Description
Recent cases of bank failure across nations have been traced largely to rising "toxic assets" in commercial banks loan portfolio. This paper analyze relationship between efficiency of credit risk management and financial health in selected Nigerian banks.
Far East Journal of Marketing and Management Vol. 2 No. 1 April 2012
39
ANALYSIS OF CREDIT RISK MANAGEMENT EFFICIENCY IN
NIGERIA COMMERCIAL BANKING SECTOR,(2004-2009)
Onaolapo A. R.
Department of Management and Accounting,
Faculty of Management Science,
Ladoke Akintola University of Technology, Ogbomoso, Nigeria
E- mail: [email protected]
ABSTRACT
Recent cases of bank failure across nations have been traced largely to rising „toxic assets? in
commercial banks loan portfolio. This paper analyze relationship between efficiency of credit
risk management and financial health in selected Nigerian banks. Data collections are mainly
secondary spanning a six- year period before and after consolidation programmme of the
Nigerian banking sector. The study hypothesized negative relationship between Efficiency of
Credit Risk Management (ECRM); bank performance and operational effectiveness. Collected
data were regressed and unit root test was conducted to verify order of integration for each time
series data employed. Findings indicate minimal causation between Deposit Exposure (DE)
(Surrogate of credit risk management) and performance but greater dependency on operational
efficiency parameters. Test of stationary properties conducted using ADF indicated all variables
were non-stationary while the pair wise Granger causality suggested that Deposit Exposure
performance influence does not hold for the Nigerian Commercial banking sector. Policy
recommendations were made on these findings.
Keywords: Credit Risk Management, Deposit Exposure, Operational Efficiency, bank credit
portfolio, Earning Factors and Bank Asset Quality
Paper Type: Research Paper
INTRODUCTION
Credit Risk Management (CRM) policies of a commercial bank comprise those decision- making
structures associated with the reduction of exposures to credit asset classification and loan loss
provisioning. According to the Basle Committee on Banking Supervision (BCBS) (2003),
management of bank risk relates to the minimization of the potential that a bank borrower or
counter-party will fail to meet its obligations in accordance with agreed terms.
In a bid to maximize shareholders wealth and ensure safety of depositors fund, banks act as
delegated monitors on behalf of lenders (depositors) using various innovations, technologies and
procedures to enforce credit contracts. These measures not with standing, banking operations are
still exposed to some inherent risks including borrowers? outright default; unwillingness or
inability to meet credit commitment due to the vagaries of business activities or other
environmental dynamics (Bidani et. al., 2004). Credit management frameworks therefore
become imperative tools in decision- making that relates to loan-pricing, delegating lending
powers, mitigating or migrating as well as managing incidences of credit risk on bank portfolio.
Far East Research Centre www.fareastjournals.com
CRM policies are designed and applied both internally as an operational tool by bank
management and externally by bank regulatory authorities to manage the financial health of the
banking sector. The focuses of such policies are the needs for asset diversification; maintainace
of balance between returns and risk, bank asset quality and ensuring safety of depositors fund.
The failure of various regulatory frameworks designed by the supervisory authorities and
inability of technological innovations to stem rising „toxic assets? in many banks constitute
matters of grave concern for stakeholders in both developed and developing nations? financial
systems; Sinkey (1998), Saunders and Cornett (2003), BCBS (2003) and Casu et al (2004).
The needs to analyse the efficiency of CRM frameworks is particularly acute in a developing
nation like Nigeria. For instance in the period 2004-2008, the Nigerian banking system reported
significant growth in its many performance parameters including rising capital adequacy (65%),
Earnings (150%), Liquidity (120%) and Asset Quality (75%); Onwumere (2005); Fatemi and
Fooladi (2006); Onaolapo (2007), Koizol and Lawrence (2008) and Hassan (2009). On the basis
of these performances, credit rating institutions such as Fitch Ratings and Augusto and Co
returned verdicts of good financial health for the sector in the financial year ended 31
st
December
2008.
Not quite a year after, another publication by the African Report (2009) Classified 11 out of the
24 Nigerian banks as either „shaken? or „stressed?; an indication of weak financial health for
about forty – six percent of the entire sector. An audit conducted by the Central Bank of Nigeria
(CBN) in May 2009 also identified eight of these banks as seriously affected by liquidity and
capitalization problems traceable to corporate misgovernance and poor credit risk management
practices (Okey) (2010). The apex bank also identified the existence of predatory debtors whose
„modus operandi? involved abandonment of debt obligations in some banks only to contract new
ones in others; thus creating a spiral of rising non-performing credit portfolios or toxic assets.
Since the completion of the Nigerian banking consolidations programme in 2005, rising
competition within the entire financial system has accounted for intensive credit disbursement as
a means of survival among banks that scaled the „recapitalization bar?. This makes perfection of
CRM practices crucial tools for ensuring safety of depositors? fund and stabilizes the soundness
of the system?s financial health. Report of rising „toxic asset? in the sector also informed the need
to undertake an analysis of the basic determinants of credit risks within the system. This paper
therefore appraises the salient credit factors affecting Nigerian commercial banks credit portfolio
in the study period as well as analyzes the efficiency of credit risk management practices adopted
by the sampled banks within the sector. The subject of the analysis is to determine the variation
in performance brought about by changing CRM approaches among the banks and how such
practices affected the financial health of the sector as a whole.
There are four more sections in this paper including section two which examines the theoretical
and empirical framework as well as identifies basic determinants of credit risk management
policies and performance in the Nigerian Commercial Banking Sector. Section three briefly
describes the research design adopted in relating the explanatory variables concerned while
section four discusses and analyzes the nexus between the variables. The last section, that is
section five, provides summary, conclusion as well as recommendations for bank management
and their implications for regulatory policy decisions.
Far East Journal of Marketing and Management Vol. 2 No. 1 April 2012
41
CONCEPTUAL FRAMEWORK
Risk according to Athanasoglou et al (2005) is a situation occasioned by internal or external
environmental factors that create hindrances in the way of achieving certain objectives of an
entity. Modern banks operate in a highly dynamic environment that exposes them to varied risks
ranging from liquidity, credit, foreign exchange to market vagaries; thus creating a source of
threat towards survival and success. Al- Thamimi and Al Mazrooei (2007). Management of a
bank therefore incorporates a trade-off between risks and returns especially its credit portfolio
which constitutes a substantial income generating asset.
Traditionally, the five C?s representing borrower?s Characters, Capacity, Collateral and
Conditions have been recommended in some literatures; Fatemi and Fooladi (2006), and
Onyiriuba (2004). More than borrowers? attributes, the nature of bank credit delivery is often
influenced by such operational antecedents as the state of the economy; market conditions or
industrial dynamics that characterized borrowers? business and financial health. In the nineties
when Nigerian banks operate in less dynamic environment, the use of „avoidance strategy? with
preferences for self liquidating short term lending was common. Today, bank lending requires a
more pragmatic approach involving identification, analysis and mitigation of damaging effects of
credit risk exposures.
Apart from deposit – mobilization, lending is considered the most significant function of a
commercial bank as loan asset created constitutes the highest income generating portfolio;
Diamond (1984), Berger et al (2003) and Casu et al (2006). The role of lending to finance the
deficit economic sector constitutes essential banks? efforts at ensuring shareholders? wealth
maximization; but this is achieved at the expense of a risk – return trade – off. Almost every
operational activities of a bank give rise to one form of risk ranging from interest, market,
technology, concentration, reputational and other forms of risk with credit risk exposure
occupying the centre stage.
Figure - 2.1 Generic Linkage Between Various Bank Operational risks.
Capital, reputational
and Solvency Risks
Interest/ Foreign Exchange
and Market Risks
Credit Risk
Concentration Risk
Balance Sheet and Off
Balance sheet Risks
Liquidity Risk
Sovereign/ Country
Risks
Far East Research Centre www.fareastjournals.com
Source: Author?s Conceptualization (2012).
Figure 2.1 above describe the centrality of the linkage between credit risk and other risks
associated with banking operation.The ripple effects of a bank credit risk aggravates such
operational risk like liquidity, concentration; market and reputational risks among others.
Findings in some other studies have also identified systemic, international and national macro-
economic variables as well as un-systemic bank specific factors as determinants of credit risk,
Demirguc-kunt (1989), Ariff and Marrisetti (2001), Cebenoyan and Strahan (2004) and
Athanasoglou et al (2005).
