Loan Loss Provisions Policy - Emerging vs. Developed Economies

1
Loan Loss Provisions Policy -Emerging vs.
Developed Economies
The Academy of Economic Studies Doctoral School
of Finance and Banking

• MSc student Irina Gabriela Bidivenciu
• Supervisor Professor Moisa Altar, PhD

Bucharest, July 2007
2
CONTENTS

• Introduction
• Literature review
• The model
• Estimation Results
• Estimations of the Loan Loss Provisions Model
  using GLS
• Estimations of the Loan Loss Provisions Model
  using GMM
• Estimations of the Loan Loss Provisions Model
  within a commercial bank
• Conclusions

3
Introduction

• The modern economies are different from those in
  the past in their ability to identify the risk,
  to measure it, to appreciate its consequences and
  in taking action accordingly.
• The loan loss provisions are a device that can
  correct the negative effects of the loan
  portfolio problems within the bank sector. The
  level of the loan loss provisions must be
  designed to cover the expected losses during the
  economic cycle.

4
Literature review

• The Basel Committee (1988) ? new method for
  evaluating the capital - assets correlation based
  on a simplified weights system algorithm and a
  minimum capital adequacy ratio of 8.
• Basel II (2004)? International Convergence of
  Capital Measurement and Capital Standards all
  the credit institutions are required to have a
  policy in relation to credit risk, arrears and
  provisioning management.
• Pérez, D., Salas-Fumás, V., Saurina, J., (2006) ?
  the banks must protect their capital from
  expected or unexpected losses through loan loss
  provisions and not to wait until the negative
  events occurred without affecting the
  transparency using the statistical provisions.
• Laeven, L. and Majnoni, G., (2002) ? banks on
  average postpone provisioning when faced with
  cyclical upturns and favorable income conditions
  until negative conditions set in (income
  smoothing practices).
• Cavallo M. and Majnoni G., (2001) ? the fiscal
  authority may affect relevant business decisions
  for banks and financial institutions.
• Fernández de Lis, S., Pagés, J. M. and Saurina
  J., (2000) ? introduction of statistical
  provisions in Spain. In good times the banks have
  to set aside provisions that might be depleted in
  bad times when the excesses of the last upturn
  appear in the form of impaired assets.

5
The Model3.1. The Model Variables

• Total Assets (A)
• Loan Loss Provisions (LLP)
• Profits Before Tax and Provisions (EBP)
• Loan Growth in real terms (?L)
• Real Growth in GDP per capita (?GDP) or Real
  Growth of Industrial Output Index (?IOI)
• Note The values of the loan loss provisions at
  time t correspond to the values of the assets at
  time t-1. ?
• Data Source
• Bankscope
• EUROSTAT
• BNR

6
3.2. The Model Hypothesis of Prudent Loan Loss
Provisioning Behavior. Data filters

• The coefficient on earnings before tax and
  provisions is negative
• The coefficient on loan growth is negative
• The coefficient on real growth rate of GDP per
  capita / the real growth of the Industrial Output
  Index is negative.
• The bank/year observations that exhibit one of
  the following features were excluded
• Ratio of loan loss provisions over lagged total
  assets gt 90 or lt10
• Earnings before provisions over lagged total
  assets gt 12
• Loan growth rate in real terms gt 56
• Loan decreasing rate in real terms gt 50

7
3.3. The Model Description

• Testing the hypothesis of imprudent behavior and
  verifying the nature of the relationship between
  banks provisions and earnings
• (1)
• The speed of adjustment of the dependent lagged
  variable is depicted through
• (2)

8
3.4. Correlation matrix
Developed Economies Income smoothing Imprudent behavior Anti-business cyclical behavior Emerging Economies No Income smoothing Imprudent behavior No anti-business cyclical behavior

9
The Estimations Results 4.1. Generalized Least
Squares
Developed Economies Emerging Economies

10
4.1. Generalized Least Squares (contd.)

• Running the GLS estimates ? the results are
  different amongst the developed and emerging
  economies
• The banks within developed countries smooth the
  income while within the emerging countries this
  is not a common practice
• The loan loss provisions follow the loan
  portfolio growth only within the emerging
  economies
• The loan loss provisions policies are correlated
  with the economic cycle.

11
4.1. Generalized Least Squares (contd.)

• Testing the stationarity ? Levin, Lin Chu
• Testing the robustness of the estimations ?
  Hausman Test (endogeneity test)
• Developed Countries the fixed effects results do
  not differ significantly from the random effects
  results
• Emerging Countries when running an auxiliary
  regression the resid term takes value of 0.001

12
4.1. Generalized Least Squares negative
earnings dummy (contd.)
Emerging Economies ? Hausman test - robustness Emerging Economies ? Hausman test - robustness
GLS negative earnings dummy GLS negative earnings dummy resid

13
4.2. Generalized Method of Moments
Developed Economies Emerging Economies

14
4.2. Generalized Method of Moments (contd.)

• Running the GMM estimates ? the results are
  different amongst the developed and emerging
  economies
• All the banks considered are slow in adjusting
  their provisions over a certain number of years
  as suggest the slow decrease of the lagged
  dependent variable coefficient.
• The banks within the developed countries smooth
  their earnings while within the emerging
  countries this is not a common practice.
• The banks within the developed countries have an
  imprudent behaviour regarding provisioning while
  the others are showed to be prudent in their
  polices
• The loan loss provisions polices follow the
  economic cycle only within the banks from Western
  Europe.

