Three Empirical Essays on Bank Accounting

Three Empirical Essays on Bank Accounting A Thesis submitted to the University of Manchester for the degree of Doctor of Philosophy In the Faculty of Humanities 2012 CHU YEONG LIM MANCHESTER BUSINESS SCHOOL List of contents ABSTRACT DECLARATION COPYRIGHT STATEMENT ACKNOWLEDGEMENTS INTRODUCTION 12 BANK ACCOUNTING CONSERVATISM AND BANK LENDING BEHAVIOUR 23 2.1 Introduction 24 2.2 Literature Review and Institutional Setting 27 2.2.1 Literature Review 27 2.2.1.1 Accounting conservatism 27 2.2.1.2 Loan loss accounting 31 2.2.1.3 Credit crisis 32 2.2.2 Institutional Setting 33 2.2.2.1 Loan syndication market 33 2.2.2.2 Loan provisioning process 34 2.3 Hypotheses Development 35 2.4 Methodology and Data 38 2.4.1 Methodology 38 2.4.2 Data 44 2.5 Empirical Results 46 2.5.1 Descriptive Statistics 46 2.5.2 Main Results 51 2.6 Conclusion 57 BANK RISK AND THE VALUE RELEVANCE OF FAIR VALUE GAINS AND LOSSES 73 3.1 Introduction 75 3.2 Institutional Setting and Prior Literature 76 3.2.1 The economic nature of banks 76 3.2.2 The demand for fair value information 77 3.2.3 The problems with fair value accounting 79 3.2.4 Accounting standards involving fair value accounting 82 3.2.5 Bank regulatory and corporate governance environment 84 3.2.6 Prior empirical literature 86 3.3 Hypotheses Development 90 3.4 Research Design 95 3.4.1 Methodology 95 3.4.2 Sample Selection 99 3.5 Empirical Results 102 3.5.1 Descriptive statistics and correlations 102 3.5.2 Main results 104 3.5.3 Additional Tests 109 3.5.3.1 Additional controls for size and bank credit default swap rates 109 3.5.3.2 Alternative Risk Proxies 109 3.5.3.3 Additional tests on the alternative possibility 112 3.5.3.4 Further Remarks on the Crisis Period 113 3.6 Conclusion 115 ARE THE LOAN LOSS AND FAIR VALUE COMPONENTS OF BANK INCOME RATIONALLY PRICED? 134 4.1 Introduction 135 4.2 Literature Review 137 4.2.1 Accrual anomaly 138 4.2.2 Bank loan provisions 143 4.2.3 Fair value accounting 146 4.3 Hypotheses Development 148 4.4 Research Design 159 4.4.1 Methodology 159 4.4.2 Sample selection 165 4.5 Empirical Results 166 4.5.1 Descriptive statistics and correlations 166 4.5.2 Main results 170 4.6 Conclusion 183 SUMMARY AND SUGGESTIONS FOR FUTURE RESEARCH 206 Total Word Count: 53,244 Abstract The University of Manchester Chu Yeong Lim Doctor of Philosophy (PhD) Three Empirical Essays on Bank Accounting November 2012 This thesis presents new empirical evidence on three important aspects of financial reporting by banks The thesis consists of an introductory chapter that explains how the three issues are related to each other, three empirical chapters and a final summary chapter The first empirical chapter studies the effects of accounting conservatism on the pricing of syndicated bank loans I provide evidence that banks more timely in loss recognition charge higher spreads for the same loan provision I go on to consider what happens to this relationship during the financial crisis During the crisis, banks more timely in loss recognition increase their spreads to a lesser extent than banks less timely in loss recognition The policy implication is that banks more timely in loss recognition exhibit more prudent and less pro-cyclical debt pricing behaviour The second empirical chapter examines the relationship between the value relevance of fair value gains and losses and bank risk in an international bank sample One possibility is that, as risk increases, the scope for subjectivity in fair value estimates increases thereby potentially rendering the numbers less useful However another possibility is that the relevance of faithfully reported fair value gains and losses increases as risk increases The study provides evidence that the value relevance of fair value gains and losses is positively associated with bank risk prior to the crisis During the crisis there is also evidence of a similar positive relationship, but it is not possible to draw firm conclusions for reasons discussed in the chapter My research also shows that the fair value gains and losses of banks that elect to use the fair value option for assets that could have been accounted for using amortized costs are more value relevant and persistent This study provides information to policy makers on the situations when fair values are most useful to investors The third empirical chapter examines if the market rationally prices the loan loss provisions, and the reported fair value gains and losses of US banks The chapter models the discretionary components of loan loss provisions and fair value gains and losses, and tests if the discretionary