All rights reserved. No part of this journal may be reproduced or utilized in any form or by any means,electronic or mechanical, including photocopying, recording or by any information storage and retrievalsystem, without the permission of the publishers.
The views expressed and facts stated in the papers contained in this volume are of the individual authorsand are in no way those of either the Editors, the institution to which they belong and the publisher.
Journal of Socialand
Management Sciences
National Institute of Bank Management
2 Prajnan
Objectives of PRAJNAN
PRAJNAN is a quarterly double-blind refereed journal that presents research papers, briefarticles and commentary on banking and finance and operations. The objectives of theJournal are (a) to encourage new thinking on concepts and theoretical frameworks in thevarious disciplines of Social Sciences, Administrative and Management Sciences, and(b) to disseminate such new ideas and research papers (with strong emphasis on realismof analysis, provision and use of empirical evidence) which have broad relevance to theworking and development of banking and other financial institutions, to help themanagement of such institutions in formulating various policies that are related both tothe short-term and long-term needs of the organizations as well as of the economy.
Submission of full Research Papers which would fulfil the above objectives are welcome.The Journal would also publish Brief Articles, Notes or Comments, etc., which conformsthe standards. It is a condition of publication, that the Research Papers, Brief Articles,Notes or Comments, etc., are original works and they have not already been published orthey are not submitted for publication elsewhere and will not be reprinted without leave ofthe Editor (i.e. the copyright of the published materials to be with the Journal).
Each paper is reviewed by the Editor and if it is judged suitable, then it enters the double-blind review process. Manuscripts may be rejected, edited or returned for specified revisionon the basis of the recommendations of the referees and decision of the editorial board.Utmost care is taken to ensure faster response to our authors.
Opinions expressed by the authors in the Papers, Notes or Comments are purely individualand no responsibility for such views is assumed by the Editor or the Publishers.
Authors whose papers are published in the Journal will be supplied with 25 reprints oftheir paper and a complimentary copy of the issue.
The Journal also reviews important books published recently in the disciplines of SocialSciences, Administrative and Management Sciences. Publishers desirous of such reviewsshould send preferably two copies of the books to the Journal.
Instructions to the Authors
We prefer electronic submission of all papers (by email). Where this is not absolutelypossible, two hard copies may be sent to the Editor. All articles submitted for considerationshould follow strict academic standards and it should be of interest to a broad audience ofpractitioners and academics. The preferred text format for papers is MS-Word. The lengthof the full Research Paper will be around 7000-8000 words. Type every portion of themanuscript double-spaced (a minimum of 6 mm between lines), including figure legends,table footnotes, and references, and number all pages in sequence, including the abstract,figure legends, and tables. In case of submission by postal service kindly send the CD-Romalong with two print-outs. The length of a brief article would be around 1500-2000 words.
The first page (Title page) of the manuscript should contain only: (a) Title of the Paper, BriefArticle, Note or Comment, (b) Name of the Author, (c) Name of the Institution to which theauthor is affiliated, (d) Brief academic bio-data and work experience of the author, and(e) complete mailing and e-mail address, and phone numbers. The second page (main text)
Bagchi & Khamrui: Modeling Working Capital Management Using Stationary Series 3
should contain an executive summary or abstract of no more than 150 words, and 3-5keywords (e.g., profitability, bank competitiveness and performance, etc.) and JELclassification code. The main text subject matter should commence from second pageonwards for anonymity (so that the referee will not know the identity of the author).
Authors are fully responsible for the accuracy of the data used in the manuscripts. If themanuscripts contain statistical analysis, the authors should provide supplementary notes(which will not be published) on the methods used in the analysis for the convenience ofthe referees. Statistical tables should be clearly titled and numbered and the reader shouldbe able to understand clearly the meaning of each row or column. Units of measurementand sources of data should be clearly stated.
The full mathematical workings necessary for justifying each step of the argument (whichwill not be published) should accompany the manuscripts of mathematical character tofacilitate the referee's work. Diagrams should be clearly drawn (or generated) and labelledso that the reader would be able to understand its meaning.
Bibliographical references should be cited using the Harvard style [name (year)] andcarefully checked for completeness, accuracy and consistency. The authors should carefullycheck and complete the reference citation in respect of:
For Journal Article: (a) Author's surname, initials as it appears on the cited works, (b) Yearof publication, (c) "Title of the paper", (d) Name of the Journal, (e) Volume number, (f) (Issuenumber), and (g) Page references.
For Books: (a) Author's surname, initials as it appears on the cited works, (b) (Year ofpublication), (c) Title of the book, (d) Publisher, (e) Place of publication, and (f) Page references.
For Book Chapters: (a) Author's surname, initials as it appears on the cited works, (b) (Yearof publication), (c) "Chapter Title", (d) Editor's surname, initials (Ed.), (e) Title of the book,(f) Publisher, (g) Place of publication, and (h) Page references.
For Working Papers: (a) Author's surname, initials as it appears on the cited works,(b) (Year of publication), (c) "Title of the paper”, (d) Working paper number, (e) Institutionor Organization, (f) Place of Organization, and (g) Date.
For Electronic Sources: If an article is available online, the full URL should be supplied atthe end of the reference along with the date when the resource was accessed.
It is important to note the papers listed in the bibliographical reference should also bediscussed in the main text. The authors should cite publications in the text as follows:using the author's surname, e.g. Merton (1973); citing both names if there are two authors,e.g. (Black and Scholes, 1974), or (Altman, et al., 1980) when there are three or more authors.The reference list in alphabetical order should be supplied at the end of the paper.
We are also on the EBSCO: Business Source Corporate Plus, Facebook & Linkedin1. http://www.facebook.com/prajnan.journal2. http://in.linkedin.com/pub/prajnan-nibm/54/982/542
INVITATION OF PAPERS FOR PUBLICATIONPRAJNAN would endeavour to publish papers from leadingacademicians, researchers and practitioners on Banking and Financeand on other disciplines of Social and Management Sciences. Wealso invite articles, notes, and comments based on operationalexperiences and supported by relevant evidences, from policy-makers, practicing senior bank executives and management expertswho would like to share their thoughts, ideas and views with others.
All communications should be addressed to:
The EditorPRAJNAN
National Institute of Bank ManagementNIBM Post Office, Kondhwe Khurd, Pune 411 048, INDIA
Bagchi & Khamrui: Modeling Working Capital Management Using Stationary Series 5
Editorial
I am very pleased to release the April-June 2018 issue of PRAJNAN, the Journal of Socialand Management Sciences. I sincerely thank our readers, authors, referees, editorialcommittee and editorial board members their continued support and guidance. Specialthanks to all reviewers who supported the journal with their expert comments on papers.Established in 1972, the journal is mainly devoted to the publication of original researchpapers that have high academic and professional standards to the readers and our endeavorwill continue. This journal is a place for exchanging original studies and research onbanking and finance and other areas of Social & Management Sciences that have relevancein building a new body of knowledge and improvement in practice of banking operations.We encourage new submissions in the areas of banking and regulation, corporate finance,financial market and operations, macroeconomic issues, and so forth. Our journal followsrigorous double blind peer-review process to ensure high standards of scholarlypublication. The journal is indexed and abstracted in EBSCO. It has now being listed inProQuest database of scholarly journals. It is also in the UGC approved list of academicjournals (reference number 41904).
In this issue we have put together a group of four excellent original articles and one bookreview.
The first paper by Shubhasree Bhadra, Bibek Ray Chaudhuri and T P Ghosh, "Efficiencyof MFIs in India during the Crisis Years: A Data Envelopment Analysis Approach",studies the efficiency performance of Indian Micro Finance Institutions for the period2008-09 to 2010-11. Using the Data Envelopment Analysis (DEA) and in the second stagethrough regression analysis, the paper attempts to assess the impact of contextual variableslike Capital-asset ratio, Debt-equity ratio, average loan balance, loan per officer, personnelallocation ratio, cost ratio and profit margin on technical efficiency. The regression resultshave empirically identified the sources of inefficiency in microfinance sector. The paperfinds there is an improvement in performance of Micro Finance Institutions (MFIs) during2010-11 and more matured NGOs are more efficient than Non-Bank Financial Institutions(NBFIs.)
In the second paper, "Determinants of Non-Performing Loans in India: A System GMMPanel Approach", Asit Ranjan Mohanty, Binay Ranjan Das and Satyendra Kumarinvestigate the key determinants of Non-Performing Loans of the Indian banking system.Using a rich panel data set of Scheduled Commercial Banks in India (public, private andforeign) over the period 2000-01 to 2015-16, the study adopts Blundell and Bond (1998)developed dynamic Generalized Moments Method (GMM) to estimate the causative factorsthat influence the generation of Non-Performing Loans in banks. The regression resultsreveal that macroeconomic variables like GDP growth, stock market index and marketcapitalization have negative impact on Gross NPL ratio. However, Gross Fiscal DeficitRatio has positive effect on GNPL ratio. The study finds that the bank specific variableslike higher credit to deposit ratio, growth in bank branches, higher return on equity andhigher CRAR can lower Gross NPL ratio. On the other hand, higher operating expenses
6 Prajnan
causes NPA increase. The study suggests that strengthening of the balance sheet of privatecorporate sectors will support the improvement of balance sheet of banks and wouldultimately reduce the NPA burdens of banks. In this context, the study highlights thathigher efficiency, credit expansion, profitability, expansion bank branches are the few keyparameters that can strengthen the balance sheet of banks.
In the third paper, "A Comparative Analysis of Loan Recovery Strategy of Indian Banks",Robin Thomas and Ram Krishna Vyas compared the loan recovery process of Indiancounterparts. The study recommends banks to use proactive monitoring, improvingportfolio mix, reducing concentration risk, deleveraging stressed assets by sales or enforcingcontractual rights as effective strategy to deal with the stock and flow of NPAs.
In the fourth article, "Dynamic Capital Adequacy Ratio for Bringing Equilibrium inLending in the Banking Industry: A Study of the Five Largest Banks in India", AkashBaruah has made an interesting attempt to demonstrate that the present capital adequacyframework which captures capital in comparison to risks weighted assets ratio does notactually capture the inherent risk of a bank. Moreover, it may erroneously lead to a sense ofsafety and high loss absorbency capacity of a bank if its CAR ratio is high. Using cases offive public sector banks, their Gross Non-Performing Assets and Net Non-PerformingAssets over the years and their Capital to Risk-weighted assets ratio, the study advocatesto use dynamic capital adequacy ratio by taking into account the marginal cost of lendingwhich will reflect the inherent risks of banks.
In this issue, we have published one book review. Shri Sunil Bakshi, Visiting Faculty,National Institute of Bank Management (NIBM) reviews the book "Connected or Disconnected-The Art of Operating in Connected World" authored by Micke Damrmell and Kapil Rampaland published by Sage.
I look forward to your active participation and cooperation. We invite new submissionsthat include topics related to banking and finance and social & management sciences forour forthcoming issues.
Dr Arindam Bandyopadhyay(Editor, PRAJNAN & Associate Professor, Finance, NIBM)
Contents
April – June 2018Volume XLVII
Number 1
Full Research Papers
Shubhasree Bhadra Efficiency of MFIs in India during the 9Bibek Ray Chaudhuri Crisis Years: A Data Envelopment AnalysisT P Ghosh Approach
Asit Ranjan Mohanty Determinants of Non-Performing Loans 37Binay Ranjan Das in India: A System GMM Panel ApproachSatyendra Kumar
Robin Thomas A Comparative Analysis of Loan Recovery 57Ram Krishna Vyas Strategy of Indian Banks
Akash Baruah Dynamic Capital Adequacy Ratio for 89Bringing Equilibrium in Lending in theBanking Industry: A Study of the FiveLargest Banks in India
Book Reviews
Sunil Bakshi Connected or Disconnected – The Art of 109Operating in Connected World– Micke Damrmell & Kapil Rampal
Efficiency of MFIs in Indiaduring the Crisis Years:
A Data Envelopment Analysis Approach
Shubhasree BhadraBibek Ray Chaudhuri
T P Ghosh
Indian microfinance sector has gone through ups and downs since itsinception. From late 90s to 2005, the growth was phenomenal. But theKrishna crisis, Kolar crisis and finally the Andhra Pradesh crisis inconsecutive years impacted the sector immensely deciding the futurepath it would traverse. There were significant changes in regulatoryframework after these turbulent years. In this paper we attempt toexamine the technical efficiency of different types of MFIs during 2008-09 to 2010-11, 2009 being the year of Andhra Crisis. The study showsthat majority of the MFIs were operating below the efficient frontier.Inefficiency slightly increased during 2009-10 but significantly reducedin the subsequent year. It is seen that NGOs were dominating the NBFIsin terms of efficient units. Among the NGOs the matured ones weremore efficient than NBFIs. Further random effect Tobit model showsthat capital structure of MFIs is a significant determinant for overalltechnical efficiency. But efficiency of NGOs are more sensitive tochanges in financial expense-asset ratio than NBFIs.
Keyword: Efficiency, NGO, NBFC
JEL Classification: G2, G21, G28
Section IIntroduction
Financial sector policy should ensure access to quality financial services forall. It is already documented that financial development and economic growthare complementary in nature. Depth and breadth of financial services enhance
Ms Shubhasree Bhadra ([email protected]), Ph.D Candidate, University of Calcutta, Kolkata.
Dr T P Ghosh ([email protected] ) Assistant Professor, Indian Institute of Foreign Trade, Kolkata.
10 Prajnan
efficiency of all firms through developed market structure and payment system.Effective, efficient and stable financial system of an economy facilitatesregulators to balance between inflation, growth and unemployment. The reportof the Rajan Committee (2009) placed "inclusion, growth, and stability as thethree objectives of any reform process".
After bank nationalization, an alternative avenue to cater to poor, especiallythe rural poor led to the origin of microfinance SEWA cooperative bankestablished in 1974 and started provision of banking services to poor womenemployed in the unorganized sector in Gujarat. During 4th Five Year Plan,group based approach was initiated through setting up of Self-Help Group(SHGs) for small, marginal farmers and agricultural labourers. Later NationalBank for Agriculture and Rural Development (NABARD), National Credit Fundfor Women/Rashtriya Mahila Kosh (RMK) under Ministry of Woman and ChildDevelopment and Small Industries Development Bank of India (SIDBI) havetaken initiatives to promote SHGs for providing credit to women members ofpoor households. From 1996 Reserve Bank of India had included financingSHGs as main stream activity of bank under priority sector lending. After thatMicro Finance Program (MFC) (2004) and Micro Finance Bill (2006) helpedthe sector to develop. Two major models of credit delivery have been the bank-based model and the MFI-based model. In bank-based model SHGs are formedby mainly by NGOs and financed by commercial banks (SHGs looking aftertheir own finances after being linked to the bank by the NGOs). In the MFI-based model groups are formed and financed by MFIs (banks being the majorsource of debt capital). Subsequently, regulatory changes led to new institutionslike NBFC and NBFC-MFIs to enter the fray.
In India three groups of players are working in microfinance sector. SHG-Bank linkage model which is accounting 58 per cent, Non-Banking FinanceCompanies 34 per cent and trust, society, etc., 8 per cent of outstanding loanportfolio( RBI Report, 2011). The latter ones are part of MFI model working asan alternative to the first one in the Indian microfinance sector.
Performance of Indian MFIs in recent years' in terms of number of clients(breath of outreach), percentage of women borrowers, gross loan portfolio andaverage loan per borrowers (depth of outreach) is shown in the following table
Average Loan per Borrower INR 11425 INR 11425 INR 10079
Source: Bharat Microfinance Report, 2015 & 2016
Bhadra, Chaudhuri & Ghosh: Efficiency of MFIs in India during the Crisis Years:... 11
The first entrants in Indian MFI sector were the NGOs supported by NABARD.Report of the Committee on Financial Inclusion (January 2008), provided themuch needed recommendations for the commercial banks, regional ruralbanks, micro-finance institutions and the NGOs to achieve the mission offinancial inclusion at the nation level. The sector has gone through a rapidgrowth phase between late 1990s and 2005. Given the state of regulation andlack of understanding the sector started facing difficulties. The first warningcame when Krishna crisis broke out in 2005-06. After the Crisis, banks werereluctant to lend to the MFIs and more capital flowed in the sector as long-term investment rather than short-tern bank loan (Mader, 2013). The KrishnaCrisis like event was also repeated within short period in Nizamabad (AP),Kolar (Karnataka) and Idukki (Kerala).
In February 2009, in Kolar borrowers refused to repay their loan. After fourmonths of Kolar crisis, similar incidents took place in Ramanagara, a districtin south Bangalore and Mysore. In Mysore there was a mass default and afterthat a communal riot broke out in April 2009. Competition and over-supply ofcredit and "a combination of irresponsible lending practices of some lenders,religious intervention and low levels of customer awareness "were alsoresponsible for Kolar crisis and its geographical spread"(IFMR). After Kolarcrisis, in October 2010, AP Government promulgated an Ordinance to saveborrowers from unethical practices to recover loan which led to suicide by theborrowers, multiple borrowing and higher interest rate charged by the MFIs.In December the Ordinance was enacted as Andhra Pradesh MicrofinanceInstitutions (regulation of money lending) Act. The operation of the MFIs wereseverely impacted by this Act.
In this paper we attempt to evaluate the performance of MFIs in India duringthis period. The paper is organized as follows. Section II looks at the erstwhilestudies in this area. Section III spells out the methodologies, data sources andresults. Section IV concludes the paper by summarizing the results anddiscussing their implications.
Section IILiterature Review
Efficiency of an MFI refers to efficient use of resources like human capital,asset, subsidy, other funds to produce output, measured in terms of grossloan portfolio and number of active borrowers. Efficiency is important for anMFI for a number of reasons. Firstly, resources like, labour, time, money, rawmaterials, etc., all are scarce and limited in supply. Again donors are alsounwilling to fund MFIs to a level where they can serve all poor clients(Rosenberg, 1994). Secondly, growth in microfinance across the world hasincreased competition among MFIs to get funds to finance their operations.The donors and funders on the other hand are interested in funding the MFIs
12 Prajnan
who are sustainable and efficient (Barres et al, 2005). The increasedcompetition among MFIs has resulted in lowering interest rate and an urge tooperate more efficiently (Hermes et al, 2009). Thirdly, commercial banks andinvestors are interested to finance MFIs to fulfil their social responsibility, andat the same time for investments which have attractive risk-return profile(Deutsche Bank Research, 2007). This increased interest of commercial playersput pressure on MFIs to be sustainable and efficient. It helps organizations insetting their target through better management and to increase performanceand profitability (Reynolds & Thompson, 2002).
For MFIs, efficiency can be divided into two components – financial efficiencyand social efficiency (Nieto et al 2009). Financial efficiency of microfinanceinstitution is based on technical efficiency, which is based on the assumptionthat efficient MFIs have higher productivity (Sanchez 1997). Two approaches –production approach and intermediate approach are prevalent methods tocalculate efficiency under this method. Social efficiency is related to the abilityof MFIs to utilize their resources to empower weaker sections through alleviationof poverty by creating opportunities to earn a living and improve their livingconditions.
But it is very costly to reach the poor because making small loans involve hightransaction costs in terms of screening, monitoring and administrative costper loan. So a unit of large loan is cheaper compare to a unit of small loan(Conning, 1999; Hulme & Mosley, 1996; Lapenu & Zeller, 2001; Paxton &Cuevas, 2002). This leads to the debate surrounding outreach vs efficiency ofMFIs. It is felt that to be an efficient and sustainable MFI, it may have tocompromise on its traditional mission – to reach poor.
A comprehensive study by Cull et al (2007), with data set of 124 MFIs across49 countries, finds evidence of trade-off between outreach and efficiency –MFIs becoming larger shift their focus to wealthier borrowers. Cull et al (2009)also provides evidence of trade-off between outreach and commercialization.Hermes et al (2009) have tried to examine the relation between outreach andefficiency obtaining a negative relationship between the two based on a sampleof 435 MFIs over 11 years from 1997 to 2007 across the globe. Further, theMFIs who have more women borrowers as clients are found to be less efficient.
Various methods have been used in the literature to calculate efficiency. A studyby Guitierrez-Nieto, Serrano-Cinca and Molinero (2006) uses a non-parametricapproach (Data Envelopment Analysis) to measure efficiency of 30 LatinAmerican MFIs. The study shows that NGOs and non-bank financial institutionsare most efficient among the group of 18 MFIs. A paper by Hassan and Tufte(2001) uses stochastic frontier analysis to show Grameen Bank's branchesmanaged by female are more efficient than those managed by male staff.Similarly study by Leon (2001) reported that productivity of resources,governance and business environment contribute positively to cost efficiency
Bhadra, Chaudhuri & Ghosh: Efficiency of MFIs in India during the Crisis Years:... 13
of the Peruvian municipal banks. Baumann (2005) has also found positiverelation between MFI efficiency and productivity. Lafourcade, Isern, Mwangiand Brown (2005) show that African MFI-staffs are highly productive as thenumber of borrowers and savers per staff member are among the highest inthe World.
Regarding determinants of efficiency a study by Farrington (2000) identifiesthat administrative expense ratio, number of loans per loan officer and loanofficers to total staff, portfolio size, loan size, lending methodology, source offunds and salary structure are the factors on which efficiency of MFIs depend.
One controversial issue in this sphere is the relationship between regulatoryrestriction and performance of MFIs in terms of reaching the poor people.Some argue that more stringent restriction reduces efficiency whereas othersfeel the opposite is true. Barth, Caprio and Levine (2001), shows that, acrosscountries tighter restrictions make banks inefficient on average and enhancethe likelihood of banking crisis. Bhattacharyya and Pal (2013) in context ofIndian banking system shows that the 1990-91 reforms which included policiesto de-regulate the sector improved their technical efficiency. Post-1997 reformsput more emphasis on bank stability and that resulted in loss of efficiency ofbanks. Similarly, Chortareas, Girardone and Ventouri (2012) notes regulatoryrestrictions on bank activities adversely affect efficiency of banks in case ofEuropean Union. A study related to Spanish banks shows that deregulationswere associated with a decrease in relative cost efficiency for commercial banksbut no change was observed for savings banks (Lozano-Vivas (1998)). Theyshow that deregulation has negative impact on overall technical efficiency. Thusthe results are mixed as far as regulation and efficiency is concerned in thebanking sector.
Section IIIMethodology
A two-step approach has been adopted in this paper. First we calculate theefficiency for individual MFIs during 2008-09 to 2010-11. In the next step welook at the determinants of the efficiency scores using a suitable regressionmodel.
Production and intermediation approach are two alternative approaches thatcould be used to measure productivity and efficiency of Indian microfinanceinstitutions. According to intermediation approach MFIs are intermediariesbetween formal financial institutions and the poor people who don't have accessto formal finance. So according to this approach deposits and loans are to beconsidered as output, and loanable funds, interest expense and labour costsare considered as inputs. Production approach considers MFIs as producers,producing financial services like, deposits accounts, loan and advances, and
14 Prajnan
others for their clients. Therefore, number of accounts, number of loans,transaction volume is considered as output produced using labour, capitaland operating cost as inputs.
The analysis of the efficiency for Indian MFIs is based on production approach.In Indian, all types of institutions providing financial services to unbankedpeople are not allowed to collect savings from savers. Especially, MFIs are notallowed to take savings deposits. So efficiency analysis of Indian MFIs is carriedout on the basis of production approach.
In this approach, gross loan portfolio and number of women borrowers areconsidered as output of the production process of the financial services.Borrowings, equity, deposit, operating expenses and number of personnel arethe inputs used to produce the services.
Data Envelopment Analysis (DEA) is a non-parametric mathematicalprogramming approach to frontier estimation. This method constructs bestpractice frontier for each time period for each technology category. Comparingeach unit to the best-practice frontier provides a measure of its catching up inefficiency to that frontier. It is based on input-output data without requiring apriori specification of the functional form. Major advantage of the measure isthat the results are not sensitive to different specification of productionfunctions.
Indian MFIs are targeting women borrowers as it is perceived that it reducesdefault risks (compared to giving to men) and to achieve desirable socialoutcomes like education of children, improved health and sanitation etc. Againtotal loan disbursed is also important from outreach point of view. DEA allowsmultiple outputs to be considered simultaneously. Thus number of borrowersand gross loan portfolio has been used as outputs in our paper. In an alternativeapproach we have replaced number of borrowers with number of womenborrowers. This gives us an additional dimension to check the differences inefficiency when loans being given to women vis-à-vis men.
In DEA, technical efficiency can be measured in two ways. One is output orientedapproach and another is input oriented approach. Under the former approachtechnical efficiency score throws light on the scope of expansion of output ofDecision Making Units (DMUs) with the given level of input. The latter, however,looks at control of inputs to produce a given amount of output. The IndianMFIs have very little scope to reduce inputs to provide financial services to agiven level of women borrowers. On the contrary expansion of financial serviceswith limited amount of inputs can be thought as a more desirable strategy. Inthis paper we have thus adopted the output approach.
