’ return on asset ratio. The bank that records higher profit ratio also keeps lower cash. Figure 8.6 shows that three PBs follow the same trend. In 2014, Turkish Lira (TRY) depreciated drastically and the CBRT had to increase its target interest ratio from 4.5 to 10%. This change affected all PBs’ profitability negatively, as can be seen at Figure 8.6. RoA (%) 100% 80% 60% 40%

0% –20%

04/07 08/07 12/07 04/08 08/08 12/08 04/09 08/09 12/09 04/10 08/10 12/10 04/11 08/11 12/11 04/12 08/12 12/12 04/13 08/13 12/13 04/14 08/14 12/14 04/15 08/15

20%

–40% BANK A

BANK B

BANK C

BANK D

Figure 8.6: Changes in Profit Ratio (PBs in Turkey, Bank Level Data).

8.4 Introducing a New Framework for Addressing Liquidity Risks It is suggested that the current Basel requirements and IFSB standards are disincentives to Islamic banks to provide risk-sharing or partnership based investments and services to their customers and depositors. Instead, Islamic banks

152

8 New Regulatory Framework for PBs’ Liquidity Management

hold higher cash than it should be, as shown in Figure 8.4. The standards set only the minimum liquidity ratio but not the maximum. Here, it is suggested a maximum liquidity ratio for Islamic banks. Applying a cap to the LCR ratio will impose discipline on Islamic banks to manage their assets appropriately as well as to encourage their financial intermediation to the real sector. In addition, it is argued that even if the cash outflows from investment account on the right side of Islamic banks’ balance sheets are included in the short-term projection, they should not be included in the denominator of the liquidity coverage ratio. These investments are long-term and have very low level of cash outflows. The ceiling must also be imposed on the cash account on the asset side. Further deterrence measures must be taken to keep Islamic banks from holding more cash than a regulated level. Arguably, capital adequacy would help in addressing liquidity risks for two reasons. Two arguments were brought forward to support this hypothesis: First, as long as an institution holds sufficient capital, it will be able to refinance itself via the market or the central bank at any time and will therefore not face excessive liquidity shortages. Second, requiring banks to hold sufficient capital for riskweighted assets directly incentivizes banks to hold more assets with lower riskweights, which usually have better liquidity quality. Considering liquidity risk to be a subcomponent of capital risk reduces the need for liquidity-specific minimum standards. However, in the case of Turkish PBs, a bank with a strong capital base still can have a liquidity problem. If it is looked at PBs capital adequacy ratio, it can be seen that PBs of Turkey have enough capital but most of their capital is in cash form. Although they began to issue Tier-2 capital, these instruments have not reached to a satisfactory level. Therefore, the capital adequacy has never been a problem for the PBs. Instead, their main issue is asset-liability and liquidity management. The arbitrage effect of regulations on banks’ activities is one of the reasons for increasing activities of shadow banking. Liquidity requirements have effects on government bonds, interbank credit lines and loans, retail deposits and net interest income riskiness, and quality of assets. Most banks perform maturity transformation and give priority to buying short-term instruments, which have higher liquidity treatment than long-term instruments. This shows that liquidity requirements are influencing portfolio choices of banks. While it is well accepted that without adequate regulatory involvement it would not be possible to control and mitigate the risks related to Islamic banking financial intermediation, there should be a balance between the growth and stability of the industry. The regulatory involvement that satisfies this balance would be welcome. For this reason, the objectives of applying prudential regulation and supervision of Islamic financial activities are to mitigate risks and to maintain financial stability by ensuring the safety and resiliency of these banks.

8.6 A Proposal for Treatment of Esham

153

8.5 Central Banks and Monetary Policy Central banks’ policies bear important implications on banks’ risk-taking or risksharing. According to Delis et al. (2017), the propensity of risk taking is negatively related to monetary policy rates. That is, as monetary policy eases, banks assume a higher percentage of riskier assets. Moreover, central bank’s treatment of assets and liquidity tools are applicable to all banks. If a central bank develops special products for these institutions, Islamic banks in its jurisdiction will manage liquidity much better because of their confidence in the lender of last resort mechanism. While the role of the central bank would be limited in a truly Islamic system, in today’s real global world, there is a need for a prudent central bank that tries to level the playing field for Islamic banks. In this context, an important issue of an asset’s liquidity is whether or not it is a central bank eligible. An asset is accepted as central bank eligible if it can be used as collateral for the central bank open market operations. The BCBS considers the harmonization of collateral frameworks as an almost impossible task (BCBS, 2013b). Hence, the BCBS deemed harmonization of the definition of central bank eligibility unfeasible. This acceptance, if done, permits the treatment of Islamic assets as eligible assets for central banks. For these reasons, central banks in Muslim countries should be receptive of Islamic assets as eligible collateral for open market operations. Since Islamic banks are small, treatment of their assets as eligible for open market operations would not cause any problems to the interbank money markets and monetary policy transmission channels.

8.6 A Proposal for Treatment of Esham As discussed in the preceding sections, one of the challenges that Islamic banks are facing is the scarcity of new products. There is still a dearth in Islamic products that are relevant to the needs of Islamic banks’ liquidity management. In this section, it is planned to discuss new products relevant for Islamic banks’ liquidity management, mainly Esham, risk-sharing instruments, GDP-linked Sukuk, CommodityBased Sukuk and Perpetual Sukuk. While Esham is an old financial product, others are new products designed and developed recently, and some of them have already been issued by sovereigns. Esham was used as a financial product during the Ottoman Empire for long-term public finance and has recently been restructured for modern usage (Çizakca, 2014a). In the Islamic world, its origins can be traced back to the year 1775. Esham is a hybrid fixed income and profit-loss sharing instrument and is designed to be traded in the secondary market. It is issued by the issuer/borrower/obligor based on tangible asset/assets yielding regular annual revenue. Since it is an

154

8 New Regulatory Framework for PBs’ Liquidity Management

asset-based instrument, the investors have full recourse to the asset in case of default. The asset is managed by the issuer, who allocates a mere fraction of the annual revenue to investors (Çizakça, 2014a). Subsequent to the issuance of Esham by an obligor, the expected revenue of assets is securitized into equal shares according to annual revenue pro rata. Although Esham was not issued for a fixed period, it could be structured for as a long-term financial asset, e.g. more than 25 years or in perpetuity. The regulatory framework of Esham should be constructed in the same way as perpetual Sukuk. Both of them should be accepted as high-quality liquid instruments since both are based on tangible (physical) assets and will have a functional secondary market if issued regularly. In some jurisdictions, the regulatory framework contains a single set of capital adequacy requirements applicable to all banks. Thus, no distinction is made between the capital requirements that apply to Islamic and conventional banks. In these countries, it would be difficult to give regulatory treatment to Esham because it is a new product, which has yet to have secondary market performance to appraise its liquidity in the market. However, it is possible to treat the product in the same way as perpetual bonds provided that their historical record is available. In other jurisdictions (e.g., Bahrain, Jordan, Malaysia, and Sudan), the regulatory capital adequacy requirements contain prescriptions that are often based on IFSB prudential standards and guiding principles on needed adjustments to the BCBS capital framework to accommodate certain Islamic banking features. In these jurisdictions, Esham can be accepted in the same way as GDP-linked Sukuk and perpetual Sukuk. In these countries, underlying assets of Esham is important. Since Esham would be a long-term paper generally longer than 25 years, it would be possible to treat these assets as liquid assets. However, the underlying assets should have a high quality of cash flow, and there would be full transparency regarding the cash, administration, and structure of the asset. The treatment of commodity-based Sukuk is more problematic because of restrictions on the type of underlying assets. For example, according to general acceptance, Sukuk cannot be issued based on gold, silver, and four other items. Additionally, the Basel Committee does not treat gold as a liquid asset. Therefore, commodity-based Sukuk should be backed by tangible assets with price stability to have relevant liquidity treatment. All Muslim countries wishing to develop Islamic finance should issue Esham or Perpetual Sukuk with a maturity longer than 25 years. These papers would be structured on Ijarah and can be easily traded on secondary markets. When these instruments are issued by the Treasury of sovereign authorities, they may well serve as an indicator of the future of Islamic finance. They will increase trust in Islamic financial institutions and be treated as high-quality assets which are eligible for open market operations. In this sense, these assets could become accepted as a benchmark instrument by the market.

8.8 The IFSB Supervisory Framework

155

8.7 EXIMS – Export-Based Musharakah Securities In this section, it is proposed a new liquid instrument for Islamic banks. EXIMS is defined as export-based Musharakah securities. It is a new product with the potential to increase partnership-based, high-quality liquid assets for Islamic banks. Many Muslim countries have fully state-owned banks, often called an Eximbank or Export Credit Agency, to facilitate and promote exports and imports of the country. The main objective of these banks is to finance short-term needs of exporters. Since these banks are governmental institutions, they can issue securities to mobilize new financial resources. These agencies are regarded as sovereign or quasi-sovereign authorities. This means they are accepted on par with treasuries in terms of their liabilities, indicating that they have implicit or explicit government guarantees. Consequently, these institutions can issue short-term Musharakah-based securities for financing short-term needs of exporters. Since most of the exporters benefit from the funds provided by these institutions, they have more incentives to be transparent regarding their income and value added to the products. A central function of export credit agencies or eximbanks is to provide insurance and guarantees for exporters. For this reason, agency problems would be minimized in the case of Musharakah securities issued by these organizations. When an export bank or agency issues new securities, it would act as a Mudarib. Hence, investors would become owners (Rab ul- Maal). Exporters would share their profits with the investors. Since these products would be short-term instruments and have the government’s explicit or implicit guarantee, they would be expected to have the same rating as ratings of other sovereign financial instruments. After issuance of the EXIMS, a central bank should consider them as eligible instruments of open market operations.

8.8 The IFSB Supervisory Framework A major challenge to the growth and expansion of Islamic finance is the lack of regulatory and supervisory standards specifically appropriate to Islamic finance. At present, the IFSB is the standard setter whose standards cover many issues related to Islamic finance. However, most of these standards are very similar to those of the conventional system. IFSB standards recommend risk-based supervision of Islamic banks. On a general level, supervisory authorities appear to apply the same generic supervisory framework to both Islamic banks and conventional banks. They also use the same approaches, methodologies, processes, and procedures. To deal with risks proactively, an appropriate stress-testing framework formulated according to the Islamic bank’s specificities needs to be developed by the IFSB with improved data focusing on the key features of the Islamic banking. On an idiosyncratic level, the IFSB should advise and apply a compliance approach to Islamic banks.

156

8 New Regulatory Framework for PBs’ Liquidity Management

It may be appropriate to apply risk management techniques equally to both Islamic banking and conventional banking. In this approach, of course, a generic framework requires the application of generic processes and methodologies (Song and Oosthuizen, 2014). However, a global standard setter for Islamic banks should be careful when setting these standards. For example, if this body focuses solely on risks, local supervisory authorities should also only have to focus on risks. It is expected that an Islamic bank would conduct sound risk management and ensure that it has the appropriate capabilities, systems, procedures, and governance to manage and mitigate risks. Accordingly, risk management prescriptions contained in the legal and regulatory framework typically apply to all banks in a jurisdiction. Thus, an Islamic bank is required to take account of the specific Islamic banking factors which may impact its risk profile. For this reason, the IFSB should develop rules that can accommodate these specific factors, analyze the markets, expand the playing field and focus on the religious aspects of products and contracts to cover extraordinary aspects of these banks. At the specific level, the IFSB can enforce different regulations appropriate to the risk profile of a typical Islamic bank rather than those applied to conventional banks. These differences would arise from the application of Shariah prescriptions and include factors such as the structure and legal form of the transactions, the assets and liabilities arising from the businesses, risks that were undertaken, and the party absorbing the risk. Consequently, the IFSB conceptual regulatory framework should go beyond risk management. Since the risk aspects are covered well by the Basel regulations, the IFSB should complement these standards by concentrating on the development of other regulations for Islamic banks.

8.9 Designing a New Liquidity Ratio and the Net Stable Funding Ratio The results found by Aldasoro and Faia (2016) provide strong evidence for differentiating the LCR requirements for different banks. They show that imposing LCR differentially across banks is highly effective in delivering a more stable system. Since this study shows that liquidity risk management poses particular challenges to Islamic banks, it is proposed that LCR be applied only to internationally active banks. It is a challenge for Islamic banks to maintain an appropriate level of resources for liquidity management purposes and to simultaneously optimize the return on such resources by finding appropriate investments for their surplus liquid resources. Since in Islamic banking, compatible liquid assets are very scarce in interbank markets (IMF, 2015), an Islamic bank with surplus often makes use of Commodity Murabahah transactions for its liquidity management. In developing a structure for liquidity management, the specifically appointed individuals in charge of managing the liquidity strategy may set liquidity risk limits

8.9 Designing a New Liquidity Ratio and the Net Stable Funding Ratio

157

conditional on the bank’s size and complexity. These limits would be reviewed by supervisors. It is advised for a bank to analyze its stress resistance under different scenarios and to have information systems in place to check whether or not it complies with the policies. Additionally, the bank is expected to carefully assess cash inflows against outflows to identify potential shortfalls under several scenarios, and to make sound assumptions about future funding needs. The bank can develop contingency plans specific to its needs charting strategies for emergency situations. Internal control systems, both conventional internal auditing and Shariah screening and auditing, are very important for ensuring a sound liquidity risk management process. These control activities should be reported to relevant bodies or stakeholders as well as Shariah advisory boards. It is recommended that stress tests are performed regularly, and that the results are actively integrated into the bank’s liquidity risk management strategies. The bank should also consider the potential behavioral response of counterparties under the assumed stress situations, and be aware of the fact that a stress event could affect its customers’ use of intraday liquidity – which may threaten the liquidity position of the bank. The qualitative information, such as the aspects of liquidity risk in which the bank is involved, the assumptions used in the measurement of risk, and the limits imposed on liquidity positions, should all be disclosed by the banks. The new framework will be based on the recommendations made so far. The main objective of the new LCR for only internationally active banks is to keep funding liquidity and market liquidity stable. In this context, it is possible to design a new LCR by protecting resilient and stable funding of the internationally active Islamic banks. CoðtÞ − wβ ≤ If t + LAt − C

(16)

Where CoðtÞ is cash outflows, If t is cash inflows, and LAt represents liquid assets. Β, added to the left side of the equation, is the coefficient of the restricted investment accounts cash outflows, and is specified as 40% of the restricted investment account. The main logic behind this coefficient is simple: It is assumed that PBs have 40% and investors have 60% profit share of investments according to the general practice among PBs in Turkey. This equation shows that the LCR has to be larger than 60% of net cash outflows. For Islamic banks, the right-hand side of the equation should be larger than the left-hand side of the equation since Islamic banks take deposits from customers for both investment and saving purposes. Saving accounts are exposed to cash outflows while investment accounts are expected to be stable. For this reason, 100% level of LCR would not be appropriate for Islamic banks. The variable C represents cash account on the asset side of the PBs’ balance sheets. Since PBs can keep cash arbitrarily, there is a need to restrict these arbitrary actions. Cash account should be limited to 20% of LCR. According to the newly designed LCR proposed here, PBs have to keep 60% LCR. Moreover, the β coefficient

158

8 New Regulatory Framework for PBs’ Liquidity Management

will be 40% of restricted investment account cash flows and would be deducted from net cash outflows. According to the discussion made so far, it is offered a change to the framework of LCR for Islamic banks. The suggestion is to divide the new LCR into three components: Cash-based LCR, highly-liquid instruments-based LCR, and net cash outflow. Cashbased LCR would cover cash and central bank reserves, and would be limited to 20% of new cash outflow. Highly-liquid instruments-based LCR are limited to an interval between 50% and 90%. The new LCR is proposed to have a ceiling of 110% and a floor of 60%. In this framework, banks have to control their cash stocks. This means that regulators will control both the floor and ceiling of the high-quality liquid instruments. There is no need to concentrate on cash or central bank reserves. The minimum and maximum threshold for the new LCR framework can be seen in Table 8.2.