components are priced differently from their non-discretionary counterparts The results provide little evidence that the market misprices operating cash flows, non-discretionary loan loss provisions, or fair value gains and losses (discretionary or otherwise) However there is evidence of significant mispricing of discretionary loan loss provisions The lack of evidence on the mispricing of fair value gains and losses is consistent with the finding on the value relevance of fair value gains and losses in the second empirical chapter Declaration No portion of the work referred to in the thesis has been submitted in support of an application for another degree or qualification of this or any other university or other institute of learning Copyright Statement (i) Copyright in text of this thesis rests with the Author Copies (by any process) either in full, or of extracts, may be made only in accordance with instructions given by the Author and lodged in the John Rylands University Library of Manchester Details may be obtained from the Librarian This page must form part of any such copies made Further copies (by any process) of copies made in accordance with such instructions may not be made without the permission (in writing) of the Author (ii) The ownership of any intellectual property rights which may be described in this thesis is vested in The University of Manchester, subject to any prior agreement to the contrary, and may not be made available for use by third parties without the written permission of the University, which will prescribe the terms and conditions of any such agreement (iii) Further information on the conditions under which disclosures and exploitation may take place is available from the Head of Manchester Business School Acknowledgements First, I would like to express my heartfelt gratitude to my supervisors Martin Walker, Edward Lee and Asad Kausar for accepting me to be their student I would like to thank them for their patient guidance and support during the Phd process Their kind guidance and support have been highly persistent throughout the process I have benefited greatly from their wealth of knowledge and expertise in academic research, which they have kindly shared with me The thesis would not have been possible without their sound advice and feedback They have also guided me in the presentations at the various conferences and seminars I also thank them for their kind understanding of my personal situation Besides my supervisors, this thesis has benefited from the feedback and comments of many people when I presented each chapter paper at various conferences and seminars I would like to thank Robert Hodgkinson, Ken Peasnell, Stephen Zeff, Geoff Whittington, Chris Nobes, Stuart McLeay, Geoff Meeks, Alfred Wagonhofer, Andrew Stark, Norman Strong, Richard Taffler, Ser Huang Poon, David Brierwood, Kevin Ow Yong, Atif Ellahie, the 2010 Manchester Business School doctoral conference participants, attendees at the December 2011 UK ASB academic panel meeting, participants at the 2011 Marie Curie mid-term conference in Berlin, the 2011 Marie Curie workshop participants, the 2012 London Business School Trans-Atlantic doctoral conference participants, attendees at the Manchester Business School Phd reviews, Pauline Weetman and an anonymous reviewer of the Accounting and Business Research for their helpful comments I also thank Stuart Turley and John O‟Hanlon for taking the time to be my examiners and for giving helpful comments to improve the thesis I also wish to thank Martin Walker, Christopher Humphrey and Stuart Turley, for granting me the Marie Curie research fellowship Christopher Humphrey has also kindly given me useful comments for this thesis At this stage, I would like to acknowledge the Marie Curie research funding from the European Community's Seventh Framework Programme FP7PEOPLE-ITN-2008 under grant agreement number PITN-GA-2009-237984 for my thesis Next, I would like to thank Claire Faichnie and Anusarin Lowe of the Manchester Business School Postgraduate Research Office for always being very helpful by going the extra mile to help me on administrative matters in relation to research I also wish to acknowledge the funding from the Singapore Management University (SMU) for my Phd research I thank Pang Yang Hoong, the Dean of the SMU School of Accountancy for her encouragement and support during my Phd I am also thankful for the constant kind words of encouragement from many friends and colleagues of the SMU School of Accountancy Among them, I specially thank Pearl Tan and Andrew Lee, who believed in me and wrote the reference letters for me in