More than systemic and macro-economic variables, bank specific factors such as the size of risky
loan making up the credit portfolio; bank internal loan policy; pre- lending assessment of
borrower and past audit analysis of financed project constitute significant factors that shape the
quality of a bank credit portfolio; Ahmed (2003) and Ahmed and Ariff (2007). Other sources of
credit risk highlighted in some studies include deficient loan appraisal process, inadequately
defined lending policies, high credit concentration, poor credit analysis skill of bank officials, as
well as a mismatch between credit monitoring system and external operating environment;
(Bidani et al 2004, and Chen et al 2005).
Credit Risk Management (CRM)
A deposit-banking firm like most financial institutions tend to hold little owner?s capital relative
to the aggregate value of its assts. The implication of this is that only a small percentage of total
loans need to turn bad to push the entire credit portfolio to the brink of failure.
According to Peter and Sylvia (2008) the probability that a deposit banking institution?s credit
portfolio will decline in value and perhaps become worthless is known as credit risk while
various attempts designed to control and protect banks against adversities associated with these
risk exposure are referred to as credit risk management processes.
The process of analyzing credit risk, ranking and quantifying them constitute a substantial aspect
of the framework and governance structures for most bank management .Among the reasons
advanced for CRM include managerial self- interest and appraisal goal; high cost of financial
distress and the existence of capital market imperfection. Other motivation for expending
managerial resources on CRM according to Meyer (2000) is the need for insolvency avoidance,
given the likelihood of poor credit risk management snowballing into financial crisis.
Onyiriuba (2009), provided some empirical evidence on how poor stock returns emanating from
under performing Nigerian bank credit portfolio fuelled negative volatilities in foreign exchange,
substantial reduction in the aggregate value of capital market and contagions in other sectors of
the Nigerian economy.
The process of sound CRM commences with identification of the existing and potential risk
inherent in a bank?s lending activities as well as designing appropriate policies to control them.
In the Nigerian banking system, individual bank management fashions out CRM system
comprising policies that;
Far East Journal of Marketing and Management Vol. 2 No. 1 April 2012
43
? Limits or reduce credit risk to certain industries, market or individuals (over- concentration)
? Ensures adequacy of asset classification (asset classification rule)
? Loan loss provisioning (prudential codes)
? Stipulates borrower?s key performance indexes (conditionality rule)
? Undertakes pre- lending assessment and post lending audit/monitoring.
In August 2009, the CBN issued guidelines for developing CRM framework for individual
banks? risk elements in line with its supervisory model. The motivation for the guideline among
others is the need to close the widespread “lacuna” in most codes and standards with respect to
CRM of individual banks in the sector. Other goals have been identified to include strengthening
the credit appraisal procedures, storage and dissemination of credit data, monitoring of over-
exposure to borrowers, facilitating consistent credit classification and affording regulators first
hand information on customer?s global debt profile; www.cenbank.org (2010).
By 2010 the CBN also floated an Asset Management Company (AMCON) specifically designed
to relief „troubled banks? of their „toxic? assets. Lessons from recent bank crisis have however
shown that removing assets from a bank balance sheet does not ultimately ensure that such bank
will be risk free in its subsequent operations. Further evidence of this was provided by AMCON
Managing director, Chike-Obi (2011) who submitted that the corporation lost about N226 billion
in three nationalized banks (about 39.00 percent of total share holders? fund) between 2009 and
2011.
On the basis of the above, this paper therefore attempts an examination of the causality between
credit management practices adopted by Nigerian banks and performance in the pre and post
consolidation era of 2004– 2009.
METHODOLOGY
The chosen units of analysis are the Nigerian commercial banks whose credit risk management
practices are related to performance in the study period. Data relating to the study variables
encompass performance parametric and surrogates of credit risk management practices. Six
banks constituting twenty–five percent of the entire commercial banking population are selected
for the study. The study employs mainly secondary (financial) data generated from Central Bank
of Nigeria (CBN) and selected banks annual accounts and reports. The proposed period was
2004 to 2009.The statistical method of analysis was in line with the Triangular Gap model
previously adopted by Cebenoyan and Strahan (2001); and Houston et al (1997). These have
been cross-sectional regression and correlation analysis with credit risk exposures as the
dependent variable. The entire variables for the study are based on book – values in line with the
argument by Meyers (2000) that book values are proxies for the values of assets in place.
Credit Risk Modeling
Credit risk model is concerned with exploring the basic relationship between the bank
performance (portfolio) and loss distribution in the study period. It is an attempt to provide
quantitative analysis of the extent to which the loss distribution (risk factors) varies with changes
in bank credit management efficiency. According to Pesaran et al (2004) and Satchidananda
(2006), this type of modeling can either be approached from the perspective of individual loans
making up the entire portfolio or by considering returns on the loan portfolio directly. The model
Far East Research Centre www.fareastjournals.com
for the study follows the latter to explain the relationship between Nigerian bank credit risk
management efficiency and credit risk factors. The model is further augumented using a Pairwise
Granger Causality (PGC) for testing Co- integration between the variables.
Panel data involving the pooling of observations on a cross-sectional form are adapted and the
models are of the form below:-
Y
i t
= ? + ?
i
X
i t
+ e
i t
-------------- Equation I
Where the subscript i represents Cross – Sectional dimension and t denotes the time series
dimension. Y
it
represent the dependent variable while X
it
stands for the explanatory variables
such as Net Interest Income (NII), Net Asset Quality (NAQ), Earning Per Share (EPS), Insider
Loan (IL), Operational Efficiency (OE) and Provision Coverage (PC) among others included in
the model. Given the ability of ? to be the same across units, Ordinary Least Squares (OLS)
provides a consistent and efficient estimate of ?
0
and ?
i-n
in the two general equations presented
below:
DEit = f (?o + ? i1 NII + ?i2 ps E + ?pc + ?i4 IL + OEi5 + e) -------- Equation 2
Where DEit = the squared differences between the sampled banks deposit exposure; that is
excess of liquid asset over deposits and short term liabilities. Explanatory variables are as
described while e represent the stochastic (error) term for the equation. In Equation 3,
Operational Efficiency (OE) for sampled banks are adopted as performance metric and related to
selected risk and earning factors.
OEit = F?o + ?1 + ?iPC + ?2PAT + ?3EPS + ?4NII + ?5NAQ + ?6IL + ?7DE +
e…Equation 3.
OE as a representative term for bank management efficiency is expressed as a function of
selected risk and performance metrics namely adequacy of Provision coverage (PC), Profit after
tax (PAT), Earning per Share (EPS), Net Interest Income (NII), Net Asset Quality (NAQ),
Insider Loan (IL) and Deposit Exposures (DE) for the sampled banks in the study period, with
(e) representing the stochastic error term.
ANALYSIS AND DISCUSSION
Relationship between CRM and Bank Performance
The first tested model undertakes an investigation into the relationship between Deposit
Exposure (DE) as surrogate for bank credit management efficiency in relation to performance
using such performance parameters as Earnings per Share (EPS), Insider Loan (IL), Net Asset
Quality (NAQ), Net Interest Income (NII), Operational Efficiency (OE), Profit after Tax (PAT)
and Provision Coverage (PC). Findings are presented in table 4.1 below:
Table - 4.1 Regression Result for model 1.
Dependent
Variable
Explanatory
Variables
Obtained
estimates
Standard Error Estimated t-value
Constant -50220.09 48390.63 -1.04
EPS -9228.52 4526.49 -2.04
Far East Journal of Marketing and Management Vol. 2 No. 1 April 2012
45
DE
IL +0.11 0.79 0.15
NAQ +767.13 646.68 1.19
NII -0.24 0.09 -2.65
OE +0.23 0.11 2.02
PAT -0.05 0.05 -0.96
PC +1.31 0.79 1.65
R
2
= 0.3285 t
0.025
= 2.o74
Adj – R
2
= 0.1185
F-stat = 1.5738 F
0.05
(7, 22) = 2.46
D-W = 0.710
Source: Computations and Out-Put of e-view based on Author?s Field study (2012).