Table 8.2: Minimum and Maximum Threshold for New LCR Framework. Minimum (%)

Maximum (%)

Cash





Central Bank Reserves





Sukuk





Risk Sharing Instruments





Perpetual Sukuk or long-term Securities





Esham





EXIMS





Other New Structured liquid instruments





Other Shariah-Compliant Instruments





Maximum Share for a Group (%) 



Highly-liquid instruments, from an Islamic perspective, are instruments that are based on real assets, have a transparent mechanism and structured to safeguard the interests of both Mudarib and Rab ul-Maal (capital owner), and have active (first and secondary) market. These products are expected to be sold very easily in the markets because they have a real right to recourse to the asset and have a secondary market. The model proposed here simplifies the LCR for Islamic banks. This new liquidity framework would classify cash and central bank statutory reserves as Level 1 assets and Sukuk, Exims, risk-sharing instruments, Esham, Perpetual Sukuk, any new partnership-based instruments and other liquid instruments backed by, or based on,

8.9 Designing a New Liquidity Ratio and the Net Stable Funding Ratio

159

tangible assets as Level 2 assets. Relevant potential haircuts based on the market demand should be applied to these assets according to their riskiness and transparency levels. Credit rating agencies should credit banks’ transparency levels and product transparency levels. Assets that are fully transparent should not have any haircut if they have deep secondary markets. An indicative haircut list for new LCR is presented in Table 8.3. Table 8.3: New LCR Hair Cut. Asset

Haircut

Level  Cash (If between ceiling and floor) Central Bank Reserves (If between ceiling and floor) Level 



Sukuk or other Shariah-compliant Securities issued by MDBs



Sovereign Sukuk (Ijarah Sukuk) had secondary market and based on tangible assets



Perpetual Sukuk or long-term Securities



Esham



Sovereign Sukuk (Wakalah Sukuk) have secondary market and based on pool of assets



Risk Sharing Instruments issued for public finance but can be sold in the secondary market.



EXIMS



Other New Structured Liquid Instruments issued by Sovereigns



Other Shariah-Compliant Instruments issued by PSEs



Corporate Sukuk based on AAA – AA rating



Corporate Sukuk based on A rating



Corporate Sukuk based on BBB – BB rating



After discussing the LCR, this section aims to provide a description of the second measure, the Net Stable Funding Ratio (NSFR). The NSFR has a 1-year horizon and aims to ensure a sustainable maturity structure of assets and liabilities. The NSFR is supposed to incentivize banks to fund their activities with more stable sources of funding, defined as follows: NSFR =Available Stable Funding/Required Stable Funding ≥ 100% (17)

160

8 New Regulatory Framework for PBs’ Liquidity Management

Available Stable Funding (ASF) is funding on which banks are likely able to rely for one year or longer. Similar to the LCR, ASF is calculated by multiplying balance sheet positions by assumed stability factors. A bank is assumed to be able to rely fully on its regulatory capital, and to a large extent, on retail deposits. Funding provided by wholesale clients, on the other hand, is assumed to be less stable. The stress testing results show that PBs need only 2% of required stable funding. Most of PBs’ NSFR is more than 100%, and only one bank needs to increase its required stable funding stock. Since the NSFR requirement will be launched after 2018, PBs can easily increase their stable funding in time. Indeed, if the recommended LCR is accepted and applied by regulators, Islamic banks would not face any problem regarding required stable funding.

8.10 Transparency and Symmetric Information There are many different applications in Muslim Majority Countries in terms of transparency and accounting standards. A few countries are appliying the AAOIFI standards. Many others conduct the IFRS or the national accounting standards. The AAOIFI accounting standards are mandatory in eight countries. AAOIFI has a program for certification of auditors, accountants, Shariah scholars, and Shariah trainers (Song and Oosthuizen, 2014). In this book, it is suggested that the transparency of Islamic banks’ liquidity management should be enhanced through compliance with capital market requirements for their investments. In this sense, all Islamic banks should be required to make available clear and direct information regarding their liquid investments to the investors who deposit cash in their account. Increasing the transparency of liquid instruments would be an effective buffer against cash outflows. Since in this structure investors have substantial information regarding an Islamic bank’s liquid investment, they would not have a right to demand their cash in stress times. Improving transparency would provide Muslim investors a better base for understanding banks’ internal strategies and risks inherent in the model. López et al. (2014) offer a disclosure regime that requires provision of sufficient information to assess the relevance of policies regarding investment and portfolio diversification, risk-weight of exposure to illiquid assets, and the corporate governance system. Stiglitz (2000) gives an example of putting a race car engine into an old car and setting off without checking the tires or the training of the driver. This interesting anology reminds us that all components should be reformed and restructured together to increase resiliency and reduce vulnerabilities. In this sense, the principles of transparency will help the process of checking other components that are not discussed in details as well as the training and disciplining. Correspondingly, the markets with low transparency may have to increase liquidity (Mousavi and Mousavinia, 2014). Increasing transparency is an important buffer

8.11 Differentiation between Local and International Active Islamic Banks

161

against liquidity runs. For this reason, the following principles for increasing transparency for investors are developed: 1. A more detailed report on profit loss sharing investment should be provided to investment account holders, capital owners and regulators every month. 2. Internal models and stress testing models’ results should be provided to investment account holders annually. 3. Counterparty risks and the methods for mitigating these risks should be transparent to customers. 4. Develop and implement good corporate governance to enhance the trust of investors and other stakeholders. In this context, financial goals of the bank, the achievement rate of objectives, and statements for future objectives should be transparent to all investors. 5. Both pre-trade (if needed) and post-trade information regarding Sukuk issuances, and information related to assets that back Sukuk or other transactions (Tawarruq, Commodity Murabahah, etc.) should be available to investment account holders, and even to the unrestricted investment account holders. More importantly, this information should be provided after each issuance or trade of instruments. 6. A cash report should be prepared regularly, in which all the reasons for staying in cash should be explained in detail. 7. Information on Shariah board members, meeting days, the material provided to them for their decision making, as well as the Shariah arguments underpinning the boards’ decision, should be provided to all stakeholders. Islamic banks can disclose this information to the public, or they can choose to provide it only to their investors, or their stakeholders. Each bank can construct a new platform via internet in where they provide all this information to investors especially restricted and unrestricted investment account holders.

8.11 Differentiation between Local and International Active Islamic Banks Basel Accords are basically designed for internationally active banks; those that have international connections, have correspondent relations, and invest in international financial centers. Small local banks are concentrated in one region or a singular country that have no international activities. Although Basel II and Basel III standards were originally designed for internationally active banks, most of the regulators apply these standards to all banks. The findings of Aldasoro and Faia (2016) have reinforced differentiating the LCR requirements for different banks because imposing LCR differentially across banks is highly effective in delivering a more stable system. Since this study shows that liquidity risk management poses

162

8 New Regulatory Framework for PBs’ Liquidity Management

particular challenges to Islamic banks, it is proposed to exempt small banks from the LCR requirements. Basel III standards give discretionary powers to the regulatory bodies for country-specific applications. It is possible to use these powers to expand the playing field for Islamic banks. However, many Muslim countries’ regulatory authorities strictly follow Basel standards without using their discretionary powers to facilitate the growth and expansion of Islamic banks. Small local Islamic banks should be excluded from Basel accords that are developed specially for internationally active banks. These small Islamic banks are proposed to keep cash, Sukuk, risk sharing infrastructure project financing and partnership-based products in their asset side of balance sheets (Table 8.4). Table 8.4: Proposed Balance Sheet for Small Islamic Banks. Assets

Liabilities

Cash and Cash Equivalents – less than % of total assets

Investment Accounts

Sukuk

Wadiah Accounts

Risk Sharing Infrastructure Financing

Tier  Sukuk

Partnership Based Financing – at least % of assets

Paid – in Capital

Others (Murabahah, Ijarah) – Less than % of total assets

Others (undistributed Profit, etc.)

It is possible to propose a fully risk-sharing based framework for Islamic banks. In this framework Islamic banks would act similarly to mutual funds, managing both asset and liability side risks with risk-sharing instruments. Since mutual funds are currently not under the Basel framework, they can quickly issue risk-sharing based instruments. However, in this model, these institutions have to manage liquidity more carefully because on the liability side they will have investment papers, and on the asset side, they will have short-term and long-term financing, including diminishing Musharakah and private equity financing or venture capital. The developed model in this book stands at the border of the current framework, proposing some changes to regulations in the existing framework to expand the playing field for Islamic banks. Such changes would allow Islamic banks to expand financial inclusion of a large and currently excluded segment of the society when a framework or model more amiable to their beliefs is adopted.

8.12 New ALA Approach There are three options within the current ALA mechanism developed by the Basel Committee and accepted and proposed to Islamic banks by IFSB (2015). The first

8.13 Summary, Conclusion and Policy Recommendations

163

option is to use contractually committed liquidity facilities from the relevant central bank with a fee. This option is not reasonable for Islamic financial institutions that have to bear the costs of high liquid assets and sacrifice profitability. This option has led to reductions in the profit of Islamic banks. According to many scholars, the mechanism accepted and proposed by the IFSB is disputable. The second option is to use foreign currency HQLA to cover currency liquidity needs. This option can be accepted as an alternative because Islamic banks have option to hold the IDB Sukuk or the IILM Sukuk, both having different mechanisms. Using this option, Islamic banks have to find US dollars to buy these securities and hedge their foreign currency risk. This option also increases the cost of capital for Islamic banks. The third option is to use additional Level 2 assets with a higher haircut. This option is also not viable for IIFS because there is a shortage of Level 2 assets in many jurisdictions. An ALA facility can therefore be designed in a different way. First of all, under the ALA mechanism, there should be no limit to Level 2A or Level 2B assets. Secondly, a contractual commitment mechanism should be offered by central banks without any fee under several circumstances, such as, if a bank has a shortterm shortage of high-quality liquid assets. The banks that have enough cash should have the right to sign a Wakalah contract with the central bank without any fee. The central bank will give high-quality liquid assets to Islamic banks by applying criteria specified in its regulations. The last option is the use of foreign currency HQLA with the hedging opportunity of central banks.

8.13 Summary, Conclusion and Policy Recommendations This book critically investigates the factors that affect liquidity risk management of Islamic banks and then develops an alternative regulatory framework appropriate for liquidity management of these banks. It includes time series and panel data analyses as well as stress testing. While there are several studies on the performance, growth, and efficiency of Islamic banks, there are limited empirical studies from the regulatory and supervisory perspectives. For this reason, this book concentrates on filling this gap by examining liquidity risk management of PBs in Turkey. The specific risk profile of an Islamic bank requires developing a new and more efficient regulatory framework. The Islamic banking system relies on risk sharing and needs to maintain symmetric information among parties. The regulatory framework should mitigate risks and concurrently protect the interests of investment account holders. This necessitates the design and implementation of new liquidity standards for Islamic banks. Protecting Islamic banks against the shocks coming from market liquidity and funding liquidity requires ensuring sound infrastructure, a well-developed surveillance system, prudential supervision, new products, and

164

8 New Regulatory Framework for PBs’ Liquidity Management

deep financial and capital markets. Moreover, the trust of investors is an important component for the growth of this sector. The only way to increase trust and confidence is to strengthen transparency, develop good corporate governance, and have a better business model. Moreover, the Basel standards and the IFSB standards do not currently provide the appropriate playing field required for Islamic banks. Although establishing the Islamic Financial Services Board was a much-needed attempt, the standards introduced by this institution are far below the expectations of Muslim investors. Indeed, these standards mimic those of the conventional system. For example, the LCR of the IFSB is almost similar to the Basel LCR (2013), except the former gives a higher treatment of several Sukuk. Nevertheless, it is recommended that the IFSB standards be applied to countries that wish to start licensing Islamic banking operations, since the IFSB has incorporated several new standards that allow expansion of the playing field for Islamic banks, especially the capital standards. Therefore, at present, these standards can be accepted as the second best option. However, it needs to be noted that liquidity risk management of Islamic banks has become highly crucial after the introduction of the LCR standards. The new liquidity framework entails the necessity to have enough eligible liquid assets, Shariah-compliant secondary markets for these instruments, and functional interbank short-term markets that have Shariah-compliant instruments. In this context, the short-term and long-term determinants of these banks’ liquidity holdings are examined using monthly data over the period of 2007–2015 using liquidity model that incorporates Sukuk, interbank market rate, required reserves, inflation rate and credit default swap rate, and these factors are compared with conventional banks’ liquidity. The long run analysis and Granger analysis show that these banks liquidity management has a direct causal relation with market liquidity and indirect relation with funding liquidity. Moreover, it is evidenced that if banks were forced to keep a higher level of liquidity, they would decrease credit given to the real sector. As PBs are exposed to market liquidity and market risks, more high-quality liquid instruments and a risk-sharing regulatory framework may provide the inner adjustment process through which any mismatch concerning maturity, risk, value or linkage with the real economy is corrected systematically. Short-term, long-term and stress testing analysis in this thesis showed that the LCR is not protecting Islamic banks against market stress and from other problems related to market liquidity, although Islamic banks have persistently held more than 100% LCR. Basel liquidity rules do not adequately protect Islamic banks against several risks, although they also reduce the profitability of Islamic banks and make them vulnerable to shocks. Islamic banks either have to keep cash or keep it in the central bank as reserves. This puts down the gains investors as well as the competitive power of Islamic banks. In the fifth chapter of this book, a time-series based liquidity model assessed liquidity management of banks in Turkey based on unit root analysis, cointegration