Under output oriented approach Technical Efficiency (TE) is measured as ratioof actual output to maximum producible output. Inefficiency of a DMU is
Bhadra, Chaudhuri & Ghosh: Efficiency of MFIs in India during the Crisis Years:... 15
measured by the divergence of actual production from the production on feasibleproduction set. The technical efficiency of a DMU is measured under threeheads – overall technical efficiency, pure technical efficiency and scale efficiency.Overall Technical Efficiency (OTE) uses Constant Returns to Scale (CRS) whichmeasures inefficiencies due to the input-output relationship and the size ofthe operation. On the other hand Pure Technical Efficiency (PTE) is measuredunder the assumption of Variable Returns to Scale (VRS). It measuresinefficiencies of DMU due to managerial under performance. And ratio of OTEto PTE measures scale efficiency of a DMU in our case an MFI.
We have used data on 53 MFIs operating in India from MIX Market Database.These 53 MFIs are selected according to maximum common number of MFIsfor whom information is available between the years 2008-2009 to 2010-2011.
ResultsFor analysis of efficiency of Indian microfinance sector, we have used one stageDEA technique to calculate different efficiency scores. To find out the sourcesof inefficiency of Indian MFIs we have calculated OTE, PTE and scale efficiency.
In 2008-09, average OTE of 53 MFIS is 0.88. It is ranging from 0.642 to 1.00,which implies considerable heterogeneity in terms of efficiency among MFIsoperating in Indian. In output oriented approach, both CCR1 and BCC2 modelanswer the basic question – by how much can output be expanded for eachMFI with existing input usage. The average OTIE [Overall Technical Inefficiency– (1-OTE)*100] result shows that, the MFIs on an average can expand theiroutput in terms of gross loan portfolio and number of women borrowers by 12per cent with same level of inputs. Alternatively, MFIs on an average can reduceinputs by 13 per cent to produce the same level of outputs.
In 2008-09, 22 MFIs are overall technically efficient. These 22 efficient MFIstogether define the efficient frontier and it becomes reference set for theinefficient MFIs. The efficient MFIs set benchmarks for inefficient MFIs toimprove their performance. In DEA terminology, these efficient MFIs are calledpeer. Peer MFIs are supposedly functioning well to avoid input wastage. Table1.1 shows details of efficient and inefficient Indian MFIs.
1. Charnes, Cooper and Rhodes (1978) proposed the model had input-oriented approach with constantreturns to scale (CRS). This is known as CCR model.
2. Banker, Charnes and Cooper (1984) proposed the model with variable returns to scale (VRS)and is known as BCC model.
16 Prajnan
Table 1.1Overall Technical Efficiency, Pure Technical Efficiency and
Scale Efficiency of India MFIs in 2008
MFI no MFI Name OTE PTE SE
1 Adhikar 0.812 0.822 0.782
2 Ajiwika 1 1 1
3 AML 0.897 1 0.897
4 Arohan 0.834 0.844 0.943
5 Asirvad 1 1 1
6 Asomi 1 1 1
7 Bandhan 0.885 1 0.885
8 BASIX 0.704 0.873 0.807
9 BISWA 1 1 1
10 BJS 0.796 1 0.796
11 BSS 0.918 0.948 0.968
12 BWDAFinance 0.749 1 0.749
13 CashporMC 1 1 1
14 Disha 1 1 1
15 Equitas 1 1 1
16 ESAF 0.904 1 0.904
17 GFSPL 1 1 1
18 GOF 0.719 0.753 0.955
19 GramaVidiyalMicrofinanceLtd. 0.686 0.979 0.701
20 GU 1 1 1
21 HiH 1 1 1
22 IDFFinancialServices 0.78 0.82 0.951
23 IndurMACS 0.693 0.71 0.976
24 Janodaya 0.667 0.692 0.964
25 KBSLAB 0.674 0.73 0.922
26 Mahasemam 1 1 1
27 Mahashakti 0.83 0.904 0.918
28 MimoFinance 0.746 0.754 0.989
29 MMFL 0.852 0.938 0.908
30 NBJK 0.922 1 0.922
31 NCS 0.894 1 0.894
(Contd.)
Bhadra, Chaudhuri & Ghosh: Efficiency of MFIs in India during the Crisis Years:... 17
Table 1.1 (Contd.)Overall Technical Efficiency, Pure Technical Efficiency and
Scale Efficiency of India MFIs in 2008
MFI no MFI Name OTE PTE SE
32 NEED 0.778 0.806 0.966
33 Pustikar 1 1 1
34 PWMACS 0.681 0.689 0.988
35 RASS 1 1 1
36 RGVN 0.776 0.782 0.993
37 Samasta 1 1 1
38 Sanghamithra 1 1 1
39 Sarala 1 1 1
40 SarvodayaNanoFinance 1 1 1
41 SCNL 0.642 0.662 0.970
42 SEWABank 1 1 1
43 SHARE 0.934 1 0.934
44 SKDRDP 1 1 1
45 SMILE 1 1 1
46 SMSS 0.798 0.82 0.973
47 Sonata 0.892 0.91 0.98
48 SU 0.717 0.782 0.918
49 Swadhaar 1 1 1
50 SWAWS 1 1 1
51 TridentMicrofinance 0.902 0.943 0.956
52 Ujjivan 0.865 1 0.865
53 VFS 0.716 0.764 0.937
In the same year OTE of inefficient MFIs ranges from 0.642 to 0.934 and averagescore of OTE is 0.796. It implies that set of inefficient MFIs can expand outputon an average by 20.4 per cent. The inefficient MFIs are categorized accordingto which quartile their efficiency values figure in. Table 1.2 shows thedistribution of inefficient MFIs in 2008. Here most inefficient MFIs are thosewhich have OTE score less than first quartile values. Similarly OTE score ofmarginally inefficient MFIs are above the fourth quartile value. The eight mostinefficient MFIs are overall technically inefficient in terms of both the CCR andBCC model. This implies that they are unable to use resources properly toreach the appropriate scale. Managerial skills are lacking in such MFIs. Onthe other hand marginally inefficient ones are very near their peers. Thus withbetter managerial skill such MFIs can move closer to their peers.
18 Prajnan
Table 1.2Distribution of Inefficient MFIs in 2008-09
Most Inefficient MFIs Below Average Above Average Marginally InefficientMFIs MFIs MFIs
In 2008-09, 58 per cent (31 MFIs out of 53) MFIs are efficient in terms of puretechnical efficiency under BCC model. 22 MFIs are operating at optimal scalesize. Among the rest 19 MFIs are operating at decreasing returns to scale (DRS)which implies that those MFIs can improve OTE by reducing scale size. Rest ofthe MFIs (12) are operating at increasing returns to scale (IRS) which impliesthat by increasing scale they can improve their overall technical efficiency.Overall DRS is observed for the sector as a whole which indicates over expansion(one of the causes of the Crisis).
Under output oriented DEA, output slacks represent under production ofrespective outputs and input slacks represent usage of inputs that can bereduced to enhance efficiency of an MFI. Here input slacks provide informationabout the areas where Indian MFIs can enhance their technical efficiency byreducing wastage in input usage. Under output oriented approach OTE scoreindicates the proportional expansion of outputs to be achieved by Indian MFIsto be efficient. For example BASIX, an inefficient MFI has an OTE score of0.704. This implies that it has scope to expand both gross loan portfolio andnumber of women borrowers by 29.6 per cent (1-OTE)*100. This equi-proportionate movement towards efficient frontier is a radial movement. Tobe efficient, the MFI needs to expand number of women borrowers and furtherreduce input usage in terms of deposits, operating expenses and number ofpersonnel. This output expansion and input reductions are known as outputslack and input slack. For all inefficient MFIs total of radial movement andnon-zero output slack indicate the total output expansion mandated for theMFI and the non-zero input slacks shows the scope of input reduction requiredfor the outcome to be Pareto Efficient.
Data Envelopment Analysis also states target values of inputs and outputs forthe MFIs. With OTE score, optimum value of slacks and observed values of
Bhadra, Chaudhuri & Ghosh: Efficiency of MFIs in India during the Crisis Years:... 19
variables the targeted value of variables are calculated. Difference betweentarget value and observed value of output indicates the amount of outputexpansion required for inefficient MFIs to move onto the efficient frontier.Similarly, difference between observed input level and target level inputs showsthe scope for input reduction required to be efficient.
The following table shows distribution of MFIs who have either output slacksor input slacks or both.
Table 1.3Output and Input Slacks in 2008-09
Output Slacks Input Slacks
GLP No of Women Borrowings Equity Deposits Operating No ofBorrower Expenses Personnel
No of MFIs 4 8 1 5 6 13 3
Total 14901724 544073 32254.42 11282176.35 12906210.6 10242854.5 1187
Average 3725431 68009 32254.42 2256425.27 2151035.10 787911.88 396
Scale efficiency is the extent to which an organization can take advantage ofreturns to scale by altering its size towards optimal scale. In Indian microfinancesector it is likely that size of MFIs will influence their ability to provide financialservices to poor people more efficiently. The scale efficiency of an MFI is theratio of the constant returns to the variable returns to scale efficiency score.Average Scale efficiency score of all 53 MFIs are 0.954. So the sector as awhole has the scope to improve scale efficiency by 4.6 per cent.
In 2009-10, 21 MFIs are overall technically efficient under CCR model (Table2.1 in the Appendix). Average OTE score is 0.876. The sector thus has thepotential to expand output by 12.4 per cent on an average. Table 2.2 shows thedistribution of inefficient MFIs in 2009-10.
Table 2.2Distribution of Inefficient MFIs in 2009-10
Most Inefficient MFIs Below Average Above Average Marginally InefficientMFIs MFIs MFIs
Among inefficient MFIs, Mimo Finance and SCNL are categorized in the firstcategory due to improper input usage and lack of proper management skills.Similarly in the second category Sonata is scale efficient. In the third categoryBandhan and Ujjivan have used efficient management skill with improper scalesize and input usage. In the fourth category, AML and Equitas are like Bandhanand Ujjivan. As we gradually move from first to fourth category, along withinput wastage reduction, scale improvement and better managerial skills areobserved for the sample. Among inefficient MFIs, 5 are efficient in terms ofpure technical efficiency and three are scale efficient. Out of 32 overalltechnically inefficient 9 are operating at IRS and 20 are operating at DRS.Similar to 2008, even in this year DRS is observed to be predominant traitimplying scale reduction to improve efficiency. This is in conflict with theoutreach objective which requires higher scale.
Total of radial and slack movement of output and inputs gives scope of outputexpansion and input reduction to be Pareto efficient. The following Table 2.3shows the slack of all inefficient MFIs in the year 2009-10.
Table 2.3Output and Input Slacks in 2009-10
Output Slacks Input Slacks
GLP No of Women Borrowings Equity Deposits Operating No ofBorrower Expenses Personnel
No of MFIs 6 10 0 4 3 18 19
Total 91346.16 728691 0 8771257.79 6465695.85 3300548.76 6756
Average 15224.36 72869 0 2192814.45 2155231.95 1833363.82 365
Comparing the values with the Table for the year 2008-09 a significant differenceobserved is the slack on borrowing. In 2009-10 the slack of borrowing is zero.This implies no scope to reduce borrowing to improve efficiency. The reasonfor this may be drastic fall in new borrowings by MFIs from Banks. The Bankswere reluctant to lend to the MFIs due the Crisis in Kolar, Krishna and otherplaces in the South of India. The technical efficiency and scale efficiency for53 MFIs for period 2010 are reported in Table 3.1 (Appendix)
In the year 2010-11, average OTE score of the sector is 0.914. Out of 53 MFIs,24 MFIs are efficient in terms of OTE. All the efficient MFIs managed to controlinput wastage with efficient managerial skill and appropriate scale size. So,average OTE score implies that in the year 2010-11, the sector as a whole canimprove its performance by 8.6 per cent on an average. Thus the sectorrecovered from the crisis very quickly. The following Table 3.2 showsdistribution of inefficient MFIs.
Bhadra, Chaudhuri & Ghosh: Efficiency of MFIs in India during the Crisis Years:... 21
Table 3.2Distribution of Inefficient MFIs in 2010-11
Most Inefficient MFIs Below Average Above Average Marginally InefficientMFIs MFIs MFIs
Adhikar BWDA Finance Arohan AML
BASIX GFSPL BISWA Bandhan
BSS GOF ESAF Indur MACS
PWMACS KBSLAB Mahashakti MMFL
Samasta SCNL Mimo Finance NEED
Sonata SHARE SMSS RGVN
SU Ujjivan Swadhaar Trident Microfinance
VFS
Among inefficient MFIs, seven are managed optimally and only one MFI isoperating at optimum scale size. Total of 19 MFIs are operating at DRS and 9are at IRS. So, during the study period the sector is dominated by DRS. Itimplies that the sector as a whole can improve overall efficiency by reducingscale size.
Table 3.3Output and Input Slacks in 2010-11
Output Slacks Input Slacks
GLP No of Women Borrowings Equity Deposits Operating No ofBorrower Expenses Personnel
No of MFIs 4 7 0 2 2 9 11
Total 3462280.97 691649 0 2160494.31 4626846.38 34566511.27 10158
Average 865570.24 98807 0 1080247.15 2313423.19 3840723.47 923
The above table shows output and input slacks of inefficient MFIs during theyear 2010-11. A comparison of the slacks across the years in presented inTable 4.1.
22 Prajnan
Table 4.1Average Output and Input Slacks between 2008-09 and 2010-11
Output Slacks Input Slacks
Average GLP No of Women Borrowings Equity Deposits Operating No ofBorrower Expenses Personnel
The above table shows that average output slack in terms of women borrowersis gradually increasing over the study period but in case of GLP, we found Vshaped movement. In case of input slacks, average borrowings and equity aregradually decreasing and average deposits are more or less same during thetime period. In case of average operating cost and number of personnel, theyhave increased remarkably in 2010-11 compare to 2008-09. So we can concludethat the MFIs have given emphasis on efficient capital usage to control financialcost of operation when they faced the crisis. Reluctance of the banks to lend tothe sector following the Crisis may have contributed to this behaviour of theMFIs to ensure sustainability. But in case of operating cost and employment ofpersonnel, decision have not been optimal. Scope of expansion of the sector interms of no of women borrowers and GLP shows that there is a scope to increaseoutreach without compromising on the efficiency provided input decisions areoptimal.
Table 4.2Efficiency Measures Across the Years
Year OTE PTE SE
2008 0.88 0.923 0.954
2009 0.876 0.904 0.969
2010 0.914 0.94 0.973
The above Table 4.2 shows that during the study period representative Indianmicrofinance sector is gradually improving its overall technical efficiency, puretechnical efficiency and scale efficiency. Crisis only slightly impacted the OTEand PTE.
Determinants of the Efficiency ScoresWe have attempted to build a suitable statistical model to identify the impactof various factors on the technical efficiency scores of the MFIs. In this paperwe are working with a balanced panel of 53 MFIs over three years. The efficiencyscore of MFIs are positive fraction lying between 0 and 1. Thus a censored
Bhadra, Chaudhuri & Ghosh: Efficiency of MFIs in India during the Crisis Years:... 23
regression model has been used to analyze the data. The outcome variable, isthe overall technical efficiency score of MFI which lies between 0 and 1. So theoutcome variable is both right-censored and left-censored. Tobit model hasthus been used to predict OTE scores of MFIs within the specified range.
A random effects model is used to estimate the relationships. The rationalebehind random effects model is that, unlike the fixed effects model, the variationacross entities is assumed to be random and uncorrelated with the predictoror independent variables included in the model. Here we assume that differencesacross MFIs have some influence on efficiency of the institutions, so we haveapplied random effect model for estimation of unknown parameters. Further,another advantage of random effects model is that, it allows the use of timeinvariant variables which is absorbed by intercept term in a fixed effects model.
The linear regression model with panel-level random effects is
Yit = βXit + ui + εit
For i = 1 to 53 panels and t = 1 to 3. The output includes the overall andpanel-level variance components (labeled sigma ε and sigma u, respectively)together with ρ (labeled rho). Xit is the vector of predictor variables. Yit is theefficiency score of the sample MFIs.
ρ = σu2/(σu
2+σε2),
which is the contribution in percentage to the total variance of the panel-levelvariance component. When rho is zero (sigma_u=0 means no variance in theunit effect, which means they all have the same intercept), the panel-levelvariance component is unimportant, and the panel estimator is not differentfrom the pooled estimator.
Table 5.1Descriptive Statistics for the Year 2008
Variables Number of Max Min Mean SDObservations
GLP 53 239270983 226360 23236332 42584932
No of Women Borrowers 53 1502418 416 171547 304602
Borrowings 53 190990185 0 20010707 38749542
Equity 53 35003279 -396646 3998281 6580834
Deposits 53 22721053 0 1643886 4546930
Operating expense 53 20903165 0 2395775 4247465
No of personnel 53 4259 19 783 1092
24 Prajnan
Table 5.2Descriptive Statistics for the Year 2009
Variables Number of Max Min Mean SDObservations
GLP 53 376593362 551901 45529217 82838867
No of women borrowers 53 2357456 2075.184 271593 490462.8
Variables Growth in Growth in2009 over 2008 2010 over 2009
GLP 95.94 27.19
No of women borrowers 58.32 24.32
Borrowings 115.53 -87.43
Equity 115.50 28.58
Deposits 93.56 36.60
Operating expense 84.06 21.40
No of personnel 33.96 31.17
Descriptive statistics (Tables 5.1 to 5.3) shows that the sector was not muchaffected in the year 2009. In fact the mean value of all the variables increased
Bhadra, Chaudhuri & Ghosh: Efficiency of MFIs in India during the Crisis Years:... 25
substantially. If we on the other hand look at the growth of the variables (Table5.4) it is apparent that growth of all the variables has reduced substantially in2010. The change in borrowings was negative in this period. This can be dueto two reasons. Following the crisis the public sector banks reduced lending tothese institutions substantially. The institutions themselves consolidated theirbalance sheets to improve efficiency.
The results of panel estimates are presented in the Appendix. Table 6 showsthat sigma_u is significantly different from zero. It implies that random effectpanel estimator's are significantly different from pooled estimators. Thusrandom effects panel estimation is the appropriate method in this case.Additionally, Breusch and Pagan Lagrangian multiplier test for random effectsupports its usage.
Table 6.1Breusch and Pagan Lagrangian Multiplier Test for Random Effects
Test: Var(u) = 0
chibar2(01) = 17.62
Prob> chibar2 = 0.0000
26 Prajnan
The Wald test statistics rejects the null hypothesis that the parameters in theregression equation are jointly equal to zero. Modified Bhargava et al. Durbin-Watson test is used to test for autocorrelation in the model. The result indicatesthat there is no autocorrelation in this model. The tests also show that there isno heteroscedasticity at the panel level. Table 7 reports the diagnostic checksfor model results reported in Table 6.
Table 7Tests for Significance of Estimated Parameters,
Autocorrelation and Heteroscedasticity
Independent Variable Coefficient Std. Error t Prob.
Constant 0.5033433 0.000
Wald (χ2(7)) 22.02 0.0025
Rho (ρ) 0.4595
Log likelihood 10.095
Number of observation 149
Levene, Brown and Forsythe Heteroscedasticity Test (panel level MFI-wise):W50 = 0.57261079df(52, 106) Pr > F = 0.98646751
Results reported in Table 8 show that capital-asset ratio, debt-equity ratio,loan per officer and personnel allocation ratio of MFIs are major determinantsof efficiency of MFIs at 5 per cent level of significance. Along with these factorsaverage loan balance per borrower/GNI per capita is another significantcontributor to efficiency of MFIs at 10 per cent level of significance. Financialexpense-asset ratio and profit margin are not significant predictors.
Table 8Random Effects Tobit Results for NBFI vs. NGOs
NBFI Coeff P>|z| NGO Coeff P>|z|
Number of obs 73 58
Wald chi2(7) 26.05 20.55
Prob>chi2 0.0005 0.0045
OTE
Capital-asset ratio 0.695 0.002 -0.010 0.958
Debt-equity ratio 0.012 0.039 0.003 0.155
Average loan balance per 0.1913 0.718 0.002 0.997borrower / GNI per capita
Loan per loan officer 0.0002 0.001 0.0001 0.021
Personnel allocation ratio 0.2913 0.044 0.582 0.007
Financial expense/asset 1.556 0.020 -2.113 0.075
Profit margin -0.046 0.517 0.063 0.262
constant 0.192 0.334 0.694 0.001
sigma_u 0.083 0.000 0.078 0.035
sigma_e 0.089 0.000 0.121 0.000
rho | 0.466 0.296
Bhadra, Chaudhuri & Ghosh: Efficiency of MFIs in India during the Crisis Years:... 27
Capital-asset ratio measure whether the MFI's capital is sufficient to supportits risk weighted asset. Higher capital asset ratio implies higher overall efficiencyscore for MFIs. It is the ratio of capital of an MFI to its risk weighted asset.Higher value of the ratio implies lower risk and it contributes positively to theefficiency score of MFIs. Similarly, debt-equity ratio or leverage ratio is ameasure of financial health of MFIs. Higher the ratio implies higher pressureto payback the debt and it leads to intensive use of debt. It tends to moreproductive use of debt capital and ultimately leads to higher technical efficiencyscore. Personnel allocation ratio is a ratio of number of loan officers to totalpersonnel of the institution. Positive sign of the coefficient implies that higherthe number of loan officers (A loan officer is a staff member who is directlyresponsible for arranging and monitoring client loans) higher is the efficiencyof MFIs. Loan officers employed in field contributes not only by increasing theoutput but also control of risk through gathering of detail information aboutcurrent and future borrowers. It also leads to intensive monitoring of projectfinanced by the MFI which ensure payback of loan taken by the borrowers.Thus higher number of loan officers helps reduce risk of default and that mayenhance the productivity of MFIs. Loan per loan officer is the ratio of numberof loans outstanding to the number of loan officers. Higher value of the ratioimplies higher number of outstanding loans handled by a loan officer. Thismay lead to reduction in loan handling costs and higher level of averageproductivity of a loan officer. As a result it is expected to enhance productivityas well as efficiency of MFIs. Average loan per borrower/GNI per capita is ameasure of depth of outreach. The result shows that depth of outreach is animportant determinant of efficiency of MFIs.
Further we have also tried to find out whether the results vary according to thelegal status of the MFIs. For this we have compared the results for the twolargest groups of MFIs the NGOs and the NBFIs (Table 8, Appendix).
Both the models are significant and random effect Tobit model is justified forboth. In both the cases personnel allocation ratio and loan per loan officer aresignificant contributors to efficiency. But coefficients of personnel allocationratio for NGOs are more than NBFI. That implies that marginal contributionof personnel allocation ratio to overall technical efficiency of institutions ismore in case of NGOs than NBFIs.
On the other hand capital-asset ratio and financial health of institution aresignificant contributors in case of NBFI but not in case of NGOs. Profit marginand depth of outreach are not significant in both the cases. But in case offinancial expense-asset ratio it positively impacts efficiency of NBFIs in case ofNGOs the impact is negative. As NGOs are more dependent on donations thanNBFIs it leads to reduction of financial expenses for them. This capital structureis one of the reasons for higher overall efficiency of NGOs. So if the financialexpenses increase, that will lead to lower efficiency for NGOs. The interceptterm is significant for NGOs but not for NBFIs. For NBFI, a non-significant
28 Prajnan
intercept term implies that efficiency of NBFIs is almost fully explained by themodel. But in case of NGOs some unexplained factors are still relevant.
Section IVConclusion
On the basis of the sample selected for the study, on an average Indian MFIsare technically inefficient during the study period of 2008-09 to 2010-11, underboth CRS and VRS. The crisis seems to have increased this inefficiency slightly.But the sector has strongly bounced back in the year 2010-11. Also the sectoris operating at sub-optimal scale in the same period. Under output orientedapproach, the sector has enough scope of expansion of operation by servingmore women borrowers with higher loan portfolio. The regression analysishelps us to find out the sources of inefficiency of Indian MFIs. Lack of adequateplanning regarding usage of different categories of capital and manpower canenhance operating costs. Higher operating costs are a major source ofinefficiency of Indian MFIs during 2008-09 to 2010-11. During the crisis non-repayment of loans coupled with increased efforts to get them back might havecaused inefficiencies in many MFIs. Further reduction in loans by banks mighthave caused reduction in output as well. The Andhra Microfinance Bill onlyadded to the misery by restricting the MFIs to reach the optimum scale ofoperation.