According to the result presented in Table 4.1 above, a unit increase in the Earnings per Share of
the sampled banks over the study period resulted in a 9228.52 per cent decline in the banks?
deposit exposure. Also, such unit changes in Insider Loan, Net Liquid Assets, Net Interest
Income, Operational Efficiency, Profit after Tax, and Provision Coverage led to 0.12 per cent
increase, 767.13 per cent increase, 0.24 per cent decrease, 0.23 per cent increase, 0.05 per cent
decrease and 1.31 per cent increase in the explanatory variables respectively.
The statistical properties however reveal a weakness in the model. For instance, the R-Squared
value of 0.3285 suggests that only about 33 per cent of the total changes in the dependent
variable are attributable to changes in the explanatory variables. The R
2
is further reduced to a
mere 12 per cent after adjustment which is even lower.
Furthermore, given that t-value (at 0.025 for a two-tailed test or 5% for a one-tailed test) the
tabulated 0.025 is not the same as the estimated t- value at the 0.05 or 5% significance level, only
the Net Interest (NII) parameter is significant as other parameters turned up with estimated t-
values lower than the tabulated t-value. The same applies to the F-statistic, whose estimated
value of 1.5738 is lower than the 5% tabulated value of 2.46, implying that the overall regression
plane is statistically insignificant. There is the presence of First-order serial correlation
(autocorrelation) also as the estimated Durbin Watson (DW) statistic of 0.710 is lower than the
Lower Limit value of the 5% tabulated Durbin Watson value of 1.07 (for at least five
explanatory variables).
Due to the observed weakness of the first model of the study in equation 2, the study concludes
that there is no significant relationship between Credit Risk Management (CRM) and bank
performance over the period under review. The study further undertakes to determine the
relationship, if any, between financial heath and CRM.
Relationship between CRM and Financial Health
The need to examine the relationship between CRM and Financial health gave rise to the
formulation of the second model of the study in Equation3. In this case, Operational Efficiency
(OE) is used as a proxy for the banks? financial health while the other seven explanatory
variables were used to capture CRM. The result obtained is presented in the following table.
Table - 4.2 Regression Result for model II.
Far East Research Centre www.fareastjournals.com
Dependent
Variable
Explanatory
Variables
Obtained
estimates
Standard Error Estimated t-value
OE
Constant +75079.58 85425.37 0.88
DE +0.69 0.3 2.02
EPS +18057.57 7752.92 2.33
IL -2.33 1.30 -1.79
NAQ -925.59 1152.99 -0.80
NII +0.64 0.13 5.04
PAT +0.05 0.10 0.53
PC 5.14 0.99 -5.18
R
2
= 0.7211 D-W = 0.909
Adj.R
2
= 0.6429 t
0.025
= 2.074
F-stat = 8.459 F
0.05
(7,22) = 2.46
Source: Out-put of e=view based on Authors Field study (2012)
According to the result in Table 4.2 above, within the period under review, a unit change in the
Deposit exposure of the banks resulted in about 0.69 per cent rise in Operational efficiency. Such
a positive relationship is seen also for the Earnings per share, Net Interest Income, and Profit
after Tax variables where unit changes raised Operational efficiency by 18057.57 percent, 0.64
percent and 0.05 percent respectively. On the other hand, there is a negative relationship between
Insider Loan and Operational Efficiency, between Net Asset Quality and Operational Efficiency
and between Provision Coverage and Operational Efficiency. In the first case, a unit change in
the volume of Insider loan decreased Operational Efficiency by about 2.33 percent, while the
same is 925.59 percent for Net Asset Quality and 5.14 percent for Provision Coverage.
The estimates obtained for the parameters of the second model are more reliable as three of the
seven explanatory variables are significant: Earnings per Share, Net Interest Income and
Provision Coverage. The R-Squared value of 0.7291 (about 73 per cent) is satisfactory as only
about 27 per cent of the total variations in the banks? operational efficiency over the review
period is due to variations in other factors not captured by the study models; even after
adjustment the R-Squared is still as high as 64 percent. The result is further attested by the F-
Statistic value of 8.459 (higher than tabulated value of 2.074), the entire regression plane is
significant.
The case of whether there is the presence of positive first order serial correlation is indeterminate
by the Durbin-Watson (DW) Criterion at the 1% significance level and maximum of five
explanatory variables. This is due to the fact that the observed DW statistic of 0.909 lies between
the tabulated Lower and Upper limit values of 0.88 and 1.61 (for maximum of five explanatory
variables)although autocorrelation is established at the 5% level of significance in which case the
observed DW statistic is clearly lower than the tabulated value of 1.07. The above
notwithstanding, due to the overall significance of the estimated regression model (II), the study
safely concludes that there is a significant positive relationship between CRM and Financial
health for the sampled banks
Unit Root Test
Far East Journal of Marketing and Management Vol. 2 No. 1 April 2012
47
It is almost a convention in time series analysis, to verify the order of integration for each series
so as to avoid perennial problem of spurious regression (see Granger and New bold 1974; Philip,
and Perron 1988). A test of stationarity property for each variable in the study is conducted using
Augmented Dickey Fuller (Dickey and Fuller, 1981) procedure.
The result of the Unit-root test in table 4.3 suggest that at both level and first difference, the unit
root hypothesis cannot be rejected at 1%;5% and 10% critical level for all the variables. This in
effect suggests that employed data series are non.-stationary and this is suitable for the purpose
A test of stationary employed is the Unit Root Test using the Augmented Dickey Fuller (ADF)
Test. All the time series data on the variables of the study are included. Part of the result obtained
is presented in table. 4.3 below.
Table - 4.3 Result of Unit Root Test
Variable ADF test
statistics
1%
Critical
Value
5%
Critical
Value
10%
Critical
Value
Order of
Integration
DE -4.3767 -4.3382 -3.5867 -3.2279 1(1)
EPS -3.9813 -3.6852 -2.9705 -2.6242 1(0)
OE -4.5526 -4.3382 -3.5867 -3.2279 1(1)
IL -4.7970 -4.3382 -3.5867 -3.2279 1(1)
NAQ --4.610 -4.3226 -3.5796 -3.2239 1(0)
NII -4.4481 -4.3382 -3.5867 -3.2279 1(1)
PAT -5.7510 -4.3382 -3.5867 -3.2279 1(1)
PC -5.3454 -4.3382 -3.5867 -3.2279 1(1)
Source: Author?s Compilation based on Output result from the E-view (2012).
As observed in table 4.3, in relation to the explanatory variables, the explained variable in
equation I, Deposit Exposure (DE), is stationary at all levels of critical values, being integrated
of order 1. Similarly the data for the dependent variable of equation II, Operational efficiency
(OE), is stationary at all observed levels of critical values, being integrated also of order 1.
However the focal point of reference is the 5% Critical Level; hence, the variable which are
stationary only after differencing are non-stationary at the 5% Critical Value. Stationarity means
that the estimated values of the regression model do not indefinitely diverge from the true
regression line.
Causality Test between Financial Health and CRM
The Pair wise Granger Causality Test was further used to augment the estimated models of the
study. A simple standard causality test that is the pair wise Granger causality test employed
examines bi-directional relationship between two variables selected at a time in the study. Our
empirical results presented in table 4.4 (see appendix 1) indicate that decomposed Bank Deposit
Exposures (DE) against indicators of performance found that the former did not Granger cause
the later. Adjustment was made for 2 lags; hence the result includes 28 observations.