8.13 Summary, Conclusion and Policy Recommendations

165

test, ARDL test and longrun and short-run analyses. Since the LCR was formulated for the short run, greater importance was given to short-run relations. In this sense, Granger analysis shows that Islamic banks’ liquidity management has a causal relation with market liquidity. There is a direct causal relation between liquidity and credit volume of PBs. Their relation, however, are indirect for conventional banks. This result can be interpreted to mean that if Islamic banks are forced to keep higher levels of liquidity after the LCR implementation, they have to decrease credit given to the real sector. Moreover, since the inter-bank money market interest rates affect the liquidity of PBs, but not vice versa, the treasury of the country should increase new issuance. Furthermore, the impulse response analysis showed that liquidity responds to impulses generated by government bond, interbank rate, capital level, CDS rate and return on equity. Thus, these variables have causal relationships with liquidity. The main result is that market liquidity influences banks’ liquidity, and banks’ liquidity responds to the profitability of PBs. In the sixth chapter, it is formulated a panel data model for PBs and CBs of Turkey to address three research questions. According to fixed-effect models results of PBs, it is not found any evidence for the significant impact of market liquidity on the liquidity of PBs. Secondly, all variables included in the base model show significant relationships with liquidity, but coefficients and signs differ. In this context, short-term assets, capital, return on asset and net cash outflows have a negative sign, but others have positive impacts on liquidity. Moreover, strong evidence is documented in support of the fact that most bank specific variables (i.e. credit, deposits, cash, asset size, and profitability) are directly correlated to the changes in liquidity risk of banks in Turkey. More specifically profitability and deposits has a negative impact on liquidity risk of CBs and PBs; while credit and cash has positive effects on the liquidity of PBs but negative effects of liquidity of CBs. Capital, required reserves, high-quality liquid assets, cash and total assets have a positive effect on liquidity risk of PBs and CBs, except cash which is negatively related to the liquidity of CBs. However, for macro factors, the only mild evidence is found in the negative impact of CDS on liquidity risk. Also, inflation and government bond are insignificant for both types of banks, while the interbank rate has positive impacts on the liquidity of CBs. Additionally, there is evidence of positive effects of net cash outflows on the liquidity of CBs, but there is a negative and significant relationship between net cash outflows and liquidity of PBs. The results of the fixed-effect models indicate that liquidity risk in PBs can be differentiated from liquidity risk in CBs concerning the impact of net cash outflows, credit, cash, CDS and interbank rate. Then, a stress testing procedure is applied to the PBs of Turkey. The results indicate that PBs would not face large liquidity shortage under several assumptions, but not as far as expected. Since PBs are small and mid-sized banks, they are less likely to add direct stress in the market. However, indirectly, both the idiosyncratic loss of reputation and the wider systemic effects should be taken into consideration

166

8 New Regulatory Framework for PBs’ Liquidity Management

because the market and customers are very sensitive to trigger fire sales in case of liquidity problems even seen in a small bank. Later, generous haircuts on liquid assets applied according to Basel liquidity rules and authorities’ regulations do not worsen the liquidity of these banks. Since only one PB may face liquidity run under the scenarios, the spillover effects of freezing credit lines on other banks should be handled very carefully. After these quantitative explorations, it is formulated the new LCR for Islamic banks and offered an alternative liquid assets approach for the treatment of Islamic products. It is argued that this new regulatory framework will increase the resilience of Islamic banks’ liquidity management against market conditions and macroeconomic factors, strengthen Islamic bank’s top management to be more rulecompliant, and will also enhance the level of trust of investment account holders. Building and maintaining the trust of investment account holders would act as a buffer against cash outflows and strengthen the rational base for cash inflows. In this context, a new regulatory framework for Islamic banks’ liquidity based on the findings of the models discussed in the book and several policy options will be outlined as policy recommendations. Besides, an alternative regulatory treatment of liquidity for Perpetual Sukuk and Esham is proposed.

8.13.1 Policy Implications for Regulators and Supervisors In this book, it is recommended that more resilient infrastructure be developed for Islamic banks. In this context, it is proposed that Islamic banks’ stable funds can be increased by applying these new formulas and financial stability would be strengthened by usage of the following tools and applications: – With the new regulatory framework, Islamic banks should be supervised differently from the current supervision system. In this new system, supervisors will concentrate not only on specific risks of the model but also on the review the Shariah aspects of the applications. Thus, it is proposed incorporating Shariah review into the regulatory supervision procedures. – In the proposed new structure, more Sukuk and risk-sharing instruments would be issued. – Shariah-compliant credit rating, transparency mechanism, special credit rating mechanisms or institutions are also very essential for strengthening liquidity and asset liability management of these banks. – International liquidity management initiatives, enhancing cooperation, trust and communication are vigorous for controlling vulnerabilities, mitigating risks and curtailing regulatory arbitrage. – Moreover, both macro-level and micro-level mitigating mechanisms should be designed.

8.13 Summary, Conclusion and Policy Recommendations

167

– Since PBs are exposed to market liquidity and funding liquidity risks, more liquid instruments within a tailored regulatory framework are required for safety and profitability of these banks. – Additionally, some policy directions and actions are worth considering for Turkish regulators. In this context, the Turkish Treasury or other authorities should consider issuing more high-quality liquid instruments for these banks, especially short-term instruments, and should issue these papers in more frequently. – Regulators should work towards shaping a risk-sharing focused regulatory framework for Islamic banks (for example, they can use the liquidity framework formulated in this book).

8.13.2 Policy Implications for International Organizations It is challenging to regulate financial systems where risks are persistently shifted rather than shared, and regulation is constrained by moral hazards and negative externalities. A new paradigm for financial regulation requires a restructuring of these relations. Risk sharing thrives on the avoidance of leverage and debt refinancing on an interest rate basis, the matching of assets and liabilities, the elimination of credit multiplier effect, and the promotion of wealth formation rather than money creation. Future regulation should thus be guided by the following principles (Mirakhor, 2014): 1. Materiality-matching structure implies that the nominal value of each financial transaction is matched with the value of real assets such that the expected payoffs to stakeholders in financial intermediaries are reflective of the rates of return to the real sector of the economy. 2. Risk-matching structure is required to ensure that no asset is associated with a riskier corresponding element on the liabilities side. 3. Maturity-matching structure imposes a strict correspondence of assets to liabilities with similar maturities. 4. The value-matching structure requires that price fluctuations leading to asset revaluation are systematically offset by appropriate changes on the liabilities side. 5. Full transparency is required for each item on both sides of balance sheets of Islamic banks. 6. Governance structure requires that all stakeholders in financial intermediaries are empowered with appropriate rights of participation in the decision-making process. 7. Key elements to achieve this objective include: i) Understanding the nature of Islamic banking activities, ii) Making appropriate changes to the existing regulatory framework for Islamic banking, and

168

8 New Regulatory Framework for PBs’ Liquidity Management

iii) Levelling the playing field between Islamic and conventional banking. Current trends indicate that specific elements relating to Islamic banking are being increasingly incorporated into the regulatory framework. These conditions provide an inner adjustment process through which any mismatch regarding maturity, risk, value or linkage with the real economy are systematically corrected. The corrective dynamics of balance sheet structures can take place without undermining the crucial functions of financial intermediaries in the economy. As a result of the inherent stability of financial systems based on risk sharing, the costs of financial regulation and supervision are also clearly diminished. There is greater scope for the humanizing and democratizing process and more effective regulation under the risk-sharing principles of Islamic finance.25

8.13.3 Shortcomings of the Study There are several shortcomings of this study. First of all, this book covers only Turkey due to data availability. There should be various researches done on other countries where Islamic finance is growing and is systemically important. If all Islamic countries are considered, it would be possible to use other techniques such as the generalized method of moments (GMM) etc. Moreover, a stress test involving the data of several Islamic banks in different countries can be very useful for evaluating the liquidity of Islamic banks. However, separate assumptions must be made for each country. Since this study evaluates the factors affecting liquidity management from the regulatory point of view, it concentrates on regulatory aspects and does not discuss practitioner’s aspects in details. The thesis focuses only on the LCR; therefore, a new impact analysis is required for the effects of incoming NSFR on Islamic banks using monthly data.

8.13.4 Direction for Future Research It is possible to develop this research in the future in three particular areas. When more comprehensive and bank level data are available for multiple countries, running the model for assessing the determinants of liquidity risk would give tutorial results. Moreover, examining the effects of Basel reforms, including capital adequacy ratio, the NSFR, counterparty risks, leverage ratio and output floor on Islamic banking

25 The concept of value-matching structure is consistent with the marking-to-market process, which underlies the fair value accounting rules by the Financial Accounting Standards Board and the organized exchange of futures contracts.

8.13 Summary, Conclusion and Policy Recommendations

169

practices would be very informative for regulators and supervisors. Another issue will be to include small Islamic banks from advanced countries to see the effects of market liquidity on determinants of Islamic banks’ applications. Furthermore, using more advanced methods that include impact analyses of each regulation on banks’ balance sheets may give us more robust results. Lastly, for robustness tests, alternative measures of market liquidity and funding liquidity could be added to the models for better evaluating the factors that affect the attitude of Islamic banks regarding liquidity management.

References Abdymomunov, A. & Gerlach, J. (2014). Stress testing interest rate risk exposure. Journal of Banking & Finance, 49 (2014), pp. 287–301. Acharya, V.V., Gromb, D. & Yorulmazer, T. (2008). Imperfect Competition in the Interbank Market for Liquidity as a Rationale for Central Banking. Working Paper. Adam, N. J. & Thomas, A. (2005). Islamic Bonds – Your Guide to Issuing Structuring and Investing in Sukuk. Euromoney Books, London, UK. Admati, A., DeMarzo, P., Hellwig, M., & Pfleiderer, P. (2014). Fallacies and Irrelevant Facts In the Discussion on Capital Regulation. In Goodhart C., Gabor D., Vestergaard J., & Ertürk I. (Eds.), Central Banking at a Crossroads: Europe and Beyond (pp. 33–50). London; New York, NY: Anthem Press. Retrieved September 21, 2020, from http://www.jstor.org/stable/j.ctt1gxpd6m.6. Adrian, T., Fleming, M., Shachar, O. & Vogt, E. (2017). Market Liquidity after the Financial Crisis. Federal Reserve Bank of New York Staff Reports, No.796, October 2016, Revised January 2017. Adrian, T. & Brunnermeier, M. K. (2011). CoVaR. NBER Working Paper No. 17454 September 2011. Later published as Federal Reserve Bank of New York Staff Reports, no. 348, revised in September 2014. Ahmed, N., Ahmed, Z. & Naqvi, I.H. (2011). Liquidity Risk and Islamic Banks: Evidence from Pakistan. Interdisciplinary Journal of Research in Business, Vol. 1, Issue. 9, (pp.99–102) September, October, 2011. Aikman, D., Alessandri, P., Eklund, B., Gai, P., Kapadia, S., Martin, E., Mora, N., Sterne G. & Willison, M. (2009). Funding Liquidity Risk in a Quantitative Model of Systemic Stability. Bank of England Working Paper, 372. Akhtar, M.F., A. K. & Sadaqat, S. (2011). Factors Influencing the Profitability of Islamic Banks of Pakistan. International Research Journal of Finance and Economics. ISSN 1450–2887 Issue 66. Akhtar, S. (2007). Building an effective Islamic financial system. Global Islamic Financial Forum Governor’s: Financial Regulators Forum in Islamic Finance, Kuala Lumpur, Malaysia. Akkizidis, I. & Khandelwal, S. K. (2008). Financial Risk Management for Islamic Banking and Finance. Palgrave Macmillan, New York. Alam, N. (2013). Impact of banking regulation on risk and efficiency in Islamic banking. Journal of Financial Reporting and Accounting, 11(1), pp. 29–50. Aldasoro, I. & Faia, E. (2016). Systemic loops and liquidity regulation. Journal of Financial Stability, Volume 27, December 2016, pp. 1–16 Alessandri, P., Gai, P., Kapadia, S., Mora, N. & Puhr. C. (2009). Towards a Framework for Quantifying Systemic Stability. International Journal of Central Banking, 5 (3), pp. 47–81. Allen, F. and Carletti, E. (2013). Systemic risk from real estate and macro-prudential regulation. Int. J. Banking, Accounting and Finance, Vol. 5, Nos. 1/2, pp.28–48. Almeida, H., Cunha, I., Miguel A. Ferreira, M. A. & Restrepo, F. (2017). The Real Effects of Credit Ratings: The Sovereign Ceiling Channel. The Journal of Finance, Volume 72, Issue 1, February 2017. Amihud, Y., Mendelson, H. & Pederson, L.H., (2005). Liquidity and asset prices. Foundations and Trends in Finance, 1, 4, pp.269–364. Amin, S.I.M. (2016). Liquidity Risk Determinants: Are Islamic Banks Different from Conventional Banks? Dissertation, INCEIF. Andries, A. M., Fischer, A. M. & Yesin, P. (2017). The Asymmetric effect of International Swap Lines on Banks in Emerging Markets. Journal of Banking and Finance, 75, pp. 215–234. Angora, A., & Roulet, C. (2011). Transformation risk and its determinants: A new approach based on the Basel III liquidity management framework. Université de Limoge Working Paper.

https://doi.org/10.1515/9783110582901-009

172

References

Arif, A. & Anees, A.N. (2012). Liquidity risk and performance of banking system. Journal of Financial Regulation and Compliance, Vol. 20 No. 2, pp. 182–195. Ariffin, N.M. (2012). Liquidity Risk Management and Financial Performance in Malaysia: Empirical Evidence From Islamic Banks. Aceh International Journal of Social Sciences, 1(2), pp. 77–84. December 2012. Ariss, R.T. (2010). Competitive conditions in Islamic and conventional banking: A global perspective. Review of Financial Economics. 19 (2010) 101–108. Arshad, M. K. (2011). Implementation of an Islamic Deposit Insurance System for Islamic Financial Services Industry. Fourth Islamic Financial Stability Forum – Strengthening Financial Safety Nets in the Islamic Financial Services Industry. Kuala Lumpur: Islamic Financial Services Board (IFSB). Ashraf, D., Rizwan, M.S. & L’Huillier, B. (2016). A net stable funding ratio for Islamic banks and its impact on financial stability: An international investigation. Journal of financial Stability. Volume 25. August 2016, pp. 47–57. Asutay, M. (2013). The Development of Islamic banking in Turkey: regulation, performance and political economy. Islamic Finance in Europe, Edited by V. Cattelan, pp.213–227. Aysan, A. F., Dolgun, M. H. and Turhan, M. I. (2013). The assessment of Participation banks in Turkey and their role on financial inclusion. Emerging Markets Finance and Trade, 49. December 2013. Azam, A. S., Mohamad, M. T., & Samsudin, M. L. (2013). How Islamic banks of Malaysia managing liquidity? An emphasis on confronting economic cycles. International Journal of Business and Social Science, 4(7), pp. 253–264. Baltagi, B. H. (2005). Econometric Analysis of Panel Data. John Wiley & Sons Ltd. Banerjee, R.N. & Mio, H. (2014). The Impact of Liquidity Regulation on Banks. BIS Working Papers, No.470. October 2014. Barth, J. R., Caprio, Jr. G. & Levine, T. (2004). Bank Regulation and supervision: what works best? Journal of Financial Intermediation, 13, pp. 205–248. Barth, J. R., Lin, C., Mac, Y., Seade, J. & Song, F. M. (2013). Do bank regulation, supervision and monitoring enhance or impede bank efficiency? Journal of Banking and Finance, 37, pp. 2879–2892. Benmelech, E., Meisenzahl, R.R. & Ramcharan, R. (2017). The Real Effects of Liquidity during the Financial Crisis: Evidence from Automobiles. The Quarterly Journal of Economics, Volume 132 Issue 1, February 2017. Berrospide, J.M. & Edge, R. M. (2009). Linkages between the Financial and Real Sectors: Bank Credit and Capital over the Crisis. October 2009. Available at: file:///C:/Users/ukdhdpe/ Downloads/Linkages_between_the_Financial_and_Real_Sectors_Ba.pdf. BIS, (Basel Committee on Banking Supervision-BCBS), (2005). Basel II: International Convergence of Capital Measurement and Capital Standards: A Revised Framework. November, 2005. BIS, (Basel Committee on Banking Supervision-BCBS), (2013a). Liquidity stress testing: a survey of theory, empirics and current industry and supervisory practices. BIS Working Paper No. 24, October. BIS, (Basel Committee on Banking Supervision-BCBS), (2013b). Literature review of factors relating to liquidity stress – extended version. BIS Working Paper No. 25, October. BIS, (Basel Committee on Banking Supervision-BCBS) (2013c). Basel III: the Liquidity Coverage Ratio and liquidity risk monitoring tools. Download: 6 March 2015, www.bis.org. BIS, (Basel Committee on Banking Supervision-BCBS) (2014). Impact and Implementation Challenges of the Basel Framework for Emerging Market, Developing and Small Economies. Download: 6 March 2015, www.bis.org. BIS, (2015). Basel III monitoring report. Download: 10 March 2015. www.bis.org.