Another finding of the paper is about the relation of regulation and efficiencyin Indian microfinance sector. We found that NGOs subject to non-prudentialregulation are efficient compare to other types of MFIs. Thus regulation is animportant concern. On one hand it secures the fate of the borrowers but onthe other hand it may severely restricts the scale of operation and optimumpricing of credit products.
B and Stephens, M (2005), "Measuring Performance of Microfinance Institutions: AFramework for Reporting, Analysis and Monitoring", USA, The Seep Network andAlternative Credit Technologies, LLC.
2. Barth J R; Caprio G and Levine R (2001), "The Regulation and Supervision of BanksAround the World: A New Database", Policy Research Working Paper 2588, WorldBank, USA , April 2001.
3. Baumman, T (2005), "Pro-poor microcredit in South Africa: Cost-Efficiency andProductivity of South African Pro-Poor Microfinance Institutions", Journal ofMicrofinance, 7(1), pp 95-118.
4. Bhattacharyya, A and Pal, Sudeshna (2013), "Financial Reform and Technical Efficiencyin Indian Commercial Banking: A Generalized Stochastic Frontier Analysis", Reviewof Financial Economics, 22(3), pp 109-117.
Bhadra, Chaudhuri & Ghosh: Efficiency of MFIs in India during the Crisis Years:... 29
5. Chortareas, Georgios E; Girardone, Claudia and Ventouri, Alexia (2012), "BankSupervision, Regulation, And Efficiency: Evidence From The European Union", Journalof Financial Stability, 8(2012), pp 292-309.
6. Coelli, T (1996), "A Guide to DEAP Version 2.1: A Data Envelopment Analysis(Computer)Program, CEPA working paper 96/08, Centre for Efficiency and Productivity Analysis,Department of Econometrics, University of New England, web:http://www.une,edu.au/econometrics/cepm.htm.
7. Conning, J (1999), "Outreach, Sustainability And Leverage In Monitored And Peer-Monitored Lending", Journal of Development Economics, 60(1), pp 51-77.
8. Cull, R; Kunt, A D and Morduch, J (2007), "Financial Performance And Outreach: AGlobal Analysis Of Lending Micro Banks", The Economic Journal, 117(1), F107-F133.
9. Cull, R; Kunt, A. D and Morduch, J (2009), "Microfinance Meets The Market", Journalof Economic Perspectives, 23(1), pp 167-192.
10. Farrington, T (2000), "Efficiency in Microfinance Institutions", Micro Banking Bulletin.
11. Guitierrez-Nieto B; Serrano-Cinca, C and Molinero, C M (2006), "MicrofinanceInstitutions and Efficiency", International Journal of Management Science, 35(2),pp 131-142.
12. Hassan, M K and Tuffe, D R (2001), "The X-Efficiency of A Group Based LendingInstitution: The Case of Grameen Bank", World Development, 29(6), pp 1071-1082.
13. Hermes, N; Lensink R and Meesters, A (2009), "Financial Development And TheEfficiency Of Microfinance Institutions". Online resource available at: http://papers.srn.com/ sol3/papers.cfm?abstract _id 1396202 (24 Sept,2012).
14. Hulme, D and Mosley, P (1996), "Finance Against Poverty" Volumes 1 and 2, London:Routledge.
15. Lafourcade A L; Isern J; Mwangi P and Brown M (2005), "Overview Of The OutreachAnd Financial Performance Of Microfinance Institutions In Africa", The MIX marketInc, Pdf available:http://www.griequity.com/resources/industryandissues/financeandmicrofinance/Africa_Data_Study.pdf
16. Lapenu, C and Zeller, M (2001), "Distribution, Growth, and Performance of TheMicrofinance Institutions In Africa, Asia And Latin America", FCND Discussion paperno 114, Food Consumption and Nutrition Division, International Food Policy ResearchInstitute, USA.
17. Leon, J V (2001), "Decentralized Efficient Organizations of Microfinance: The Case ofThe Peruvian Municipal Banks", Working Paper Series. Wittenberg University, Ohio.
18. Lozano-Vivas, A (1998), "Efficiency And Technical Change for Spanish Banks", AppliedFinancial Economics, 8, pp 289-300.
19. Microfinance: An Emerging Investment Opportunity, Deutsche Bank Research (2007).
20. Nieto, B; Serrano- Cinca, C and Mar Molinero, C (2009), "Social Efficiency inMicrofinance Institutions", Journal of the Operational Research Society, 60,pp 104-119.
21. Paxton, J and Cuevas, C (2002), "Outreach and Sustainability of Member- Based RuralFinancial Intermediaries". in: M Zeller and R L Meyer (eds), "The Triangle ofMicrofinance: Financial Sustainability, Outreach, And Impact", Baltimore and London:Johns Hopkins University Press, 2002.
30 Prajnan
22. Reynolds, D and Thompson, G (2002), "Multiunit Restaurant Productivity Assessment:A Test of Data-Envelopment Analysis", The Center for Hospitality Research at CornellUniversity Report, Ithaca, NY: Cornell University.
23. Rosenberg, R (1994), "Beyond Self Sufficiency: Licensed Leverage and MicrofinanceStrategy", Mimeo: Consultative Group to Assist the Poorest.
24. Sanchez, R (1997), "Financial Efficiency and Economic Growth: The Case of Spain",International Advances in Economic Research, 3, pp 333-351.
Appendix
List of MFIs
MFIs Full Name of MFIs
Adhikar Adhikar Microfinance Private Ltd
Ajiwika Ajivika Society
AML Asmitha Microfin Ltd
Arohan Arohan Financial Services Private Ltd
Asirvad Asirvad Microfinance Private Ltd
Asomi Asomi Finance Private Ltd
Bandhan Bandhan Financial Services Private Ltd
BASIX Bhartiya Samruddhi Investments and ConsultingServices (BASICS Ltd) - holding company
Determinants of Non-Performing Loans inIndia: A System GMM Panel Approach
Asit Ranjan MohantyBinay Ranjan DasSatyendra Kumar
The objective of the resent study is to investigate the determinants ofNon-Performing Loans (NPLs) of the Indian banking system for theperiod 2000-01 to 2015-16. This study utilized the system-GMM panelestimation method. This method reduces finite sample bias and anyother imprecision by regressing levels and changes in NPLs of its lagsand other explanatory variables using lagged levels as instruments.The major findings of the study are as follows; (a) amongmacroeconomic variables; economic growth, stock market index andmarket capitalization ratio have negative impact on the Gross NPLsratio, whereas, expansionary fiscal policy escalates the Gross NPL ratio.(b) Corporate specific variables; net sales growth (SLGC) and net profitmargin (NPMC) have statistically negative impact on the Gross NPLratio. (c.) Bank specific variables; higher credit deposit ratio, growthin bank branches, higher return on equity and higher CRAR will lowerGross NPL ratio. Higher operating expense ratio has significant positiveimpact on the NPLs, which is indicative of inefficiency of the banks. Byexamining the impact of corporate specific variables (e.g., SLGC andNPMC) on NPLs in India, the study has established a link between thebalance sheet of banks and corporate sectors for the first time. It isestablished that strengthening the balance sheet of private corporatesectors will strengthen the balance sheet of banks by lowering the NPLs.Therefore, it resolves the twin balance sheet problem in India.
Keywords: Non-Performing Loans, Macroeconomic variables, Bank specificvariables, Corporate specific variables, System-GMM approach,Indian banking system
JEL Classification: C23, G21, G28, O16
Prof Asit Ranjan Mohanty ([email protected]) is a Professor in Finance, Xavier UniversityBhubaneswar, Odisha.
Binay Ranjan Das ([email protected]) is a Research Associate in Centre of Excellencein Fiscal Policy and Taxation (CEFT), Xavier University Bhubaneswar, Odisha.
Satyendra Kumar ([email protected]) is a Research Associate in Centre of Excellence inFiscal Policy and Taxation (CEFT), Xavier University Bhubaneswar, Odisha.
38 Prajnan
Section IIntroduction
Financial stability is one of the key fundamentals which endorses rapid andsustained economic development. Besides ensuring the most productiveallocation of investible funds in the economy, it is also thwarts any externalthreats that an economy may face any time (Dutta et.al. 2013). The globaleconomic recession in 2008-09 was one of such threats that worsened thefinancial stability across the world. Among the various financial stabilityindicators, non-performing loans (NPLs) is considered as most critical indicator,because it is closely linked with problems pertaining to the financial system.In this context, Us (2017) found a strong correlation between the higher levelof NPLs and banking crises. NPLs not only adversely affects the asset qualityof the banks but also the efficiency in resource allocation and the credit riskmanagement which further depreciate their profitability, liquidity of banks,and creditworthiness of the borrowers (Michael, et al, 2006; Ghosh, 2015;Us, 2017).
NPLs may increase deposit liabilities of the banks and reduce the availabilityof bank credit for the private sector, and thereby, hampering the privateinvestment, Fofack (2005). According to Global Financial Stability Report (April,2017)1, increasing non-performing loans are reflecting various economicchallenges, for instance, economic weakness (specifically, in Russia and Brazil),sector-specific slumps in India and corporate leverage growth in China. Inresponse, banks have also raised provisioning level, but that was not enoughto keep pace with the increasing of bad loans.
In India, the concept of NPL emerged as one of the prime issue after therecommendation of the Narasimham Committee Report (1991)2 because thecommittee was highlighted its influences on the financial health of the banks.The committee reported that the high level of NPL has been the proximatecause of the low profitability levels of the banks. According to the report, themajor causes of high level NPL are poor credit decisions in the banks'management, inconsistency in the recovery management of the banks, andcyclical and structural changes in the economic environment.
In 1996-97, the gross NPLs of all scheduled commercial banks was Rs. 473billion (i.e., 15.7 per cent of their gross advances) in which the contribution ofpublic sector banks was Rs. 435.77 billion (i.e., 92 per cent of total gross NPLof all scheduled commercial banks). However, due to implementation of therecommendations of the Narsimham committee and other corrective measures
1. International Monetary Fund, (2017), Global Financial Stability Report: Getting the Policy MixRight, Washington, DC, April.
2. Government of India, Narasimhan Committee Submits its report to FM, retrieved fromhttp://pib.nic.in/focus/fomore/narsim.html.
Mohanty, Das & Kumar: Determinants of Non-Performing Loans in India 39
taken by RBI, the gross NPL declined to its lowest level at 2.26 per cent (i.e.,Rs. 566.06 billion) of gross advances in 2007-08. After the global economiccrisis, all bank-groups witnessed a sharp increase in their gross NPL ratio. Inthe year 2016, the gross NPL ratio was 6.4 per cent for SBI and its associates;10.7 per cent for Nationalised banks; 2.8 per cent for private sector banks;and 4.2 per cent for foreign banks. According to RBI's Financial Stability Report(June, 2017), deterioration of asset quality and profitability of banks hasworsened the banking stability indicator between September 2016 to March2017. The worsening financial health of banks may because of minimum capitalrequirement as stipulated in Basel norms. In addition, it indicates that NPLsof all scheduled commercial banks may surge to 10.2 per cent (this may riseto 11.2 per cent in severe stress scenario) by March 2018 from its correspondingvalue of 9.6 per cent in March 2017.
The Economic Survey of India (2016-17) has attributed "Twin Balance Sheet(TBS) Problem" for the worsening of asset quality of Banks. 'TBS' is a situationwhere both the banking and corporate sectors are under stress. Over leveragedcorporates are unable to service their debts and invest more. Because of thehuge bad loans in the balance sheet, banks are unwilling to lend more andstruggle to keep up their business. TBS problem, essentially, is due to theweak balance sheet of both banks and private corporate sectors that has ledsubdued supply and demand for loans. As a result, the economy has furtherslowed down.
Under this backdrop, an attempt has been made to reinvestigate the keydeterminants of NPLs in the Indian banking system for the period 2000-01 to2015-16. The Asian financial crisis in 1997 and nuclear blasts in 1998 havebrought structural changes in the Indian economy (Sengupta and Vardhan,2017). In the 2000s, the banking sector also witnessed structural changes.During 2003-07, the bank credit has grown at a staggering rate of 25 per centalong with the higher volume of NPLs, which posed alarming concern for thebanking sector in India. According to Us (2017), the global financial crisisalso played a role of the catalyst for enhancement of the NPLs in the bankingsystem. Aforementioned circumstances have motivated us to take-up the studyperiod 2000-01 to 2015-16. Apart from macroeconomic and bank specificvariables, corporate specific variables are also introduced in the study, namely,net sales growth (SLGC) and net profit margin (NPMC). This is because higherSLGC and NPMC generates more cash flows for the private corporate sectorsand thereby, increases loan repayment capacity during the scheduled time.Therefore, SLGC and NPMC as major corporate specific variables may havesignificant impact on NPLs.
The system-GMM estimation developed by Arellano and Bover (1995), Blundelland Bond (1998) is used in the present study, because the study constitutes ofmicro panel data. The system-GMM estimator has the superiority over otherestimators when the number of individuals are small and largely unknown(Soto, 2009).
40 Prajnan
Rest of the paper is divided in to four sections. While Section II deals with theliterature review, data and methodology used in the study is described in theSection III. Section IV explains the empirical results, whereas Section Vconcludes the finding of the study along with policy implication.
Section IILiterature Review
The existing economic literature on NPL suggests that the determinants ofNPL may be categorised in two groups, viz., macroeconomic and bank specificdeterminants (Louzis et al., 2012; Klein, 2013; Makri et al. 2014; Ekanayakeand Azeez, 2015; Ghosh, 2015; Dimitrios et al., 2016; Patra and Padhi, 2016;Rajha, 2016; Amuakwa-Mensah et al., 2017; Kjosevshi and Petkovski, 2017).Under the macroeconomic variables, economic growth (GDP), interest rate,fiscal deficit ratio, stock market index, market capitalization ratio and othervarious macroeconomic factors (e.g., inflation, exchange rate, unemploymentrate and public debt, etc.) are vital determinants of NPL. In addition, underthe bank specific variables, credit deposit ratio, capital adequacy ratio,operating expenses ratio, returns on equity, ratio of priority sector advancesto total advances, growth in bank branches and various other bank specificindicators (e.g., net interest margin, bank size and returns on assets, etc.) areimportant determinants of NPL.
Higher GDP growth usually deciphers more income, which will boost debtservicing ability of borrowers, and thereby, contributes to the reduction ofNPLs and vice versa (Nkusu, 2011; Khemraj and Pasha, 2009). There aresubstantial amount of empirical findings that approved the inverse relationshipbetween economic growth and NPLs (Castro, 2013; Bofondi and Ropele, 2011;Tanaskovi? and Jandri?, 2015; Abid et al., 2014; Louzis et al., 2012; Ghosh,2015; Makri et al., 2014; Bhattarai, 2014; Haniifah, 2015; Ekanayake andAzeez, 2015; Das and Ghosh, 2007; Reddy, 2015; Nkusu, 2011; Messai andJouini, 2013; Tucker, 2013; Rajha, 2016; Kjosevski and Petkovski, 2017;Prasanna et al., 2014; Roy, 2014).
The lending rate can be used as a proxy for interest rates charged againstloans and advances (Rajha, 2016). Higher interest rate weakens repaymentability of borrowers and escalates their debt burden further. Therefore, lendingrate and NPLs may have a positive association. There are numerous empiricalevidences that suggest a positive relation between lending rate and NPLs(Khemraj and Pasha, 2009; Adebola et al., 2011; Farhan et al. 2012; Abid etal., 2014; Ekanayake and Azeez, 2015; Ghosh, 2015; Patra and Padhi, 2016).
It is expected that higher gross fiscal deficit adds to the existing borrowingsresulting into accumulation of debt and rise in outstanding debt of thegovernment. According to Dimitrios et al. (2016) expansionary fiscal policy
Mohanty, Das & Kumar: Determinants of Non-Performing Loans in India 41
may alleviate or worsen the NPLs problem of the economy. Patra and Padhi(2016) found a positive correlation between fiscal deficit and NPL. An analysisof the euro-area banking system has suggested that output gap and personalincome tax are key drivers of NPLs (Dimitrios et al., 2016).
Apart from the above findings, Nkusu (2011) scrutinized the nexus betweenNPLs and macroeconomic variables of 26 advanced economies over the period1998 to 2009. In his panel analysis, he found that unemployment rate andpolicy rate have positive impact on NPLs, whereas house prices and equityprices have negative impact on the same. Bofondi and Ropele, (2011) alsoconcluded that house prices and NPLs of the Italian banking system areinversely related and unemployment rate is positively related to NPLs for theperiod 1990:Q1 to 2010:Q2. Louzis et al. (2012) have found that NPLs in theGreek banking sector to be majorly determined by GDP, interest rate,unemployment rate, public debt, and management quality. Castro (2013) foundthat housing price and share price indices are inversely related to the bankingcredit risk, whereas, real exchange rate, unemployment rate and credit growthare positively related for the Sub-Saharan African countries. Makri et al. (2014)used difference GMM approaches and found that return on equity and bankcapital ratio have negative impact on NPLs, while lag of NPL ratio, public debtratio and unemployment have positive impact in Eurozone's banking system.
In the Tunisian banking system, the key determinant of the NPLs are GDP,inflation, interest rate and bad management (Abid et al., 2014). An empiricalanalysis observed that the foreign currency loan ratio and exchange rate havepositive impact on NPLs for Central and Eastern and South-Eastern Europe(CESEE) economies (Tanaskovi? and Jandri? 2015). In his analysis on thebanking system of 50 US states and the District of Columbia, Ghosh (2015)concluded that among bank specific variables; liquidity risks, highercapitalization, poor credit quality, banking industry size, and greater costinefficiency have positive influence on NPLs, whereas higher bank profitabilityhas negative impact. Under macroeconomic specific variables; real personalincome growth and state housing price index are negatively related to NPLs,while inflation, US public debt, and state unemployment rates are positivelyrelated to NPLs.
Rajha (2016) found that inflation has negative impact on NPLs, whereas globalfinancial crisis has positive impact on NPLs in Jordanian Banks over the period2007 to 2012. For the period 2005 to 2014, Kjosevski and Petkovski (2017)analyzed an unbalance panel data of 27 banks of the Baltic States. They foundthat return on assets and return on equity are negatively related to NPLs, whileinflation, unemployment, and gross loan growth are positively related to NPL.Net interest margin and cost-income ratio are vital indicators of bank distresswhere the higher cost-income ratio indicates a poor quality of managementand net interest margin designates profitability (Gerhardt and Vennet, 2017).The findings of Gerhardt and Vennet (2017), implicitly imply that the higher
42 Prajnan
cost-income ratio and low net interest margin lead to higher NPLs. In theTurkish banking system, ownership of banks is a major determinant of NPLs(Us, 2017). Caporale et al. (2017) analyzed a panel data of 400 Italian banksto examine the key determinants of loan-loss provision for the period 2001 to2015. They used discretionary components (e.g., income smoothing and capitalmanagement) and non-discretionary components (which were related tobusiness cycle). They found that non-discretionary components andmacroeconomic shocks are the major drivers of loan loss provision in Italianbanks.
For the period 1995-96 to 1999-2000, Rajaraman and Vasishtha (2002) founda significant bivariate relationship between the operating efficiency indicatorsand NPLs of 27 public sector banks of India. Ranjan and Dhal (2003) foundthat terms of credit has a significant impact on NPLs in Indian banking systemby considering the bank size and presence of macroeconomic shocks.Additionally, bank size has negative impact on NPL, if it is measured in termsof assets. However, if it is measured in terms of capital, it has significant positiveeffect on NPL. Reddy et al. (2006) evaluated the nexus of NPL and prioritysector lending of public sector banks for the period 1992-2004. By using chi-square technique they found that priority sector lending may not be keydeterminants of NPL. The finding of Das and Ghosh (2007) show that the keydeterminant of the loan problems are real loan growth, GDP growth, bank sizeand operating expenses. An empirical results indicates that priority sectorlending is negatively associated with NPLs in Indian banking sector, Vallabh etal. (2007). Additionally, capital adequacy ratio is positively related with NPLof public sector banks but it is negatively related with NPL of private sectorbanks.
For the period 1995 to 2009, Misra and Dhal (2010) analyzed 26 public sectorbanks of India and observed that share of term loans in total advances, interestrate, share of unsecured loans, and credit-deposit ratio have significant adverseimpact on the NPLs in the presence of macroeconomic shocks. In panel analysis,Swamy (2012) found that per capita income, bank credit growth rate, credit-deposit ratio, index of industrial production and lower cost of fund arenegatively related with NPLs, however, higher operating expense ratio and lowreturn on assets are positively associated with NPL. Prasanna et al. (2014)examined the determinants of the NPLs for 31 Indian banks for the period2000 to 2012 and concluded their finding that economic growth, stock marketindex, per capita income growth, foreign exchange reserves, constructionexpenditure, bank size and performances are inverse relationship with non-performing advances. However, repo interest rate, exchange rate and inefficiencyratio are positively related with non-performing advances in Indian bankingsystem. By utilizing fixed effect model in panel analysis, it is found that economicgrowth, real effective exchange rate and volatility index are inversely related toNPLs (Roy, 2014).
Mohanty, Das & Kumar: Determinants of Non-Performing Loans in India 43
In the case of India, Dhar and Bakshi (2015) examined the effect of bank specificfactors upon NPLs of public sector banks for the period 2001 to 2005. Byusing panel regression technique, they found that quantum of sensitive sectoradvances has positive impact on NPL, however, capital adequacy ratio and netinterest margin have negative impact. Reddy (2015) suggested that return onassets, capital adequacy ratio, priority sector advances, growth in advances,total assets, growth in GDP, are negatively associated with NPL while operatingcost ratio is positively related.
In the view of the above literature, GDP growth, average lending rate (as aproxy of interest rate), gross fiscal deficit ratio, market capitalization ratioand stock market index are examined as macroeconomic determinants in thepresent study; and credit deposit ratio, capital adequacy ratio, operatingexpenses ratio, returns on equity, priority sector advances ratio, and growthin bank branches are inspected as bank specific determinants in Indian bankingsystem. In addition, net sales growth and net profit margin of private corporatesector are introduced as corporate specific variables in this study.
Section IIIData and Methodology
According to Reserve Bank of India, when an asset (including a leased asset) isunable to generate income to the bank, it is classified as non-performing asset.In addition, when interest and/or installment of principal of a loan or an advanceremain overdue for a period of more than 90 days, is called Gross Non-Performing Loan (GNPL).
The objective of the present study to re-investigate the determinants of theNPLs in the Indian banking system for the period 2000-01 to 2015-16. Tofulfill the objective, the study have taken into account four bank-groups of allscheduled commercial banks,4 viz., (i) SBI and its associate banks-group,(ii) nationalized banks-group, (iii) private sector banks-group, (iv) foreignbanks-group. Hence, total ninety five banks are covered in the study, where,SBI and its associate banks-group includes six banks, nationalized banks-group contains twenty one banks, private sector banks-group involves twentytwo banks and foreign banks-group comprises forty six banks. Apart fromthese four bank-groups, the present study also includes three major categoriesof variables. The description of these variables are given in Table 1.
4. All scheduled commercial banks excluding Regional Rural Banks (RRBs).
44 Prajnan
Table 1Description of Variables Included in the Study
Variables Description
Macroeconomic Specific Variables
GDP Annual Growth of Real GDP
FDR Gross Fiscal Deficit to GDP
MCR Market Capitalization Ratio (Market Capitalization as a percentage of GDP)
NIFTY Stock Market Index of National Stock Exchange of India (CNX Nifty)
Bank Specific Variables
GNPLR Gross Non-Performing Loans as a percentage of Gross Advances
CDR Total Credit as a percentage of Total Deposit
LR Average Lending Rate for the Borrowers
ROE Return on Equity ( Net Profit as a percentage of Total Inside Liability)
PSAR Priority Sector Advances as a percentage of Net Advances
CRAR Capital to Risk Weighted Asset Ratio
OPR Ratio of Operating Expenses to Total Assets
GBO Growth in Banks Branches
Corporate Specific Variables
NPMC Net Profit Margin of Private Corporate Sector(Net Profit as a percentage of Net sales)
SLGC Net Sales Growth of Private Corporate Sector
The data utilized in this study are sourced from various sources as follow;(i) Statistical Tables Relating to Banks in India (ii) Handbook of Statistics onIndian Economy (iii) RBI Bulletins, all three are published by RBI (iv) EPWResearch Foundation (v) Annual Reports of Securities and Exchange Board ofIndia (SEBI) (vi) Historical data of National Stock Exchange of India Ltd.