my applications to the Phd programmes There are many friends who accompanied me during this Phd journey and I appreciate their friendships They include my Phd classmates Colin Cheng Zeng, Jirada Petaibanlue, Chanchai Tangruenrat, Chiwamit Pimsiri and Agyenim Cletus In particular, I would like to thank my flat-mate Colin and his wife Qianghui for their warmth and friendship Outside the academia, I thank many friends (too numerous to name) for giving me the morale support Augustine Law is one close friend whom I could always turn to in difficult times and I have special thanks for him Now, I come to the most special group of people in my life My family is a pillar of support for me during this Phd process I wish to thank my wife for taking care of our daughter while I am away I would like to thank her for patiently waiting for my return My daughter has provided me with the comfort and joy whenever I am home during breaks, for which I am thankful My father and mother have always played the most special and important roles in my life journey They are always there for me, encouraging me and constantly giving me the full support I am eternally indebted to them I also thank my brothers for their brotherly love and their morale support Finally, I thank God for making the completion of my Phd thesis possible 10 Table 4.13: Yearly abnormal returns for portfolios ranked by discretionary loan provisions = lowest, 10 = highest discretionary LL) 10 Deciles & 10 Deciles & Deciles & Deciles & Deciles & Deciles & Deciles & Deciles & Deciles & Deciles & 10 Obs 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 -0.102 0.059 -0.006 -0.070 -0.180 -0.276 -0.008 -0.093 0.122 -0.128 (0.848) (0.363) (0.711) (0.656) (0.311) (0.536) (0.152) (0.491) (0.224) (0.167) 128 -0.403 -0.391 -0.443 -0.428 -0.447 -0.480 -0.507 -0.466 -0.417 -0.592 (0.082) (0.915) (0.532) (0.848) (0.840) (0.728) (0.733) (0.624) (0.606) (0.057) 188 0.010 -0.216 -0.025 -0.207 -0.155 -0.252 -0.146 -0.191 -0.203 -0.150 (0.252) (0.094) (0.056) (0.066) (0.347) (0.105) (0.191) (0.549) (0.823) (0.399) 231 -0.226 -0.131 -0.010 -0.127 -0.302 -0.223 -0.006 -0.186 -0.271 -0.075 (0.338) (0.544) (0.350) (0.414) (0.169) (0.278) (0.057) (0.121) (0.295) (0.103) 262 0.063 0.112 0.303 0.086 0.189 0.068 0.127 -0.022 -0.069 0.019 (0.639) (0.626) (0.130) (0.059) (0.333) (0.312) (0.546) (0.066) (0.484) (0.273) 304 0.030 0.183 0.192 0.123 0.053 0.058 0.042 0.055 0.068 0.052 (0.751) (0.019) (0.936) (0.497) (0.252) (0.921) (0.815) (0.868) (0.836) (0.809) 322 0.102 0.146 0.006 -0.036 0.072 -0.032 0.010 -0.019 0.006 0.131 (0.790) (0.634) (0.127) (0.543) (0.135) (0.187) (0.578) (0.634) (0.697) (0.241) 349 0.021 -0.003 0.055 -0.094 -0.004 -0.091 -0.101 -0.100 0.005 -0.013 (0.677) (0.784) (0.473) (0.006) (0.064) (0.077) (0.788) (0.988) (0.169) (0.840) 344 0.125 0.085 -0.042 0.049 -0.058 -0.060 -0.013 0.012 -0.065 -0.110 (0.003) (0.709) (0.160) (0.188) (0.098) (0.954) (0.342) (0.655) (0.204) (0.399) 364 -0.135 -0.119 -0.098 -0.162 -0.102 -0.131 -0.177 -0.163 -0.190 -0.186 (0.888) (0.767) (0.826) (0.477) (0.201) (0.564) (0.270) (0.689) (0.391) (0.888) 385 -0.118 -0.139 -0.128 -0.078 0.015 -0.065 -0.009 -0.076 -0.015 -0.090 (0.591) (0.697) (0.837) (0.339) (0.238) (0.341) (0.334) (0.154) (0.452) (0.359) 387 -0.180 -0.200 -0.066 -0.155 -0.140 -0.213 -0.231 -0.236 -.1955 -0.321 (0.165) (0.830) (0.071) (0.137) (0.800) (0.282) (0.775) (0.948) (0.597) (0.147) 385 -0.122 -0.121 -0.047 -0.282 -0.225 -0.193 -0.193 -0.304 -0.202 -0.279 (0.327) (0.994) (0.541) (0.037) (0.626) (0.824) (0.995) (0.305) (0.428) (0.568) 385 -0.063 0.122 0.037 -0.052 -0.118 -0.117 -0.105 -0.203 -0.108 -0.342 (0.087) (0.273) (0.605) (0.492) (0.610) (0.998) (0.916) (0.374) (0.449) (0.136) 442 This table shows the yearly abnormal returns for each portfolio ranked by discretionary loan provisions, a yearly breakdown of table 4.12 panel A 200 Table 4.14: Abnormal returns for portfolios ranked by discretionary and nondiscretionary fair value gains and losses Panel A Portfolios ranked by discretionary fair value Abnormal returns at t+1 gains and losses (1 = lowest, 10 = highest discretionary fair value gains and losses) -0.092 -0.094 -0.110 -0.110 -0.079 -0.081 -0.127 -0.079 -0.072 10 -0.041 Difference bet deciles and 10 p value 0.051 (0.068) Difference bet deciles and p value -0.002 (0.940) Difference bet deciles and p value -0.016 (0.543) Difference bet deciles and p value 0.000 (0.999) Difference bet deciles and p value 0.031 (0.296) Difference bet deciles and p value -0.002 (0.948) Difference bet deciles and p value -0.046 (0.104) Difference bet deciles and p value 0.048 (0.072) Difference bet