Table - 4.4: Result of Granger Causality Test
Null Hypotheses F-Statistics Probability Status
Far East Research Centre www.fareastjournals.com
OE does not Granger Cause DE
DE does not Granger Cause OE
2.39966
0.06409
0.11310
0.93808
Accepted
Accepted
EPS does not Granger DE
DE does not Granger Cause EPS
0.76100
0.38469
0.47861
0.68496
Accepted
Accepted
IL does not Granger Cause DE
DE does not Granger Cause IL
0.30309
1.49583
0.74144
0.24506
Accepted
Accepted
NAQ does not Granger Cause DE
DE does not Granger Cause NAQ
0.60167
0.55495
0.55630
0.58160
Accepted
Accepted
NII does not Granger Cause DE
DE does not Granger Cause NII
1.34902
0.18877
0.27927
0.82925
Accepted
Accepted
PAT does not Granger Cause DE
DE does not Granger Cause PAT
0.02820
0.16146
0.97222
0.85186
Accepted
Accepted
PC does not Granger Cause DE
DE does not Granger Cause PC
2.18232
0.14283
0.13558
0.86766
Accepted
Accepted
EPS does not Granger Cause OE
OE does not Granger Cause EPS
0.07146
2.91814
0.93124
0.07423
Accepted
Rejected
IL does not Granger Cause OE
OE does not Granger Cause IL
0.37147
0.14463
0.69378
0.86613
Accepted
Accepted
NAQ does not Granger Cause OE
OE does not Granger Casus NAQ
0.28790
0.75002
0.75250
0.48356
Accepted
Accepted
NII does not Granger Cause OE
OE does not Granger Casus NII
1.57293
3.55585
0.22895
0.04512
Accepted
Rejected
PAT does not Granger Cause OE
OE does not Granger Cause PAT
0.52752
0.95377
0.59703
0.40000
Accepted
Accepted
PC does not Granger Cause OE
OE does not Granger Cause PC
0.69415
7.67791
0.50966
0.00279
Accepted
Rejected
IL doe not Granger Cause EPS
EPS does Granger Cause IL
0.64533
0.69932
0.53372
0.50718
Accepted
Accepted
NAQ does Granger Casus EPS
EPS does Granger Cause NAQ
1.05816
3.35684
0.36339
0.05259
Accepted
Rejected
NII does not Granger Cause EPS
EPS does not Granger Cause NII
3.16423
5.80692
0.06110
0.00909
Rejected
Rejected
PAT does not Granger Cause EPS
EPS does not Granger Cause PAT
0.07416
0.92242
0.92875
0.91206
Accepted
Accepted
PC does not Granger Cause EPS
EPS does not Granger Cause PC
0.42311
2.20529
0.66000
0.13299
Accepted
Accepted
NAQ does not Granger Cause IL
IL does not Granger Cause NAQ
0.33492
0.26524
0.71882
0.76934
Accepted
Accepted
NII does not Granger Cause IL
IL does not Granger Cause NII
0.29495
0.16603
0.74734
0.84802
Accepted
Accepted
PAT does not Granger Cause IL
IL does not Granger Cause PAT
0.04054
0.36042
0.96034
0.70125
Accepted
Accepted
PC does not Granger Cause IL
IL does not Granger Cause PC
0.29319
0.00020
0.74863
0.99980
Accepted
Accepted
NII does not Granger Cause NAQ
NAQ does not Granger Cause NII
1.94855
0.38967
0.16529
0.68167
Accepted
Accepted
Far East Journal of Marketing and Management Vol. 2 No. 1 April 2012
49
Source: Result Generated from e-view output (2012).
Table 4.4 testifies that the Granger Causality Test reveals a satisfactory situation. The Observed
or empirical F – statistic was checked against the 5% tabulated values of 2.46 to ascertain cases
of two-way causality. Of the twenty-eight-pair result obtained, only in a single case do we have a
two-way causation in which case both null hypotheses are rejected. The reported case is between
Net Interest Income (NII) and Earnings per share (EPS). The null hypotheses were rejected,
meaning that either one Granger-cause, or give rise to the other. In other paired results the null
hypotheses were accepted for each pair or for an item in the pair.
CONCLUSION AND POLICY IMPLICATION OF FINDINGS
The application of annual data for the period 2004-2009 has examined the validity of credit risk
management efficiency in selected Nigerian commercial bank. For this purpose empirical
investigation of the extent of relationship between dependent and explanatory variables was
undertaken using regression analysis.
Results of the regression analysis found minimal causation between Deposit exposure (DE) as
surrogate of credit management efficiency and performance indicators but greater dependency
Operational Efficiency (OE) parameters.
Empirical investigation of the stationary properties and the order of integration of the employed
variables are conducted using ADF. The results show that all the variables were non-stationary at
both level and first differences.
Evidence from the pair wise Granger causality tests suggest that Deposit Exposure –
performance influence does not hold for he Nigerian commercial banks. These results could be
explained due to the influence of several factors such as the nature of technology; cost structure,
managerial effectiveness considered as significant Key Performance Indicators (KPI) which have
not been fully factored into decision- making process in the Nigerian Banks, Ogubunka (2011),
Woherem (2001) and Onaolapo (2007).
These empirical findings have significant implications for bank strategic planners and regulatory
authorities. Operational activities of a typical Nigerian commercial bank are not sufficiently
optimized to ensure maximum earnings from credit creation cum loans and advances. This
explains why loan interest earnings are sufficiently eliminated by rising cost of credit given the
high incidence of non-performing credit. A more pragmatic credit management and screening
PAT does not Granger Cause NAQ
NAQ does not Granger Cause PAT
0.52322
0.87512
0.59949
0.43024
Accepted
Accepted
PC does not Granger Cause NAQ
NAQ does not Granger Cause PC
1.63380
0.06262
0.21704
0.93946
Accepted
Accepted
PAT does not Granger Cause NII
NII does not Granger Cause PAT
0.57077
1.63148
0.57289
0.21748
Accepted
Accepted
PC does not Granger Cause NII
NII does not Granger Cause PC
0.22700
1.95306
0.79869
0.16466
Accepted
Accepted
PC does not Granger Cause PAT
PAT does not Granger Cause PC
0.62326
0.59690
0.54500
0.55882
Accepted
Accepted
Far East Research Centre www.fareastjournals.com
procedure are necessary attempt needed to enhance earnings from credit as the single largest
activity of a commercial bank.
Furthermore undue emphasis of regulatory authorities on bail-out; recapitalization and merger of
banks often designed as exit-attempts to correct incidences of financial ill- health in Nigeria
might not after all be necessary where policy structure are designed to reduce high mortality of
business and worrisome infrastructural decay, currently the bane of Nigerian manufacturing
sector.
Perhaps more often than policies and infrastructural development is the need for personnel
training and retraining in Nigeria banks. These will enhance articulate research and development
in new product and market, as well as cost minimization efforts.
REFERENCES
African Report Group Jeune African (2009); The African Reports On Nigerian Bank.
Ahmed N.H (2003); Credit Risk Determinants: BY Institutional Type Proceedings of Malaysian
Finance Association Conferences (2003).
Ahmed N.H and Ariff M. (2001); Multi-country study of Bank Credit Determinants
“International Journal of Banking and Finance Vol.5 Issue I Article 6 (2007).
Athanasoglou P; Brissimis S. and Lelis M. (2005); “Bank – Specific, industry specific and
Macroeconomic Determinants of Bank Profitability”. Bank of Greece Working paper No. 25.
Al-Thamimi H. and Al-Mazrooei F. (2007); Bank Risk Management. Comparison Study Of
UAE National And Foreign Banks “The Journal Of Risk Finance Vol.8 Issue 4 pp.394-409.
Ariff M. and Marrisetti V.B (2001); A New Approach To Modeling Multi-country Risk
Premium Using Panel Data Test Method “Proceeding of MFS Conference In Cuprus (2003).
Basle Committee on Banking Supervision (2003) Risk Management Principles for Electronic
Banking, July. http//www.bis.org/Publ/bcbs98.pdf.
Berger, A. N. Daiq : Ongena S. and Smith D.C (2003); “To what Extent is the Banking Industry
Globalized?. A study of Banks, National Banking and Research in 20 European nations. Journal
of Banking and Finance Vol. 27 (3) March 2003 pp: 383-415.
Bidani S.N Mitra P.K and Praniod Kumar (2004); Taxmann?s Credit Risk Management.
Taxmann Allied Services Ltd New Delhi, India.
Casu B; Claudia G. and Philip M. (2006) Introduction to Banking. Pearson Education Ltd
Prentice Hall 100-273-69302- pp 258-269.
www.cenbank.org.ng (2010).
Cebenoyan A.S and Strahan P.E (2004); Risk Management, Capital Structure and Lending at
banks “Journals of Banking And Finance Vol. 28 pp: 19-43.
Far East Journal of Marketing and Management Vol. 2 No. 1 April 2012
51
Chen B, Cheng X and Wu L. (2005); Dynamic Interactions Between Interest Rate Credit and
Liquidity Risk “Theory and Evidence From Term Structure Of Credit Default and Swap spreads.
Working Paper Series Aug. 8 2005.
Central Bank of Nigeria (Various Years); Annual Reports and Statement of Accounts CBN.