Econometric Estimation ProcedureTo achieve the objective of this study the panel data analysis has been utilized.A time series analysis is carried out when economic agents are homogeneousin nature. However, a cross section analysis provides meaningful insightregarding interlinkages between financial and economic variable by consideringheterogeneity of economic agents and their behaviour (Ranjan and Dhal, 2003).Four bank groups i.e., (i) SBI and its associate banks-group, (ii) nationalizedbanks-group, (iii) private sector banks-group, (iv) foreign banks-group,assessed in the study may be considered homogeneous in nature from theinstitutional perspective, but the functional behaviour (e.g., cost structure,loan portfolio and performance) of each bank across different bank groups
Mohanty, Das & Kumar: Determinants of Non-Performing Loans in India 45
can be heterogeneous in nature (Misra and Dhal, 2010). In this context, apanel analysis would serve the purpose because it incorporates individualcharacteristics along with regularity/continuity in the cross-section.
Blundell-Bond Linear Dynamic Panel ModelSince the study constitutes of micro panel, the system-GMM estimationapproach developed by Arellano and Bover (1995), Blundell and Bond (1998)is considered in this study. According to Blundell and Bond (1998) and Blundellet al. (2000) the system GMM estimator has the superiority over otherestimators, when the number of individuals are small and largely unknown(Soto, 2009). This method reduces finite sample bias and any other imprecisionby regressing levels and changes in NPLs of its lags and other explanatoryvariables using lagged levels as instruments (Amuakwa-Mensah et al., 2017).Blundell and Bond (1998) assumes no autocorrelation in the idiosyncraticerror terms. The initial requirement of this method is that the panel-level effectsto be uncorrelated from the first difference of the first observation of thedependent variable5.
The Blundell-Bond linear dynamic panel model can be defined as:
Where, αj are the p parameters to be estimated, xit are the exogenous variablesand β1 are the coefficient for the exogenous variable, wit are the endogenousvariables and β2 are their coefficients., vi are the panel level effects and εit isthe identically independently distributed for the whole sample with varianceσ2
ε. There is an assumption that vi and the εit are independent for each i andover all t.
Section IVEmpirical Results
This study used Gross NPLs as compared to Net NPLs because it adequatelycaptures the default risk of the borrowers as well as loan default faced by thebanks. Net NPLs is Gross NPLs net of provisions, hence, captured the actualloan default in the balance sheet of the banks. Whereas, Gross NPLS reflectsabsolute value of defaulted loans availed by the borrowers. Therefore, GrossNPLs is a better indicator than Net NPLs for measuring defaulted loans of thebanks.
5. xtdpdsys - Arellano-Bover/Blundell-Bond linear dynamic panel-data estimation, retrieved fromhttps://www.stata.com/manuals13/xtxtdpdsys.pdf.
46 Prajnan
The growth in real GDP (GDP at factor cost at constant prices) is indicative ofrise in income of the economy, hence, reflects the performance of the economy.The rise in income of the economy improves the debt servicing ability ofborrowers.
Fiscal Deficit ratio is taken as one of the macroeconomic variables. Throughhigher fiscal deficit, Government takes away resource available in the marketleaving less resources for corporates. As a result, the volume of business ofcorporates is adversely affected. This in turn, reduces the capacity of thecorporates to service the existing loans taken from the banks.
Rise in the growth in real GDP improves the financial performances of privatecorporate sectors. It induces higher sales growth and net profit margin of theprivate corporate sectors.
When the economy is growing and the financial performances of the privatecorporates are improving, then the stock market remains buoyant. For, stockmarket index, CNX Nifty Index is taken as a proxy.
Market Capitalization at the exchange level shows the size of the corporatesectors listed in that Exchange. Since, National Stock Exchange (NSE) is thebiggest stock market in India, we have taken market capitalization from NSE.The market capitalization indicates the value of the listed firms (corporates)and valuation of the firms as perceived by the investors. The marketcapitalization relative to GDP of India (MCR) is reflective of relative valuationof the listed firms derived from their performances relative to the performanceof the economy. When financial performances of the corporate sectors or firmsimprove, their share price goes up. Besides, the firms mobilize funds by issuingequity instruments. This leads to higher MCR both at firm level and exchangelevel.
Hence, a high collinearly exists between macroeconomic specific variables (GDP,FDR, MCR, and NIFTY) and corporate specific variables (NPMC and SLGC).
Therefore, to capture the impact of macroeconomic variables, corporate specificvariables and bank specific variables on GNPLR separately, we have built sevenmodels using Blundell-Bond Dynamic Panel Estimation. Bank specific variablesare same across all seven models. However, in order to avoid collinearityproblem, macroeconomic and corporate specific variables are usedinterchangeably across all seven models (see for details Appendix-I). Thedetailed estimation is presented in Table 2.
Mohanty, Das & Kumar: Determinants of Non-Performing Loans in India 47
Table 2Blundell-Bond (1998) System-GMM Estimation
Variables Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Model 7
Analysis of Macroeconomic Specific VariablesSince, growth in real GDP (RGDP), stock market index (Nifty) and MarketCapitalization Ratio (MCR) at exchange level are performance indicators of theeconomy, they have a negative effect on GNPLR. When the economy performs,the rise in income enables the borrower to repay the loan due to the banks instipulated time (90 days norms), therefore, recognizing the loan as standardloan (Model 1, Model 4 and Model 5).
However, Gross Fiscal Deficit Ratio (GFDR) as one of the macroeconomicvariables has positive effect on GNPLR (Model 6). Expansionary fiscal policythrough higher Gross Fiscal Deficit by the Government takes away the availablefunds in the economy through borrowings. As a result, fewer funds are availablefor the private corporate sector for their economic activities. This finding is intune with the study made by Dimitrios et al. (2016) and Patra and Padhi (2016).This adversely impacts the business performance of the private borrower; as aresult, they are constrained to service their loans during the stipulated time,leading to higher GNPLR. Marki et al. (2014) established a significantrelationship between higher sovereign borrowings and GNPLR. DaCosta andFoo (2002) examined efforts of China to transform its financial system in orderto support its growing economy. Their findings suggest that a financialframework will not be effectual to support an emerging market economy, iffinancial functions are not distinct from fiscal functions and banks are notindependent in terms of deficit financing and policy directives. The findings ofthis study support this view that an expansionary fiscal policy will induce highernon-performing loans, as a result, the reform measures in banking sector willnot be effective.
Analysis Corporate Specific VariablesBetter performance of the private corporate sectors is captured by the financialindicators such as higher sales growth and higher net profit margin. HigherNet Sales Growth (SLGC) and Net Profit Margin (NPMC) generates more cashflows for the private corporate sectors. This increases the loan repaymentcapacity, and repayment due, is paid during the scheduled time. Therefore,rise in SLGC and NPMC as corporate specific variables reduce GNPLR (Model2 and Model 3).
It is pertinent to note that among macroeconomic and corporate specificvariables, RGDP has a significant negative effect on GNPLR, which impliesthat GNPLR declines with rise in RGDP and vice versa. Therefore, RGDP is asignificant determinant of GNPLR. The results of this study supports thefindings of Salas and Saurina (2002), Ranjan and Dhal (2003), Dash and Kabra(2010). Likewise, during downturn of the economy, the business performancesof the borrowers are adversely affected and borrowers face difficulties in debtservicing due to the banks, resulting into higher NPLs.
Mohanty, Das & Kumar: Determinants of Non-Performing Loans in India 49
Banking Specific VariablesThe lagged dependent variable (GNPLR) is indicative of legacy effect of NPLsmanagement. In all our models, the size of the coefficient is less than 0.5reflecting better NPLs management by Banks (Model 1 to Model 7). Since, theprevious year's GNPLR affects the present year GNPLR, a shock to GNPLRwould take time to reduce across the banking sector.
Priority Sector includes housing, education, social infrastructure, micro, smalland medium enterprises (MSME), agriculture, export credit, renewable energyand others. The Reserve Bank of India data reveals that the contribution toGNPL by the priority sector constitutes 35 per cent, whereas, the share of non-priority sector is 66 per cent6. The priority sector lending is socio-economicresponsibility of the banks. Therefore, banks serve as instruments ofdevelopment for the poorer sections of people. In priority sector lending,national government and many sub national governments in India provideinterest subvention scheme so as to reduce the interest burden of the borrowers.Sometimes, the government waive off the loan availed by the borrowers duringbad crops and repay the debt to the banks on behalf of the borrowers. As aresult, higher Priority Sector Advance ratio (PSAR) has negative effect on GNPLR(Model 7, Table 2).
Return on Equity (ROE) indicates financial performance of banks, because itmeasures the profitability of banks over inside liability (net worth). Profitablebanks are not interested in taking exposure on low credit worthy borrowers. Ifthe credit portfolio has more standard loans as compared to toxic loans, thenROE will have negative impact on GNPLR (Godlewski, 2004). In our analysis,ROE has negative impact on GNPLR in all eight models. Additionally, withhigher ROE, banks provide loans to borrowers with low creditworthiness, whichin turn reduces GNPLR. This suggests that better performing banks managetheir net worth more efficiently. Banks with higher profitability may not takelow rated borrowers in their loan portfolio. Results of the study supports thefindings of Louzis et al., (2012). Besides, Fan and Shaffer (2004) has establisheda link between high credit risk and low profitability.
Growth in Bank Branches (GBO) facilitate the banks to effectively screening ofthe loan applications, and the managers get a space to make backgroundverification of the potential use of funds and repaying capacity of the borrowersthrough proper credit appraisal. GBO, essentially, decentralizes the decisionmaking process of selecting potential borrowers. Hence, in our analysis, risein GBO has a negative impact on GNPLR across all eight models.
Higher operating expenses (non-interest expenses) relative to total assets (OPR)indicates the inefficiency of banks in managing their Income and Expenditure
6. Reserve Bank of India, (2016), Basic Statistical Returns of Scheduled Commercial Banks in India,March.
50 Prajnan
statement. Higher operating expenses as compared to operating income of theinefficient banks induce more GNPLR (Berger and De Young, 1997). High OPRhave fewer resources available to monitor the risk associated with the borrowersduring post disbursement period. Banks inefficiency because of high OPR isleads to poor skills in credit management, post disbursement follow-upmechanism, etc. In all eight models of our analysis, OPR not only has positiveimpact on GNPLR but also the magnitude of coefficient is very high. It supportsbad management hypothesis (Gerhardt and Vennet, 2017).
Higher Credit to Deposit ratio (CDR) indicates higher credit expansion by thebanks as compared to the deposit mobilization. Higher CDR may lead tostressed loans (Caprio and Klingebiel, 1996). However, more expansion of creditas compared to deposit mobilization may not lead bad loans if it is accompaniedby better screening and lending to borrower with low probability of default(low PD). In our analysis, CDR reduces GNPLR in all our eight models. This isattributable to better credit management and post disbursement loan reviewmechanism adopted by the banks.
Lending rate (LR) is the rate of interest charged to the borrowers for availingloans from banks. It is not the lending rate, but the spread, which is crucialfor the banks to generate profit. Spread is the difference between yield onloans and cost of deposits. Along with spread, the higher credit offtakestrengthen the financial performance of the banks. The demand for credit ismainly due to macroeconomic factors and the investment opportunities forthe private corporate borrowers. As long as the economy is growing andcorporate sectors have the opportunity to invest, they will have the repaymentcapacity irrespective of the level of interest rate. In our empirical analysis, thecoefficient of lending rate is statistically insignificant in all eight models. Infact, the Reserve Bank of India uses the policy rate to tackle inflation and tospur GDP growth. The findings of this study contradicts that rise in interestrate worsens the NPL according to the findings of Espinoza and Prasad (2010),and Bofondi and Ropele (2011).
The strength of the bank is repented by Capital (both Tier I and Tier II) to theRisk Weighted Asset Ratio (CRAR). CRAR absorbs unexpected credit loss. Thehigh NPLs level in banks is due to low CRAR (Mukherjee, 2003). High CRAR isalso reflective of low leverage of the banks. High CRAR banks may expand thecredit base after deciding the risk-return profile of the borrowers. Our empiricalresults reveal that higher CRAR has negative effect on GNPLR in all eight models,expect the fourth model, where the coefficient is statistically insignificant. TheNPL management is taken care of by provisioning, instead of depleting thecapital base of the banks. Our result confirms to the studies made by Poghosyanand Cihk, (2011) for the European Union. The banks with low CRAR respondsto low credit worthy borrowers, thereby, increasing risks associated with theircredit portfolio, which further results into higher GNPLR. Our results supportthe findings of Berger and DeYoung (1997) that there is inverse relationship
Mohanty, Das & Kumar: Determinants of Non-Performing Loans in India 51
between CRAR and GNPLR, which is based on moral hazard hypothesis. Thisis because of capitalization in domestic and foreign banks in Oman, Bahrain,Dubai and Abu Dhabi, performance of banks has fared well for the past severalyears (Islam, 2017).
Section VConclusion and Policy Implication
The dynamic panel model approach provides an idea how the macroeconomicspecific variables, corporate specific variables and bank specific variablesdetermine GNPL ratio of different types of banks in India during the periodbeginning in FY 2000-01 and ending in FY 2015-16. In fact, our empiricalframework has not been about why the growth of the NPLs happens. Instead,it examines various variables in explaining the deterioration of asset qualityand rising NPLs. Among macroeconomic specific variables, the economic growthhas greater impact in reducing the Gross NPLs ratio, whereas, expansionaryfiscal policy escalates Gross NPLs ratio. Other macroeconomic specific variablessuch as stock market index and market capitalization ratio of the stock markethave statistically significant inverse relationship with Gross NPLs ratio.
Besides, corporate specific variables such as net sales growth and net profitmargin have statistically negative impact on Gross NPLs ratio. This impliesthat higher economic growth, rise in stock market index and better performancein the private corporate sector will scale down Gross NPLs ratio so that bankscan supply more credit in order to meet the credit demand in the economy.
Bank specific variables such as higher credit deposit ratio, growth in bankbranches, higher return on equity and higher CRAR will lower Gross NPLsratio. On the contrary, higher operating expense ratio will raise Gross NPLsratio. Higher operating expense ratio is indicative of inefficiency of the banks.Operating expenses includes expenses on employee wages and administrativeexpenses. It does not include interest expenses. Hence, higher operatingexpenses as per cent of total assets is indicative of inefficiency of the banks.However, the impact of lending rate on Gross NPLs ratio is not statisticallysignificant. When both macroeconomic specific variables and corporate specificvariables are performing along with other robust bank specific variables asanalyzed above, the impact of lending rate is not significant on NPLs. Ourfindings also belie the fact that priority sector advances generates non-performing loans.
Expecting that Gross NPLs ratio will increase to 10.2 per cent by March 2018,the Government of India is contemplating to infuse Rs.95,000 crore to bolsterthe capital base of the public sector banks7. From our analysis, capital infusion
7. Ministry of Finance, Government of India (2017).
52 Prajnan
is one of the factors that impede the problem of NPLs. However, there areother several factor which have been emerged from our study, are to be factoredinto while addressing deteriorating asset quality of the banks. From policyperspective, investment climate is to be created in order to augment demandfor credit so as to have higher capital formation and economic growth. Withrise in economic growth, the borrower's performance will also improve in termssales growth, net profit margin, etc. As a result, the balance sheet of thecorporates will be strengthened. Simultaneously, the bank's balance sheet willbe strengthened to overcome twin balance sheet problem. Higher efficiency,credit expansion, profitability, expansion of bank branches are the few keyparameters that can strengthen the balance sheet of banks. At the same time,to meet the socio-economic objectives, bank can provide adequate credit topriority sector.
Reference1. Abid, L; Ouertani, M N and Zouari-Ghorbel, S (2014), "Macroeconomic and Bank-
Specific Determinants of Household's Non-Performing Loans in Tunisia: A DynamicPanel Data", Procedia Economics and Finance, Vol. 13, pp 58-68.
2. Ali, A and Daly, K (2010), "Macroeconomic Determinants of Credit Risk: Recent Evidencefrom a Cross Country Study", International Review of Financial Analysis, Vol. 19,pp 165-171.
3. Amuakwa-Mensah, F; Marbuah, G and Marbuah, D A A (2017), "Re-Examining theDeterminants of Non-Performing Loans in Ghana's Banking Industry: Role of the 2007-2009 Financial Crisis", Journal of African Business, Vol. 18, No. 3, pp 357-379.
4. Arellano, M and Bond, S (1991), "Some Tests of Specification for Panel Data: MonteCarlo Evidence and an Application to Employment Equations", The Review of EconomicStudies, Vol. 58, No. 2, pp 277-297.
5. Berger, A N and DeYoung, R (1997), "Problem Loans and Cost Efficiency in CommercialBanks", Journal of Banking & Finance, Vol. 21, No. 6, pp 849-870.
6. Bloem, A M and Gorter, C N (2001), "The Treatment of Nonperforming Loans inMacroeconomic Statistics", International Monetary Fund Working Paper, WP/01/209,pp 1-17.
7. Blundell, R and Bond, S (1998), "Initial Conditions and Moment Restrictions in DynamicPanel Data Models", Journal of Econometrics, Vol. 87, No. 1, pp 115-143.
8. Blundell, R; Bond, S and Windmeijer, F (2001), "Estimation in Dynamic Panel DataModels: Improving on the Performance of the Standard GMM Estimator", In Non-Stationary Panels, Panel Cointegration and Dynamic Panels, Emerald Group PublishingLimited, pp 53-91.
9. Bofondi, M and Ropele, T (2011), "Macroeconomic Determinants of Bad Loans: Evidencefrom Italian Banks", Questioni di Economia e Finanza (Occasional papers), pp 1-40.
10. Caporale, G M; Alessi, M; Colli, S D and Lopez, J S (2017), "Loan Loss Provisions andMacroeconomic Shocks: Some Empirical Evidence for Italian Banks during the Crisis",Finance Research Letters, available at https://doi.org/10.1016/j.frl.2017.10.031.
Mohanty, Das & Kumar: Determinants of Non-Performing Loans in India 53
11. Caprio, G and Klingebiel, D (1996), "Bank Insolvencies: Cross Country Experience",The World Bank Policy Research Working Paper Series No. 1620, pp 1-52.
12. Castro, V (2013), "Macroeconomic Determinants of the Credit Risk in the BankingSystem: The Case of the GIPSI", Economic Modelling, Vol. 31, pp 672-683.
13. DaCosta, M M N and Foo, J P N (2002), "China's Financial System: Two Decades ofGradual Reforms", Managerial Finance, Vol. 28, No. 10, pp 3-18.
14. Das, A and Ghosh, S (2007), "Determinants of Credit Risk in Indian State-OwnedBanks: An Empirical Investigation". Munich Personal RePEc Archive, Paper No. 17301,pp 1-21.
15. Dash, M K and Kabra, G (2010), "The Determinants of Non-Performing Assets inIndian Commercial Bank: An Econometric Study", Middle Eastern Finance andEconomics, Vol. 7, No. 2, pp 94-106.
16. Dhar, S and Bakshi, A (2015), "Determinants of Loan Losses of Indian Banks: A PanelStudy", Journal of Asia Business Studies, Vol. 9, No. 1, pp 17-32.
17. Dimitrios, A; Helen, L and Mike, T (2016), "Determinants of Non-Performing Loans:Evidence from Euro-Area Countries", Finance Research Letters, Vol. 18, pp 116-119.
18. Dutta, S; Gupta, N and Rao, P H (2013), "Determinants of Return on Assets of PublicSector Banks in India: An Empirical Study", Pacific Business Review International,Vol. 5, No. 11, pp 23-28.
19. Ekanayake, E M N N and Azeez, A A (2015), "Determinants of Non-Performing Loansin Licensed Commercial Banks: Evidence from Sri Lanka", Asian Economic andFinancial Review, Vol. 5, No. 6, pp 868-882.
20. Espinoza, R and Prasad, A (2010), "Non-Performing Loans in the GCC Banking Systemand their Macroeconomic Effects", International Monetary Fund Working Paper,WP/10/224, pp 1-24.
21. Fan, L and Shaffer, S (2004), "Efficiency versus Risk in Large Domestic US Banks",Managerial Finance, Vol. 30, No. 9, pp 1-19.
22. Fofack, H (2005), "Non-Performing Loans in Sub-Saharan Africa: Causal Analysis andMacroeconomic Implications", World Bank Policy Research Working Paper, WPS 3769,pp 1-36.
23. Gerhardt, M and Vennet, R V (2017), "Bank Bailout in Europe and Bank Performance",Finance Research Letters, Vol. 22, pp 74-80.
24. Ghosh, A (2015), "Banking-Industry Specific and Regional Economic Determinants ofNon-Performing Loans: Evidence from US States", Journal of Financial Stability,Vol. 20, pp 93-104.
25. Godlewski, C J (2004), "Capital Regulation and Credit Risk Taking: Empirical Evidencefrom Banks in Emerging Market Economies", available at SSRN: https://ssrn.com/abstract=588163 or http://dx.doi.org/10.2139/ssrn.588163.
26. International Monetary Fund, (2015), Global Financial Stability Report: NavigatingMonetary Policy Challenges and Managing Risks, Washington, DC, April.
27. International Monetary Fund, (2017), Global Financial Stability Report: Getting thePolicy Mix Right. Washington, DC, April.
54 Prajnan
28. Khemraj, T and Pasha, S (2009), "The Determinants of Non-Performing Loans: AnEconometric Case Study of GUYANA", Munich Personal RePEc Archive, Paper No.53128, pp 1-25.
29. Kjosevski, J and Petkovski, M (2017), "Non-Performing Loans in Baltic States:Determinants and Macroeconomic Effects", Baltic Journal of Economics, Vol. 17,No. 1, pp 25-44.
30. Klein, N (2013), "Non-Performing Loans in CESEE: Determinants and Impact onMacroeconomic Performance", International Monetary Fund Working Paper, WP/13/72, pp 1-26.
31. Lokare, S M (2014), "Re-Emerging Stress in the Asset Quality of Indian Banks: Macro-Financial Linkages", Reserve Bank of India Working Paper Series, WPS (DEPR): 03/2014, pp 1-43.
32. Louzis, D P; Vouldis, A T and Metaxas, V L (2012), "Macroeconomic and Bank-SpecificDeterminants of Non-Performing Loans in Greece: A Comparative Study of Mortgage,Business and Consumer Loan Portfolios", Journal of Banking & Finance, Vol. 36,pp 1012-1027.
33. Mahajan, P (2014), "Non Performing Assets: A Study of Public, Private & Foreign SectorBanks in India", Pacific Business Review International, Vol. 7, No. 1, pp 09-16.
34. Makri, V; Tsagkanos, A and Bellas, A (2014), "Determinants of Non-Performing Loans:The Case of Eurozone", Panoeconomicus, Vol. 2, pp 193-206.
35. Islam, M M (2003), "Development and Performance of Domestic and Foreign Banks inGCC Countries", Managerial Finance, Vol. 29, No. 2/3, pp 42-72.
36. Messai, A S and Jouini, F (2013), "Micro and Macro Determinants of Non-PerformingLoans", International Journal of Economics and Financial Issues, Vol. 3, No. 4,pp 852-860.
37. Misra, B M and Dhal, S (2010), "Pro-Cyclical Management of Banks' Non-PerformingLoans by the Indian Public Sector Banks", BIS Asian Research Papers, pp 1-23.
38. Mukherjee, P (2003), "Dealing with NPAs: Lessons from International Experience",Money & Finance ICRA Bulletin, January-March, pp 64-90.
39. Nkusu, M (2011), "Nonperforming Loans and Macrofinancial Vulnerabilities in AdvancedEconomies", International Monetary Fund Working Paper, WP/11/161, pp 1-27.
40. Ötker-Robe, I and Podpiera, J (2010), "The Fundamental Determinants of Credit DefaultRisk for European Large Complex Financial Institutions", International Monetary FundWorking Paper, WP/10/153, pp 1-31.
41. Patra, B and Padhi, P (2016), "Determinants of Nonperforming Assets-Bank-Specificand Macroeconomic Factors: A Panel Data Analysis of Different Group of CommercialBanks Operating in India", Theoretical & Applied Economics, Vol. 23, No. 4,pp 215-236.
42. Poghosyan, T and Cihak, M (2011), "Determinants of Bank Distress in Europe: Evidencefrom a New Data Set", Journal of Financial Services Research, Vol. 40, No. 3,pp 163-184.
43. Prasanna, P K; Thenmozhi, M and Rana, N (2014), "Determinants of Non-PerformingAdvances in Indian Banking System", Banks and Bank Systems, Vol. 9, No. 2,pp 65- 77.
Mohanty, Das & Kumar: Determinants of Non-Performing Loans in India 55
44. Rajan, R and Dhal, S C (2003), "Non-Performing Loans and Terms of Credit of PublicSector Banks in India: An Empirical Assessment", Reserve Bank of India OccasionalPapers, Vol. 24, No. 3, pp 81-121.