deciles and p value 0.007 (0.783) Difference bet deciles and 10 p value 0.031 (0.266) Obs 4476 Panel B Portfolios ranked by non-discretionary fair Abnormal returns at t+1 value gains and losses (1 = lowest, 10 = highest non-discretionary fair value gains and losses) -0.113 -0.033 -0.117 -0.079 -0.102 -0.112 -0.066 -0.096 -0.098 10 -0.068 Deciles and 10 p value 0.045 (0.104) Deciles and p value 0.080 (0.005) Deciles and p value -0.084 (0.004) Deciles and p value 0.038 (0.151) Deciles and p value -0.023 (0.381) Deciles and p value -0.010 (0.711) Deciles and p value 0.046 (0.120) Deciles and p value -0.030 (0.313) Deciles and p value -0.002 (0.923) Deciles and 10 p value 0.030 (0.245) Obs 4476 Figures in brackets are p values in tests of significant difference from zero and difference between portfolio deciles 201 Table 4.15: Abnormal returns for portfolios based on discretionary loan provision and bank risk Low DLP - Low returnvol -1 High returnvol -5 High minus Low P value High returnvol/ Low DLP minus Low returnvol/High DLP P value Abnormal returns at t+1 High DLP - -0.116 -0.046 -0.132 -0.124 -0.148 High minus Low -0.032 -0.122 -0.034 -0.009 -0.000 -0.069 -0.096 -0.059 0.012 -0.121 -0.099 -0.116 -0.071 -0.108 -0.119 -0.089 -0.124 -0.140 -0.117 -0.126 -0.099 -0.018 -0.083 -0.117 -0.099 0.116 0.058 0.061 0.000 0.049 (0.053) 0.148 (0.237) (0.184) (0.997) (0.331) P value (0.667) (0.069) (0.015) (0.068) (0.496) (0.004) This table shows the abnormal returns for each discretionary (DLP) quintile and return volatility quintile The column „High minus Low‟ shows the difference in returns between the highest and lowest discretionary loan provision portfolios for each risk (return volatility) quintile The row „High minus Low‟ shows the difference in returns between the highest and lowest risk (return volatility) portfolios for each discretionary loan provision quintile Figures in brackets are p values in tests of significant difference from zero 202 Table 4.16: Mimicking portfolios for Khan (2008), Chan and Chen (1991) tests Panel A: Descriptive statistics Variable Obs Mean Std Dev Min Max Median Skewness DLPdif Riskdif 14 14 0.036 -0.020 0.043 0.104 -0.041 -0.260 0.127 0.122 0.034 0.006 0.344 -1.057 Panel B: Correlation matrix DLPdif Riskdif DLPdif 1.000 -0.527 Riskdif 1.000 This table shows the descriptive statistics and correlation of the returns to two mimicking portfolios for the 14 years from 1996 to 2009 DLPdif is the return on low discretionary loan provision minus high discretionary loan provision portfolios Riskdif is the return on high risk (return volatility) and high discretionary loan provision minus low risk and low discretionary loan provision portfolios 203 Fig 4.1: Composition of mimicking portfolios L, H, LL and HH in Khan (2008), Chan and Chen (1991) tests Portfolio L: Bottom Discretionary Loan Provision quintile Portfolio HH: High discretionary loan provision and high risk Low risk Discretionary Loan Provision quintile Low risk Discretionary Loan Provision quintile Discretionary Loan Provision quintile High risk Portfolio H: Top Discretionary Loan Provision quintile Portfolio LL: Low discretionary loan provision and Low risk High risk 204 Appendix A: Movements in sample size across years 1996 C/f total Drop Survive (C/f total minus Drop) New Total Drop breakdown Bankruptcy Forced delisting e.g low price, insufficient equity, stop trading Delisting at company request e.g gone private Missing data in Compustat24 Merger 24 128 1997 128 1998 188 22 1999 231 30 2000 262 33 2001 304 25 2002 322 27 2003 349 46 2004 344 29 2005 364 31 2006 385 39 2007 387 31 2008 385 36 2009 385 36 128 166 201 229 279 295 303 315 333 346 356 349 349 60 188 65 231 61 262 75 304 43 322 54 349 41 344 49 364 52 385 41 387 29 385 36 385 93 442 1 1 1 2 1 14 2 7 10 14 19 18 24 23 20 18 33 22 28 28 14 10 This excludes cases where missing returns are set to zero based on Kraft et al (2006) 205 Chapter Summary and Suggestions for Future Research In this thesis, I examine three major bank accounting topics: bank accounting conservatism, fair value accounting and loan loss accounting Summary of findings and their implications In chapter 2, I find that discretionary loan provisions can be used in a positive manner when banks are conservative by recognizing losses on a timely basis Banks which recognise losses on a more timely basis make sounder and more prudent economic decisions, shown in their debt pricing behaviour The banks that recognize losses on a timely basis also exhibit less pro-cyclical lending behaviour I find that there is a relation between banks‟ debt pricing behaviour and their