Dickey D. and Fuller W. A (1998); Likelihood Ratio Statistics for Auto Regressive Time Series
with Unit Root. Econometrica: 0574972.
Diamond D.W (1984); Financial Intermediation and Delegated Monitoring. Review of
Economics Studies 51(3)393-414.
Demirguk – Kunt (1989); Deposit Institution Failures: “A Review Of the Empirical Literature”
Federal Reserves Of Cleveland Economic Review Qtr 4, 1998.
Fatemi A. and Fooladi I. (2006); Credit Risk Management. A Survey Of Practices “Managerial
Finance Vol. 32, Issue 3, pp:227-233.
Granger C.W.J. and Newbold P. (1974); Spurious Regression In Econometrics Pp:111-120.
Hassan M.K (2009); Risk Management Practices of Islamic Banks of Brunci Darusallam “The
Journal Of Risk Finance, Vol. 10, pp:23-37.
Koizol C. and Lawrence J. (2008); What Makes a Bank Risk? Insights from the Optimal Capital
Structure Of Banks “Journal of Banking and Finance. Vol. 33 pp:861-873.
Meyer L. (2000); Why Risk Management is Important for Global Financial Institutions. Risk
Management of financial Institutions; UN Conference Center Bangkok, Thailand.
Onyiriuba L. (2004) Analysing and Managing Risk of Banks Lending. Malthouse press Ltd,
Lagos Nigeria.
Okey N. (2010); Sanusi: Still fighting one year after. Nigeria Business World Intelligence, June
6-13 (2010).
Ogubunka U.M (2011); An Analytical Review of Cost Trend in The Banking Industry as a
reflection of the Cost Pressure in the Nigerian Economy. The Nigerian Banker, March 2011,
ISSN 0197-6679.
Onaolapo A (2007) “An Evaluation of the Effects of Recapitalization On the Financial Heath of
Nigerian Commercial Banks. A Ph.D thesis submitted to the school of Postgraduate Studies,
LAUTECH, Nigeria.
Onyiriuba L. (2004) Analysing and Managing Risks of Bank Lending. Malthouse Press Ltd,
Lagos, Nigeria.
Onwumere J.U.J (2005) Banking Industry Reforms and Developing the Nigerian Economy. The
Nigerian Bankers Journal of the CIBN, Oct. – Dec. (2005) ISSN 0197-6679.
Peter S. R. and Sylvia C. H. (2008) Bank Management and Financial Services, 7
t h
Edition,
McGraw Hill, 978-007-125938-4.
Far East Research Centre www.fareastjournals.com
Phillp P. C and Peron P. (1988); Testing for Unit Root in Time Series Regression biometrica 75
335-346.
Satchidananda S.S (2006); Multi-Step Method for Analysis of Credit Risk of Bank Loan, Center
for Banking IT Working Paper WP. (2006).
Saunders A. and Cornett M.; (2003); Financial Institutions Management. A Risk Management
Approach, 4
t h
Edition New York McGraw Hall Irwin.
Sinkey J. (1998); Commercial Bank Financial Management, 5
t h
Edition, London Prentice Hall
International.
Woheren E.E (2001); Information Technology in the Nigerian Banking Industry. Spectrum
Books Ltd; Ibadan Nigeria.
doc_317263103.pdf
Recent cases of bank failure across nations have been traced largely to rising "toxic assets" in commercial banks loan portfolio. This paper analyze relationship between efficiency of credit risk management and financial health in selected Nigerian banks.
Far East Journal of Marketing and Management Vol. 2 No. 1 April 2012
39
ANALYSIS OF CREDIT RISK MANAGEMENT EFFICIENCY IN
NIGERIA COMMERCIAL BANKING SECTOR,(2004-2009)
Onaolapo A. R.
Department of Management and Accounting,
Faculty of Management Science,
Ladoke Akintola University of Technology, Ogbomoso, Nigeria
E- mail: [email protected]
ABSTRACT
Recent cases of bank failure across nations have been traced largely to rising „toxic assets? in
commercial banks loan portfolio. This paper analyze relationship between efficiency of credit
risk management and financial health in selected Nigerian banks. Data collections are mainly
secondary spanning a six- year period before and after consolidation programmme of the
Nigerian banking sector. The study hypothesized negative relationship between Efficiency of
Credit Risk Management (ECRM); bank performance and operational effectiveness. Collected
data were regressed and unit root test was conducted to verify order of integration for each time
series data employed. Findings indicate minimal causation between Deposit Exposure (DE)
(Surrogate of credit risk management) and performance but greater dependency on operational
efficiency parameters. Test of stationary properties conducted using ADF indicated all variables
were non-stationary while the pair wise Granger causality suggested that Deposit Exposure
performance influence does not hold for the Nigerian Commercial banking sector. Policy
recommendations were made on these findings.
Keywords: Credit Risk Management, Deposit Exposure, Operational Efficiency, bank credit
portfolio, Earning Factors and Bank Asset Quality
Paper Type: Research Paper
INTRODUCTION
Credit Risk Management (CRM) policies of a commercial bank comprise those decision- making
structures associated with the reduction of exposures to credit asset classification and loan loss
provisioning. According to the Basle Committee on Banking Supervision (BCBS) (2003),
management of bank risk relates to the minimization of the potential that a bank borrower or
counter-party will fail to meet its obligations in accordance with agreed terms.
In a bid to maximize shareholders wealth and ensure safety of depositors fund, banks act as
delegated monitors on behalf of lenders (depositors) using various innovations, technologies and
procedures to enforce credit contracts. These measures not with standing, banking operations are
still exposed to some inherent risks including borrowers? outright default; unwillingness or
inability to meet credit commitment due to the vagaries of business activities or other
environmental dynamics (Bidani et. al., 2004). Credit management frameworks therefore
become imperative tools in decision- making that relates to loan-pricing, delegating lending
powers, mitigating or migrating as well as managing incidences of credit risk on bank portfolio.
Far East Research Centre www.fareastjournals.com
CRM policies are designed and applied both internally as an operational tool by bank
management and externally by bank regulatory authorities to manage the financial health of the
banking sector. The focuses of such policies are the needs for asset diversification; maintainace
of balance between returns and risk, bank asset quality and ensuring safety of depositors fund.
The failure of various regulatory frameworks designed by the supervisory authorities and
inability of technological innovations to stem rising „toxic assets? in many banks constitute
matters of grave concern for stakeholders in both developed and developing nations? financial
systems; Sinkey (1998), Saunders and Cornett (2003), BCBS (2003) and Casu et al (2004).
The needs to analyse the efficiency of CRM frameworks is particularly acute in a developing
nation like Nigeria. For instance in the period 2004-2008, the Nigerian banking system reported
significant growth in its many performance parameters including rising capital adequacy (65%),
Earnings (150%), Liquidity (120%) and Asset Quality (75%); Onwumere (2005); Fatemi and
Fooladi (2006); Onaolapo (2007), Koizol and Lawrence (2008) and Hassan (2009). On the basis
of these performances, credit rating institutions such as Fitch Ratings and Augusto and Co
returned verdicts of good financial health for the sector in the financial year ended 31
st
December
2008.
Not quite a year after, another publication by the African Report (2009) Classified 11 out of the
24 Nigerian banks as either „shaken? or „stressed?; an indication of weak financial health for
about forty – six percent of the entire sector. An audit conducted by the Central Bank of Nigeria
(CBN) in May 2009 also identified eight of these banks as seriously affected by liquidity and
capitalization problems traceable to corporate misgovernance and poor credit risk management
practices (Okey) (2010). The apex bank also identified the existence of predatory debtors whose
„modus operandi? involved abandonment of debt obligations in some banks only to contract new
ones in others; thus creating a spiral of rising non-performing credit portfolios or toxic assets.
Since the completion of the Nigerian banking consolidations programme in 2005, rising
competition within the entire financial system has accounted for intensive credit disbursement as
a means of survival among banks that scaled the „recapitalization bar?. This makes perfection of
CRM practices crucial tools for ensuring safety of depositors? fund and stabilizes the soundness
of the system?s financial health. Report of rising „toxic asset? in the sector also informed the need
to undertake an analysis of the basic determinants of credit risks within the system. This paper
therefore appraises the salient credit factors affecting Nigerian commercial banks credit portfolio
in the study period as well as analyzes the efficiency of credit risk management practices adopted
by the sampled banks within the sector. The subject of the analysis is to determine the variation
in performance brought about by changing CRM approaches among the banks and how such
practices affected the financial health of the sector as a whole.