45. Rajha, K S (2016), "Determinants of Non-Performing Loans: Evidence from theJordanian Banking Sector", Journal of Finance and Bank Management, Vol. 4, No.1, pp 125-136.
46. Rawlin, R; Shanmugam, R and Bhat, V (2014), "A Comparison of Key Determinants onProfitability of India's Largest Public and Private Sector Banks", European Journal ofBusiness and Management, Vol. 6, No. 34, pp 62-68.
47. Reddy, K S (2015), "Non-Performing Loans in Emerging Economies – Case Study ofIndia", Asian Journal of Finance & Accounting, Vol. 7, No. 1, pp 183-206.
48. Reserve Bank of India, (2017), Financial Stability Report, Issue No. 15, Mumbai, June.
49. Roy, S G (2014), "Determinants of Non-Performing Assets in India – Panel Regression",Eurasian Journal of Economics and Finance, Vol. 2, No. 3, pp 69-78.
50. Salas, V and Saurina, J (2002), "Credit Risk in Two Institutional Regimes: SpanishCommercial and Savings Banks", Journal of Financial Services Research, Vol. 22,No. 3, pp 203-224.
51. Soto, M (2009), "System GMM Estimation with a Small Sample", Barcelona EconomicsWorking Paper Series, WP No. 395, pp 1-27.
52. Tanaskovic, S. and Jandric, M. (2015), "Macroeconomic and Institutional Determinantsof Non-Performing Loans", Journal of Central Banking Theory and Practice, Vol. 1,pp 47-62.
53. Us, V (2017), "Dynamics of Non-Performing Loans in the Turkish Banking Sector byan Ownership Breakdown: The Impact of the Global Crisis", Finance Research Letters,Vol. 20, pp 109-117.
A Comparative Analysis ofLoan Recovery Strategy of Indian Banks
Robin ThomasRam Krishna Vyas
In the aftermath of the Global Financial Crisis of 2008, the Non-Performing Assets (NPAs) of Indian banks grappled with unprecedentedheights with striking deterioration in banks' asset quality. The assetquality performances of various bank groups, however, registerdiscernible divergences, with bulk of the problem skewed towards thePublic Sector Banks (PSBs). Empirical evidence hints at inadequateappraisal and lax monitoring by the Indian banks leading to assetquality impairments. The need of the study arises mainly because thereis evidence that some bank groups have been better in dealing withthe challenge of rising NPAs compared to others. The study, therefore,compares the loan recovery strategy of various bank groups in Indianbanking sector using the McKinsey 7S model as a qualitative,analytical-comparative tool. The study concludes that the private sectorbanks in India employ loan recovery strategy which has resulted insignificant lower NPAs (for the period 2005-2017) compared to StateBank of India (SBI) and Associates, PSBs and foreign banks operatingin India.
In the aftermath of the global financial crisis, the credit growth trajectory ofIndian banks registered a sharp decline. Statistical data from the Reserve BankIndia (RBI) divulge that the annual average credit growth of ScheduledCommercial Banks (SCBs) in India declined from 19 per cent during 2005-10
Shri Robin Thomas, ([email protected]), Ph.D. Research Scholar in Management, DeviAhilya University, Indore.
Dr Ram Krishna Vyas ([email protected]), Professor and Ex-Dean (Management),International Institute of Professional Studies, Devi Ahilya University, Indore.
58 Prajnan
to 13 per cent during 2011-15. The Indian economy registered a slowdown;the average annual GDP growth rate during the period 2005-10 was at 8.3 percent which declined to 6.74 per cent during the period 2011-15.
It is noteworthy that during the period 2005-10 the annual average credit growthrate of Public Sector Banks (PSBs) and State Bank of India (SBI) and Associateswas 19.3 per cent while that of the private sector banks and foreign banksoperating in India was 14.6 per cent. In a stark contrast the annual averagecredit growth rate of PSBs and SBI & Associates declined to 12.1 per centduring the period 2011-2015 while that of private sector banks and foreignbanks declined to 14.4 per cent.
It is inferred from the above discussion that the PSBs and SBI & Associateswere lending at a higher rate during the expansionary phase of the economycompared to private sector banks and foreign banks operating in India. Thistrend reversed when the Indian economy witnessed a slowdown; the creditgrowth rate of PSBs and SBI & Associates declined more sharply as comparedto private sector banks and foreign banks operating in India.
At the end of 2015 the Gross Non-Performing Assets Ratio (GNPA Ratio - GrossNon Performing Assets to Total Assets) of SCBs stood at 4.3 per cent, while atthe end of 2010 the GNPA ratio of all SCBs was 2.5 per cent. The Gross NPAsduring the period 2005-15 grew at a Compound Annual Growth Rate (CAGR)of 18.9 per cent. While during the period 2005-10 the GNPA CAGR was 7.3 percent, in contrast to the period 2010-15 where the GNPA CAGR was 31.63 percent.
The above discussion outlines the build-up of the NPA problem in Indianbanking sector after the world economic crises of 2008. Above data surmisesthat the Indian banks found themselves in a huge stock of sticky assets thatcollected in their balance sheets over the years; with severity of the problemheightened as the banks were unable to arrest the flow of new NPA accretionsatisfactorily.
Rajan (June, 2016) argues that at times of strong economic growth and rapiddeposit growth banks make mistakes. They extrapolate past growth andperformance to the future. So they are willing to accept higher leverage inprojects, and less promoter equity. Sometimes banks extend loans based onproject reports prepared by promoter's investment bank, without doing theirown due diligence. This is the historic phenomenon of irrational exuberance,common across countries at such a phase in the cycle. Rajan (June, 2016)further elaborates the years of strong global growth before the global financialcrisis were followed by a slowdown, which extended even to India. Demandprojections for various projects proved to be unrealistic, project cost overrunsescalated for projects stalled due to various reasons and they becameincreasingly unable to service debt. Subsequently, the banks were saddled with
Thomas & Vyas: A Comparative Analysis of Loan Recovery Strategy of Indian Banks 59
huge NPAs in their books; challenged with the problem of recovery of debtunder default and fear of default. Rajan's (2016) discussion elucidates thebroad and general reasons of NPA build-up and the challenge of loan recoveryin the Indian banking sector.
Thus the present paper is an attempt to answer the question – How and whatwas the response of the various banks towards the problem of NPA build-upand recovery of debt under default and fear of default?
Lokare (2014) in his RBI working paper titled 'Emerging Stress in IndianBanking Sector' emphasizes that deteriorating asset quality could be contagious,insidious and they prey on the weak. The contagious nature of loan lossesemanates from the fact that their downside impact can quickly transmit toearnings, capital, and liquidity. They are insidious in the sense that it is oftendifficult to know that there is a problem until it's too late. Moreover, theseproblems prey on the weak banks, which are vulnerable and have relativelysmall amounts of capital to absorb unanticipated losses. NPAs generate avicious cycle of effects on the sustainability and growth of the bankingsystem, and if not managed properly could lead to bank failures.
Lokare's (2014) view on the possibility of a 'Bank Run' caused due to massivebuild-up of NPAs in banks' balance sheets strengthens the research enquiry soas to further understand the management by the banks with respect to theirloan recovery efforts.
Chakrabarty (2013) argues that post the onset of the Global Financial Crisisin 2008, the NPA ratios of Indian banks started increasing, indicating a markeddeterioration in asset quality of the banking system. He observes that a closerscrutiny of the asset quality, however, reveals considerable divergence betweenthe performances of various bank groups. Chakrabarty (2013) further affirmsthat this divergent trend clearly indicates that the ability to manage asset qualityacross banks varies markedly and, in the post crisis years in particular, theconcerns on asset quality are largely confined to the PSBs.
Lokare (2014) concludes that the bank group-wise assessment of NPA trends(2001 to 2013) reveal that though PSBs contribute to the bulk of NPAs, theshare of private sector banks and foreign banks in the total NPAs has goneup in the post-crisis period. Nonetheless, PSBs and foreign banks have mainlycontributed to the rise in NPAs. Lokare (2014) also concludes that inadequateappraisal and lax monitoring by the Indian banks has resulted in asset qualityimpairments.
The aforementioned arguments as presented by Chakrabarty (2013) and Lokare(2014) can be surmised to conclude that different bank groups in Indianbanking sector have responded differently to the challenge of rising NPAs intheir balance sheets. It can also be inferred that some bank groups were better
60 Prajnan
in dealing with the challenge, compared to other bank groups. Thus abovepreliminary discussions suggest that some bank groups recognised the problemof rising NPAs in its early stages and initiated corrective measures in the rightearnest while some others either responded late or did not respond until itwas too late. Aforesaid literature, therefore, solicits a detailed comparativereview of the loan recovery strategy of various bank groups in Indian bankingsector.
Section IIReview of Literature and Theoretical Framework
The analysis of asset quality of a bank in particular and bank balance sheetsin general is challenging due to the presence of information asymmetry and ausual delay in reporting of information. Simons and Cross (1991) in theirpaper published in the New England Economic Review; and titled; 'Do CapitalMarkets Predict Problems In Large Commercial Banks?' argue that the healthof a bank or banking system is difficult to determine because depositors,regulators, and other outsiders are unable to see through the veil surroundingbanks' balance sheets until the credit quality of assets is severely deteriorated.
Elaborating on the information delay problem which Simons and Crossemphasized; Capiro & Klingebiel (1996) in their World Bank paper titled 'BankInsolvency: Bad Luck, Bad Policy, Or Bad Banking?', argue that informationdelay problem, coupled with banks' demandable debt and sequential servicingfeatures, makes banking inherently fragile and susceptible to runs.
A 'Bank Run' is a condition when a large number of customers of a bank oranother financial institution withdraw their deposits simultaneously due toconcerns about the bank's solvency. As more people withdraw their funds, theprobability of default increases, thereby prompting more people to withdrawtheir deposits. In extreme cases, the bank's reserves may not be sufficient tocover the withdrawals.
Elaborating on the manifestation of banking system insolvency, Capiro &Klingebiel (1996) further argue that since banks are part of the paymentssystem, contagion could lead to a halting of payments and a return to barter,to the detriment of overall economic activity. The possibility of contagion meansthat a single bank failure has not only a direct negative effect on GDP associatedwith the loss of the bank's profits and wages (as with any bankruptcy) but alsoan indirect and potentially larger effect to the extent that the bank failure leadsto or is associated with other bank failures and the shut-down of the paymentssystem.
Expounding more on systemic crises in banking system and its links withpoor asset quality and subsequent default in payments; Demirguc-Kunt &
Thomas & Vyas: A Comparative Analysis of Loan Recovery Strategy of Indian Banks 61
Detragiache (1997) in their Word Bank paper titled 'The Determinants ofBanking Crises: Evidence from Developed and Developing Countries' arguethat default risk cannot be entirely eliminated from the banking system but ifloan losses exceed a bank's compulsory and voluntary reserves as well as itsequity cushion, then the bank is insolvent. When a significant portion of thebanking system experiences loan losses in excess of their capital, a systemiccrisis occurs.
Underscoring the importance of asset quality for banks, Capiro (1998) in hisanother World Bank paper titled 'Banking on Crises: Expensive Lessons fromRecent Financial Crises' argue that financial crises regularly originate in orinduce insolvency in the banking system and that the banks and other financialintermediaries usually do not get into trouble if borrowers can easily servicetheir debt. If borrowers find it hard to service their debts, probability of defaultmay culminate.
Default is a condition owing to borrowers' failure to pay interest or principalon a loan or security when due. Default occurs when a debtor is unable tomeet the legal obligation of debt repayment. Extensive literature is availablestressing out on the central role of poor asset quality as a predictor of bankfailures as suggested by Demirguc-Kunt(1989); Whalen, (1991); Barr and Siems(1994) and Berger and DeYoung (1997).
In reference to Indian banking sector also, considerable literature points atthe importance of asset quality for banks. Ranjan and Dhal (2003) in theirReserve Bank of India paper titled 'Non-Performing Loans and Terms of Creditof PSBs in India: An Empirical Assessment' argue that amongst the variousindicators of financial stability, banks' non-performing loan assumes criticalimportance since it reflects on the asset quality, credit risk and efficiencyin the allocation of resources to productive sectors.
One very important contributor to macroeconomic stability is healthy banks(Rajan, Jan, 2016). The financial strength and viability of banks should reallybe measured in terms of the quality of their assets and the good rate of growthof their deposits year after year (Narasimhan Committee-I on Financial Systems-1991). It is evident thus that asset quality is one of the parameters that definethe strength and resilience of the banking sector (Gandhi, Feb, 2016) andmanaging asset quality is always very important and becomes a prominentobjective especially during a period of economic downturn (Gandhi, Sep, 2015).Banks require that delinquency risks and quality of their portfolio are carefullymanaged. Adherence to the provisioning requirement, building countercyclicalprovisions and holding countercyclical capital buffers in good times can,to some extent, insulate banks from excessive default stress during crisissituations (Sinha, Nov, 2011).
62 Prajnan
The manifestation of a sound and effective loan recovery strategy for bankstakes utmost importance as suggested by Capiro & Klingebiel (1996); arguingthat the output and production processes of nonfinancial firms often are moretransparent than those of banks, reflecting both the information-intensivenature of banking and its inter-temporal quality-most bank products or servicesinclude a promise to pay in the future, meaning that it can take time for abank's inability to fulfil its contracts to become evident. Banks can concealproblems by rolling over bad loans or by raising more deposits and increasingthe size of their balance sheets. Especially when depositors enjoy explicit orimplicit protection, banks often can attract new depositors with the promiseof high interest rates so as to increase their bets with current clients; look fornew, high-risk, high-return areas; or work a Ponzi scheme. The information-sensitive and inter-temporal nature of banking business requires that bankshave a loan recovery strategy in place to take-on the challenge posed by theincreasing non-performing loans.
The need for, a bank by bank, portfolio by portfolio strategy for sustainedreduction in non-performing loans across the banking system is suggested byDonnery (Oct, 2016) (Deputy Governor of the Central Bank of Ireland) in herlecture at the Peterson Institute for International Economics and titled ''Non-Performing Loans: Workout And Resolution In The Euro Area'. The authoremphasises that the effective workout and the resolution of non-performingloans is central to both bank viability and macroeconomic performance.Furthermore, that the structure of profit and loss accounts across banks inIreland is quite heterogeneous and for some banks, impairment charges(provision) continue to be a drag on profitability. Coupled with continuedpressure from the low interest rate environment, and squeezed lending margins,these dynamics can threaten long-term viability. Therefore, banks potentiallyneed to adapt their business models to the changing environment and risks. Afully functioning euro area banking sector is needed to support investment,growth, and employment. Removing impediments to these critical economicfunctions underpins the rationale for reducing NPAs and supervisory actionsto accomplish this. Given these challenges, high levels of NPAs require todeliberate and sustained action.
The importance of debt recovery is underlined as one of the core elementsof financial system stability as suggested by Jayamaha (2005) (DeputyGovernor of the Central Bank of Sri Lanka) in her speech at a conclave onproposed law on restructuring of sick industries and other legal issuesat the Centre for Banking Studies, Rajagiriya, 1 February 2005. The authoremphasizes that promoting credit culture and expediting the process of debtrecovery is important for central banks given the fact that loan recovery formsa pillar of financial system stability.
In reference to Indian banks, Jalan (1999) discusses in detail about thechallenges facing the Indian banking sector and the importance of timely
Thomas & Vyas: A Comparative Analysis of Loan Recovery Strategy of Indian Banks 63
recovery and NPA resolution. He argues that vigorous effort has to be made bybanks to strengthen their internal control and risk management systems, andto set up early warning signals for timely detection and action for addressingthe stock of NPAs. The resolution of the NPA problem also requires greateraccountability on the part of corporates, greater disclosures in the case ofdefaults, and an efficient credit information system. Furthermore, the authordiscusses the issue of legal reform and NPA resolution.
In Indian context, Muniappan (2002) emphasizes the need for banks andfinancial institutions to recover bad debts and arrest fresh accretion of NPAs.The author, a Deputy Governor of the Reserve Bank of India in his lecture atCII Banking Summit, at Mumbai on April 1, 2002 argues that the high level ofNPAs in banks and financial institutions has been a matter of grave concern tothe public as bank credit is the catalyst to the economic growth of the countryand any bottleneck in the smooth flow of credit, one cause for which is themounting NPAs, is bound to create adverse repercussions in the economy.NPAs are not therefore, the concern of only lenders but of borrowers andpotential borrowers also.
The need for resolution and recovery of bad loans is underlined by Sinha (2011)(Deputy Governor of the Reserve Bank of India) at BANCON-2011, organizedby the Indian Banks' Association and Indian Overseas Bank, in Chennai, on 6November 2011. Sinha argues that aggressive lending stance of banks andinadequate due diligence and laxity in monitoring of the loan accounts arealso responsible for deterioration in the asset quality. Banks need to, not onlyutilize effectively, the various measures such as CDR mechanism, One TimeSettlement schemes, Debt Recovery Tribunals, provisions of the SARFAESIAct put in place by RBI and the Government of India, but also have to strengthentheir due diligence, credit appraisal and post sanction loan monitoring systemsto minimize and mitigate the problem of increasing NPAs.
More recently, Rajan (2016) discourses in Indian context that, a number ofbanks in the system have taken the necessary action to recognize and resolvestressed loans in a timely fashion. But some others need to take more proactiveaction.
Singh & Brar (2016) conclude that there has been a substantial rise in stressedassets in Indian banking sector, mainly in PSBs and argue that Indian banksshould take this opportunity to introspect and analyse the causes of suchdeterioration in asset quality while taking remedial measures. The authorsfurther suggest that in addition to global factors, there could be many morefactors like deficiencies in procedures and practices followed in PSBs comparedto other bank groups. It is imperative from the conclusion of Singh & Brar(2016) that an analysis into the levels of non-performing assets in Indian banksis made and more importantly an analytical enquiry be conducted to studywhat strategy the bank groups are following to resolve and recover stressedand non-performing assets.
64 Prajnan
The above elaborate literature survey surmises that the information-sensitiveand inter-temporal nature of banking business requires that banks have aloan recovery strategy in place to take-on the challenge posed by the increasingnon-performing loans. The purpose of the current study therefore, is tocomparatively analyse the loan recovery strategy adopted and followed bycommercial banks in India.
Figure 1Conceptual Framework
Source: Literature Review
ObjectivesTo comparatively analyse the loan recovery strategy of scheduled commercialbanks in India through application of a qualitative strategic model.
To compare the effectiveness of loan recovery strategy of scheduled commercialbanks in India.
HypothesesH0:1 There is no significant difference in the effectiveness of loan recovery strategy
of various bank groups in India.
H1:1 There is significant difference in the effectiveness of loan recovery strategy ofvarious bank groups in India.
Section IIIResearch Methodology
The present study is an analytical comparative research, aimed at criticalevaluation of the loan recovery strategy in scheduled commercial banks inIndia. Analytical studies use facts or information already available and analyse
Thomas & Vyas: A Comparative Analysis of Loan Recovery Strategy of Indian Banks 65
these to make a critical evaluation of the material Kothari (1985). Moen et. al.(1998) explains that methods for designing and learning from analytical studiesinvolve methods that are graphical and focus on prediction of futureperformance. Provost (2011) explains that analytical studies assume that theactions taken as a result of the study will be on the process or causal systemthat produced the frame studied, rather than the initial frame itself. The aimis to improve future performance. The author also explains that the aim of ananalytical study is to enable prediction about how a change in a system willaffect that system's future performance, or prediction as to which plans orstrategies for future action on the system will be superior.
For comparative analysis the scheduled commercial banks are sub-dividedinto four groups namely; 1) SBI & Associates, 2) PSBs, 3) Private Sector Banksand 4) Foreign Banks operating in India. The McKinsey 7S model is used toanalyse and compare the loan recovery strategy of the four bank groups inIndian banking sector. To determine which of the bank group's loan recoverystrategy has been more effective, a comparison of the movement of Gross Non-Performing Assets amongst the various bank groups has been performed.
Data SourcesThe study employs secondary data collected from various sources as describedbelow:
Statistical Data on NPA, bank-specific and economic indicators during 2004-05 to 2016-17, collected from RBI website and Indian Banks Association. Inaddition to the above data, information is obtained from individual banks'websites; through their Annual Reports and Investor Presentations.
Research reports, published articles, news reports and conference proceedingsavailable in both national and international level related to loan recovery andNPA management. The information obtained from these sources is used forcritical evaluation of the subject and identify research gap in the area of study.
Sampling DesignFor the purpose of comparison, the banks in Indian banking sector (scheduledcommercial banks) were divided into four groups; namely (1) SBI & Associates,(2) PSBs, (3) Private Sector Banks and (4) Foreign Banks. The study excludesthe Regional Rural Banks and Co-operative Banks from the purview of analysis.To study the loan recovery strategy of bank groups, judgemental sampling isused to select a sample of banks representing each group.
66 Prajnan
Sample Units Reasons/Judgement for considering thesample
SBI & Associates (Two banks from the groupwere selected) – (a) SBI and (b) State Bank ofTravancore
State Bank of India is the largest and foremostentity within the SBI group and hence is obviouschoice for being in the sample. State Bank ofTravancore is arbitrarily selected to representthe practises of Associate Banks.
PSBs (Seven banks from the group wereselected) – (a) Bank of Baroda, (b) Bank ofIndia, (c) Union Bank of India, (d) Canara Bank,(e) Indian Overseas Bank, (f) Oriental Bank ofCommerce, (g) and United Bank of India.
The PSBs can be largely divided into threegroups based on business size and/or numberof branches; namely Large Banks, Mid-sizebanks and Small Banks. Bank of Baroda andBank of India represent the Large PSBs whereasUnion Bank of India, Canara Bank and IOBrepresent the Mid-Size PSBs. Oriental Bank ofCommerce and United Bank of India representthe Small Size PSBs.
Private Sector Banks – (Six banks from thegroup were selected) – (a) ICICI Bank, (b) HDFCBank, (c) Axis Bank, (d) Kotak Mahindra Bank,(e) Federal Bank and (f) Karnataka Bank.
ICICI Bank, HDFC Bank and Axis Bankrepresent the three largest private players inIndian Banking Sector. Also they belong to thefirst branch of private sector banks posteconomic l iberal isation of 1991. KotakMahindra Bank represents the neo -new privatesector banks whereas Federal Bank andKarnataka Bank represent the old privatebanks in India.
Foreign Bank (Operating in India) – (Threebanks from the group were selected) – (a) CitiBank, (b) Standard Chartered Bank and(c) HSBC Bank.
The sample is selected based on mostprominent Foreign Bank players operating inIndia based on both business size and numberof branches.
Section IVTools of Analysis
Analytical ToolsMcKinsey 7S model is employed to study and qualitatively compare the loanrecovery strategy of four bank groups.
Ratio analysis, Percentage analysis, Compound Annual Growth Rate andAverage Annual Growth Rate are also used in analysis.
The McKinsey 7s Model
The 7S model was developed at McKinsey & Company Consulting Firm in1980. It's a model of organization effectiveness that postulates that thereare seven internal factors, that needs to be aligned and reinforced in order tobe successful. Channon & Caldart (2015) describe that the model was developedas a conceptual framework; useful in diagnosing the causes of organizational
Thomas & Vyas: A Comparative Analysis of Loan Recovery Strategy of Indian Banks 67
malaise and in formulating programs for improvement. The model constitutedan attempt to provide a response to the widespread frustration experienced byexecutives at the time of dealing effectively with general management problemsrelated to strategic and organizational factors. The framework surpasses theclassic simplistic notion that 'structure follows strategy' as it links strategy notonly with structure but also with other five elements. In addition to Strategy,Structure & System (Three Hard elements) the other elements or variables ofthe framework are the following: Shared Values, Style, Skills & Staff (The FourSoft Elements). Hard elements are easily identified and influenced bymanagement while soft elements are more intangible and are influenced bycorporate culture.
Kaplan (2005) in his paper titled 'How The Balanced Scorecard ComplementsThe Mckinsey 7S Model' elaborate as under, on the seven 'S' of the model asadapted from Waterman, Peters & Phillips (1980), paper titled 'Structure IsNot Organization,' and Clued (1996) Harvard case study titled 'OrganizationalAlignment: The 7S Model,' (HBS Case 9-497-045)
1. Strategy – The positioning and actions taken by an enterprise, in response toor anticipation of changes in the external environment, intended to achievecompetitive advantage.
2. Structure – The way in which tasks and people are specialized and divided,and authority is distributed.
3. System – The formal and informal procedures used to manage theorganization, including management control systems, performancemeasurement and reward systems.
4. Shared Values – The core or fundamental set of values that are widely sharedin the organization and serve as guiding principles of what is important; vision,mission, and values statements that provide a broad sense of purpose for allemployees.
5. Style – the leadership style of managers - how they spend their time, whatthey focus attention on, what questions they ask of employees, how they makedecisions; also the organizational culture.