levels of accounting conservatism even though the banks use a similar incurred loan loss accounting standard This suggests that accounting standard setters should also focus on the incentives and the accounting practice dimensions particularly conservatism while setting the standards The finding of this research raises questions as to whether the switch to an expected loan loss accounting method can really induce the less conservative banks to engage in prudent accounting and lending practices In chapter 3, my research shows that fair values become more useful to the equity market when banks become more risky and when banks elect to use fair value option This result provides inputs to the policy makers on the situations when fair value gains and losses are most value-relevant 206 In chapter 4, I find that loan loss provisions, in particular the discretionary components are mispriced by the equity market This is consistent with the notion that discretionary loan provisions are subjective and can be used for earnings and capital management evidenced in prior literature (Kanagaretnam et al., 2004) As a result, the market finds it difficult to understand and to price the discretionary loan provisions correctly This is an important piece of finding to the standard setters, which suggests they should place more emphasis in their policy making on discretionary loan provisions The implications of the research to policy makers and standard setters are summarized as follows: The incurred loan loss accounting model is generally criticized for being pro-cyclical (Turner, 2010) This gives impetus to the standard setters to propose the expected loan provisioning model in the new IFRS Chapter shows that banks which exhibit conditional conservatism (in recognising timelier losses than gains) are more prudent and less volatile in their loan pricing While all the banks in the sample are applying the incurred loan loss provisioning model, not all the banks show pro-cyclical loan pricing behaviour Banks with different loan provisioning practices in terms of timeliness in loss versus gains recognition display different levels of prudence and volatility in their loan pricing Standard setters should also consider bank loan provisioning practices beyond standards There have been a lot of debates in both academic and practitioner literature on the pros and cons of fair value accounting Chapter documents that during stable economic times prior to the crisis the value relevance of fair value gains and losses is 207 positively related to bank risks In addition, the fair value gains and losses of banks which adopt fair value option accounting are more value relevant than those banks which not adopt This study provides inputs to standard setters and policy makers on the situations when fair values are more value relevant and contributes to the debates on fair value accounting Chapter provides evidence that discretionary loan provisions are mispriced by the equity market but less evidence of non-discretionary loan provisions and fair value gains and losses (discretionary or otherwise) being mispriced This indicates to the standard setters the importance of increasing the transparency of discretionary loan provisions in their standard setting Chapters and results show the negative and positive aspects of discretionary loan provisions The negative aspect is that discretionary loan provisions can be mispriced by the equity market The positive aspect is that banks which use their discretion by being timelier in recognising losses than gains are more prudent and stable in their loan pricing These results provide inputs to the policy makers in weighing the pros and cons of providing discretion to the banks on loan provisioning Limitations and Qualifications of the Findings The limitation of chapter is that although I find a relation between bank accounting conservatism and bank lending behaviour, there may be unobserved bank characteristics that drive both bank timeliness in loss recognition and bank lending behaviour I have attempted to control for bank characteristics such as corporate governance and bank risk and added bank fixed effects However, it may still be 208 possible for unobserved bank characteristics to influence the bank timeliness in loss recognition and the bank debt pricing On chapter 3, a major limitation is that it is difficult to disaggregate the effects of various factors at work during the crisis One factor is the cost of capital, which is likely to have increased during the crisis compared to pre-crisis The increase in costs of capital, which are also measures of bank risks, may change