15
4.2. Generalized Method of Moments negative
earnings dummy (contd.)
Emerging Economies ? Hausman test - robustness Emerging Economies ? Hausman test - robustness
GMM negative earnings dummy GMM negative earnings dummy resid

16
4.3. A Commercial Bank / monthly data

• Similar model for a commercial Romanian bank ?
  the bank behavior between 1st of June 2004 and
  31st of March 2007
• Test the stationarity ? Augmented Dickey Fuller
• The estimates results
• Prudent behavior of the bank management regarding
  provisioning
• The relation with the economic cycle Industrial
  Output Index ?the overall portfolio exposure with
  the industrial sector represents about 25 percent
  of the total loan portfolio exposure
• No income smoothing

17
4.3. A Commercial Bank / monthly data (contd.)
OLS estimates GMM estimates

18

19
Developed Economies Loans
20
Developed Economies Loan Loss Provisions
21
Emerging Economies Loans
22
Emerging Economies Loan Loss Provisions
23
Conclusions

• The banks within the developed countries
  provision less during high GDP growth, suggesting
  an undesirable anti-business cyclical behavior.
  On the contrary, the banks behavior from the
  emerging countries does not follow the economical
  cycle. The reason of this behavior is related
  with the economy development and the boom of the
  bank sector within all those countries
• The banks from developed countries smooth their
  income through the loan loss provisioning
  policies. This might result in lower earnings
  quality since net income does not
  representatively portray the economic performance
  of the business entity for the period. The banks
  from the emerging countries do not smooth their
  income

24
Conclusions (contd.)

• The amounts allocated to the loan loss provisions
  in the emerging countries follow the growth of
  the loan portfolio showing a prudent behavior of
  the banks managers accordingly with the new
  fiscal and prudential requirements
• Credit risk is a normal part of banking. However,
  where the amount of risk is excessive or where
  this is not properly monitored and controlled, it
  can produce a significant threat to the credit
  institution and its earnings.

25
References

• Basel Committee on Banking Supervision (2006),
  Sound credit risk assessment and valuation for
  loans, Bank for International Settlements
• Cavallo, M and Majnoni, G (2001), Do Banks
  Provision for Bad Loans in Good Times, Empirical
  Evidence and Policy Implications, World Bank
  Research Working Paper No. 2619
• Crouhy, M., Galai, D. and Mark R. (2006), The
  Essentials of Risk Management, The McGraw-Hill
  Companies, Inc, New York
• Cossin, D. and Pirotte H. (2001), Advanced
  Credit Risk Analyses Financial Approaches and
  Mathematical Models to Assets, Price and Manage
  Credit Risk, John Wiley Sons, Inc, New York
• Epstein, B. J. and Mirza A.A. (2002), IAS 2002
  Interpretation and Application of International
  Accounting Standards, John Wiley Sons, Inc,
  New York
• Fernández de Lis, S. F., Pagés, J. M., Saurina,
  J., (2000), Credit growth, problem loans and
  the credit risk provisioning in Spain, Banco de
  España Servicio de Estudios, Documento de
  Trabajo No. 0018
• Fisher, S., (2003), Implications of the Basel II
  for Emerging Market Countries, The William
  Taylor Memorial Lectures No. 7, Group of trinity,
  Washington, DC
• Hansen, B.E. and West, K.D. (2002), Generalized
  Method of Moments and Macroeconomics, Journal of
  Business Economic Statistics
• Laeven, L and Majnoni, G (2002), Loan Loss
  Provisioning and the Economic Slowdowns Too
  Much, Too Late?, Conference Series, Federal
  Reserve Bank of Boston
• Levine, A and Lin, C-F (1992), Unit Root Tests
  in Panel Data Asymptotic and Finite-sample
  Properties, University of California, San Diego,
  Department of Economics, Discussion Paper 92-23

26
References (contd.)

• Mazararu, E, (2005), The New Basel Accord, The
  Corporate Development Sector, Raiffeisen Bank
• Pérez, D., Salas-Fumás V., Saurina, J., (2006),
  Earnings and capital management in alternative
  loan loss provision regulatory regimes, Banco de
  España, Documento de Trabajo No. 0614
• Pynnonen, S. (2007), A Short Introduction to the
  Generalized Method of Moments Estimation,
  University of Vaasa, Department of Mathematics
  and Statistics, Finland
• Keller G. and Warrack B. (2001), Statistics for
  Management and Economics, Fifth Edition, Duxbury
  Thomson Learning
• Yaffee, R., (2003), A primer for Panel Data
  Analysis, Connect Information Technology at New
  York University, Information Technology Services
• Wooldridge, J. M., (2001), Econometric Analysis
  of Cross Section and Panel Data, The MIT Press,
  Cambridge, Massachusetts, London, New England
• Credit Policy (2007), Raiffeisen Bank
• Annual Reports (2004, 2005, 2006),
  Raiffeisen Bank
• EUROSTAT, General and regional statistics,
  Economy and finance indicators
• BANKSCOPE, Bureau Van Duk
• National Bank of Romania, Annual Reports and
  Monthly Bulletins
• National Bank of Romania, Regulation No. 5 (2002)