6. Staff – the people, their backgrounds and competencies.
7. Skills – The distinctive competencies of the organization; what it does bestalong dimensions such as people, management practices, processes & systems.
Statistical Tools
Averages, Ratios and ANOVA & Levene's & Welch Statistics are the measuresutilised for statistical analysis.
MS Excel and SPSS 22 has been employed for statistical analysis.
68 Prajnan
Section VResults & Discussion
Qualitative Comparative Analysis of Loan Recovery Strategy of Bank Groupsin IndiaBased on the McKinsey 7S Framework, the 7 elements of the four bank groupswere analysed and compared with respect to their loan recovery and resolutionstrategy. For comparative evaluation the 7 elements of the four bank groupsare presented together so that an analysis and contrast may be made at aglance. Seven tables have been generated for each of the 7 elements of McKinsey7S framework. The compilation of the analysis is presented below.
1. Superordinate Goals
SBI and Associates Public Sector Banks Private Sector Banks Foreign Banks inIndia
1. Honesty,Transparencyand Ethics
2. To ensureaccountability forperformance andcustomer service andto achieve excellenceat all levels
1. Quality BankingService
2. To create value forall stakeholders
3. To be a responsivecorporate socialcitizen
4. To Observe higheststandards ofcorporategovernance andcorporate socialresponsibilities
1. Maintain a healthyfinancial profileand diversify ourearnings acrossbusinesses andgeographies
Thomas & Vyas: A Comparative Analysis of Loan Recovery Strategy of Indian Banks 69
3. Structure
SBI and Associates Public Sector Banks Private Sector Banks Foreign Banks inBanks
Separate Vertical forStressed AssetsManagement (SAMG)headed by a Dy.Managing Director.
Separate recoveryvertical withadditional seniorofficers posted in allzones.
Independent RiskManagement Groupformed. A separateCredit MonitoringGroup formed forproactive resolution ofrecovery in SMEsegment.
Impairedaccountsmanaged byEarly Alert,Retail and GroupSpecial AssetsManagement. Thegroup reports toCountry RiskCommittee
4. Systems
SBI and Associates Public Sector Banks Private Sector Banks Foreign Banks inBanks
1. Digital Technologybased Early WarningSystem installed.
2. Litigation ManagementSystem and NPA Portalfor monitoring andcontrol
1. Credit Audit/Loan ReviewMechanism (LRM)digital tool forconstantlyevaluating thequality of loanbook.
2. Diversifiedportfolio of riskyassets ismaintained and asystem to conductregular analysisof the portfolio soas to ensureongoing effortsfor recovery andup-gradation.
1. Use of Predictivemodels for stress-identification.
2. CentralizedDelinquentDatabase to reviewthe borrower'sprofile beforedisbursements
3. Independentaudit of alldecisionsoriginatingcredit risk
5. Style
SBI and Associates Public Sector Banks Private Sector Banks Foreign Banks inBanks
Recovery monitoredunder BoardSupervision includingChairman and all MDs
Direct boardsupervision undervarious committeesand sub-committeesof the board
The Retail Credit andPolicy Group is anindependent unitfocusing on policyformulation andportfolio tracking andmonitoring andreports to a group ofExecutive Directors
CommitteeGovernanceStructureensures that risk-taking authorityand policies arecascaded downfrom Chief RiskOfficer to theappropriatefunctional anddivisionalcommittees.
70 Prajnan
6. Staff
SBI and Associates Public Sector Banks Private Sector Banks Foreign Banks inBanks
Various Regional Officesfunctioning underSAMG; various StressAssets ManagementBranches and StressedAssets RecoveryBranches createdespecially for NPArecovery.
1. Setting- up assetrecovery branches.
2. SlippagePrevention Teamsformed at all Zonaland RegionalOffices for thepurpose ofarresting slippagesand also forinitiating necessarycorrective actionplan at an earlystage in a timebound manner.
Independent Groupsdeployed to identify,assess and monitor thebank group's principalrisks
Dedicatedspecialistrecovery unit ofemployeesindependent ofthe mainbusinesses.
7. Skills
SBI and Associates Public Sector Banks Private Sector Banks Foreign Banks inBanks
1. Mandatory trainingfor BCs for DebtRecovery.
2. CertificationProgram on Creditfor loan life cyclemanagement
1. Training ofemployeesthrough e-learningand mentoring ofthe younger staff.
2. Mass sensitizationof employeesabout recoveryand NPAmanagement.
Section VIComparative Analysis of the Effectiveness of
Loan Recovery Strategy of Bank-Groups in India
Analysis using Gross NPA RatioThe Non-Performing Assets (NPAs) are an important prudential indicator toassess the financial health of the banking sector. Besides asset quality, NPAsepitomize the credit risk management and efficacy in the allocation ofresources as suggested by Lokare (2014).
The quantum of a bank's loan performance has generally been measured bythe level of NPAs of that bank. Gross NPA Ratio (defined as - Gross NonPerforming Assets to Total Assets) has traditionally been used as a measure ofthe health of asset quality of a bank (Swamy, 2015 & Samir & Kamra, 2013).The Gross NPA Ratio of a bank therefore, also reflects the measures, means
Thomas & Vyas: A Comparative Analysis of Loan Recovery Strategy of Indian Banks 71
and strategies adopted by the bank to arrest and recover the flow and stock ofNPAs in its balance sheet. Gross NPA Ratio thus is employed as a measure ofthe effectiveness of loan recovery strategy of a bank.
Table 1Bank Group-Wise Gross Non-Performing Assets, Gross Advances and
GNPA Ratio (Amount in Rs. Crores)
Year SBI And Associates Public Sector Banks
Gross Gross GNPA Gross Gross GNPANPA Advances Ratio NPA Advances Ratio
DescriptionTable 1 represents bank group-wise GNPA, Gross Advances and GNPA Ratiofor the four bank groups, namely SBI & Associates, PSBs, private sector banksand foreign banks.
The Gross NPA of SBI & Associates at the end of FY 2017 stand at 1.78 LakhCrores, that of PSBs stand at 5.00 Lakh Crores, private sector banks standsat 91.90 Thousand Crores and foreign banks stands at 13.60 Thousand Crores.
The GNPA ratio for SBI & Associates at the end of FY 2017 is 9.11 per cent,PSBs at 12.95 per cent, private banks at 4.06 per cent and foreign banks at3.96 per cent.
Chart 1Movement of GNPA Ratio
Source: Reserve Bank of India
Thomas & Vyas: A Comparative Analysis of Loan Recovery Strategy of Indian Banks 73
DescriptionChart 1 graphically represents the trend in movement of GNPA Ratio for thefour bank-groups for the period 2005-2017. It is noteworthy that during 2010-2017, has been a considerable rise in the GNPA Ratio of SBI & Associates (2.8in 2010; 9.11 in 2017) and PSBs (2.0 in 2010; 12.95 in 2017). However, theprivate banks registers comparatively flattish increase in their GNPA Ratio(3.0 in 2010; 4.06 in 2017) while Foreign Banks register a staggering GNPARatio over the period 2010-2017 (4.36 in 2010; 3.96 in 2017), though thetrend in movement of their GNPA Ratio is just higher than the private banks.
Chart 2Growth Rate in GNPA
Source: Reserve Bank of India
DescriptionChart 2 graphically represents the CAGR (Compound annual growth rate) ofGNPA for the four bank-groups for the period 2005 to 2017. For comparisonand analysis the CAGR of GNPA for the four bank-groups for the period 2005to 2010 and 2010 to 2017 is also computed. The GNPA CAGR of SBI &Associates during the period 2005-2010 is registered at 5.74 per cent whileduring the period 2010-2017 the GNPA CARG is registered at 29.98 per cent.
Similarly the GNPA CAGR of PSBs during the period 2005-2010 is registeredat 2.28 per cent while during the period 2010-2017 the GNPA CARG is registeredat 39.44 per cent.
74 Prajnan
The private banks register a GNPA CAGR of 12.44 per cent in 2005-2010 whilethis increases to 23.21 per cent in 2010-2017. The foreign banks register areverse trend in GNPA CAGR. They register GNPA CAGR of 21.29 per cent in2005-2010 and 8.46 per cent in 2010-2017.
Chart 3Ratio Analysis: Average Gross NPA Ratio Bank Group Wise
(2005-2017)
Source: Reserve Bank of India
The above chart 3 represents the average GNPA Ratio for the four bank-groupsfor the period 2005 to 2017. For analysis and comparison the Mean GNPARatio of the bank-groups for the period 2005-2010 and 2010-2017 is alsocomputed.
For the period 2005-2017, the average GNPA Ratio is the least for private sectorbanks at 2.61, while it is highest for PSBs at 4.51. The SBI & Associates andforeign banks in India have registered average GNPA ratio of 4.32 and 3.18respectively in 2005-2017.
For the period 2005-2010, the average GNPA Ratio is the least for private sectorbanks at 2.80, while it is highest for SBI & Associates at 3.23. The PSBs andforeign banks in India have registered average GNPA ratio of 2.95 and 2.96respectively in 2005-2010.
For the period 2010-2017, the average GNPA Ratio is the least for private sectorbanks at 2.52, while it is highest for PSBs at 5.37. The SBI & Associates andforeign banks in India have registered average GNPA ratio of 4.94 and 3.50respectively in 2005-2010.
Thomas & Vyas: A Comparative Analysis of Loan Recovery Strategy of Indian Banks 75
Hypothesis TestingHypothesis Testing - Significance of difference in the movement of GNPA Ratioamongst various bank groups (Testing the Null Hypothesis – There is nosignificant difference in the movement of GNPA Ratio across bank-groups forthe period 2005-2017)
Table 2Hypothesis Testing – Movement of GNPA Ratio
Test Applied: Levene's Test of Equality of Error Variances; Welch Test.
Variable: GNPA Ratio (Amongst Bank Group)
Levene's Statistic 5.556 (Significant at 0.05)
Welch Statistic 4.219** (p=0.5)
Note: A double (**) asterisk indicates the coefficients are significant at the 05 per cent level ofsignificance.
Based on Levene's statistics (Table 2), since the significance is less than 0.05(Sig =0.002 <0.05), it is inferred that homogeneity of variances does not exist,hence Welch Test is applied. The Welch Statistic is significant with respect tomovement of GNPA Ratio among various bank groups, thus it is inferred thatthere exist significant differences in movement of Gross Non-Performing AssetsRatio between different bank groups.
GNPA Ratio Multiple Comparisons
Table 3Multiple Comparison Amongst GNPA Ratio of Bank Groups
(Post-Hoc Analysis Using Games-Howell Test for unequal variances)
(I) Bank Group (J) Bank Group Mean Difference (I-J) Std. Error Sig.
SBI and Associates PSBs -.19462 1.09819 .998
Private Banks 1.70193* .55418 .034
Foreign Banks 1.13385 .57629 .238
PSBs SBI and Associates .19462 1.09819 .998
Private Banks 1.89654 .98692 .267
Foreign Banks 1.32846 .99950 .561
Private Banks SBI and Associates -1.70193* .55418 .034
PSBs -1.89654 .98692 .267
Foreign Banks -.56808 .31637 .301
Foreign Banks SBI and Associates -1.13385 .57629 .238
PSBs -1.32846 .99950 .561
Private Banks .56808 .31637 .301
*. The mean difference is significant at the 0.05 level.
76 Prajnan
Post-hoc analysis (Table 03) was done using Games Howell test for unequalvariances.
Post-hoc analysis divulges the following significant differences between theGross Non-Performing Assets indicators of-
There exists no significant difference between SBI & Associates and PSBs andforeign banks with respect to movement of GNPA Ratio.
There exists no significant difference between Private Sector Banks and Foreignbanks and PSBs with respect to movement of GNPA Ratio.
There exists no significant difference between PSBs and SBI & Associates,private banks and foreign banks with respect to movement of GNPA Ratio.
Analysis Using Recovery RatioThis paper employs a novel ratio for computation of Recovery efforts of Bank-groups. The Ratio is computed as -
Recovery Ratio=R/(GNPAe+W-A)
Where,
1. R = GNPA Recovery During the Year
2. GNPAe = Gross NPA at the end of the Year
3. W = GNPA Write-Off During the Year
4. A= New NPA Accretion During the Year
The Ratio sheds light on the recovery performance of the NPA stock which thebanks carry in their balance sheets. This ratio indicates the NPA reduced bybanks owing to their recovery measures. The current study is concerned tooutline and quantify the Recovery efforts and performance of bank-groups thusthe above ratio is computed. For the computation of the denominator term,the New NPA Accretion is subtracted from the total GNPA at the end of the yearas New NPA Accretion is a measure of banks' ex-ante NPA prevention measures.Similarly, GNPA Write-Off during the year is added to the denominator as itrepresents unfruitful recovery efforts for the banks resulting into total loss.The higher the ratio, the better are the recovery efforts of the banks.
Thomas & Vyas: A Comparative Analysis of Loan Recovery Strategy of Indian Banks 77
Table 4Recovery Ratio (2005-2017)
Year SBI & Associates Public Sector Banks Private Sector Banks Foreign Banks
2005 0.38 0.62 0.60 0.75
2006 1.12 0.86 1.02 1.51
2007 0.96 1.23 0.77 1.07
2008 1.11 1.53 0.70 3.12
2009 1.77 1.99 0.67 -14.94
2010 0.73 0.83 0.53 0.77
2011 0.78 0.76 0.38 0.54
2012 0.95 0.87 0.45 0.53
2013 0.78 0.55 0.59 0.34
2014 0.80 0.65 0.80 0.30
2015 0.51 0.49 0.69 0.34
2016 0.43 0.27 0.68 0.20
2017 0.30 0.21 0.73 0.30
Source: Computed and compiled from the NPA Movement Data for the period 2005-2017 fromReserve Bank of India.
Chart 4
Source: Reserve Bank of India.
78 Prajnan
Table 4 represents the computed Recovery Ratio (2005-2017) for the four bank-groups, and the Chart 3 graphically represents the trend in movement of bank-group wise Recovery Ratio for the period 2005-2017. It is noteworthy that thewhile the recovery ratio of private banks show a rising trend, the recoveryratio of SBI & Associates and PSBs register a downward trend. The foreignbanks course below other three bank-groups in their recovery ratio.
Hypothesis Testing - Significance of difference in the movement of RecoveryRatio amongst various bank groups. (Testing the Null hypothesis that thereexists no significant difference in the movement of Recovery Ratio across bank-groups for the period 2005-2017 and 2012-2017)
Testing the Null hypothesis that there exists no significant difference in themovement of Recovery Ratio across bank-groups for the period 2005-2017
Table 5Hypothesis Testing – Movement of Recovery Ratio (2005-2017)
Test Applied: Levene's Test of Equality of Error Variances and Welch Test
Variable: Recovery Ratio 2005-2017 (Amongst Bank Group)
Levene's Statistic 3.579 (Significant at 0.05)
Welch Statistic 1.137* (p=0.5)
Note: A Single (*) asterisk indicates the coefficients are not significant at the 05 per cent levelof significance.
Based on Levene's statistic (Table 5), since the significance is less than 0.05(Sig =0.02 <0.05), it is inferred that homogeneity of variances does not exist,hence Welch Test is applied. The p value = .355 (For Welch Test) is greaterthan p=.05; thus the null hypothesis cannot be rejected. Hence there exists nosignificant difference in the movement of Recovery Ratio of the bank-groupsfor the period 2005-2017.
Testing the Null hypothesis that there exists no significant difference in themovement of Recovery Ratio across bank-groups for the period 2012-2017
Table 6Hypothesis Testing – Movement of Recovery Ratio (2012-2017)
Test Applied: Levene's Test of Equality of Error Variances and Welch Test
Variable: Recovery Ratio 2012-2017 (Amongst Bank Group)
Levene's Statistic 3.123 (Significant at 0.05)
Welch Statistic 7.476** (p=0.5)
Note: A double (**) asterisk indicates the coefficients are significant at the 05 per cent level ofsignificance.
Thomas & Vyas: A Comparative Analysis of Loan Recovery Strategy of Indian Banks 79
Based on Levene's statistic (Table 6), since the significance is less than 0.05(Sig =0.049 <0.05), it is inferred that homogeneity of variances does not exist,hence Welch Test is applied. The p value = .006 (For Welch Test) is less thanp=.05; thus the null hypothesis cannot be accepted. Hence there existssignificant difference in the movement of Recovery Ratio of the bank-groupsfor the period 2012-2017.
Table 7Comparison of Means
Recovery Ratio (2012-2017)
Bank Groups Mean N Std. Deviation
SBI & Associates .62 6 .25
PSBs .50 6 .24
Private Banks .65 6 .12
Foreign Banks .33 6 .10
Total .53 24 .22
Source: Author calculations
Post the findings of Welch Test (Table 7 – Recovery Ratio – 2012-2017), theMean Recovery Ratio for the period 2012-2017 of the bank-groups are computedand analysed. It is evident from the comparison of means that during the period2012-2017 that the private banks have registered higher mean recovery ratiothan the other three bank-groups with comparatively lower standard deviationthan SBI & Associates and PSBs.
Analysis Using Other Measures – (Comparative analyses of various otherasset quality ratios for the three bank groups, viz Public Sector Banks-including the SBI Group, the Private Sector Banks and the Foreign Banks)Non-parametric Statistical Analysis (Kruskal Wallis Test ) is further applied tocertain other asset quality ratios amongst three bank groups (Public SectorBanks, Private Sector Banks and Foreign Banks), namely (i) Sub-StandardAssets Ratio, (ii) Doubtful Assets Ratio and (iii) Loss Assets Ratio (ReferTable 8). The sources of this ratio data is RBI Statistical Data from RBI website.The data from RBI has been prearranged in three groups, public sector banks(Which includes the SBI and Associates), the private sector banks and theforeign banks. Therefore, the additional tests on asset quality ratios have beenperformed on three bank groups. These additional tests are done to elucidatethe divergences in the asset quality indicators of the bank groups.
80 Prajnan
Table 8Other Asset Quality Ratios of the Three Bank-Groups
Year Public Sector Banks* Private Sector Banks Foreign Banks
Sub- Doubtful Loss Sub- Doubtful Loss Sub- Doubtful LossStandard Advances Advances Standard Advances Advances Standard Advances AdvancesAdvances Ratio Ratio Ratio Ratio Ratio Ratio Ratio Ratio
Ratio
2005 1.24 3.47 0.66 0.98 2.50 0.40 0.93 1.34 0.40
2006 0.99 2.17 0.49 0.76 1.39 0.30 0.96 0.68 0.30
2007 0.97 1.35 0.33 1.04 0.94 0.22 1.07 0.47 0.22
2008 0.95 1.06 0.22 1.39 0.85 0.24 1.20 0.47 0.24
2009 0.89 0.90 0.18 1.81 0.86 0.23 3.46 0.57 0.23
2010 1.05 0.93 0.21 1.37 1.02 0.34 2.94 0.86 0.34
2011 1.05 0.99 0.19 0.56 1.33 0.35 0.94 1.06 0.35
2012 1.58 1.24 0.15 0.52 1.06 0.30 0.89 0.95 0.30
2013 1.79 1.67 0.15 0.56 0.96 0.27 1.07 1.02 0.27
2014 1.84 2.33 0.19 0.63 0.84 0.30 1.44 1.43 0.30
2015 1.88 2.90 0.18 0.67 1.10 0.33 0.69 1.62 0.33
2016 3.44 5.55 0.28 0.94 1.57 0.31 1.67 1.59 0.31
2017 2.95 8.36 0.36 1.37 2.29 0.40 1.17 2.38 0.40
Note: *Includes SBI and AssociatesSource: Reserve Bank of India
Hypothesis Testing – Results
Table 9Hypothesis Testing for Other Asset Quality Ratios
Null Hypothesis Time-Period Test Sig Decision
There exists no significant 2005-2017 Independent Samples .047 Reject thedifference in the movement of Kruskal Wallis NullSub-Standard Assets Ratio Test hypothesisacross bank-groups.
There exists no significant 2005-2017 Independent Samples .049 Reject thedifference in the movement of Kruskal Wallis NullDoubtful Assets Ratio across Test hypothesisbank-groups.
There exists no significant 2005-2017 Independent Samples .001 Reject thedifference in the movement of Kruskal Wallis NullLoss Assets Ratio across Test hypothesisbank-groups.
Thomas & Vyas: A Comparative Analysis of Loan Recovery Strategy of Indian Banks 81
The Hypothesis testing performed through the Kruskal-Wallis Test (Table 9)statistically substantiates the existence of divergences in the asset quality ofthe three bank groups for the period 2005-2017. The comparison of the meansof the Sub-Standard Assets Ratio, Doubtful Assets Ratio and Loss Assets Ratioillustrates that Public sector banks record the highest mean Sub-StandardAssets Ratio and Doubtful Assets Ratio for the period 2005-2017. While meanLoss Assets Ratio for Foreign banks is recorded as the highest for the period2005-2017.
Table 10Mean Asset Quality Ratios
Bank-Group Wise Mean Asset Quality Ratios (2005-2017)
Public Sector Banks Private Sector Banks Foreign Sector BanksMean Mean Mean
Sub-Standard Assets 1.59 .97 1.42
Doubtful Assets Ratio 2.53 1.29 1.11
Loss Assets Ratio .28 .31 .61
Source: Author calculations based on data from RBI
Section VIIFindings
The finding from the comparative analysis of the loan recovery strategy ofthe four bank groups is itemized as under. Findings are deliberated uponpoint-wise under the 7S elements.
Super-ordinate goals
SBI & Associates place their emphases on corporate governance and ethicalstandards, and focus on accountability and performance in customer service.The PSBs include an important shared value of being a responsive corporatecitizen as one of their superordinate goals. This is reminiscent of the socialresponsibility motto with which the nationalised banks in India function.
Private sector banks on the other hand have a clearly defined shared value ofmaintaining a healthy financial profile and diversifying earnings acrossbusinesses and geographies. This is reminiscent of the profit motive with whichthe private sector banks in India function. Whereas, the foreign banks operatingin India also place their focus on delivering relevant and timely solutions forclients and customers; a superordinate goal which clearly defines foreign banksfocus and strength in retail banking segment.
82 Prajnan
Strategy
SBI & Associates have a broad base two-pronged strategy for loan recoveryand resolution. The umbrella strategy considers arresting fresh NPAs andfocusing on faster resolution of existing NPAs. The two pronged strategy focuson both; the stock and flow of NPAs. The flow of NPAs is being restrainedthrough close monitoring of loan accounts with installation of Early WarningModel and the stock of NPAs is being restrained by boosting the resolutionmechanism of loans under stress.
PSBs, largely have responded to the threat of rising NPAs with four-prongedstrategy. The strategic response of PSBs towards the rising NPAs threat is tobecome more risk averse by lowering the risk appetite. The second and thirdstrategies are closely interconnected; reduction of the overall concentrationrisk in their asset portfolio and rebalancing the loan portfolio in favour ofretail assets. The fourth strategy is to increase the emphasis on sector-wisefollow-up to recover NPAs.
In the years when the Indian economy was registering GDP growth rate inexcess of 8 per cent annually, the PSBs catered credit to new infrastructureand commercial real-estate projects and other industries. The credit portfolioof PSBs is skewed towards the corporate loans. When economy startedregistering slack with growth slowing down, these loans started becoming stickyfor the PSBs. The strategic response of the PSBs therefore, is to lower the riskappetite and to cater to high credit rated corporates only. Also, since theportfolio of the PSBs was skewed towards corporate loans, the usual alternativewas to address the concentration risk by shifting the focus towards retailsegment which is less risky and has lower NPAs.
The private sector banks have a four-pronged strategy for loan recovery. Theirresponse is to increase the monitoring of the overall loan portfolio andparticularly of those accounts that are showing incipient signals of stickiness.Improving portfolio mix and reducing concentration risk are two sides of thesame coin; and the private banks share similarity in strategy with their publicpeers with regard to this strategic response. The private banks also have anincreased focus towards a balance sheet clean-up by deleveraging stressedassets by sales or enforcing contractual rights.
The foreign banks operating in India, have an urban contour to their businessand they cater largely to retail segment of the 'emerging affluent'. Foreign banksstrength is their technologically sound and savvy management which reflectsin their strategy of stress-testing and scenario analysis of the loans portfolios;which is aimed at assessing banks' ability to maintain operations during periodsof stress. While the first strategy is largely focused on overall operations of thebanks considering their respective positions; the second strategy of takingremedial actions including exposure reduction, security enhancement orstrategic exit is oriented towards recovery of loans.