References

173

BIS, (2015). Making supervisory stress tests more macroprudential: Considering liquidity and solvency interactions and systemic risk. November 2015. https://www.bis.org/bcbs/publ/ wp29.pdf. Bissoondoyal, B. E. & Treepongkaruna, S. (2011). An Analysis of the Determinants of Bank Ratings: Comparison across Ratings Agencies. Australian Journal of Management, Vol. 36, No. 3, (December 9, 2011). Boissay, F. & Collard, F., (2016). Macroeconomics of Bank Capital and Liquidity Regulations. BIS Working Paper, No.596. December 2016. Bonfim, D. & Kim, M., (2012). Liquidity risk in banking: is there herding? European Banking Center Discussion Paper No. 2012–024, available at SSRN: http://ssrn.com/abstract=2163547. Bonner, C., Lelyvelda, I. V. & Zymekd, R. (2013). Banks’ liquidity buffers and the role of liquidity regulation. DNB Working Paper No. 393 / September 2013. Bonner, C. (2014). Liquidity regulation and bank behaviour. Tilburg: CentER, Center for Economic Research. Borio, C., Drehmann, M. & Tsatsaronis, K. (2012). Stress Testing Macro Stress Testing: Does it Live Up to Expectations? BIS Working Paper No. 369. Boss, M., Krenn, G., Puhr, C. & Summer, M. (2006). Systemic Risk Monitor: A Model for Systemic Risk Analysis and Stress Testing of Banking Systems. Financial Stability Report, No.11, OeNB, pp. 83–95. Boudt, K., Paulus, E. C. S. & Rosenthal, D. W. R. (2013). Funding liquidity, market liquidity and TED spread: A two-regime model. National Bank of Belgium Working Paper, Download: 9 November 2014, 244. Bourkhis, K. & Mahmud, N. S. (2013). Islamic and conventional banks’ soundness during the 2007–2008 financial crises. Review of Financial Economics, 22, pp. 68–77. Brunnermeier, M. K. & Pedersen, L. H. (2008). Market Liquidity and Funding Liquidity. Oxford University Press on behalf of The Society for Financial Studies, December 10, 2008. Calluzzo, P. & Dong, G. N. (2015). Has the financial system become safer after the crisis? The changing nature of financial institution risk. Journal of Banking and Finance, 53 (2015) pp. 233–248. Carlson, M, H Shan & Warusawitharana, M. (2013). Capital ratios and bank lending: A matched bank approach. Journal of Financial Intermediation, vol 22, no 4, pp 663–87, ac.els-cdn.com/ S1042957313000284/1-s2.0-S1042957313000284-main.pdf?_tid=f2b970c8-6cfd-11e5-9aad00000aab0f02&acdnat=1444227546_56f2c5c50afe6372dc759a2028f3b828. Cathcart, L., El-Jahel, L. & Jabbour, R. (2015). Can regulators allow banks to set their own capital ratios? Journal of Banking and Finance, 53 (2015), pp. 112–123. Cerutti, E & Schmieder, C. (2014). Ring fencing and consolidated banks’ stress tests. Journal of Financial Stability, 11 (2014), pp. 1–12. Çizakca, M. (2014a). Can there be innovation in Islamic finance? Case study: ESHAM. 11th IFSB Summit, May 20, 2014. Mauritius. Çizakca, M. (2014b). Risk sharing and risk shifting: An historical perspective. Borsa Istanbul Review, Volume 14, Issue 4, December 2014, pp. 191–195 Čihák, M. (2007). Introduction to Applied Stress Testing. IMF Working Papers, Vol., pp. 1–74, 2007. SSRN: https://ssrn.com/abstract=973989. Čihák, M. (2014). Stress Tester: A Toolkit for Bank- by- Bank Analysis with Accounting Data. A Guide to IMF Stress Testing: Methods and Models, 17. Chapra, M. U. (1992). Towards a Just Monetary System. The Islamic Foundation, Liecester. Chen, T. H., Chou, H. H., Chang, Y., & Fang, H. (2015). The effect of excess lending on bank liquidity: evidence from China. International Review of Economics & Finance, 36, pp. 54–68.

174

References

Chong, B.S. & Liu, M. L. (2009). Islamic banking: Interest-free or interest-based? Pacific-Basin Finance Journal, 17 (2009) pp. 125–144. Cornett, M. M., McNutt, J. J., Strahan, P.E. & Tehranian, H. (2011). Liquidity risk management and credit supply in the financial crisis. Journal of Financial Economics, 101, pp. 297–312. Covas, F. & Driscoll, J. C. (2014). Bank liquidity and capital regulation in general equilibrium. Finance and Economics Discussion Series, Federal Reserve Board. Covitz, D., Liang, N. & Suarez, G. (2013). The evolution of a financial crisis: collapse of the assetbacked commercial paper market. Journal of Finance, Vol. 68, no 3, pp. 815–48. Cucinelli, D. (2013). The Relationship between Liquidity Risk and Propability of Default: Evidence from the Euro Area. Risk Governance and Control: Financial Markets and Institutions, Vol.3, Issue 1. Delis, M.D., Hasan, I. & Mylonidis, N. (2017). The Risk-Taking Channel of Monetary Policy in the US: Evidence from Corporate Loan Data. Journal of Money, Credit and Banking, Volume 49, Issue 1, February 2017. Demekas, D.G., (2015). Designing Effective Macroprudential Stress Tests: Progress So Far and the Way Forward. IMF Working Paper, No. WP/15/146. June 2015. Deming, W. & Hong, H., (2012). Liquidity risk, market valuation and bank failures. Download: 27 December 2014, SSRN, http://ssrn.com/abstract=2177583. Demirgüç-Kunt, A., Laeven, L. & Levine, R. (2003). The Impact of Bank Regulations Concentration and Institutions on Bank Margins. World Bank Policy Research Working Paper, No: 3030. Demirgüç-Kunt, A. & Detragiache, E. (2011). Basel core principles and bank soundness: Does compliance matter? Journal of Financial Stability, 7, pp. 179–190. de-Ramon, S, Iscenko, Z, Osborne, M, Straughan, M and Andrews, P (2012). Measuring the impact of prudential policy on the macroeconomy. Financial Services Authority Occasional Paper 42, available at: http://webarchive.nationalarchives.gov.uk/20130106140043/http://www.fsa. gov.uk/static/pubs/occpapers/op42.pdf. DeYoung, R. & Jang, K. Y. (2016). Do Banks Actively Manage Their Liquidity? Journal of Banking and Finance. Vol. 66, (2016) pp. 143–161. Diamond, D.W., & Rajan, R.G. (2005). Liquidity shortages and banking crises. Journal of Finance, Vol. 60(2), pp. 615–647. Distinguin, I., Roulet, C., & Tarazi, A. (2013). Bank Regulatory Capital and Liquidity: Evidence from US and European publicly traded banks. Journal of Banking & Finance, 37(9), pp. 3295–3317. Drechsler, I., Drechsel, T., Marques-Ibanez, D., & Schnabl, P. (2016). Who Borrows from the Lender of Last Resort? Journal of Finance. Vol. 71, Issue 5. First published: 14 September 2016. Drehmann, M. & Nikolaou, K. (2009). Funding Liquidity Risk, Definition and Measurement. ECB Working Paper Series, No:1024, March 2009. Dusuki, A. W. (2007). Commodity Murabahah programme (CMP): An innovative approach to liquidity management. Journal of Islamic Economics, Banking and Finance, Vol. 3, Issue 1. Engle, R. F. & Granger, C. W. J. (1987). Co-Integration and Error Correction: Representation, Estimation, and Testing. Econometrica, Vol. 55, No. 2. (March 1987), pp. 251–276. Erden, L., & Ozkan, I. (2014). Determinants of international transmission of business cycles to Turkish economy. Economic Modelling, 36, pp. 383–390. Ergec, E. H. & Arslan, B. G., (2013). Impact of interest rates on Islamic and conventional banks: the case of Turkey. Applied Economics, 45, 17, pp. 2381–2388. Farooqi, F. M. & O’Brien, J. (2015). A Comparison of the Impact of the Basel Standards upon Islamic and Conventional Bank Risks in the Gulf State Region. Electronic copy available at: http://ssrn. com/abstract=2647562. Fraisse, H. Mathias, L. & David, T. (2015). The Real Effects of Bank Capital Requirements. Secretariat General De L’autorite De Controle Prudentiel et De Resolution. available at: https://acpr.ban que-france.fr/sites/default/files/medias/documents/def_numero_8_bis.pdf.

References

175

Fuhrer, L.M., Müller, B. & Steiner, L. (2017). The Liquidity Coverage Ratio and Security Prices. Journal of Banking and Finance, 75 (2017) 292–311. Fukuyama, H. and Matousek, R. (2011). Efficiency of Turkish banking: Two-stage network system. Variable returns to scale model. Journal of International Financial Markets, Institutions and Money, 2011, vol. 21, issue 1, pp. 75–91. Gauthier, C. & Souissi, M. (2012). Understanding systemic risk in the banking sector: A macrofinancial risk assessment framework. Bank of Canada Review, spring 2012, pp.29–37. Gauthier, C., He, Z. & Souissi. M. (2010). Understanding Systemic Risk: The Trade-Offs Between Capital, Short-Term Funding and Liquid Asset Holdings. Bank of Canada Working Paper No. 2010–29. Gehde-Trapp, M., Gündüz, Y. & Nasev, J. (2015). The liquidity premium in CDS transaction prices: Do frictions matter? Journal of Banking and Finance, 61, pp. 184–205. Giannotti, C., Gibilaro, L. & Mattarocci G. (2010). Liquidity Risk Exposure for Specialised and unspecialized Real Estate Banks: Evidence from the Italian Market. Journal of Property Investment and Finance, Vol. 29, No. 2, pp. 98–114. Goodhart, C.A.E., Sunirand, P. & Tsomocos, D.P. (2006). A model to analyse financial fragility. Economic Theory, Springer, Vol. 27(1), pp.107–142, 01. Gorton, G. & Winton, A. (2000). Liquidity provision, bank capital, and the macroeconomy. University of Minnesota, Mimeo. Granger, C. W. J. (1969). Investigating Causal Relations by Econometric Models and Cross-spectral Methods. Econometrica, Vol. 37, No. 3. August, 1969, pp. 424–438. Greenlaw, D., Kashyap, A. K., Schoenholtz, K., & H. S. Shin, (2012). Stressed Out: Macroprudential Principles for Stress Testing. Booth School of Business Working Paper No. 71. Gujurati, D. N. (2011). Basic Econometrics, Fifth Edition. The McGraw-Hill Companies. Haan, L.D. & End, J. W., (2013). Bank liquidity, the maturity ladder, and regulation. Journal of Banking and Finance, 37(10), pp. 3930–3950. Haque, N. & Mirakhor, A. (1999). The Design of Instruments For Government Finance in an Islamic Economy. Islamic Economic Studies, Vol. 6, No. 2, May 1999. Hardy, L. (2012). The Evolution of Participation Banking in Turkey. Al Nakhlah: Online Journal on Southeast Asia and Islamic Civilization, Winter: 2–15. Available from: http://fletcher.tufts. edu/~/media/Fletcher/Microsites/al%20Nakhlah/archives/Winter2012/Hardy_Final.pdf. Hasan, M. & Dridi, J. (2010). The effects of the global crisis on Islamic and conventional banks: A comparative study. Download: 24 October 2014, IMF Working Paper, WP/10/201. Hesse, H., Andreas, A.J. & Sole, J. (2008). Trend and Changes in Islamic Finance. World Economics. Vol. 9, No: 2. Hilscher, J. & Nosbusch, Y. (2010). Determinants of sovereign risk: macroeconomic fundamentals and the pricing of sovereign risk. Rev. Finance, 14 (2010), pp. 235–262 Hoque, H., Andriosopoulos, D., Andriosopoulos, K. & Douady, R. (2015). Bank regulation, risk and return: Evidence from the credit and sovereign debt crises. Journal of Banking and Finance, 50 (2015), pp. 455–474. Hong, H., Huang, J.Z. & Wu, D. (2014). The information content of Basel III liquidity risk measures. Journal of Financial Stability, 15 (2014), pp. 91–111 Hunt, A. K. (2013). Contemporary Islamic Finance: Innovations, Applications and Best Practices. John Wiley & Sons, 14 Oca 2013. IFSB, (2012). Guiding Principles on Stress Testing For Institutions Offering Islamic Financial Services. IFSB-13. March 2012. IFSB, (2014). Islamic financial services industry stability report. Download: 6 November 2014, www. ifsb.org. IFSB, (2015a). Islamic Financial Services Industry Stability Report 2015. Kuala Lumpur.