There are four more sections in this paper including section two which examines the theoretical
and empirical framework as well as identifies basic determinants of credit risk management
policies and performance in the Nigerian Commercial Banking Sector. Section three briefly
describes the research design adopted in relating the explanatory variables concerned while
section four discusses and analyzes the nexus between the variables. The last section, that is
section five, provides summary, conclusion as well as recommendations for bank management
and their implications for regulatory policy decisions.
Far East Journal of Marketing and Management Vol. 2 No. 1 April 2012
41
CONCEPTUAL FRAMEWORK
Risk according to Athanasoglou et al (2005) is a situation occasioned by internal or external
environmental factors that create hindrances in the way of achieving certain objectives of an
entity. Modern banks operate in a highly dynamic environment that exposes them to varied risks
ranging from liquidity, credit, foreign exchange to market vagaries; thus creating a source of
threat towards survival and success. Al- Thamimi and Al Mazrooei (2007). Management of a
bank therefore incorporates a trade-off between risks and returns especially its credit portfolio
which constitutes a substantial income generating asset.
Traditionally, the five C?s representing borrower?s Characters, Capacity, Collateral and
Conditions have been recommended in some literatures; Fatemi and Fooladi (2006), and
Onyiriuba (2004). More than borrowers? attributes, the nature of bank credit delivery is often
influenced by such operational antecedents as the state of the economy; market conditions or
industrial dynamics that characterized borrowers? business and financial health. In the nineties
when Nigerian banks operate in less dynamic environment, the use of „avoidance strategy? with
preferences for self liquidating short term lending was common. Today, bank lending requires a
more pragmatic approach involving identification, analysis and mitigation of damaging effects of
credit risk exposures.
Apart from deposit – mobilization, lending is considered the most significant function of a
commercial bank as loan asset created constitutes the highest income generating portfolio;
Diamond (1984), Berger et al (2003) and Casu et al (2006). The role of lending to finance the
deficit economic sector constitutes essential banks? efforts at ensuring shareholders? wealth
maximization; but this is achieved at the expense of a risk – return trade – off. Almost every
operational activities of a bank give rise to one form of risk ranging from interest, market,
technology, concentration, reputational and other forms of risk with credit risk exposure
occupying the centre stage.
Figure - 2.1 Generic Linkage Between Various Bank Operational risks.
Capital, reputational
and Solvency Risks
Interest/ Foreign Exchange
and Market Risks
Credit Risk
Concentration Risk
Balance Sheet and Off
Balance sheet Risks
Liquidity Risk
Sovereign/ Country
Risks
Far East Research Centre www.fareastjournals.com
Source: Author?s Conceptualization (2012).
Figure 2.1 above describe the centrality of the linkage between credit risk and other risks
associated with banking operation.The ripple effects of a bank credit risk aggravates such
operational risk like liquidity, concentration; market and reputational risks among others.
Findings in some other studies have also identified systemic, international and national macro-
economic variables as well as un-systemic bank specific factors as determinants of credit risk,
Demirguc-kunt (1989), Ariff and Marrisetti (2001), Cebenoyan and Strahan (2004) and
Athanasoglou et al (2005).
More than systemic and macro-economic variables, bank specific factors such as the size of risky
loan making up the credit portfolio; bank internal loan policy; pre- lending assessment of
borrower and past audit analysis of financed project constitute significant factors that shape the
quality of a bank credit portfolio; Ahmed (2003) and Ahmed and Ariff (2007). Other sources of
credit risk highlighted in some studies include deficient loan appraisal process, inadequately
defined lending policies, high credit concentration, poor credit analysis skill of bank officials, as
well as a mismatch between credit monitoring system and external operating environment;
(Bidani et al 2004, and Chen et al 2005).
Credit Risk Management (CRM)
A deposit-banking firm like most financial institutions tend to hold little owner?s capital relative
to the aggregate value of its assts. The implication of this is that only a small percentage of total
loans need to turn bad to push the entire credit portfolio to the brink of failure.
According to Peter and Sylvia (2008) the probability that a deposit banking institution?s credit
portfolio will decline in value and perhaps become worthless is known as credit risk while
various attempts designed to control and protect banks against adversities associated with these
risk exposure are referred to as credit risk management processes.
The process of analyzing credit risk, ranking and quantifying them constitute a substantial aspect
of the framework and governance structures for most bank management .Among the reasons
advanced for CRM include managerial self- interest and appraisal goal; high cost of financial
distress and the existence of capital market imperfection. Other motivation for expending
managerial resources on CRM according to Meyer (2000) is the need for insolvency avoidance,
given the likelihood of poor credit risk management snowballing into financial crisis.
Onyiriuba (2009), provided some empirical evidence on how poor stock returns emanating from
under performing Nigerian bank credit portfolio fuelled negative volatilities in foreign exchange,
substantial reduction in the aggregate value of capital market and contagions in other sectors of
the Nigerian economy.
The process of sound CRM commences with identification of the existing and potential risk
inherent in a bank?s lending activities as well as designing appropriate policies to control them.
In the Nigerian banking system, individual bank management fashions out CRM system
comprising policies that;
Far East Journal of Marketing and Management Vol. 2 No. 1 April 2012
43
? Limits or reduce credit risk to certain industries, market or individuals (over- concentration)
? Ensures adequacy of asset classification (asset classification rule)
? Loan loss provisioning (prudential codes)
? Stipulates borrower?s key performance indexes (conditionality rule)
? Undertakes pre- lending assessment and post lending audit/monitoring.
In August 2009, the CBN issued guidelines for developing CRM framework for individual
banks? risk elements in line with its supervisory model. The motivation for the guideline among
others is the need to close the widespread “lacuna” in most codes and standards with respect to
CRM of individual banks in the sector. Other goals have been identified to include strengthening
the credit appraisal procedures, storage and dissemination of credit data, monitoring of over-
exposure to borrowers, facilitating consistent credit classification and affording regulators first
hand information on customer?s global debt profile; www.cenbank.org (2010).
By 2010 the CBN also floated an Asset Management Company (AMCON) specifically designed
to relief „troubled banks? of their „toxic? assets. Lessons from recent bank crisis have however
shown that removing assets from a bank balance sheet does not ultimately ensure that such bank
will be risk free in its subsequent operations. Further evidence of this was provided by AMCON
Managing director, Chike-Obi (2011) who submitted that the corporation lost about N226 billion
in three nationalized banks (about 39.00 percent of total share holders? fund) between 2009 and
2011.
On the basis of the above, this paper therefore attempts an examination of the causality between
credit management practices adopted by Nigerian banks and performance in the pre and post
consolidation era of 2004– 2009.
METHODOLOGY
The chosen units of analysis are the Nigerian commercial banks whose credit risk management
practices are related to performance in the study period. Data relating to the study variables
encompass performance parametric and surrogates of credit risk management practices. Six
banks constituting twenty–five percent of the entire commercial banking population are selected
for the study. The study employs mainly secondary (financial) data generated from Central Bank
of Nigeria (CBN) and selected banks annual accounts and reports. The proposed period was
2004 to 2009.The statistical method of analysis was in line with the Triangular Gap model
previously adopted by Cebenoyan and Strahan (2001); and Houston et al (1997). These have
been cross-sectional regression and correlation analysis with credit risk exposures as the
dependent variable. The entire variables for the study are based on book – values in line with the
argument by Meyers (2000) that book values are proxies for the values of assets in place.
Credit Risk Modeling
Credit risk model is concerned with exploring the basic relationship between the bank
performance (portfolio) and loss distribution in the study period. It is an attempt to provide
quantitative analysis of the extent to which the loss distribution (risk factors) varies with changes
in bank credit management efficiency. According to Pesaran et al (2004) and Satchidananda
(2006), this type of modeling can either be approached from the perspective of individual loans
making up the entire portfolio or by considering returns on the loan portfolio directly. The model
Far East Research Centre www.fareastjournals.com
for the study follows the latter to explain the relationship between Nigerian bank credit risk
management efficiency and credit risk factors. The model is further augumented using a Pairwise
Granger Causality (PGC) for testing Co- integration between the variables.