Thomas & Vyas: A Comparative Analysis of Loan Recovery Strategy of Indian Banks 83
Structure
The SBI & Associates has a separate vertical for Stressed Assets Management(SAMG) headed by a Deputy Managing Director. The PSBs also have a separatevertical for recovery with senior officials specially posted for managing recoverymeasures in all zone offices. Both the bank-groups follow a hierarchicalmanagement structure.
The private sector banks have created an independent risk management groupfor overseeing recovery efforts. Also some private banks have separate groupsfor resolution & recovery in various loan segments such as Small & MediumEnterprises and Retail. These groups have a centralised reporting and actionmechanism. The foreign banks operating in India have a separate group formanagement of impaired accounts. The management structure is similar toprivate sector banks; it is centralised and flat.
Systems
The SBI & Associates employ an Early Warning System for incipient stressidentification and appropriate response for the flow of NPAs. The group alsohas digitalised Litigation Management System and NPA Portal for monitoringand control of NPA accounts. The integrated system is utilised in legal recoverymeasures employed by the group and tracking and controlling stock of NPAs.
The PSBs also employ digitalised tool for Credit Audit/ Loan Review Mechanism(LRM) for evaluation of asset quality and arresting the flow of NPAs. The PSBsalso utilise digital systems to conduct regular analysis of the portfolio so as toensure ongoing efforts for recovery and up-gradation.
The private banks with their better digital strength compared to other groupsuse predictive models for identification of incipient stress. The system is amore sophisticated form of an Early Warning System. Private banks also utilisevarious databases to review the borrower's profile before sanction and at everydisbursement.
Foreign banks operating in India employ systems similar to their private peersthat include periodic performance reviews of accounts for detection of incipientstress and additional review and follow-up for delinquent and stressed accounts.
Style
The recovery efforts monitoring and management style in all the bank groupsis found to be on similar lines. The recovery management teams work underdirect supervision of either a committee of the board, or a sub-committee ofthe board. The recovery teams in all the bank groups function with completeindependence with no other tasks or assignments at hand except the recoveryof NPAs.
84 Prajnan
Staff
The bank groups have set-up independent verticals for recovery and resolutionof stressed assets. While the SBI & Associates and PSBs deploy recovery officialsto their specially created offices/branches for recovery and resolution; theprivate sector banks and foreign banks create independent sub-groups forrecovery. These are functional groups and have a flat reporting structure.Officials are deployed to these groups as and when required. All the bankgroups deploy help from recovery agencies and recovery agents under the extantguidelines of IBA and RBI.
Skills
The foreign banks operating in India and the private banks usually employofficials with expertise in recovery and resolution; while due to the nature ofrecruitment and hierarchy in SBI & Associates and PSBs; they employ officersfor recovery who have a varied experience of general banking and credit/loans.The focus of all the bank-groups however, is on imparting and enhancingrecovery and resolution skills of the officials.
The findings from the comparison of effectiveness of Loan RecoveryStrategy of Bank - Groups in India are itemized as under.There exist significant differences in movement of Gross Non-Performing AssetsRatio between different bank groups. The study determined that there existsignificant differences in movement of Gross Non-Performing Assets Ratiobetween different bank groups. Statistically significant differences were however,only established between SBI & Associates and private sector banks.
The GNPA Ratio of PSBs has been found to be the highest among bank groupsover the entire period of study (Average GNPA Ratio recorded at 4.51 per cent).Followed by SBI & Associates and foreign banks (Average GNPA recorded at4.32 per cent and 3.18 per cent respectively). The private sector banks reportedlowest GNPA Ratio for the duration under study (Average GNPA Ratio at 2.61per cent).
It is noteworthy however, that there exist no significant difference in the trendof movement of GNPA Ratio between PSBs and private banks & foreign banks.Propinquity in the movement of GNPA Ratio among PSBs, private banks andforeign banks is therefore, established. Propinquity in movement of GNPA Ratiois also established for SBI & Associates and PSBs & foreign banks.
The Recovery Ratio of the four bank groups reveal noticeable differences whichcan be directly attributed to banks' recovery efforts. The private sector banksrecord increasing Recovery Ratio for the period 2011-2016. For the period2014-2017 the private banks Recovery Ratio is recorded higher than otherthree bank groups.
Thomas & Vyas: A Comparative Analysis of Loan Recovery Strategy of Indian Banks 85
Statistically significant difference could not be established in the movement ofRecovery Ratio of the bank-groups for the period 2005-2017; howeverstatistically significant difference in the movement of Recovery Ratio of thebank-groups for the period 2012-2017 is established. Analysis of Mean RecoveryRatio for the period 2012-2017 reveals that private banks record highest MeanRecovery Ratio, followed by SBI & Associates and PSBs.
Additional non-parametric statistical test is performed on (i) Sub-StandardAssets Ratio, (ii) Doubtful Assets Ratio and iii) Loss Assets Ratio of the threebank groups – (Public Sector Banks, including the SBI & Associates; PrivateSector Banks and Foreign Banks) for the period 2005-2017. The statisticaltest finds significant differences in the movement of the other additional assetquality indicator amongst three bank groups. Further analysis of mean revealsthat the public sector banks Sub-Standard Assets Ratio and Doubtful AssetsRatio are higher than those of private banks and foreign banks.
Section VIIIConclusions
The study concludes that the NPA indicators of the private sector banks havebeen significantly lower than those of the public sector banks and the SBI &Associates. The study further finds evidence that the NPA recovery efforts ofthe private sector banks have been more successful than the other three bankgroups in the Indian banking sector. Thus empirical evidence exists for thedivergences in the asset quality indicators and NPA recovery indicators of privatesector banks and the other bank-groups. The study attributes such divergencesto the Loan Recovery Strategy being employed by the bank-groups. A descriptiveand qualitative comparison of the Loan Recovery Strategy of the bank-groupsis conducted through the McKinsey 7S model. The elements of the 7S Modelwith respect to loan recovery strategy of banks is further elaborated indiscussion and analysis.
The study surmises and infers that – (i) That private banks Asset QualityIndicators and Recovery Ratio is better than the other bank groups; (ii) Thatthe loan recovery strategy of private sector banks has resulted in significantlylower NPAs compared to the other three bank groups in India; (iii) That theloan recovery strategy adopted and implemented by private sector banks aremore effective than other bank groups.
The strategic choice of private banks i.e. (1) Proactive Monitoring; (2) Improvingportfolio mix; (3) Reducing Concentration Risk and (4) Deleveraging stressedassets by sales or enforcing contractual rights thus have been found to bemore effective in successfully dealing with the stock and flow of NPAs.
86 Prajnan
Proactive monitoring is a part of loan recovery strategy which focuses onpreventing NPA rather than recovering the asset once it becomes NPA. ImprovingPortfolio Mix is a healthy risk-management strategy which reduces the portfolioconcentration risk by balancing the total exposure to sensitive sectors. Theprivate banks focus on faster balance-sheet cleanup is also evident from theirstrategy. A faster NPA resolution either by sell-off or by enforcement, is effectivein reducing NPA carry costs and detoxifying the balance sheet thus creatingfresh appetite for lending.
The other noticeable pattern in the loan recovery strategy of private banks isthe considerable independence and autonomy in the Structure, Style and Staffdeployed by them. A significant degree of autonomy and decentralization inspecialized Loan Recovery teams/branches is a key outcome from McKinsey7S Analysis.
The private banks also employ a centralized database to review the borrower'sprofile before disbursement. The private banks thus focus on preventingslippages in the loan account through continuous monitoring during the variousphases of loan disbursement.
The Gross NPA Ratio of the bank groups show a considerable decline in thefirst half of period under study (2005-2017); however, post 2010 the GNPARatio has been consistently on the rise. The delinquency is more prominent inPSBs (GNPA Ratio at 12.95 per cent at the end of 2017) and SBI & Associates(GNPA Ratio at 9.11 per cent at the end of 2017). Private banks (GNPA Ratio at4.06 per cent at the end of 2017) & foreign banks (GNPA Ratio at 3.96 per centat the end of 2017) have noticeably lower GNPA ratio compared to SBI &Associates and PSBs. High degree of non-performing assets and poor recoveriesseverely impairs the already excessively leveraged banking business and willaggravate a lazy banking scenario where the productive sectors will be deprivedof credit thus leading to a slack in the economy. The slowdown in credit off-take is already evident as discussed earlier in the introduction of this study.
Loan recovery is thus the prime concern of the entire economy in general andbanking sector in particular. The responsibility to reduce the burgeoning ofstress in the lending portfolio lies with the banks itself. The banks have tocontain successfully the factors that lead to stress in portfolio. The adverseselection problem has to be dealt with more stringent pre-sanction measuresby respective banks. Stringent and transparent credit appraisal mechanism isthe key factor in arresting impairment of assets.
Thomas & Vyas: A Comparative Analysis of Loan Recovery Strategy of Indian Banks 87
References1. Rajan, Raghuram (2016), "Resolving Stress in the Banking System" (Dr. Raghuram G.
Rajan, Governor - June 22, 2016 - at ASSOCHAM - Interactive Meet with Industry &Trade, Bengaluru). Available at: https://www.rbi.org.in/SCRIPTS/BS_SpeechesView.aspx?Id=1009.
2. Lokare, S M (2014), "Re-emerging Stress in the Asset Quality of Indian Banks: Macro-Financial Linkages", RBI Working Paper Series No. 03.
3. Misra, R and Rajmal, R V (2016), "Determinants of Recovery of Stressed Assets inIndia", Economic & Political Weekly, 51(43), pp 63.
4. He, D (2002), "Resolving Non-Performing Assets of the Indian Banking System",Published in: India: Selected Issues and Statistical Appendix No. IMF Country ReportNo. 02/193 (September 2002)
5. Chakrabarty, K C (2013), "Two Decades of Credit Management in Banks : LookingBack And Moving Ahead", (Address by Dr K C Chakrabarty, Deputy Governor, ReserveBank of India at BANCON 2013). [ONLINE] Available at: https://www.rbi.org.in/SCRIPTS/BS_SpeechesView.aspx?Id=857.
6. Simons, K and Cross, S (1991), "Do Capital Markets Predict Problems in LargeCommercial Banks?", New England Economic Review, May, pp 51-56.
7. Caprio, G and Klingebiel, D (1996), "Bank Insolvency: Bad Luck, Bad Policy, or BadBanking?" Annual World Bank Conference on Development Economics (Vol. 79).
8. Demirgüç, A and Enrica, D (1997), "The Determinants Of Banking Crises: Evidencefrom Developed and Developing Countries", IMF Working Papers, 4325, pp 1-25.
9. Caprio, G (1998), "Banking on Crises: Expensive Lessons from Recent Financial Crises",World Bank, Development Research Group, Finance.
10. Demirguc-Kunt, A (1989), "Deposit-Institution Failures: A Review Of EmpiricalLiterature", Economic Review-Federal Reserve Bank of Cleveland, 25(4), pp 2.
11. Whalen, G (1991), "A Proportional Hazards Model of Bank Failure: An Examination ofits Usefulness as an Early Warning Tool", Economic Review-Federal Reserve Bank ofCleveland, 27(1), pp 21.
12. Barr, R S; Seiford, L M and Siems, T F (1994), "Forecasting Bank Failure: A Non-Parametric Frontier Estimation Approach". Recherches Économiques de Louvain/Louvain Economic Review, 60(04), pp 417-429.
13. Berger, A N and DeYoung, R (1997), "Problem Loans and Cost Efficiency in CommercialBanks", Journal of Banking & Finance, 21(6), pp 849-870.
14. Ranjan, R and Dhal, S C (2003), "Non-Performing Loans and Terms of Credit of PublicSector Banks in India: An Empirical Assessment", Reserve Bank of India OccassionalPapers, 24 (3) Winter 2003.
15. Rajan, Raghuram (2016), "Financial Reforms - Past and Present (Dr. Raghuram Rajan,Governor - January 29, 2016 - at NCAER, New Delhi). [ONLINE] Available at: https://www.rbi.org.in/SCRIPTS/BS_SpeechesView.aspx?Id=986.
16. Gandhi R (2016), "Financial Stability - Issues and Concerns: Are We Barking up AllRight Trees?", (Shri R. Gandhi, Deputy Governor - February 5, 2016 - at Great LakesInstitute of Management, Chennai). [ONLINE] Available at: https://www.rbi.org.in/SCRIPTS/BS_SpeechesView.aspx?Id=988.
88 Prajnan
17. Gandhi R (2015), "Asset Reconstruction and NPA Management in India", (Shri R.Gandhi, Deputy Governor - September 15, 2015 - ReModel in India - AssetsReconstruction and NPA Management Summit organized by Economic Times, Mumbai).[ONLINE] Available at: https://www.rbi.org.in/SCRIPTS/BS_SpeechesView.aspx?Id=974.
18. Sinha, Anand (2011), "Indian Banking - Journey Into The Future", Remarks by MrAnand Sinha, Deputy Governor of the Reserve Bank of India, at Bancon- 2011, organizedby the Indian Bankers Association and Indian Overseas Bank, Chennai, 6 November2011.[ONLINE] Available at: http://www.bis.org/review/r120201c.pdf.
19. Donnery, Sharon (2016), "NPL Workout And Resolution In The Euro Area", Addressby Ms Sharon Donnery, Deputy Governor of the Central Bank of Ireland, at the PetersonInstitute for International Economics, Washington DC, 6 October 2016.. [ONLINE]Available at: http://www.bis.org/review/r161010a.htm.
20. Jayamaha, Ranee (2005), "Promoting Credit Culture and Expediting the Process ofDebt Recovery", Speech by Dr Ranee Jayamaha, Deputy Governor of the Central Bankof Sri Lanka, at a workshop on proposed law on restructuring of sick industries andother legal issues at the Centre for Banking Studies, Rajagiriya, 1 February 2005. .[ONLINE] Available at: http://www.bis.org/review/r050211d.pdf. [Accessed 24 January2017].
21. Jalan, Bimal (1999), "Finance and Development - Which Way Now?", (Bimal Jalan,atthe Administrative Staff College of India, Hyderabad on December 6, 1999). [ONLINE]Available at: https://www.rbi.org.in/scripts/BS_SpeechesView.aspx?Id=40. [Accessed24 January 2017].
22. Muniappan, G P (2002), "The NPA Overhang - Magnitude, Solutions, Legal Reforms",(Shri G. PPMuniappan,at CII Banking Summit 2002 at Mumbai on April 1, 2002.).[ONLINE] Available at: https://www.rbi.org.in/scripts/BS_SpeechesView.aspx?Id=104.[Accessed 24 January 2017].
23. Singh, C and Brar, J S (2016), "Stressed Assets and Banking in India", IIM BangaloreResearch Paper No. 507.
24. Kothari, C R (2004), "Research Methodology: Methods and Techniques", New AgeInternational.
25. Moen, R D; Nolan, T W and Provost, L (1999), "Quality Improvement Through PlannedExperimentation", McGraw Hill Professional.
26. Provost, L P (2011), "Analytical Studies: A Framework for Quality Improvement Designand Analysis", BMJ Quality & Safety, 20 (Suppl 1), pp 192-196.
27. Channon, D F and Caldart, A A (2015), McKinsey 7S model. Wiley Encyclopaedia ofManagement.12:1.
28. Kaplan, R S (2005), "How the Balanced Scorecard Complements the McKinsey 7-SModel", Strategy & Leadership, 33(3), pp 41-46.
29. Waterman, R H; Peters, T J and Phillips, J R (1980), "Structure is Not Organization",Business Horizons, 23(3), pp14-26.
30. Clued, J (1996), "Organizational Alignment: The 7-S Model", Harvard Business SchoolCase 9-497-045. (Nov,19,1996).
31. Swamy, V (2015), "Modelling Bank Asset Quality and Profitability: An EmpiricalAssessment", (No. 2015-27). Economics Discussion Papers.
32. Samir, D and Kamra, D (2013), "A Comparative Analysis of Non-Performing Assets(NPAs) of Selected Commercial Banks in India".
Dynamic Capital Adequacy Ratio forBringing Equilibrium in Lending in
the Banking Industry: A Study of theFive Largest Banks in India
Akash Baruah
The problem of Non-Performing Assets in banks have always been aconcern, mainly because of the far reaching impact that banks haveon the economy as a whole. There has always been a trade off in banksin terms of tighter regulation and higher risk taking. Arguments havebeen made in the past that banks tend to make riskier loans withtighter regulations. But it can be noticed that not all banks are efficientof the same degree and thus not all banks should be regulated on thesame scale. This paper does a study of the five largest Indian banksand comes to the conclusion that the capital adequacy ratio is incapableof reducing risks in banks in terms of growing NPAs. Data suggeststhat non-performing assets continued to rise regardless of the banksmaintaining high capital adequacy ratios. In the paper, we develop aratio termed as the Marginal Cost of Lending and compare it acrossbanks to find out that not all banks are similarly efficient. Thus havingthe same capital requirement does not incentivize banks to performbetter neither does it penalize them. This led to the formulation of adynamic capital adequacy ratio, by taking into account the MarginalCost of Lending, which will reflect the inherent risks of banks. Thepurpose of the study is to create a self-correcting mechanism whichprevents low quality lending and help reduce the problem of NPAs.Traditionally a higher capital ratio reflected the safety of the banks interms of its loss absorbing capacity, but a higher dynamic capital ratiowill reflect its inherent risks.
Keywords: Capital Adequacy, Bank Failure, Banking Regulation, Risk inLending, Non-Performing Assets
JEL Classification: G21; G28; G38; E58
Akash Baruah ([email protected]) Assistant Professor of Finance, Pune Institute of BusinessManagement, Ph D Scholar (Finance), Assam Don Bosco University, Assam.
90 Prajnan
Section IIntroduction
General OverviewThe Indian banking sector has been recently grappling with the problem ofmounting bad assets. The Reserve Bank of India (RBI) has issued manyguidelines in order to help the banks solve this problem but the progress hasbeen very slow. The severity of the problem can be gauged from the fact thatthe balance sheet of the five largest lenders of India, which are State Bank ofIndia, Punjab National Bank, ICICI Bank, Canara Bank and Bank of Baroda,accounted for 32.27 per cent of India's Gross Domestic Product (GDP), as of2017. The basic issue which the regulators face is that all the measures toresolve bad assets are ex-post and therefore their actions are generally behindthe curve. We found that the banks enjoyed a very high Capital to Risk-WeightedAsset Ratio (CRAR) all throughout the study period and yet was not able toresolve the issue of rising bad assets. The reason for this is that the bankswere adjusting to the capital requirements mostly through the numerator whichis made up of Tier I and Tier II capital. According to (Estrella et al. 2000) thebanks were able to set aside large amounts of profits as reserves and alsoraise capital from the market, mostly on favorable terms. Thus, even thoughthe Gross Non-Performing Assets (GNPAs) were rising, lending kept onincreasing with the fallacy of higher CRARs (Acharya, 2003). The basic featureof a capital requirement is to make banks safer, but in spite of the high capitalratios GNPAs were added at an unsustainable pace. This increased the inherentrisks in banks and posed threats towards a systemic slowdown of the economy.
This fundamental problem of the CRAR not being reflective of the inherentrisks in banks in terms of mounting bad assets led to the formulation of thedynamic CRAR. Firstly, a particular bank fails if it is not being able to honourits liabilities. It defaults on its liabilities when the assets financed throughthose it goes bad. Hence, creation of bad assets is positively correlated withbank failure. Therefore, there should be more responsible creation of assetsfor greater financial stability.
Secondly, there should be some mechanism to encourage efficiency in banksby incentivizing them for undertaking better lending practices (Cohen &Scatigna, 2014). Banks should be incentivized for aiming at higher profitsthrough lower risks and not for higher profits through higher risks. Profitabilityof a bank should not be a criterion for allowing it to lend more, instead theunderlying quality of its assets should be. We have seen to the run up of theFinancial Crisis of 2007, that how profitability can be a driver of irresponsiblelending and what its consequences can be. Thus, the dynamic CRAR wouldallow banks to have different CRARs based on their efficiency, which is areflection of their asset quality and NPA levels. This will incentivize banks toearn profits through responsible lending and factor in the efficiency of thebanks (Dio, 2003).
Baruah: Dynamic Capital Adequacy Ratio for Bringing Equilibrium in Lending... 91
Thirdly, we need to bring equilibrium in lending in the banking industry througha dynamic CRAR. With a constant CRAR not all the banks are lending at theiroptimum level, some are over lending and others are under lending (Allen etal. 2000). This is because both efficient and inefficient banks are subject tothe same set of regulations and are lending with the same amount of freedom.Thus good banks are not incentivized by giving them the capacity to lend moreand bad banks are not penalized by allowing them to lend at the same rate.Thereby the banking sector is not in equilibrium. We try argue for a situationin which each bank can lend at its optimum level by a discriminatory dynamicCAD and thereby leading to an equilibrium in the whole banking system. Thesame can be seen in a financial market where different coupon rates are paidon different bonds, which is also a reflection of the underlying efficiency of thecompanies issuing them. It can be further stated that as the company improvesits efficiency, it can borrow at better market rates and thereby increase itsleverage. The dynamic coupon rates keep a check on the standard of thecompany and provides it with an incentive to keep improving.
Model OverviewWe first develop a ratio by taking Addition of GNPA/Δ Advances and establishthe Marginal Cost of Lending. The MCL gives us a very good insight on theamount of assets going bad as a percentage of amount of assets being createdand gives us an understanding of the efficiency of the bank. This ratio wouldappraise us about the bank's deteriorating lending quality and call for preventivemeasures.
We then define a bankruptcy model; in which we prove that the CRAR isincapable of reducing lending in situations of mounting bad assets. It is onlybecause that the MCL has a steeper slope than the Marginal Revenue fromLoans (MRL) (Δ Revenue/ΔAdvances), banks reduce lending in response togrowing NPAs.
Further the MCL ratio is added to the base capital requirement to come upwith the dynamic CRAR. It is argued that a rising dynamic CRAR is a reflectionof higher risks in banks and therefore they try to converge towards the baserate by becoming more efficient.
In this paper we do an empirical analysis of the top five largest banks of Indiathrough their financials. All the data used in the study has been taken fromtheir Annual Reports and RBI database, starting from 2007 till 2017. The restof the paper is structured as follows, in Section II we do an asset quality reviewof the five banks and state certain specific reasons for the growing NPAs. InSection III we define the Marginal Cost of Lending Ratio and explain howimportant it is for identifying potential risks in banks. In Section IV we developthe bankruptcy model to validate that the CRAR was unable to reduce themounting bad assets problem and in Section V we explain the dynamic CRAR,
92 Prajnan
its derivation and provide a simulation to explain how the banks CRAR wouldhave looked like if they were dynamic. In Section VI we provide a final conclusionand sum up the results of our findings.
Section IIAsset Quality Review
Impact AnalysisGenerally, the Reserve Bank of India (RBI) inspects the books of accounts ofbanks annually because of its yearly financial survey. In accordance with this,an exceptional inspection was organized during the month of August in FY2015-16 and was termed as Asset Quality Review (AQR). During a normalroutine financial inspection, RBI scrutinizes a small sample to examine if theclassification of assets were done in accordance with the repayment schedule.
This time however, the AQR was done on a much larger sample size, whichincluded most of the large borrower accounts, to assess if the classificationwas done keeping the prudential norms in mind. Through this almost 200accounts were recognized and were directed to be classified as non-performing.
As a result of this exercise the GNPAs of the five banks in 2015 amounted to `126,816 Crores (Cr) and net NPAs to ` 66,052 Cr. The addition of new GNPAswas also an astonishing high number of ` 73,485 Cr. These figures had a verystrong momentum and witnessed high growth rates in the following year of2016, with GNPAs, net NPAs and addition of new GNPAs amounting to` 252,371 Cr, 144,432 Cr and ` 175,714 Cr. The Net NPAs had a growth of118.66 per cent in 2016, which was a reflection of the deteriorating asset qualitysituation of the Indian banks. This was because most of the banks were reluctantto classify their loans as non-performing and were resorting to ever – greeningof accounts. However, in 2017 the growth in the Net NPAs were low at 7.96 percent due to the efforts from the RBI through Various Schemes such as theStrategic Debt Restructuring Scheme (SDR), Joint Lenders Forum (JLF), 5/25Scheme etc.
The ever – greening of accounts, which amounted to the banks not taking anysteps to resolve the issue, allowed the NPAs to keep on growing. This resultedin the addition of new GNPAs growing from ` 11,232 Cr in 2007 to 175,713 Crin 2016. Therefore, the banks were required to undertake some tough measuresfor providing permanent solution to the problems at hand, which resulted inthe write offs increasing from ` 38,483 Cr in 2015 to ` 62,169 Cr in 2017.Even though the banks profit margins were hit by this step, it was necessary tosend a strong signal to the markets and the world community.