176

References

IFSB, (2015b). Guidance Note on Quantitative Measures for Liquidity Risk Management in Institutions Offering Islamic Financial Services. Kuala Lumpur. IFSB-15. April 2015. IMF, (2008). Global Financial Stability Report. April 2008. IMF, (2013). Modifications to the Current List of Financial Soundness Indicators. November 2013. https://www.imf.org/external/np/pp/eng/2013/111313.pdf IMF, (2015a). Islamic Finance: Opportunities, Challenges, and Policy Options. IMF Discussion Note, SDN/2015/5. April 2015. IMF, (2015b). World Financial Stability Report. IMF report, www.imf.org. Ismal, R. (2010). Assessing the demand and supply of liquidity in Islamic banking (The Case of Indonesia). ISRA International Journal of Islamic Finance, 2, 1. Johansen, S. & Juselies, K. (1990). Maximum Likelihood Estimation and Inference on Cointegration– with Applications to the Demand for Money. Oxford Bulletin of Economics and Statistics, Vol. 52, No. 2, pp. 169–210. Jones, M. T., Hilbers, P. & Slack, G., (2004). Stress testing financial systems: What to do when the governor calls. IMF Working Paper, no: 04/127. Kara, G.I. & Ozsoy, S.M., (2016). Bank regulation under fire sale externalities. Finance and Economics Discussion Series 2016–026. March 2016. Washington: Board of Governors of the Federal Reserve System. https://www.federalreserve.gov/econresdata/feds/2016/files/2016026pap.pdf Kassimatis, K. & Clark, E. (2015). Macroeconomic effects on emerging-markets sovereign credit spreads. Journal of Financial Stability, Volume 20, October 2015, pp. 1–13. Khediri, K. B., Charfeddined, L. & Youssef, S.B. (2015). Islamic versus conventional banks in the GCC countries: A comparative study using classification techniques. Research in International Business and Finance. 33, pp. 75–98. Kerry, W. (2008). Measuring financial market liquidity. Journal of Risk Management in Financial Institutions, 1, 81–90. Krasicka, O. & Nowak, S. (2012). What’s in it for me? A Primer on Differences between Islamic and Conventional Finance in Malaysia. IMF Working Paper, WP/12/151, June 2012. Laštůvková, J. (2016). Liquidity Determinants of the Selected Banking Sectors and Their Size Groups. Acta Universitatis Agriculturae et Silviculturae Mendelianae Brunensis, Vol. 64, Number 3. 2016. Lee, S. (2016). REITs and the Taper Tantrum. Journal of Property Investment & Finance, Vol. 34 Iss: 5, pp. 457–464. Leykun, F. (2016). Determinants of Commercial Banks’ Liquidity Risk: Evidence from Ethiopia. Research Journal of Finance and Accounting, ISSN 2222–1697 (Paper) ISSN 2222–2847 (Online), Vol.7, No.15, 2016. Litterman, R. (1986). A statistical approach to economic forecasting. Journal of Business and Economic Statistics, 4 (1986) 1–4, and later in International Journal of Forecasting, Elsevier, vol. 2(4), pp. 497–498. López, M.A., Aljabrin, S., Awad, R., Norat, M. & Song, I. (2014). Regulation and Supervision of Islamic Banks. IMF Working paper No. 14/219, Washington. Maghrebi, N. & Mirakhor, A. (2015). Risk Sharing and Shared Prosperity in Islamic Finance. Islamic Economic Studies, Vol. 23, No. 2, November, 2015, pp. 85–115. Mansor, H. I. & Shah, M. E. (2012). Bank lending, macroeconomic conditions and financial uncertainty: Evidence from Malaysia. Review of Development Finance, Volume 2, Issues 3–4, July–December 2012, pp. 156–164. Mirakhor, A. (2010). Whither Islamic finance? Risk sharing in an age of crises. The Inaugural Securities Commission Malaysia (SC) – Oxford Centre for Islamic Studies (OCIS) Roundtable “Developing a Scientific Methodology on Sharī`ah Governance for Positioning Islamic Finance Globally, 15 March 2010.

References

177

Mirakhor, A., Askari, H., Iqbal, Z. & Krichene, N. (2012). Risk Sharing in Finance: The Islamic Finance Alternative. Wiley, February 2012. Mirakhor, A. (2012). Islamic Finance, Risk Sharing and Macroeconomic Policies. MPRA Paper No. 56338 posted 31. May 2014. Mirakhor, Abbas. (2014). Regulatory Framework for Islamic Finance. Islamic Banker Asia. (11), pp. 50–53. Mohamad, A. A. S., Mohamad, M. T. & Samsudin, M. L. (2013). How Islamic banks of Malaysia managing liquidity? An emphasis on confronting economic cycles. International Journal of Business and Social Science, 4, 7. June 2013. Mohammad, S. (2015). Liquidity Risk Management in Islamic Banks: A Survey. Afro Eurasian Studies, Vol. 1, Issue 2, Spring & Fall 2013, pp. 215–230. Mongid, A. (2015). Liquidity Risk Management in the Islamic Rural Banking: Evidence from Indonesia. Banks and Bank systems, Vol.10, Issue 3, pp. 61–67. Montes, C.P., Artigas, C.T., Cristofoli, M. E. & Segundo, N.L.S. (2016). The impact of the IRB approach on the risk weights of European banks. Journal of Financial Stability, 27 November 2016. Morgan, P.J. & Pontines, V. (2013). An Asian perspective on global financial reforms. ADBI Working Paper Series, No. 433, August 2013. Mousavi, Z. & Mousavinia, S. R. (2014). Information transparency and its role in capital market. EuroAsian Journal of Economics and Finance, vol. 2. Issue: 1. (January 2014), pp.66–71. Munteanu, I. (2012). Bank liquidity and its determinants in Romania. Procedia Economics and Finance, Elsevier, 3 (2012), pp. 993–998. Obiyathulla, I. P. (2008). The Islamic inter bank money market and a dual banking system: the Malaysian experience. International Journal of Islamic and Middle Eastern Finance and Management, 1, (3), pp. 210–226. Ogilo, F. & Mugenyah, L. O. (2015). Determinants of Liquidity Risk of Commercial Banks in Kenya. The International Journal Of Business & Management (ISSN 2321 –8916). Vol. 3, Issue 9. September 2015. Quijano, M. (2014). Information asymmetry in US banks and the 2009 bank stress test. Economics Letters, 123 (2014) 203. Pagratis, S. Topaloglou, N. & Tsionas, M. (2016). System Stress Testing of Bank Liquidity Risk. Journal of International Money and Finance, doi: http://dx.doi.org/10.1016/j.jimonfin.2017.02.001 Pesaran, M. H., Shin, Y. Smith & R. J. (2001). Bounds Testing Approaches to the Analysis of Level Relationships. Journal of Applied Econometrics, Volume16, Issue:3, Special Issue: In Memory of John Denis Sargan 1924–1996: Studies in Empirical Macroeconometrics. Petrella, G. & Resti, A. (2013). Supervisors as information producers: Do stress tests reduce bank opaqueness? Journal of Banking and Finance, 37 (2013), pp. 5406–5420. Poorman, F. & Blake, J. (2005). Measuring and modelling liquidity risk: New ideas and metrics. Financial Managers Society Inc., Working Paper. Praet, P. & Herzberg, V. (2008). Market liquidity and banking liquidity: linkages, vulnerabilities and the role of disclosure. Banque de France, Financial Stability Review, Special Issue on Liquidity, 11. February 2008. Rashid, A. & Jabeen, S. (2016). Analysing Performance Determinants: Conventional versus Islamic Banks in Pakistan. Borsa Istanbul Review. Volume 16, Issue 2, June 2016, pp. 92–107. Roman, A., & Sargu, A. C. (2015). The Impact of bank-specific factors on the commercial banks liquidity: Empirical evidence from CEE countries. Procedia Economics and Finance, 20 (15), pp. 571–579. Rosenthal, R. (2016). Emerging Market Outflows Raise Specter of Taper Tantrum Redux. The Wall Street Journal. December 1, 2016. Rosly, S.A. & Bakar, M.A.A. (2003). Performance of Islamic and mainstream banks in Malaysia. International Journal of Social Economics, Vol. 30 Iss 12 pp. 1249–1265.

178

References

Rosman, R., Wahab, N. A. & Zainol, Z. (2014). Efficiency of Islamic banks during the financial crisis: An analysis of Middle Eastern and Asian countries. Pacific-Basin Finance Journal, 28 (2014) pp. 76–90. Saeed, M. & Izzeldin, M. (2014). Examining the relationship between default risk and efficiency in Islamic and conventional banks. Journal of Economic Behavior and Organization. Schich, S. (2008). Financial Turbulence: Some Lessons. Financial Market Trends. OECD web site. www.oecd.org/finance/insurance/41420525.pdf Schmieder, C., Hesse, H., Neudorfer, B., Puhr, C. & Schmitz, S. (2012). Next Generation SystemWide Liquidity Stress Testing. IMF Working Paper, WP/12/3. Siddiqi, S. H. (2001). Islamic Banking: True Modes of Financing. New Horizon, 109 (May-June). Song, I. & Oosthuizen, C. (2014). Islamic Banking Regulation and Supervision: Survey Results and Challenges. IMF Working Paper, WP/14/220. Sims, C. A. (1980). Macroeconomic and reality. Econometrica. Vol.48. No:1 (January 1980). Slovik, P. & Cournède, B. (2011). Macroeconomic impact of Basel III. OECD working paper, 844, 14 February, 2011. Soylu, A. & Durmaz, N. (2013). Profitability of interest‐free versus interest‐based banks in Turkey. The Australian Economic Review, 46, 2, pp. 176–88. Sutorova, B. & Teply, P. (2013). The Impact of Basel III on Lending Rates of EU Banks. Czech Journal of Economics and Finance (Finance a uver), Charles University Prague, Faculty of Social Sciences, Vol. 63(3), p.226–243, July 2013. Stiglitz, J. E. (2000). Capital Market Liberalization, Economic Growth, and Instability. World Development. Vol. 28, No: 6, pp.1075–1086. Thomson Reuters, IRTI, CIBAFI, A. (2014). Turkey Islamic Finance Report 2014: Fundamentals and the Promise of Growth. Trad, N., Trabelsi, M. A. & Goux, J. F. (2017). Risk and profitability of Islamic banks: A religious deception or an alternative solution? European Research on Management and Business Economics 23 (2017), pp. 40–45. Turkey Country Report, (2016). Turkey Islamic Finance Report. www.concec.org. Yildirim, B. D. (2009). The Assessment and analyze of market liquidity. Specialization Thesis. Ankara: Central Bank of the Republic of Turkey. Valla, N. & Saes-Escorbiac, B. (2006). Bank liquidity and financial stability. Financial Stability Review, Banque de France (2006), pp. 89–104 Van den End, J.W. (2012). Liquidity stress-tester: do Basel III and unconventional monetary policy work? Applied Financial Economics, Taylor & Francis Journals, vol. 22(15), August 2012, pp. 1233–1257. Vazquez, F. & Federico, P. (2015). Bank funding structures and risk: Evidence from the global financial crisis. Journal of Banking and Finance, 61, pp. 1–14. Verhoef, B., Azahaf, S. & Bijkerk, W. (2008). Islamic Finance and Supervision: an Exploratory Analysis. DNB Occational Studies Vol.6/No.3 (Amsterdam, de Netherlandsche Bank). Vodova, P. (2011). Liquidity of Czech commercial banks and its determinants. International Journal of Mathematical Models and Methods in Applied Sciences, 5(6), pp. 1060–1067. Vodova, P. (2013). Determinants of Commercial Bank Liquidity in Hungary. Financial Internet Quarterly, e-Finanse, vol. 9, N.3. November 2013. Wagner, W. (2007). The liquidity of bank assets and banking stability. Journal of Banking & Finance, Volume 31, Issue 1, January 2007, pp. 121–139. Williams, R. (2015). Panel Data: Very Brief Overview. University of Notre Dame, http://www3.nd. edu/~rwilliam/ Wong, J., Wong, T. C. & Leung, P. (2010). Predicting banking distress in the EMEAP economies. Journal of Financial Stability, pp:169–179. Zangari, P. (1996). Statistical and Probability Foundations. J. P. Morgan, Risk Metrics Technical Document, Fourth Edition, 1996.

Appendices Appendix A: PBs’ Selected Items .. (Million TL)

Total Assets

Cash+Banks+ Securities

Credits

Deposits

Equity

Net Profit

Albaraka

,

,

,

,

,



Bank Asya

,

,

,

,

,

–

Kuveyt Turk

,

,

,

,

,



Turkey Finance

,

,

,

,

,



Source: Annual Reports, 2014.

Appendix B: PBs’ Return on Equity ROA 3.00% 2.50% 2.00% 1.50% 1.00% 0.50%

https://doi.org/10.1515/9783110582901-010

01 –1 5

01 –1 4

01 –1 3

01 –1 2

01 –1 1

9 01 –1 0

8

–0 01

–0

7 01

6

–0 01

–0 01

01

–0 5

0.00%

180

Appendices

ROE 30.00% 25.00% 20.00% 15.00% 10.00% 5.00%

8 01 –0 9 01 –1 0 01 –1 1 01 –1 2 01 –1 3 01 –1 4 01 –1 5

01 –0

–0 7

6

01

01 –0

01 –0

5

0.00%

Source: CBRT.

Appendix C: PBs’ Capital Adequacy Ratio (CAR) CAR 18.00 16.00 14.00 12.00 10.00 8.00 6.00 4.00 2.00

Source: BRSA of Turkey, CBRT EVDS.

01 –1 4 01 –1 5

13 01 –

1 01 –1 2

–1 01

–1 0 01

09 01 –

08 01 –

–0 7 01

–0 6 01

01 –

05

0.00

,

,

−,

,

,

,

−,

−,

−,

−,

,

,

,

,

,

,

−,

,

−,

,

,

LINF

LCRE

LDEP

RR

BOND

INTER

CAR

CDS

ROE

SUKUK

LINF

LLCR

LLCR

−,

−,

,

−,

−,

−,

,

,

,

LCRE

,

,

,

−,

−,

−,

,

,

LDEP

,

−,

−,

−,

−,

−,

,

RR

,

−,

,

−,

,

,

BOND

Appendix D: Correlation of PBs (For Time-Series Analysis)

,

,

,

−,

,

INTER

,

,

−,

,

CAR

,

−,

,

CDS

−,

,

ROE

,

SUKUK

Appendix D: Correlation of PBs (For Time-Series Analysis)

181

,

−,

−,

,

,

,

,

,

−,

,

,

−,

−,

,

,

−,

−,

−,

−,

,

,

LINF

LCRE

LDEP

RR

BOND

INTER

CAR

CDS

ROE

SEC

LINF

LLCRE

LLCR

−,

,

,

−,

,

,

,

,

,

LCRE

−,

,

,

−,

,

,

,

,

LDEP

−,

,

,

−,

−,

−,

,

RR

−,

−,

,

−,

,

,

BOND

Appendix E: Correlation of CBs (For Time-Series Analysis)

−,

,

,

−,

,

INTER

,

,

−,

,

CAR

−,

−,

,

CDS

−,

,

ROE

,

SEC

182 Appendices

LINF . . . . . . . . . ,. . 

LCR

.

.

.

.

.

−.

.

.

.

.

.



Mean

Median

Maximum

Minimum

Std. Dev.

Skewness

Kurtosis

Jarque-Bera

Probability

Sum

Sum Sq. Dev.

Observations



.

,.

.

.

.

−.

.

.

.

.

.

LCRE

LDEP



.

,.

.

.

.

−.

.

.

.

.

.

Appendix F: Descriptive Statistics of PBs (For Time-Series Analysis)



,,

,.

.

.

.

.

.

.

,.

.

.

RR

(continued)



,.

,.

.

.

.

.

.

.

.

.

.

BOND Appendix F: Descriptive Statistics of PBs (For Time-Series Analysis)

183

CAR . . . . . −. . . . ,. . 

INTER

.

. .

.

.

.

.

.

.