Panel data involving the pooling of observations on a cross-sectional form are adapted and the
models are of the form below:-
Y
i t
= ? + ?
i
X
i t
+ e
i t
-------------- Equation I
Where the subscript i represents Cross – Sectional dimension and t denotes the time series
dimension. Y
it
represent the dependent variable while X
it
stands for the explanatory variables
such as Net Interest Income (NII), Net Asset Quality (NAQ), Earning Per Share (EPS), Insider
Loan (IL), Operational Efficiency (OE) and Provision Coverage (PC) among others included in
the model. Given the ability of ? to be the same across units, Ordinary Least Squares (OLS)
provides a consistent and efficient estimate of ?
0
and ?
i-n
in the two general equations presented
below:
DEit = f (?o + ? i1 NII + ?i2 ps E + ?pc + ?i4 IL + OEi5 + e) -------- Equation 2
Where DEit = the squared differences between the sampled banks deposit exposure; that is
excess of liquid asset over deposits and short term liabilities. Explanatory variables are as
described while e represent the stochastic (error) term for the equation. In Equation 3,
Operational Efficiency (OE) for sampled banks are adopted as performance metric and related to
selected risk and earning factors.
OEit = F?o + ?1 + ?iPC + ?2PAT + ?3EPS + ?4NII + ?5NAQ + ?6IL + ?7DE +
e…Equation 3.
OE as a representative term for bank management efficiency is expressed as a function of
selected risk and performance metrics namely adequacy of Provision coverage (PC), Profit after
tax (PAT), Earning per Share (EPS), Net Interest Income (NII), Net Asset Quality (NAQ),
Insider Loan (IL) and Deposit Exposures (DE) for the sampled banks in the study period, with
(e) representing the stochastic error term.
ANALYSIS AND DISCUSSION
Relationship between CRM and Bank Performance
The first tested model undertakes an investigation into the relationship between Deposit
Exposure (DE) as surrogate for bank credit management efficiency in relation to performance
using such performance parameters as Earnings per Share (EPS), Insider Loan (IL), Net Asset
Quality (NAQ), Net Interest Income (NII), Operational Efficiency (OE), Profit after Tax (PAT)
and Provision Coverage (PC). Findings are presented in table 4.1 below:
Table - 4.1 Regression Result for model 1.
Dependent
Variable
Explanatory
Variables
Obtained
estimates
Standard Error Estimated t-value
Constant -50220.09 48390.63 -1.04
EPS -9228.52 4526.49 -2.04
Far East Journal of Marketing and Management Vol. 2 No. 1 April 2012
45
DE
IL +0.11 0.79 0.15
NAQ +767.13 646.68 1.19
NII -0.24 0.09 -2.65
OE +0.23 0.11 2.02
PAT -0.05 0.05 -0.96
PC +1.31 0.79 1.65
R
2
= 0.3285 t
0.025
= 2.o74
Adj – R
2
= 0.1185
F-stat = 1.5738 F
0.05
(7, 22) = 2.46
D-W = 0.710
Source: Computations and Out-Put of e-view based on Author?s Field study (2012).
According to the result presented in Table 4.1 above, a unit increase in the Earnings per Share of
the sampled banks over the study period resulted in a 9228.52 per cent decline in the banks?
deposit exposure. Also, such unit changes in Insider Loan, Net Liquid Assets, Net Interest
Income, Operational Efficiency, Profit after Tax, and Provision Coverage led to 0.12 per cent
increase, 767.13 per cent increase, 0.24 per cent decrease, 0.23 per cent increase, 0.05 per cent
decrease and 1.31 per cent increase in the explanatory variables respectively.
The statistical properties however reveal a weakness in the model. For instance, the R-Squared
value of 0.3285 suggests that only about 33 per cent of the total changes in the dependent
variable are attributable to changes in the explanatory variables. The R
2
is further reduced to a
mere 12 per cent after adjustment which is even lower.
Furthermore, given that t-value (at 0.025 for a two-tailed test or 5% for a one-tailed test) the
tabulated 0.025 is not the same as the estimated t- value at the 0.05 or 5% significance level, only
the Net Interest (NII) parameter is significant as other parameters turned up with estimated t-
values lower than the tabulated t-value. The same applies to the F-statistic, whose estimated
value of 1.5738 is lower than the 5% tabulated value of 2.46, implying that the overall regression
plane is statistically insignificant. There is the presence of First-order serial correlation
(autocorrelation) also as the estimated Durbin Watson (DW) statistic of 0.710 is lower than the
Lower Limit value of the 5% tabulated Durbin Watson value of 1.07 (for at least five
explanatory variables).
Due to the observed weakness of the first model of the study in equation 2, the study concludes
that there is no significant relationship between Credit Risk Management (CRM) and bank
performance over the period under review. The study further undertakes to determine the
relationship, if any, between financial heath and CRM.
Relationship between CRM and Financial Health
The need to examine the relationship between CRM and Financial health gave rise to the
formulation of the second model of the study in Equation3. In this case, Operational Efficiency
(OE) is used as a proxy for the banks? financial health while the other seven explanatory
variables were used to capture CRM. The result obtained is presented in the following table.
Table - 4.2 Regression Result for model II.
Far East Research Centre www.fareastjournals.com
Dependent
Variable
Explanatory
Variables
Obtained
estimates
Standard Error Estimated t-value
OE
Constant +75079.58 85425.37 0.88
DE +0.69 0.3 2.02
EPS +18057.57 7752.92 2.33
IL -2.33 1.30 -1.79
NAQ -925.59 1152.99 -0.80
NII +0.64 0.13 5.04
PAT +0.05 0.10 0.53
PC 5.14 0.99 -5.18
R
2
= 0.7211 D-W = 0.909
Adj.R
2
= 0.6429 t
0.025
= 2.074
F-stat = 8.459 F
0.05
(7,22) = 2.46
Source: Out-put of e=view based on Authors Field study (2012)
According to the result in Table 4.2 above, within the period under review, a unit change in the
Deposit exposure of the banks resulted in about 0.69 per cent rise in Operational efficiency. Such
a positive relationship is seen also for the Earnings per share, Net Interest Income, and Profit
after Tax variables where unit changes raised Operational efficiency by 18057.57 percent, 0.64
percent and 0.05 percent respectively. On the other hand, there is a negative relationship between
Insider Loan and Operational Efficiency, between Net Asset Quality and Operational Efficiency
and between Provision Coverage and Operational Efficiency. In the first case, a unit change in
the volume of Insider loan decreased Operational Efficiency by about 2.33 percent, while the
same is 925.59 percent for Net Asset Quality and 5.14 percent for Provision Coverage.
The estimates obtained for the parameters of the second model are more reliable as three of the
seven explanatory variables are significant: Earnings per Share, Net Interest Income and
Provision Coverage. The R-Squared value of 0.7291 (about 73 per cent) is satisfactory as only
about 27 per cent of the total variations in the banks? operational efficiency over the review
period is due to variations in other factors not captured by the study models; even after
adjustment the R-Squared is still as high as 64 percent. The result is further attested by the F-
Statistic value of 8.459 (higher than tabulated value of 2.074), the entire regression plane is
significant.
The case of whether there is the presence of positive first order serial correlation is indeterminate
by the Durbin-Watson (DW) Criterion at the 1% significance level and maximum of five
explanatory variables. This is due to the fact that the observed DW statistic of 0.909 lies between
the tabulated Lower and Upper limit values of 0.88 and 1.61 (for maximum of five explanatory
variables)although autocorrelation is established at the 5% level of significance in which case the
observed DW statistic is clearly lower than the tabulated value of 1.07. The above
notwithstanding, due to the overall significance of the estimated regression model (II), the study
safely concludes that there is a significant positive relationship between CRM and Financial
health for the sampled banks
Unit Root Test
Far East Journal of Marketing and Management Vol. 2 No. 1 April 2012
47
It is almost a convention in time series analysis, to verify the order of integration for each series
so as to avoid perennial problem of spurious regression (see Granger and New bold 1974; Philip,
and Perron 1988). A test of stationarity property for each variable in the study is conducted using
Augmented Dickey Fuller (Dickey and Fuller, 1981) procedure.