Baruah: Dynamic Capital Adequacy Ratio for Bringing Equilibrium in Lending... 93
Figure 1Alarming Ratios
The severity of the problem can be understood from the figure above, whichshows the trend of two important ratios of the five banks calculated on aconsolidated basis. We can see that the Provisions/GNPAs ratio, also known asthe Provision Coverage Ratio, started to decline from 2012 and continued tofall until 2016. It was only in 2017 that it had risen, after the dual impact ofhigher provisioning requirements and declining GNPAs. Another ratio is theNet NPA/Total Eligible Capital, in which the denominator is the aggregate ofTier 1 and Tier 2 capital. This ratio gives an idea about the loss in eligiblecapital if the whole of Net NPA would have to be written off. A rise in the ratiois an indication of the deteriorating asset quality of the bank and is a reflectionof the underlying inherent risks (Cao and Chollete, 2014). We can see that theratio started to increase from 2012 and kept on increasing until 2016 andpeaked at 29.91 per cent. One thing which can be interpreted from Figure 1 isthat both the lines in the graph should diverge from each other in order toindicate financial stability and the convergence of the same would indicatedeteriorating financial health of the banks.
Reasons for Deteriorating Asset QualityA thorough analysis of the data, collected from the sample of five banks, revealedcertain patterns which explained why the banks were unable to recognize thesevere mounting bad assets problem. We have classified our findings into threebroad categories in accordance with what the data had to reveal.
It can be seen from Figure 2 that the increase in addition of GNPAs startedfrom 2008 whereas the write offs remained constant. This led to the start ofthe widening of the gap between the GNPAs and the write offs. In fact, in 2012there was a dip in the write offs, whereas the addition in GNPAs continued itsexponential rise.
94 Prajnan
Moral Hazard:
Figure 2Moral Hazard
The increase in write offs from 2013 was not enough to compensate the growingbad assets problem and the gap was at its peak in 2016. The diverging trendwas a matter of serious concern for the bankers and the regulators, to whichthey reacted considerably very late. But 2017 brought in some hope as boththe figures started to converge, with addition of bad assets declining and writeoffs picking up. The reason why writing off is so difficult because it bringsabout a reduction in the bank's balance sheet along with huge negative publicity.
Solvency Problem:The risk in a bank increases as the bad assets or GNPAs expand. The banksneed to provision for these bad assets. But we know that a bank's profits arelimited and thus if the level of bad assets increases beyond a certain point, itwould not be able to provide for them. This leads to a situation where a bank'sreserves are used up and poses a threat to its solvency. Later on when wedevelop the Bankruptcy Model, this issue will be dealt with in detail. It can beseen from Figure 3 that initially in 2009, the gap between the Addition of GNPAsand Addition in Reserves was very low and hence any rise in bad assets couldhave been provided for without much difficulty. But the situation started tochange after 2011 and the gap between the two extended. As the GNPAs divergefrom the PAT, bank's ability to provide for bad assets through profits diminishes.
Baruah: Dynamic Capital Adequacy Ratio for Bringing Equilibrium in Lending... 95
Figure 3Solvency Problem
Therefore, probability of provisions entering into the reserves of banks increasesand thereby makes them riskier (Andrle et al. 2017). The risk is in terms oflosses eroding away all the eligible capital (Tier1 and Tier 2) and forcing thebank into liquidation. The situation worsened from 2015 and reached its peakin 2016. Though we have seen some trends of reversal in 2017, there is still along way to go before we can think to relax.
Recognition Problem:It's a normal practice to recognize growing bad assets in the system and takeappropriate measures to tackle it. One of the ways of resolving a bad assetsproblem is reducing the amount of lending and becoming more cautious inone's approach. We can see from Figure 4 that the GNPAs started to increasefrom 2011 and had a steady growth until 2014. Then it had a very steep riseand reached its peak in 2016. But the addition in advances kept on happeningdespite the evident rise in bad assets. It was only after 2014, that the banksslowed down its lending and also started to take other remedial measuressuch as write offs.
Figure 4Recognition Problem
96 Prajnan
The constant increase in advances is a clear sign of lack of acceptance andcomplacency on the part of the bankers in terms of assessing the bad assetssituation. The Net NPA was also diverging from the Gross NPA, which showsthat write off as a measure was not adequately adopted. Serious steps regardingrecognizing the situation was taken only after 2015 when the addition inadvances started to decline and the Net NPAs started to rise.
Thus these three reasons give us an understanding as to why the five banksresulted in the situation in which they are now and sheds light on the probablereasons of their mismanagement. The threats to bank failure are very evidentand needs to be dealt upon with sincerity for a proper resolution.
Section IIIMarginal Cost of Lending
Derivation of the RatioEvery firm needs to be efficient by making maximum utilization of theirresources. The banks need to make the best possible use of their capital andsee to it that they generate the least amount of NPAs. What we have seen fromour analysis is that the CRAR is not an adequate reflection of the inherentrisks in banks and neither does it enlighten us on the degree of efficiency inbank lending. This led to the derivation of a ratio by taking Addition ofGNPA/Δ Advances, which would appraise us on the amount of assets going badper rupee of advances.
We do not take the negative values of the ratio into consideration while doingour analysis, as a negative number implies that the bank has actually contractedits lending. The ratio should have a downward trend with positive values. Anefficient bank will have the ratio decreasing, with the denominator increasingand the numerator decreasing, but never approaching zero. The numeratorcan be zero only if a bank does not have any of its loans turning bad and ishighly efficient.
Marginal Cost of Lending (MCL)
= Addition of GNPAtn
ΔAdvancesn
= ΔGNPAn + Write Offtn
ΔAdvancesn
= (GNPAtn – GNPAtn-1) + Write Offtn
ΔAdvancestn
Baruah: Dynamic Capital Adequacy Ratio for Bringing Equilibrium in Lending... 97
= (GNPAtn-1 – GNPAtn-2) + (Addition of GNPAtn-1 – Addition of GNPAtn) +
Write Offtn
ΔAdvancestn
= ΔGNPAtn-1 + ΔAddition of GNPAtn + Write Offtn-1
ΔAdvancestn
*The calculation of the above equation can be checked from annexure 1
(a) Addition of GNPAs – Amount of bad assets added each year during time periodtn
(b) ΔAdvancestn – Amount of advances added each year during time period tn
(c) ΔGNPAtn – Amount of GNPAs added each year during time period tn
(d) Write Offtn – Amount of bad assets written off each year during time period tn
This ratio tells us about the amount of assets going bad each year as apercentage of the amount of assets created. If we find the ratio increasing, itcan be implied that there are issues with the productivity and efficiency of thatparticular bank in terms of its lending. An analysis and breakdown of the ratioreveals that the previous period's ΔGNPA and Write off also affects the currentperiods MCL. Thus this allows us to identify trends that might be accumulatingfrom several previous years and thereby help us to prevent further building upof risks in the system.
Inefficiency of the CRARWe can see from the figure above that the consolidated MCL (MCL of the fivebanks) started to increase from 2011 and had a steep rise from 2014 when theAQR was called in by the RBI. But the inefficiency of the banks in terms ofrising bad assets was not fully reflected through the CRAR as it remained wellabove the Basel III requirements. A high CRAR implied that the bank was wellcapitalized and inherently less risky, but it did not reflect its lendingproductivity. Thus a rising MCL appraises us of the loss in productivity andthe rise in bad assets, which is lacking through the CRAR.
The rising MCL of the five banks from 2014 meant that the GNPAs per rupeeof advances were increasing and the adjustment of the CRAR should havehappened through the denominator, which is by reducing lending. We cananalyze from Figure 5 that the adjustment of the CRAR happened through thenumerator, as the advances kept on increasing. This phenomenon continuedfrom 2011 till 2014, but from 2014 the adjustment finally started to happenthrough the denominator, as the reduction in advances can be seen. If the MCLwas made public, investors would have factored this in their valuations andasked for higher returns. Subsequently, higher dividends and interest paymentswould have required the banks to adjust through the denominator as raisingcapital would become costly. A rising MCL indicates that something is notcorrect with the lending function of the bank and hence calls for a check onthe credit expansion. This is one of the main reasons why the CRAR has notbeen able to reduce risks in terms of growing NPAs, as the market does notfactor in the efficiency and the CRAR is adjusted through easy raising of capital.
Section IVThe Bankruptcy Model
Derivation of the ModelWe first define another ratio called the Marginal Revenue of Lending (MRL)which tells us the revenue earned per unit of lending. The ratio is as follows:
Baruah: Dynamic Capital Adequacy Ratio for Bringing Equilibrium in Lending... 99
Marginal Revenue of Lending (MRL)
= ΔReservestn
ΔAdvancestn
MRL tells us how productive a bank is in terms of its lending. This ratioincreasing signifies that the assets created by the banks are generating revenuesand resonate efficiency of the bank.
The Bankruptcy Model:Let X be the amount of GNPAs
Let Y be the amount of provisions required
Let Z be the amount of profits earned in a particular period tn
Let time = t1, t2…….tn
→ Ytn = β1* GNPAtn β1 → {0, 1}
→ β1* GNPAtn ≤ Ztn →
The above equation can only hold true when Ztn> Ytn,
Let Q be the reserves at the time tn-1
Now, if Ytn > Ztn, the following condition will hold true at any time tn for avoidingbankruptcy
→ β1* GNPAtn ≤ (Ztn + Q) →
Profit is appropriated to various reserve accounts, Reservetn (R) = Profittn +Reservestn-1
Equation can be re - written as → β1* GNPAtn ≤ Rtn →
Now we divide both sides by equation by ΔAdvancestn
β1* GNPAtn ≤ Rtn →
ΔAdvancestn ΔAdvancestn
We know that, GNPAtn = GNPAtn-1 + Addition in GNPAtn – Write Offtn and Rtn =Rtn-1 + ΔR
* The bankruptcy ratio is presented in Figure 6 calculated with the model developed above.The calculations can be checked from Annexure 1 & 2.
Equation (5) should hold true at all points in time for the bank to remainsolvent. IF the Left hand side (LHS) of the equation becomes greater than theRight Hand Side (RHS), the provisioning requirement will be greater than theprofit for the current period and previous period's reserves combined. Thiswould imply that the going concern capital of the bank has eroded away andwould call for liquidation. Thus if the GNPA in the previous period increases,the write off in the next period should also increase. Now when the GNPA onthe LHS increases, the reserves on the RHS should decrease. This assumptioncomes from the fact that when the GNPAs are on the rise, the banks wouldreduce their lending to curtail future accumulation of bad assets. Therefore,the RHS cannot be adjusted at will and the adjustment has to be done on theLHS through the increase of write offs.
But the banks can take another decision, which is to increase lending. Whenthe advances increase, the revenues also increase. Therefore, an increase on
Baruah: Dynamic Capital Adequacy Ratio for Bringing Equilibrium in Lending... 101
the LHS can be adjusted through increased revenues on the RHS, revenuesbeing part of reserves, which is contrary to solving the main problem of risingGNPAs. In theory, advances can keep on growing and so also revenues, evenwhen GNPAs are on the rise. But the only phenomenon which keeps a checkon the banks in terms of lending when the GNPAs are on the rise, is the slopeof the MRL and the MCL. As the MCL has s steeper slope than the MRL, anincrease in lending might lead to a situation in which the LHS increases at amuch faster rate than the RHS. Therefore, this is the natural process throughwhich a bank is made to put a check on its lending. But the banks can alwaysdelay the process by raising easy capital, as equity forms a part of the reserves.Hence the RHS is adjusted by other ways which allows the banks to keep onlending. We can look at Figure 7 for understanding the steepness of the MRLand MCL and Figure 6 for understanding the delaying process.
Figure 7MCL and MRL
Figure 8Delay Process
102 Prajnan
In Figure 7, we can see that from 2015 there is a steep rise in the MCL incomparison with the MRL. This reflects the fact that whenever recognition ofbad assets happens and there is a steep rise in the GNPAs, the slope of theMCL will rise at a much faster rate than the slope of the MRL. Hence theconvexity of MCL is much higher than the MRL.
Figure 8 helps us to understand the delaying process to some extent. As wecan see that there was an increase in addition of reserves and addition ofadvances from 2011 till 2014, even if the GNPAs were on the rise. It was onlyafter 2014, when the AQR was called in by the RBI that the reserves andadvances started to decline. The GNPAs also had a steep rise from 2014, mostlyon account of the intervention of the RBI and its push to make the banksrecognize the bad assets problem. This tells us the delaying mechanism of thebanks, who delayed the recognition of bad assets from 2011 till 2014 andperhaps allowed the problem to grow bigger (Acharya 2001). During this time,it kept on its lending drive and hence adjusted to the RHS with higher reserves.
Section VThe Dynamic CRAR Model
Derivation of the ModelThe CRAR has been unable to reduce risks in banks in terms of lower GNPAsand also to reflect its efficiency and productivity, as established earlier from3.2 – Inefficiency of CRAR. We make the CRAR reflective of the efficiency andproductivity of the bank and address the fundamental problem of adjustmentof the CRAR through the numerator. We have to make certain assumptions fordeveloping the model.
1. All the banks start of from the same CRAR as set by the regulator
2. No bank has the tendency to pile up reserves. Hence profits earned are lentout and not transferred to reserves
3. The dynamic CRAR cannot go below the standard base CRAR, unless alteredby the regulatory authority
4. 50 per cent of (MCLtn*Base CRARtn) will be added to Base CRARtn, to make itdynamic
5. The adjustment from dynamic CRARtn-1 to dynamic CRARtn will be done in theratio 90:10, where 90 per cent will be adjusted through the numerator and 10per cent through the denominator
The General CRAR as per BASEL III recommendations are:
Baruah: Dynamic Capital Adequacy Ratio for Bringing Equilibrium in Lending... 103
We have concentrated more on the credit risk part of the Risk Weighted Assets(RWA), because it accounts for the largest portion of the RWAs. On average itaccounted for 86.5 per cent of the total RWAs of all the five banks combinedfrom 2008 till 2017 (Figure 9).
This adjustment will be done in the ratio 90:10, where in the 90 per cent of theadjustment will be done through the numerator and 10 per cent through thedenominator. Thus from equation 1 we can say that the MCL in Yeartn is goingto decide how much the CRAR changes from Yeartn-1 to Yeartn and therefore,makes the CRAR dynamic. As the MCL increases the CRAR also increases,similarly when the MCL decreases your CRAR reduces. But the CRAR cannotgo below the threshold, which is set by the regulator. This implies that whenthe firm will be operating at optimum efficiency, it will have the threshold limitas its CRAR. Thus an increase in the dynamic CRAR would not only amount tolocking up of funds but will also result in direct reduction of lending, throughthe mandatory 10 per cent reduction in the denominator. This structure willbring in competition among the banks to operate at the optimum level and
104 Prajnan
have the maximum amount of funds at disposal for lending. A self-correctingdynamic CRAR will help better monitor, manage and control risk in terms ofgrowing NPAs. Another feature is that, a rising dynamic CRAR signifies problemswithin the firm and calls for thorough internal investigation. Whereas a normalrising CRAR will reflect a stable and robust firm and fails to reflect the inherentinefficiency.
Interpreting Results from the Model
Figure 10Dynamic CRAR
Figure 11MCL Trend
We have done a simulation on the five banks of our study, in which we calculatedthe dynamic CRAR of each bank from 2008 till 2017, using the model describedin section 5.1. In Figure 10 we can observe how the dynamic CRAR changesfrom 2008 and starts to rise from 2013, pushing the banks away from theiroptimal level of functioning. The only bank which moves back towards theoptimum level of 9 per cent is Bank of Baroda in 2016, mainly because of its
Baruah: Dynamic Capital Adequacy Ratio for Bringing Equilibrium in Lending... 105
steep cut in lending. These changes in the dynamic CRAR can be justified bylooking at the MCL trends in Figure 11. The MCL of the five banks started torise from 2013 and took a very steep upward turn from 2014 onwards. Thisrising MCL resulted in the change of the dynamic CRARs across all the banks,signifying inherent problems. Only Bank of Baroda was able to reverse backtowards the base rate because of the steep reduction in its MCL in 2016.
Figure 12Static CRAR
The static CRAR was well above the BASEL III base rate of 9per cent for all thebanks. Figure 12 depicts a very optimistic scenario in which everything looksin order. The banks seem to be capitalized well above the required amountand thus poses no threat to the banks in terms of their solvency. What itcompletely fails to capture is their productivity and efficiency in terms of lending.The main reason for this is that the banks were able to adjust the CRAR throughits numerator, as argued before in Section 3. Hence the latent problem ofgrowing NPAs remain unattended.
Section VIConclusion
The dynamic CRAR brings in equilibrium in the banking industry through thefollowing ways:
It will reflect the inherent risk of the banks and differentiate them on thebasis of efficiency
Due to the classification on the basis of risk and performance, there would bebetter allocation of resources as savings will shift from relatively bad torelatively good banks. Thus the good banks would have more funds at disposalfor lending.
Dynamic CRAR will initiate competition among the banks in terms of bringingthe CRAR as close as possible towards the base rate and hence operate at the
106 Prajnan
optimal level. Banks will try to free up as much reserves as possible andincrease their lending.
This competition among the banks to operate at the optimum level will bringin equilibrium in lending in the banking industry.
References1. Acharya, V (2001), “A Theory of Systemic Risk and Design of Prudential Bank
Regulation”, Working Paper Stern School of Business, New York University.
2. Acharya, V (2003), “Is the International Convergence of Capital Adequacy RegulationDesirable?. Journal of Finance, 58 (6).
3. Allen, F and Gale, D (2000), “Bubbles and Crises”, The Economic Journal, 110,pp 236-255.
4. Andrle, M; Tomšík, V. and Vlcek, J (2017), “Banks' Adjustment to Basel III Reform: ABank-Level Perspective for Emerging Europe”, International Monetary Fund WP/17/24.
5. Bénétrix, A; Lane, P and Shambaugh, J (2015), “International Currency Exposures,Valuation Effects and The Global Financial Crisis”, Journal of International Economics,96, pp 98-109.
6. Cao, J and Chollete, L (2014), “Capital Adequacy and Liquidity in Banking Dynamics”,US Working Papers in Economics and Finance.
7. Cohen, H B and Scatigna, M (2014), “Banks and Capital Requirements: Channels ofAdjustment”, Bank for International Settlements Working Papers No 443.
8. Dio, G (2003), “The Foundations of Banks' Risk Regulation: A Review of the Literature”,HEC Montréal CIRPÉE and CREF Working Paper, 03-08.
9. Estrella, A and Park, S (2000), “Capital Ratios as Predictors of Bank Failure”, FRBNYEconomic Policy Review.
Baruah: Dynamic Capital Adequacy Ratio for Bringing Equilibrium in Lending... 107
Connected or Disconnected –The Art of Operating in Connected World
Micke DamrmellKapil Rampal
New Delhi, Sage Publications India Pvt. Ltd., 2018, xvii + 158 pp., Rs. 395.00.
Reviewed by Shri Sunil Bakshi, Visiting Faculty, National Institute of BankManagement, Pune.
Connected or Disconnected is a book written by two authors - one from Swedenand another from India, both sharing same concerns based on their ownexperience researched on the topic. The book focuses on the common concerntoday shared by many people around the world, "are we becoming slaves oftechnology and missing the personal relationships?"
The book is not a technical book, so if someone is looking for technical knowledgethis book is not for them, but it is for everyone who is having smart phone or tabthat is always connected and feel that their very existence depends upon theirbeing connected. They are restless to look at the screen at every ping of the device,and happy to ignore surroundings.
The book consists of 14 small chapters; each chapter is not more that 5-6 pages,with only 2-3 pages to read. The matter included is pointed and specific full ofanecdotes, quotes and data supporting the topic of the chapter. The chapters areorganized systematically and takes the reader through digital evolution and itsimpact on coming generations. As quoted by a psychiatrist that constantbombardment of information through connected devices is making next generationless intelligent.
Although it is true that there are many adverse effects of being connected, it hasalso opened lots of avenues where being connected makes people aware of thesurroundings. Employees have freedom to be available from anywhere andcontribute, family though physically apart can be still in touch with each other.
110 Prajnan
Printed and Published by Dr K L Dhingra Director, National Institute of Bank Management, NIBM PostOffice, Kondhwe Khurd, Pune 411 048. Typeset and Designed by Publications Department, NIBM,Pune and Printed at United Multicolour Printers Pvt. Ltd,, Shaniwar Peth, Pune 411 030.
The book discusses in detail positive and negative impact of being connectedcontinuously, it also brings out advantages of being connected and in the end itprovides a good suggestions about what do at the work place and what to do athome and as parent.
Chapter One covers the proliferation of digital connectivity and hence revolutionacross the world by looking at the data from few select countries. Chapter Twodiscusses a concept called 'Hereness' that refers to the mental presence. When weare communicating our mental presence and attention makes communicationmeaningful and focused. Being connected provides distractions to thecommunications.
Chapter Three talks about the new culture of being connected. Interestingly manyof those who are associated may really do not know how to use this connectivityeffectively, they stay connected because others are, and if they stay disconnected,the fear of being labelled as regressive and not in keeping with times. The resultof this is that 94 per cent people are happy to forgo their other essentials needsof human life rather than mobile phone. (Chapter 4). Fifth Chapter discusses theimpact of constant use of mobile devices on our body based on the experimentsperformed by scientists.
Sixth Chapter explains positive and negative effects of digital revolution on humanbrain. Although being connected helps a person to develop multi-tasking skills,they also affects the concentration and skills in performing the task if done oneat a time. Chapter Seven highlights the psychological impact of being constantlyconnected. Often times people get addicted and thereby increase their stress levels,because they are constantly connected, which in turn takes its toll on their health.Chapter Nine illustrates the connected people and disconnected world byconsidering proliferation of technologies and use of social media by politicians.
Chapter 10 and 11 are most important chapters that discusses about 'What toDo at Workplace?' and 'What to Do At Home?' The well-researched suggestionsare worth reading that will help organizations and parents in bringing disciplinewhile staying connected. Chapter Eleven also provides guidelines about how to beconnected with children at school and at home. It gives a brief list of rules forparents that are worth following. Last three chapters briefly bring out theconclusions by the authors about a happy life.
The book is well written and provides brief glimpses of what we are witnessingand also provides guidance on how best to use digital revolution while stayingconnected. The authors suggest a diet plan for connectivity which is similar todiet plan for healthy body.
PRAJNANAnnual Subscription Information
In case you are not a subscriber to this international standard double-blind refereedjournal you are welcome to become one of it. Or if you have not yet renewed yourpast subscription kindly do so to patronize it and to benefit from it academicallyand professionally. You may also kindly circulate this information to your friendsand colleagues.Subscription Rates (inclusive of postage)
InlandOne year (2018-2019) Rs. 300.00Two years (2018-2020) Rs. 550.00Three years (2018-2021) Rs. 825.00
Foreign: One year (2018-2019) US $ 50.00 (air-mail)Single Copy (Inland): Rs.75.00 (back issues are available for sale)Publishing Period: Quarterly – April-June to January-MarchMode of Subscribing/Renewing
Write to us with full address(es) to where the copy(ies) of PRAJNAN is/areto be dispatched along with subscription amount and mentioning period ofsubscription.The subscription amount may preferably be transferred by RTGS/NEFT/ECSto our Current A/c No. 20002400021 with Bank of Maharashtra, NIBM Branch,Pune (IFSC Code MAHB0001124).For all electronic remittances, kindly send a confirmatory e-mail at:[email protected] giving transaction UTR No. and other details of thepayments.In case the amount is remitted by Demand Draft or at par cheque, it shouldbe drawn in favour of "National Institute of Bank Management" and payableat Pune.Please remember to mention your old subscription number for renewal.
Address all your subscription related communication to:Executive Officer (Publications)
National Institute of Bank ManagementNIBM PO, Kondhwe Khurd, Pune 411 048 (India)
Telephone : 0091-20-26716000 (EPABX) Direct : 26716216/26716410Fax : 0091-20-26834478
Note: Claims for missing issues should be made within three months from the last month ofeach quarter. All claims should clearly mention the subscription number, postal address,and email ID.
PRAJNANPRAJNANPRAJNANPRAJNANPRAJNANJournal of Social and Management SciencesJournal of Social and Management SciencesJournal of Social and Management SciencesJournal of Social and Management SciencesJournal of Social and Management Sciences
Subscription/Renewal FormSubscription/Renewal FormSubscription/Renewal FormSubscription/Renewal FormSubscription/Renewal Form
(Please tick ( ) in the appropriate box)
I wish to subscribe for Prajnan
I wish to renew my Prajnan Subscription (Sub. No._________)
An Agency discount of 10% is offered to book sellers/journaldistributors
Mailing Address and Details of the SubscriberMailing Address and Details of the SubscriberMailing Address and Details of the SubscriberMailing Address and Details of the SubscriberMailing Address and Details of the Subscriber