,.

,.



Mean

Median Maximum

Minimum

Std. Dev.

Skewness

Kurtosis

Jarque-Bera

Probability

Sum

Sum Sq. Dev.

Observations

(continued )



,.

,.

.

.

.

.

.

.

. .

.

CDS



,.

.

.

.

.

.

.

.

. .

.

ROE



.E+

,.

.

.

.

.

,.

.

,. ,.

,.

SUKUK

184 Appendices

. . . . . . . . . ,. .

.

.

.

.

.

.

.

.

.

.

.



Mean

Median

Maximum

Minimum

Std. Dev.

Skewness

Kurtosis

Jarque-Bera

Probability

Sum

Sum Sq. Dev.

Observations



LINF

LCR



.

.

.

.

.

−.

.

.

.

.

.

LCRE



.

,.

.

.

.

−.

.

.

.

.

.

LDEP

Appendix G: Descriptive Statistics of CBs (For Time-Series Analysis)



.E+

,,.

.

.

.

.

,.

,.

,.

.

,.

RR

(continued)



,.

,.

.

.

.

.

.

.

.

.

.

BOND Appendix G: Descriptive Statistics of CBs (For Time-Series Analysis)

185

. . . . . . . . . ,. .

.

. .

.

.

.

.

.

.

,.

,.



Mean

Median Maximum

Minimum

Std. Dev.

Skewness

Kurtosis

Jarque-Bera

Probability

Sum

Sum Sq. Dev.

Observations



CAR

INTER

(continued )



,.

,.

.

.

.

.

.

.

. .

.

CDS



,.

.

.

.

.

.

.

.

. .

.

ROE



.E+

,,

.

.

.

−.

.

.

. .

.

SEC

186 Appendices

.

.

−.

.

−.

−.

−.

.

.

.

.

.

−.

.

.

−.

−.

−.

.

−.

.

RR

ROE

CAR

LINF

CDS

INTER

BOND

LCRE

LDEP

SAFS

RR

LCR

LCR

.

.

.

−.

−.

−.

−.

−.

.

ROE

−.

−.

−.

.

.

−.

−.

.

CAR

.

.

.

−.

−.

−.

.

LINF

−.

−.

−.

.

.

.

CDS

Panel A: All Banks

−.

−.

−.

.

.

INTER

−.

−.

−.

.

BOND

.

.

.

LCRE

.

.

LDEP

(continued)

.

SAFS

This table, Pairwise Correlation Matrix, show the coefficient correlations between dependent and independent variables for the overall pooled all banks, CBs and PBs respectively.

Appendix I: Pair-Wise Correlation Matrix− All Banks for Panel Data Analysis (Selected Variables) Appendix I: Pair-Wise Correlation Matrix− All Banks for Panel Data Analysis

187

. .

−.

.

−.

−.

−.

.

.

.

.

. −.

.

.

−.

−.

−.

.

.

.

RR ROE

CAR

LINF

CDS

INTER

BOND

LCRE

LDEP

SAFS

RR

LCR

LCR

(continued )

.

.

.

−.

−.

−.

−.

−.

.

ROE

−.

−.

−.

.

.

−.

−.

.

CAR

.

.

.

−.

−.

−.

.

LINF

−.

−.

−.

.

.

.

CDS

Panel B: CBs

−.

−.

−.

.

.

INTER

−.

−.

−.

.

BOND

.

.

.

LCRE

.

.

LDEP

.

SAFS

188 Appendices

.

−.

−.

.

−.

−.

−.

.

.

.

.

.

−.

−.

.

−.

−.

−.

.

.

.

RR

ROE

CAR

LINF

CDS

INTER

BOND

LCRE

LDEP

SAFS

RR

LCR

LCR

−.

−.

−.

.

.

.

−.

−.

.

ROE

−.

−.

−.

.

.

−.

−.

.

CAR

.

.

.

−.

−.

−.

.

LINF

−.

−.

−.

.

.

.

CDS

Panel C: PBs

−.

−.

−.

.

.

INTER

−.

−.

−.

.

BOND

.

.

.

LCRE

.

.

LDEP

.

SAFS Appendix I: Pair-Wise Correlation Matrix− All Banks for Panel Data Analysis

189

,

.

−.

.

.

.

.

.

.

.

.

.

.

,

.

.

.

.

.

.

.

.

.

−.

.

.

−.

sta

hqla

car

cash

safs

rr

lata

lcre

las

ldep

cap

np

roe

sta

llcr

llcr

.

.

.

.

.

.

.

.

.

.

−.

,

hqla

−.

−.

−.

−.

−.

−.

.

−.

−.

−.

,

car

Panel A: All Banks and all Variables.

(all Variables)

.

.

.

.

.

.

.

.

.

,

cash

.

.

.

.

.

.

.

.

,

safs

.

.

.

.

.

.

.

,

rr

.

.

.

.

.

.

,

lata

.

.

.

.

.

,

lcre

.

.

.

.

,

las

.

.

.

,

ldep

.

.

,

cap

.

,

np

,

roe

roa

Appendix J: Pair-Wise Correlation Matrix− All Banks for Panel Data Analysis

cds

linf

inter

bond

nco

190 Appendices

.

−.

.

−.

−.

.

−.

−.

.

−.

−.

.

roa

cds

linf

inter

bond

nco

.

−.

−.

.

−.

.

−.

.

.

−.

−.

.

.

−.

−.

.

−.

.

.

−.

−.

.

−.

.

.

−.

−.

.

−.

−.

.

−.

−.

.

−.

.

.

−.

−.

.

−.

.

.

−.

−.

.

−.

.

.

−.

−.

.

−.

.

.

−.

−.

.

−.

.

.

−.

−.

.

−.

.

.

−.

−.

−.

−.

.

.

.

.

−.

−.

,

−.

.

.

−.

,

.

−.

−.

,

−.

.

,

−.

, ,

Appendix J: Pair-Wise Correlation Matrix− All Banks for Panel Data Analysis

191

,

.

−.

.

.

.

.

.

.

.

.

.

−.

−.

−.

.

,

.

.

−.

.

.

.

.

.

.

.

.

.

−.

−.

−.

.

sta

hqla

car

cash

safs

rr

lata

lcre

las

ldep

cap

np

roe

roa

cds

linf

sta

llcr

llcr

.

−.

−.

−.

.

.

.

.

.

.

.

.

.

−.

,

hqla

car

−.

−.

−.

−.

−.

−.

−.

−.

−.

−.

−.

−.

−.

,

Panel B: PBS and all Variables.

.

−.

−.

−.

.

.

.

.

.

.

.

.

,

cash

.

−.

−.

−.

.

.

.

.

.

.

.

,

safs

.

−.

−.

−.

.

.

.

.

.

.

,

rr

.

−.

−.

−.

.

.

.

.

.

,

lata

.

−.

−.

−.

.

.

.

.

,

lcre

.

−.

−.

−.

.

.

.

,

las

.

−.

−.

−.

.

.

,

ldep

.

−.

−.

−.

.

,

cap

.

−.

.

.

,

np

−.

.

.

,

roe

−.

.

,

roa

−.

,

cds

,

linf

inter

bond

nco

192 Appendices

−.

−.

.

−.

−.

.

inter

bond

nco

.

−.

−.

−.

.

.

.

−.

−.

.

−.

−.

.

−.

−.

.

−.

−.

.

−.

−.

.

−.

−.

.

−.

−.

.

−.

−.

.

−.

−.

−.

.

.

−.

.

.

−.

.

.

.

−.

−.

−.

.

,

−.

, ,

Appendix J: Pair-Wise Correlation Matrix− All Banks for Panel Data Analysis

193

,

.

−.

.

.

.

.

.

.

.

.

.

.

.

−.

.

,

−.

−.

−.

.

−.

.

.

−.

−.

−.

.

−.

−.

−.

−.

.

sta

hqla

car

cash

safs

rr

lata

lcre

las

ldep

cap

np

roe

roa

cds

linf

sta

llcr

llcr

.

−.

.

.

.

.

.

.

.

.

.

.

.

−.

,

hqla

car

−.

−.

.

−.

−.

−.

−.

−.

−.

−.

−.

−.

−.

,

Panel C: CBS and all Variables.

.

−.

.

.

.

.

.

.

.

.

.

.

,

cash

.

−.

.

.

.

.

.

.

.

.

.

,

safs

.

−.

.

.

.

.

.

.

.

.

,

rr

.

−.

.

.

.

.

.

.

.

,

lata

.

−.

.

.

.

.

.

.

,

lcre

.

−.

.

.

.

.

.

,

las

.

−.

.

.

.

.

,

ldep

.

−.

.

.

.

,

cap

.

−.

.

.

,

np

−.

−.

.

,

roe

−.

−.

,

roa

−.

,

cds

,

linf

inter

bond

nco

194 Appendices

−.

−.

.

−.

−.

−.

inter

bond

nco

.

−.

−.

−.

.

.

.

−.

−.

.

−.

−.

.

−.

−.

.

−.

−.

.

−.

−.

.

−.

−.

.

−.

−.

.

−.

−.

.

−.

−.

.

−.

−.

.

.

.

−.

.

.

.

−.

−.

−.

.

,

−.

, ,

Appendix J: Pair-Wise Correlation Matrix− All Banks for Panel Data Analysis

195

196

Appendices

Appendix K: Estimator Selection – All Banks & Pbs & Cbs Hausman test shows the consistency of an estimator when compared to an alternative. In this test fixed-effect is compared to random-effect for all banks, PBs and CBs. The results show that fixedeffect is preferable for all banks and PBs, whereas random-effect is more efficient for CBs. Heteroscedasticity is defined as the violation of homoscedasticity in variables. It is assumed when the size of the error term differs across values of an independent variable then homoscedasticity is rejected and heteroscedasticity is accepted. In these tests, homoscedastiticity is rejected for all banks, PBs and CBs. This means heteroscedasticity is true for variables. ALL BANKS RE Hausman Heteroskedasticity

PBs FE . .

RE

CBs FE

RE

. .

. .

FE

Appendix L: Summary of Variables (PBs, Bank-Level Data)

197

Appendix L: Summary of Variables (PBs, Bank-Level Data) The following table shows a summary of variables for participation banks (bank-level data). These variables are used for panel data analysis in Chapter 5 (1.000, TRY).

Variable

Obs

Mean

Std.Dev.

Min

Max

STA



,,

,,

,

,,

HQLA



,,

,,

,

,,

LCR (%)











CAR (%)



.

.

.

.

CASH



,

,

,

,,

SAFS



,

,



,,

RR



,,

,,



,,

LATA (%)





.

.



CRE



,,

,,

,,

,,

AS



,,

,,

,,

,,

DEP



,,

,,

,,

,,

CAP



,

,

,

,,

NP



,

,

−,

,

ROE (%) ROA (%)

 

. .

. .

−. −.

. .

CDS (bps)



,

,

,

,

INF (Index)



.

.

.

.

INTER (%)



.

.

.

.

BOND (%)



.

.

.

.

NCO



,.

,,

,

,,

LLCR



−.

.

−.

.

LAS



.

.

.

.

LINF



.

.

.

.

LCRE



.

.

.

.

LDEP



.

.

.

.

198

Appendices

Appendix M: Summary of Variables (CBs, Bank-Level Data) The following table shows a summary of variables for conventional banks (bank-level data). These variables are used for panel data analysis in Chapter 5 (1.000, TRY).

Variable

Obs

Mean

Std.Dev.

Min

Max

STA

.

,,

,,

,

,,

HQLA

.

,,

,,

,

,,

LCR (%)

.



.





CAR (%)

.

.

.

.

.

CASH

.

,

,



,,

SAFS

.

,,

,,



,,

RR

.

,,

,,

,

,,

LATA (%)

.









CRE

.

,,

,

,,

,,

AS

.

,,

,,

,

,,

DEP

.

,,

,,

,

,,

CAP

.

,,

,,

,

,,

NP

.

,,

,,

−,

,,

ROE (%) ROA (%)

. .

. .

. .

−. −.

. .

CDS (Bps)

.

.

.

.

.

INF (Index)

.

.

.

.

.

INTER (%)

.

.

.

.

.

BOND (%) NCO

. .

. ,,

. ,,

. ,

. ,,

LLCR

.

.

.

−.

.

LAS

.

.

.

.

.

LINF

.

.

.

.

.

LCRE

.

.

.

.

.

LDEP

.

.

.

.

.

Appendix N: Mean and Variance Test

199

Appendix N: Mean and Variance Test The following table provides an overview of the mean and variance test of all variables (dependent and independent) for all banks, PBs and CBs. The z-test results show whether both sample means are significantly different from each other. The test statistic is assumed to have normal distribution. Referring to Table 5.3, the normality test suggests and concludes that the residuals are close to a normal distribution. x1 − x2 Z = pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi σx12 + σx22 Z-values > 1.96 represent a significant difference, whereas Z-values < 1.96 represent insignificance. All means (x) and their respective standard errors are retrieved from Table 5.3.

Variable

All Banks

PBs

CBs

RR

.

.

.

ROE

.

.

.

CDS

.

.

.

INTER

.

.

.

BOND

.

.

.

CRE

.

.

.

SAFS

.

.

.

DEP

.

.

.

STA HQLA

. .

. .

. .

CASH

.

.

.

LATA

.

.

.

CAP

.

.

.

NP

.

.

.

ROA

.

.

.

NCO

.

.

.

−.*** (−.) .*** (.)

−. (−.)

. (.)

. (.)

. (.)

−. (−.)

.*** (.)

.*** (.)

. (.)

−. *** −.*** (−.) (−.)

.*** (.)

−. (−.)

. *** (.)

−. (−.)

.*** (.)

.*** (.)

ROE

CAR

LINF

CDS

INTER

BOND

LCRE

SAFS

.*** (.)

−.** (−.)

−. (−.)

. (.)

.* (.)

−. (−.)

.*** (.)

.*** (.)

.*** (.)

RR

()

()

()

.*** (.)

−.** (−.)

−. (−.)

.*** (.)

−.*** (−.)

. (.)

.* (.)

. (.)

.*** (.)

()

−. (−.)

.*** (.)

−.** (−.)

. (.)

.* (.)

. (.)

.*** (.)

()

.*** (.)

.*** (.)

−.** * −.** (−.) (−.)

−. (−.)

.*** (.)

−** (−.)

. (.)

.* (.)

. (.)

.*** (.)

()

All banks: Dependent variable LRR

.*** (.)

−.** (−.)

−. (−.)

.*** (.)

−.*** (−.)

. (.)

.** (.)

. (.)

.*** (.)

()

Appendix O: Random-Effect GLS Regression (all Banks – PBs and CBs)

.*** (.)

−.* (−.)

−. (−.)

.*** (.)

−.*** (−.)

. (.)

.*** (.)

−.* (−.)

.*** (.)

()

−.*** (−.)

.*** (.)

−. (−.)

. (.)

−. (−.)

. (.)

−. (−.)

−.*** (−.)

.*** (.)

()

−.*** (−.)

.*** (.)

−. (−.)

. (.)

−. (−.)

. (.)

−. (−.)