the response coefficient of fair values Since my study involves the interaction between fair values and bank risks, when bank risks also change during crisis, it is very difficult to disentangle the effects of changes in bank risks from changes in value-relevance of fair values Another factor that may be at work during the crisis and limits my ability to determine the relationship between the value relevance of fair values and bank risk during the crisis is an amendment to the fair value accounting standard during crisis, which allows banks to reclassify some assets from fair values to amortised costs The amendment to the standard as well as the decreased liquidity during the crisis may either make fair values more value-relevant or less and the limitation is that the effect of these factors cannot be disaggregated from other factors Fourth, investors may become excessively pessimistic and less rational during crisis, which creates noise in the model Finally, there may be a change in the relation between the value-relevance of fair value gains and losses and bank risk I use different research designs and find some evidence of the positive relation between fair value value-relevance and bank risk in certain models but not in others, thus the inference for the crisis period is not conclusive 209 In chapter 4, the discretionary loan provision equation is based on prior literature (Wahlen, 1994; Kanagaretnam et al., 2004) The discretionary component is derived from the residual after determining the non-discretionary element The latter is assumed to be driven by the loan loss allowances, non-performing loans, the loan loss write-offs and the outstanding loans based on prior literature My research is limited to the US because many non-US banks either not disclose their loan loss write-offs or their non-performing loans (or both) Prior research using international bank samples typically have to modify their models or focus on questions that involve available data (Fonseca and Gonzalez, 2008; Gebhardt and Novotny-Farkas, 2011) I choose to follow the full model in prior literature and restrict my study to the US bank sample The accounting standard setters plan to switch the loan loss accounting method from the incurred loan loss method to the expected loan loss method in the new accounting standard IFRS IFRS has not been implemented yet and the data available for empirical research is still limited to that under the existing standards When IFRS is eventually implemented, readers need to exercise caution before extrapolating the results of chapters and to the expected loan loss accounting regime It is not certain how the loan loss accounting standard changes would affect the results (if any) Additional data and empirical tests need to be carried out under the expected loan loss accounting regime to understand the effects of the changes in accounting method Studies of bank accounting often include regulatory-capital-related variables There are data restrictions in particular for non-US banks, which make it more difficult to use regulatory-capital data Chapter can be enhanced further to link the conditional 210 conservatism measure with the regulatory capital with a regression of regulatory capital on conditional conservatism This will provide a stronger case to validate Beatty and Liao (2011) and to support hypothesis that banks which exhibit greater conditional conservatism built up their capital prior to the crisis and face lesser regulatory capital constraints during crisis In chapter 3, regulatory capital data could be added as additional control variables in the main regression of stock returns on bank earnings components Regulatory capital may be interacted with the bank earnings components to determine if the stock market reacts differently to banks with different capital levels In chapter 4, regulatory capital could be added to the discretionary loan provision and discretionary fair value equations as additional control variables There are difficulties arising from the pooling of data on banks from different countries for chapters and In general, US banks use US GAAP while UK, European and Australian banks apply IAS (international accounting standards) Canadian banks use Canadian GAAP, which are close to US GAAP The practice of applying the IAS may differ among the IAS countries given that IAS is principlesbased and judgement is required in their applications The institutional setting (in terms of legal and regulatory frameworks) also differs across countries When handcollection of data particularly for chapter is required, the differing disclosure practices across countries