−.*** (−.)

.*** (.)

()

200 Appendices

CONS

NCO

ROA

NP

CAP

LAS

LATA

CASH

HQLA

STA

LDEP

−.*** (−.)

−. ** (−.)

−.*** (−.)

.*** (.)

−.*** (−.)

−.*** (−.)

.*** (.)

−.*** (−.)

−.*** (−.)

−.*** (−.)

−. (−.)

.*** (.)

−.*** (−.)

−.*** (−.)

−.*** (−.)

−. (−.)

.*** (.)

−.*** (−.)

−. (−.)

.*** (.)

−.*** (−.)

−.*** (−.)

. (.)

−. (−.)

−.*** (−.)

. (.)

−. (−.)

. (.)

.*** (.)

−.*** (−.)

−.*** (−.)

(continued)

−. (−.)

. (.)

. (.)

.*** (.)

−.*** (−.)

.*** (.)

.*** (.)

.*** (.) .*** (.)

−.*** (−.)

. (.)

.*** (.)

−.*** (−.)

−.*** (−.)

. (.)

.*** (.)

−.*** (−.)

.*** (.)

−.*** (−.)

.*** (.)

−.*** (−.)

−.*** (−.)

−. (−.)

.*** (.)

−.*** (−.)

.*** (.)

.*** (.) −.*** (−.)

−.*** (−.)

−.*** (−.)

−.*** (−.)

−.*** (−.)

−.*** (−.)

−.*** (−.)

Appendix O: Random-Effect GLS Regression (all Banks – PBs and CBs)

201

.

.

.

.

Theta

Overall R-sq

.

.



.

.

()

.

.



.

.

()

.

.



.

.

()

.

.



.

.

()

All banks: Dependent variable LRR

.

.



.

.

()

.

.



.

.

()

.

.



.

.

()

.

.



.

.

()

Notes: The dependant variable is LLR which represents the banks’ liquidity requirement ratio calculated according to BIS (2013c). 1) *, **, *** indicate respectively 10%, 5% and 1% significance level. The figures in parenthesis are the standard errors. 2) Specification (2) adds an indicator variable LAS (logarithm of total assets). Specification (3) adds an indicator variable CASH (total cash) for robustness check. Specification (4) adds an indicator variable NP (net profit). Specification (5) adds a variable STA (short-term assets). Specification (6) adds a variable HQLA (high-quality liquid assets). Specification (7) adds a variable ROA (return on assets). Specification (8) adds a variable CAP (total shareholders’ equity). Specification (9) adds a variable LATA (short-term assets to total assets ratio). Specification (10) adds a variable NCO (net cash outflows).





Obs.

.

.

.

Prob>chi .

Wald chi(n)

()

()

(continued )

202 Appendices

SAFS

LCRE

BOND

INTER

CDS

LINF

CAR

ROE

RR

()

.***

(.)

−.

(−.)

.***

(.)

.***

(.)

.

(.)

.**

(.)

−.**

(−.)

−.***

(−.)

−.***

(−.)

()

.***

(.)

−.

(−.)

.***

(.)

.***

(.)

.

(.)

.**

(.)

−.**

(−.)

−.***

(−.)

−.***

(−.)

(−.)

−.***

(−.)

−.

(−.)

−.

(.)

.*

(−.)

−.

(.)

.***

(.)

.***

. (.)

(.)

.***

()

(−.)

(−.)

(−.)

(−.)

−.***

−.***

−.***

−.

−.***

(−.)

(−.)

(−.)

−.***

(−.)

−.

−.

−. (−.)

.***

(−.)

−.

(.)

.***

(.)

.**

(−.)

−.

(.)

.***

()

(.)

.**

(−.)

−.

(.)

.***

(.)

.**

(−.)

−.

(.)

.***

()

(.)

. (.)

(−.)

−.

(.)

.***

(.)

.***

. (.)

(.)

.***

()

PBs: Dependent variable LRR

Appendix P: Random-Effect GLS Regression (PBs)

(−.)

−.***

(−.)

−.***

(−.)

−.

(.)

.***

(−.)

−.

(.)

.***

(.)

.**

. (.)

(.)

.***

()

(−.)

−.***

(−.)

−.***

(−.)

−.

(.)

.**

(−.)

−.

(.)

.***

(.)

.**

(−.)

−.***

(.)

.***

()

(−.)

−.***

(−.)

−.***

(−.)

−.

(.)

.**

(−.)

−.

(.)

.***

(.)

.

(−.)

−.***

(.)

.***

()

(continued)

(−.)

−.***

(.)

.***

(−.)

−.***

(.)

.

(−.)

−.***

(.)

.***

(.)

.***

(.)

.

(−.)

−.

()

Appendix P: Random-Effect GLS Regression (PBs)

203

NCO

ROA

NP

CAP

LAS

LATA

CASH

HQLA

STA

LDEP

(continued )

(.)

(.)

. (.)

(−.)

(.)

.**

(.)

(.)

.

(.)

.

(.)

.***

.***

.***

.***

(−.)

(−.)

(−.)

(.)

(.)

(−.)

(.) . (.)

−. (−.)

(.)

.

(.)

.*

−.*** (−.)

−.***

.**

(.)

(−.)

−.***

(−.)

−.

(.)

.

(.)

.

(−.)

−.

.***

. ***

.* *

(−.)

−.***

(.)

.***

(.)

.*

(−.)

−.***

()

(.)

.

(.)

.***

(−.)

−.***

(.)

.*

()

(.)

.***

(.)

.***

(.)

.***

(−.)

−.***

(.)

.

()

(.)

.

(.)

.

(.)

.***

−.***

−.***

−.***

(.)

.***

()

(.)

.***

()

(.)

.**

()

−.

(.)

.

(.)

(.)

.***

(.)

(.)

.

()

.***

(.)

.

.

(.)

(.)

(.)

.

.***

()

.

()

()

PBs: Dependent variable LRR

204 Appendices

(−.)

. .

(−.)

. .

.

.

Overall R-sq

.





. .

(−.)

–.***

.





. .

(−.)

−.***

.





. .

(−.)

−.***

.





. .

(−.)

−.***

.





. .

(−.)

−.***

.





.



. .

(−.)

−(.) . .

−.***

−.****

.



. .

(−.)

−.***

Notes: 1) The dependent variable is LLR which represents the banks’ liquidity requirement ratio calculated according to BIS (2013c). 2) *, **, *** indicate respectively 10%, 5% and 1% significance level. The figures in parenthesis are the standard errors. Specification (2) adds an indicator variable LAS (logarithm of total assets). Specification (3) adds an indicator variable CASH (total cash). Specification (4) adds an indicator variable NP (net profit). Specification (5) adds a variable STA (short-term assets). Specification (6) adds a variable HQLA (high-quality liquid assets). Specification (7) adds a variable ROA (return on assets). Specification (8) adds a variable CAP (total shareholders’ equity). Specification (9) adds a variable LATA (short-term assets to total assets ratio). Specification (10) adds a variable NCO (net cash outflows).





Theta



−.***

−.***

Observations 

Wald chi(n) Prob>chi

CONS

Appendix P: Random-Effect GLS Regression (PBs)

205

.*** (.)

. (.). (.). (.) . (.)

−.*** −.*** −.*** −.*** −.*** −.** (−.) (−.) (−.) (−.) (−.) (−.)

−. (−.) −.* (−.)

.*** (.)

. (.). (.)−. (−.)

.*** (.)

LINF

CDS

INTER

BOND

LCRE

−. (−.)

.*** (.)

−.** (−.)

. (.). (.) . (.)

.*** (.)

−.* (−.)

.*** (.) .*** (.)

.*** (.)

CAR

.*** (.)

.*** (.)

()

.*** (.)

()

−. (−.)

−. (−.)

−.*** −.*** (−.) (−.)

.*** (.)

()

−. (−.)

−. (−.)

−. (−.) −.* (−.)

.*** (.)

.*** (.)

−. (−.)

−. (−.)

−. (−.)

−. (−.)

.*** (.)

−.** (−.)

−. (−.)

.*** (.)

.*** (.)

. (.) . (.)

. (.) . (.)

−. (−.)

. (.) . (.) . (.) . (.) . (.) . (.)

. (.) . (.) . (.) .** (.)

. (.) . (.) . (.) −.** (−.)

.*** (.)

. (.). (.) . (.)

.*** (.)

−. (−.)

()

ROE

.*** (.)

()

.*** (.)

()

.*** (.)

()

RR

()

()

()

CBs: Dependent variable LRR

Appendix R: Fixed-Effect Regressions (CBs)

206 Appendices

NCO

ROA

NP

CAP

LAS

LATA

CASH

HQLA

STA

LDEP

SAFS

.*** (.) −.*** (−.)

.*** (.)

.*** (.) −.*** (−.)

.*** (.) −.*** (−.)

.*** (.) −.*** (−.)

.*** (.)

. (.) −. (−.)

.*** (.) −.*** (−.)

.*** (.) −.*** (−.)

−.*** (−.) −.*** (−.)

.*** (.)

.*** (.)

.** (.)

. (.) . (.) .*** (.)

.*** (.) −.*** (−.)

.* (.)

.*** (.)

−.*** (−.) −.*** (−.)

.*** (.)

−. (−.)

.*** (.)

−. (−.)

.*** (.)

−. (−.)

.*** (.)

.*** (.)

. (.) .** (.)

−.*** .*** (−.) (−.)

.*** (.)

.*** (.)

−.*** (−.)

.*** (.)

.*** (.)

(continued)

.*** (.)

. (.) . (.) . (.) . (.)

−. (−.)

.*** (.)

.*** (.)

−.*** −.*** −.*** −.*** −.*** −.*** −.*** −.*** (−.) (−.) (−.) (−.) (−.) (−.) (−.) (−.)

.*** (.) −.*** (−.)

Appendix R: Fixed-Effect Regressions (CBs)

207

.

.

 .

 .



.

.



.

.



.

.

.



.

.

.



.

.

.



.

.

.



.

.

−. (−.)

()

Notes: 1) The dependent variable is LLR which represents the banks’ liquidity requirement ratio calculated according to BIS (2013c). 2) *, **, *** indicate respectively 10%, 5% and 1% significance level. The figures in parenthesis are the standard errors. 3) Specification (2) adds an indicator variable LAS (logarithm of total assets). Specification (3) adds an indicator variable CASH (total cash). Specification (4) adds an indicator variable NP (net profit). Specification (5) adds a variable STA (short-term assets). Specification (6) adds a variable HQLA (high-quality liquid assets). Specification (7) adds a variable ROA (return on assets). Specification (8) adds a variable CAP (total shareholders’ equity). Specification (9) adds a variable LATA (short-term assets to total assets ratio). Specification (10) adds a variable NCO (net cash outflows).

Within R-sq

Observations 

.

.

.

.

.

.

Prob>F

()

.

()

.

()

F(n,)

()

−.*** −.*** −.*** −.*** −.*** −.*** −.*** −. (−.) (−.) (−.) (−.) (−.) (−.) (−.) (−.)

()

−.* (−.)

()

CONS

()

()

CBs: Dependent variable LRR

()

(continued )

208 Appendices

List of Abbreviations AAOIFI ADF AIC ALA ARDL BCBS BIST BRSA BIS CAR CBs CBRT CCPs CDS CIBAFI CMP CoVAR DLCR D-SIBs EBA EMDEs ECAI EU ESRB FE FSB FSI GDP GLS GSIBs HQLA IAH ICFA IDB IFSA IFSB IILM IIMM IMF INCEIF INIs IRF IRTI ISRA KLIBOR LCR LOLR

Accounting and Auditing Organization for Islamic Financial Institutions Augmented Dickey-Fuller Akaike Information Criterion Alternative Liquidity Approach An Autoregressive Distributed Lag Modelling Approach Basel Committee on Banking Supervision Borsa Istanbul Banking Regulatory and Supervisory Agency Bank for International Settlements Capital Adequacy Ratio Conventional Banks Central Bank of the Republic of Turkey Central Counterparties Credit Default Swaps Council for Islamic Banks and Financial Institutions Commodity Murabahah Programme Conditional VaR Dutch Liquidity Coverage Ratio Domestic Systemically Important Banks European Banking Authority Emerging and Developing Economies External Credit Assessment Institutions European Union European Systemic Risk Board Fixed-Effect Financial Stability Board Financial Sector Indicators Gross Domestic Product Generalized Least Squares Globally Systemic Important Banks High-Quality Liquid Asset Investment Account Holder Implied Cash Flow Analysis Islamic Development Bank Islamic Financial Services Act Islamic Financial Service Board International Islamic Liquidity Management Islamic Interbank Money Market International Monetary Fund International Centre for Education in Islamic Finance Islamic Negotiable Instruments of Deposits Impulse Response Function Islamic Research and Training Institute International Shariah Research Academy Kuala Lumpur Interbank Rate Liquidity Coverage Ratio Lender of Last Resort

https://doi.org/10.1515/9783110582901-011

210

List of Abbreviations

LRM LSDV MDBs NCO NSFR OMO PBs PDP PILLAR 2 PLS PP PSEs RE ROA ROE RWAs SC SNB SPV UK UK FSA VaR VAR VDC VECM

Liquidity Risk Management Least Square Dummy Variable Multilateral Development Banks Net Cash Outflows Net Stable Funding Ratio Open Market Operations Participation Banks Public Disclosure Platform Supervisory Review Process Profit-Loss Sharing Phillip-Perron Public-Sector Entities Random-Effect Return on Assets Return on Equity Risk Weighted Assets Schwarz Criterion Swiss National Bank Special Purpose Vehicle United Kingdom UK Financial Services Authority Value at Risk Vector Auto Regression Vector Decomposition Vector Error Correction Model

List of Figures Figure 4.1 Figure 4.2 Figure 4.3 Figure 4.4 Figure 5.1 Figure 7.1 Figure 8.1 Figure 8.2 Figure 8.3 Figure 8.4 Figure 8.5 Figure 8.6

PBs’ Cash Assets 47 Total Assets, Deposits, Credits and Equity of PBs (2005–2015, US$ Mio) PBs Products Share 48 Liquidity Requirement Ratio (million TL) – Up to 1 Month 51 Impulse Response Results for PBs 75 Response of LCR and Interbank Rate 129 Bank A’s Cash and Net Profit (left) and CAR and LCR ratio (right) 147 Bank B’s Cash and Net Profit (left) and CAR and LCR ratio (right) 148 Bank C’s Cash and Net Profit (left) and CAR and LCR ratio (right) 149 Bank D’s Cash and Net Profit (left) and CAR and LCR ratio (right) 150 Cash to Asset Ratio (PBs in Turkey, Bank Level Data) 151 Changes in Profit Ratio (PBs in Turkey, Bank Level Data) 151

https://doi.org/10.1515/9783110582901-012

48

List of Tables Table 2.1 Table 3.1 Table 3.2 Table 3.3 Table 3.4 Table 4.1 Table 4.2 Table 4.3 Table 4.4 Table 5.1 Table 5.2 Table 5.3 Table 5.4 Table 5.5 Table 5.6 Table 5.7 Table 5.8 Table 6.1 Table 6.2 Table 6.3 Table 6.4 Table 6.5 Table 6.6 Table 6.7 Table 6.8 Table 6.9 Table 6.10 Table 6.11 Table 7.1 Table 7.2 Table 7.3 Table 7.4 Table 7.5 Table 7.6 Table 7.7 Table 7.8 Table 7.9 Table 7.10 Table 8.1 Table 8.2 Table 8.3 Table 8.4