and within countries among banks of different sophistication levels pose challenges Suggestions for future research 211 There can be further research on the economic consequences of timeliness in loss recognition Do industrial firms and banks implement their investment decisions differently depending on their levels of timeliness in loss recognition? The investment decisions of banks differ from industrial firms It may be interesting to carry out further research into how the banks‟ investment decisions are influenced by their timeliness in loss recognition, compared to industrial firms In addition, the fair value empirical chapter can be improved further by establishing the links between bank risk and risky financial instruments, between risky financial instruments and subjectivity of fair value estimates, as well as between the subjectivity of fair value estimates and the value relevance of fair values I believe the implementation of the expected loan loss method under IFRS would provide fertile grounds for future research The proposed IFRS mandates disclosure of additional information on expected loan provisions The additional information may become useful in improving the discretionary loan provision model Further research can be carried out on the expected loan loss provisions In theory, the discretionary loan provisions may become more significant under IFRS because IFRS requires greater judgement and discretion than the existing method If this leads to greater difficulties for the market in interpreting the loan provision information, the market mispricing of discretionary loan provisions may become more significant On the other hand, if the banks are able to explain clearly in their disclosures how their loan loss provisions are determined, the market may be able to price the discretionary loan provisions more accurately Given these opposing possibilities, it is an empirical question if the market mispricing of discretionary loan 212 provisions reported in my chapter increases, decreases or remains the same under IFRS The market mispricing tests such as the Mishkin tests and the hedge portfolio tests used in Sloan (1996) and Xie (2001) can be applied to the expected loan loss provisions to determine if the market can better understand and price the expected loan loss provisions compared to the current incurred loan loss provisions The reason for accounting standard setters to change the loan loss accounting standard from an incurred loan loss method to an expected loan loss method is to increase the timeliness in loan loss recognition Further research after the implementation of the new loan loss accounting standard can provide inputs to the accounting standard setters on whether their objective has been achieved Another possible research area is whether the bank accounting conservatism relationship to lending behaviour changes after the expected loan loss method has been implemented If the timeliness in loan loss recognition increases after the new standard relative to the current standard, how would this change the relationships of timeliness in loan loss recognition to debt pricing and loan volume reported in my chapter and Beatty and Liao (2011) respectively? Another area for future research relates to securitization accounting There have been allegations of bad assets being taken off balance sheet onto special purpose vehicles so as to free up capital under the Basel capital requirements It may be interesting to investigate further how the equity markets price these securitized assets or how these securitization activities affect the banks‟ lending activities 213 References Beatty, A., Liao, S., 2011 Do delays in expected loss recognition affect banks‟ willingness to lend? Journal of Accounting and Economics 52(1), 1-20 Fonseca, A.R., Gonzalez, F., 2008 Cross-country determinants of bank income smoothing by managing loan-loss provisions Journal of Banking & Finance 32(2), 217-228 Gebhardt, G., Novotny-Farkas, Z., 2011 Mandatory IFRS Adoption and Accounting Quality of European Banks Journal of Banking & Finance 38 (3) & (4), 289333 Kanagaretnam, K., Lobo, G.J., Yang, D., 2004 Joint Tests of Signalling and Income Smoothing through Bank Loan Loss Provisions Contemporary Accounting Research 21(4), 843-884 Sloan, R.G 1996 Do Stock Prices Fully Reflect Information in Accruals and Cash Flows About Future Earnings? The Accounting Review 71(3), 289-315 Wahlen, J.M., 1994 The Nature of Information in Commercial Bank Loan Loss Disclosures The Accounting Review 69(3), 455-478 Xie, H., 2001 The Mispricing of Abnormal Accruals The Accounting Review 76(3), 357-373 214