The LCR Implementation Deadline 26 Islamic Banks HQLA Stocks (World) 30 Comparing Selected Liquidity Tools 35 Comparison between Sovereign Sukuk and the IILM Sukuk 35 Comparison between Commodity Murabahah and the IILM Sukuk 37 Total Asset (Thousand TRY), Asset Growth Rate and Share 46 Overview of Turkish Participation Banking Sector (%) 47 Asset of Participation Banks 50 The Utilization of CBRT Facilities by PBs 53 List of Variables 58 Unit Root Test Results 63 Cointegration Test Results 64 ARDL Cointegration Tests 66 Long Run Coefficients 67 VECM Based Granger Results (PBs) 70 VECM Based Granger Results (CBs) 73 Summarizing Results 77 List of Banks included for Panel Data Analysis 82 List of Variables 85 Summary of Variables 87 Panel Data Regressions Results for All Turkish Banks 92 Panel Data Regressions Results for PBs 96 Panel Data Regressions Results for CBs 101 Robustness Check (All banks – PBs and CBs) 107 Robustness Checks (PBs) 111 Robustness Checks (CBs) 114 Liquidity Regulation and Bank’s Financing 117 Comparing Findings with Other Studies 118 High-Quality Liquid Assets for Assumptions 131 Outflows of Funds 132 Wholesale Funding 133 Secured Funding 134 Required Stable Funding 135 Implied Cash Flow Test (5 Days) 136 Results of LCR based Stress Testing 136 Results of LCR based Stress Testing-HQLA Ratios 137 Results of Liquidity Shortfall 137 Net Stable Funding Ratio (NSFR) 138 Correlation of Cash and Profit and Correlation of LCR and CAR 146 Minimum and Maximum Threshold for New LCR Framework 158 New LCR Hair Cut 159 Proposed Balance Sheet for Small Islamic Banks 162

https://doi.org/10.1515/9783110582901-013

Index AAOIFI 50, 160 Accounting 160, 168 Accounting-based 7 Advanced countries 15, 121, 169 Advanced economies 29 Akaike information criterion (AIC) 65 ALA approach 162, 163 Alternative 37–39, 50, 54 Alternative regulatory framework 145, 166 Al-Wadiah 37 Amanah 132 Approach 69, 76, 122 Arbitrage 56, 69, 152, 166 Arbitrary 157 ARDL 57, 76, 143, 165 Asset 2, 3, 5, 8, 9, 12, 14, 18, 21, 22, 152–154, 157, 159, 160, 165, 167 Asset management 17, 32 Asset-backed commercial paper 14, 15, 124 Asset-based 154 Asymmetric information 13, 15, 20, 141 Augmented Dickey-Fuller 57 Bai’ al-inah 36, 39, 41, 42 Balance sheet 4, 5, 8, 12, 16, 24, 27, 31, 45, 46, 106, 120, 126, 127, 130, 135, 143, 152, 157, 160, 162, 167, 169 Bank Negara Malaysia 34, 36, 39,41 Bankruptcy 133 Banks’ resilience 3, 125 Basel Accords 161, 162 Basel Committee (BCBS) 2, 8, 25, 50, 56, 106, 127, 154, 162 Basel III 3, 5, 11–13, 16, 46, 56, 61, 129, 141, 142, 161, 162 Best linear unbiased estimator 62 BIS 1–4, 7, 130 Bloomberg terminal 9, 58, 85 Bond 6, 9, 25, 30, 51, 58, 59, 152, 154, 165 Bounds test 65 BRSA 50, 58, 59 Cagamas 30, 41 Capital 1, 5, 6, 11–13, 15, 16, 18, 31, 44, 45, 49, 50, 58, 68, 75, 78, 83, 84, 90, 95, 99, 104, 110, 119, 123, 126, 128, 135, 139, 141, 145, 152, 154, 158, 160, 161, 163, 165 https://doi.org/10.1515/9783110582901-014

Capital Adequacy Ratio (CAR) 6, 13, 35, 50, 58, 59, 60, 77, 82, 85, 86, 89, 91, 105, 118, 143, 146, 180 Capital conservation buffer 3 Capital market 28, 34, 124 Capital requirement 2, 11, 14–16, 49, 89, 145, 154 Capital standards 3, 12, 14 Cash 3, 16, 17, 25, 26, 30, 32, 33, 37, 41, 46, 55, 60, 82, 83, 87, 88, 90, 93, 95, 100, 104, 106, 110, 117, 119, 123, 126, 134, 135, 142, 143, 145, 146, 151, 152, 159, 160–162, 179 Cash flow 6, 55, 76, 84, 135, 136, 154, 158 Cash inflow 8, 25, 26, 61, 129, 130, 137, 157 Cash outflow 5, 8, 19, 22, 25–27, 31, 45, 58, 59, 83–85, 91, 95, 99, 105, 120, 125, 127, 130, 137–139, 141, 144, 152, 165, 166 Cash to asset ratio 146 Causality 57, 64, 72 CBRT 52, 53, 60, 85 CBRT EVDS 58, 85, 180 CDS 9, 58–60, 65, 67, 68, 75, 77, 84, 124, 165 Central bank 10, 13, 21, 24, 27, 28, 34, 35, 39, 42, 55, 121, 131, 134, 139, 153, 163 Central counterparties 125 Centrally cleared 28 Cholesky 75 Coefficient 117, 128 Cointegration 57, 62, 64–66, 76, 164 Colleteral 153 Commercial banks 12, 13, 32, 43, 90 Commodity Murabahah 53, 126, 156, 161 Commodity-based Sukuk 4, 153 Conventional banks 2, 4, 6, 7, 16, 17, 19, 32, 45, 49, 53, 57, 59, 62, 64, 68, 69, 72, 81, 99, 100, 104, 125, 140 Corporate bond 30, 68 Correlation 119, 121, 143, 146 Cost of funding 4, 59, 126 Countercyclical capital buffer 3 Credit allocations 14 Credit rating 13,16, 159, 166 Credit supply 14 Currency 24, 28, 163 Current account deposits

216

Index

Debt 38, 40, 41, 53, 125, 167 Debt securities 51 Debt trading 39 Debt-based 20 Debt-like instrument 16 Dependent variable 7, 8, 57, 58, 77, 82, 83, 89, 92, 99, 102, 200 Deposit 104–106, 110, 118, 124, 126, 129, 132, 136, 143, 157, 165 Deposit banks 45, 61 Deposit insurance 31, 33, 132, 140 Deposit management 31 Determinant 27, 55, 58, 60 Developing 121, 123, 129, 143 Diagnostic 124, 129 Domestic systemically important banks (D-SIBs) 1, 55 Downside effects 141 Dummy 82, 84, 85, 117, 120 Dutch Central Bank 57 Dynamic Stochastic General Equilibrium (DSGE) 23 Efficiency 2, 3, 9, 12, 17–19, 27, 47, 52, 80, 83, 104, 117, 163 Emerging 13–15, 125, 126 Empirical studies 7, 163 Endogenous 22, 128 Equation 62, 67, 81, 84 Esham 144, 153, 154, 166 Estimator 62, 80, 81, 86, 94, 97, 105, 196 EU banks 15, 126 Exchange-traded 28, 53, 135 Exims 155 Exogenous 64, 135 Externalities 14 Financial friction 16 Financial intermediation 17, 152 Financial liabilities 6 Financial stability 1, 5, 6, 7, 12, 14, 56, 76, 120, 123, 126, 137, 143–145, 152, 166 Financing 76, 78, 91 Finding 7, 12, 13, 68, 76, 89, 95, 99 Fire sales 124 Fixed-effect 9, 105 Foreign banks 33 French banks 15 FSB 1, 11

Functional fence 140 Funding liquidity 22, 128, 135, 157, 164 GDP-linked Sukuk 4, 154 Generalized method of moments (GMM) 168 Global financial crisis 14, 21, 24, 45 Global systemically important banks (G-SIBs) 11 Government investment issue Granger 56, 57, 64, 67, 69, 72, 76, 78, 143, 164, 165 Growth 2, 14, 15, 17, 23, 27, 45, 50, 56, 69, 76, 78, 90, 127, 152, 162, 164 Haircut 130–132, 134, 139, 159 Hausman 80, 105, 118, 196 High-quality liquid assets (HQLA) 119, 163, 167 Homoscedasticity 62, 105, 196 IDB Sukuk 163 Idiosyncratic 13 IFSB 4, 130, 163, 164 IIFS 163 IILM 3, 27, 34, 35 IILM Sukuk 28, 29, 34, 35, 37, 52, 59, 131, 132, 163 Ijarah 48, 50, 126, 154 Illiquid 124, 136 Implied cash flow analysis 9 Implied cash flow test 136 Impulse response 57, 76 Independent variable 59, 60, 65, 70, 73, 74, 80, 82, 187, 196 Indonesian banks 138 Indonesian Islamic banks 17 Inflation 59, 61, 67, 68 Inflation rate 60, 65, 68 Interbank deposit 22, 30, 33 Interbank market rate 164 Interbank money market 53, 165 Intermediation cost 16, 27 Internationally active banks 11, 25, 141, 145, 161 Investment banks 45, 61 Investment deposits 16 Investment reserves 50 Islamic accepted bills 39 Islamic banks 1–3, 5, 8–10, 16–20, 24, 26, 27, 29, 30, 32, 34, 36, 38, 41, 42, 44, 49, 55, 57, 61, 76, 121–126, 132, 139, 140, 143–146, 151, 152, 155, 156, 163

Index

Islamic financial institution 154, 163 Issuance 49, 76, 125, 131 Johansen-Juselies 57, 62, 64 Jurisdiction 2, 11, 20, 28, 34, 53, 55, 131, 153, 154, 156, 163 Lender of last resort 2, 15, 24, 54, 78, 130, 138, 140 Level 2A 131–133, 163 Level 2B 131, 163 Leverage 3, 11, 14, 17 Leverage ratio 47, 49 Liability 125, 127, 129, 162 Liability management 2, 121, 166 Liquid asset 8, 13 Liquidity 55, 57 Liquidity Coverage Ratio (LCR) 3, 138, 144, 156, 162, 164, 168 Liquidity management 119, 121 Liquidity requirement ratio 57, 61, 82, 84 Liquidity risk 57, 90 Liquidity run 117 Logarithm 13, 57 Long-run 1, 5, 16, 143, 164 Long-term 37 Low-yielding assets 142 Macroprudential 3, 21, 123, 125 Macroprudential tools 1, 5, 13, 14, 144 Malaysia 33, 34 Malaysian banks 18 Malaysian Islamic banks 8, 30, 42, 49, 60 Manager 140 Market liquidity 2, 13, 163–165 Maturity 6, 16, 31, 168 Maturity-matching 167 Mean 62, 199 Measure 140 Metric 57, 139 Minimum level 25, 27 Mismatch 15, 121, 164 Modus operandi 122 Monetary Notes 39 Moral hazard 2, 167 Mudarabah 4, 20, 29, 31, 34, 37, 39 Multilateral development banks (MDBs) 25 Multinational investment banks 59, 83 Murabahah 41, 48, 50 Musharakah 4, 32, 48

217

Net Profit 58, 60, 83, 146 Net Stable Funding Ratio (NSFR) 3, 138, 168 Non-performing loans 45, 46, 60, 123, 124, 126 Objectives 1, 5, 10 Observation 62, 65, 84, 88 Open market operations 53, 121 Operational 125, 126 Outflow 8, 23 Pair-wise 7, 187,193 Panel data 79, 80, 82, 86 Participation banks (PBs) 45, 50, 69, 165, 167 Partnership-based 162 Performance 78 Period 136, 154, 164 Perpetual Sukuk 166 Phillip-Perron 57 Playing field 143 Pro-cyclically 122 Profit equalization 50 Profit-and-loss sharing 49 Qualitative 157 Quantitative 7, 9, 11, 21, 144, 166 Quarterly 61, 85, 86, 129 Random-effect 105 Regulation 14, 18, 19, 25, 27, 45, 49, 50, 56, 57, 82, 84, 110, 117, 122, 124, 139, 142, 145 Regulatory 125, 131 Regulatory arbitrage 13, 14 Regulatory authorities 82, 132 Regulatory capital 14, 83 Regulatory framework 123 Required reserves 59, 85 Resilience 1, 3, 9, 17, 20, 24, 25, 55, 123–125, 143, 166 Retail deposits 132, 133 Return on asset 59, 83 Return on equity 59 Risk-sharing 153 Risk-taking 16, 153 Risk-weight 3, 16, 60, 85, 126 Robustness 83, 84, 104, 105 Rule-compliant 166 Safe havens 121 Salam 42, 50

218

Index

Sale and buy-back 38, 125 Saving deposits 16 Schwarz Criterion 57 Secondary markets 164 Sensitivity 136 Serial auto-correlation 105 Shariah-compliant 55, 78, 130, 133, 164 Shariah-compliant securities 28 Shock 57, 72, 75 Shortfall 137, 138, 157 Short-run 165 Short-term 1, 3, 7, 11, 13, 15, 18, 19, 25, 31–38, 42, 52, 53, 56, 58, 62, 69, 72, 165, 167 Sovereign bond 52 Spillover effects 166 Stable deposits 132, 133 Stakeholders 121, 141, 157 Standard 57 Standard deviation 61, 62 Standards 69, 79 Stationarity 57, 62 Stress testing 121, 122, 144 Sudan 33 Sukuk 144, 164, 166 Summary 86, 104, 163 Supervision 10, 12, 49, 55, 122, 139, 141, 152, 155, 163, 166, 168 Supervisory 9, 11, 13, 17, 21, 27, 33, 50, 123, 125, 142, 155, 156, 163 Surveillance 141, 163 Swiss banks 100 Systematic 13

Tier-II capital 58, 85, 152 Tier-II Sukuk 49 Tightening 8, 27, 57 Time-series 164 TKBB Database 58, 85 Total asset 13, 14 Total loans 13, 57 Transparency 15, 145, 154 Treasury bond 59, 60, 85 Turkish banks 12, 68, 86, 89, 90, 92 Turkish Treasury 52, 76, 125, 167

Tawarruq 36, 42, 161 Theorem 16 Tier-I capital 11, 58, 85

Z-test 88, 199 Z-values 199

UK banks 15 Underlying asset 52 Unit root 57 Unsecured 130 US banks 23 Validity 57 VAR analysis 57 Variable 7, 165, 198, 199 Variance test 199 Venture capital 162 Wa’ad-based 52 Wadiah acceptance 37, 162 Wakalah 29 Weighted 45, 80, 81, 137 Wholesale 130, 132 Withdrawal 32, 50, 130 Yield curve 22, 31, 35