The result of the Unit-root test in table 4.3 suggest that at both level and first difference, the unit
root hypothesis cannot be rejected at 1%;5% and 10% critical level for all the variables. This in
effect suggests that employed data series are non.-stationary and this is suitable for the purpose
A test of stationary employed is the Unit Root Test using the Augmented Dickey Fuller (ADF)
Test. All the time series data on the variables of the study are included. Part of the result obtained
is presented in table. 4.3 below.
Table - 4.3 Result of Unit Root Test
Variable ADF test
statistics
1%
Critical
Value
5%
Critical
Value
10%
Critical
Value
Order of
Integration
DE -4.3767 -4.3382 -3.5867 -3.2279 1(1)
EPS -3.9813 -3.6852 -2.9705 -2.6242 1(0)
OE -4.5526 -4.3382 -3.5867 -3.2279 1(1)
IL -4.7970 -4.3382 -3.5867 -3.2279 1(1)
NAQ --4.610 -4.3226 -3.5796 -3.2239 1(0)
NII -4.4481 -4.3382 -3.5867 -3.2279 1(1)
PAT -5.7510 -4.3382 -3.5867 -3.2279 1(1)
PC -5.3454 -4.3382 -3.5867 -3.2279 1(1)
Source: Author?s Compilation based on Output result from the E-view (2012).
As observed in table 4.3, in relation to the explanatory variables, the explained variable in
equation I, Deposit Exposure (DE), is stationary at all levels of critical values, being integrated
of order 1. Similarly the data for the dependent variable of equation II, Operational efficiency
(OE), is stationary at all observed levels of critical values, being integrated also of order 1.
However the focal point of reference is the 5% Critical Level; hence, the variable which are
stationary only after differencing are non-stationary at the 5% Critical Value. Stationarity means
that the estimated values of the regression model do not indefinitely diverge from the true
regression line.
Causality Test between Financial Health and CRM
The Pair wise Granger Causality Test was further used to augment the estimated models of the
study. A simple standard causality test that is the pair wise Granger causality test employed
examines bi-directional relationship between two variables selected at a time in the study. Our
empirical results presented in table 4.4 (see appendix 1) indicate that decomposed Bank Deposit
Exposures (DE) against indicators of performance found that the former did not Granger cause
the later. Adjustment was made for 2 lags; hence the result includes 28 observations.
Table - 4.4: Result of Granger Causality Test
Null Hypotheses F-Statistics Probability Status
Far East Research Centre www.fareastjournals.com
OE does not Granger Cause DE
DE does not Granger Cause OE
2.39966
0.06409
0.11310
0.93808
Accepted
Accepted
EPS does not Granger DE
DE does not Granger Cause EPS
0.76100
0.38469
0.47861
0.68496
Accepted
Accepted
IL does not Granger Cause DE
DE does not Granger Cause IL
0.30309
1.49583
0.74144
0.24506
Accepted
Accepted
NAQ does not Granger Cause DE
DE does not Granger Cause NAQ
0.60167
0.55495
0.55630
0.58160
Accepted
Accepted
NII does not Granger Cause DE
DE does not Granger Cause NII
1.34902
0.18877
0.27927
0.82925
Accepted
Accepted
PAT does not Granger Cause DE
DE does not Granger Cause PAT
0.02820
0.16146
0.97222
0.85186
Accepted
Accepted
PC does not Granger Cause DE
DE does not Granger Cause PC
2.18232
0.14283
0.13558
0.86766
Accepted
Accepted
EPS does not Granger Cause OE
OE does not Granger Cause EPS
0.07146
2.91814
0.93124
0.07423
Accepted
Rejected
IL does not Granger Cause OE
OE does not Granger Cause IL
0.37147
0.14463
0.69378
0.86613
Accepted
Accepted
NAQ does not Granger Cause OE
OE does not Granger Casus NAQ
0.28790
0.75002
0.75250
0.48356
Accepted
Accepted
NII does not Granger Cause OE
OE does not Granger Casus NII
1.57293
3.55585
0.22895
0.04512
Accepted
Rejected
PAT does not Granger Cause OE
OE does not Granger Cause PAT
0.52752
0.95377
0.59703
0.40000
Accepted
Accepted
PC does not Granger Cause OE
OE does not Granger Cause PC
0.69415
7.67791
0.50966
0.00279
Accepted
Rejected
IL doe not Granger Cause EPS
EPS does Granger Cause IL
0.64533
0.69932
0.53372
0.50718
Accepted
Accepted
NAQ does Granger Casus EPS
EPS does Granger Cause NAQ
1.05816
3.35684
0.36339
0.05259
Accepted
Rejected
NII does not Granger Cause EPS
EPS does not Granger Cause NII
3.16423
5.80692
0.06110
0.00909
Rejected
Rejected
PAT does not Granger Cause EPS
EPS does not Granger Cause PAT
0.07416
0.92242
0.92875
0.91206
Accepted
Accepted
PC does not Granger Cause EPS
EPS does not Granger Cause PC
0.42311
2.20529
0.66000
0.13299
Accepted
Accepted
NAQ does not Granger Cause IL
IL does not Granger Cause NAQ
0.33492
0.26524
0.71882
0.76934
Accepted
Accepted
NII does not Granger Cause IL
IL does not Granger Cause NII
0.29495
0.16603
0.74734
0.84802
Accepted
Accepted
PAT does not Granger Cause IL
IL does not Granger Cause PAT
0.04054
0.36042
0.96034
0.70125
Accepted
Accepted
PC does not Granger Cause IL
IL does not Granger Cause PC
0.29319
0.00020
0.74863
0.99980
Accepted
Accepted
NII does not Granger Cause NAQ
NAQ does not Granger Cause NII
1.94855
0.38967
0.16529
0.68167
Accepted
Accepted
Far East Journal of Marketing and Management Vol. 2 No. 1 April 2012
49
Source: Result Generated from e-view output (2012).
Table 4.4 testifies that the Granger Causality Test reveals a satisfactory situation. The Observed
or empirical F – statistic was checked against the 5% tabulated values of 2.46 to ascertain cases
of two-way causality. Of the twenty-eight-pair result obtained, only in a single case do we have a
two-way causation in which case both null hypotheses are rejected. The reported case is between
Net Interest Income (NII) and Earnings per share (EPS). The null hypotheses were rejected,
meaning that either one Granger-cause, or give rise to the other. In other paired results the null
hypotheses were accepted for each pair or for an item in the pair.
CONCLUSION AND POLICY IMPLICATION OF FINDINGS
The application of annual data for the period 2004-2009 has examined the validity of credit risk
management efficiency in selected Nigerian commercial bank. For this purpose empirical
investigation of the extent of relationship between dependent and explanatory variables was
undertaken using regression analysis.
Results of the regression analysis found minimal causation between Deposit exposure (DE) as
surrogate of credit management efficiency and performance indicators but greater dependency
Operational Efficiency (OE) parameters.
Empirical investigation of the stationary properties and the order of integration of the employed
variables are conducted using ADF. The results show that all the variables were non-stationary at
both level and first differences.
Evidence from the pair wise Granger causality tests suggest that Deposit Exposure –
performance influence does not hold for he Nigerian commercial banks. These results could be
explained due to the influence of several factors such as the nature of technology; cost structure,
managerial effectiveness considered as significant Key Performance Indicators (KPI) which have
not been fully factored into decision- making process in the Nigerian Banks, Ogubunka (2011),
Woherem (2001) and Onaolapo (2007).
These empirical findings have significant implications for bank strategic planners and regulatory
authorities. Operational activities of a typical Nigerian commercial bank are not sufficiently
optimized to ensure maximum earnings from credit creation cum loans and advances. This
explains why loan interest earnings are sufficiently eliminated by rising cost of credit given the
high incidence of non-performing credit. A more pragmatic credit management and screening
PAT does not Granger Cause NAQ
NAQ does not Granger Cause PAT
0.52322
0.87512
0.59949
0.43024
Accepted
Accepted
PC does not Granger Cause NAQ
NAQ does not Granger Cause PC
1.63380
0.06262
0.21704
0.93946
Accepted
Accepted
PAT does not Granger Cause NII
NII does not Granger Cause PAT
0.57077
1.63148
0.57289
0.21748
Accepted
Accepted
PC does not Granger Cause NII
NII does not Granger Cause PC
0.22700
1.95306
0.79869
0.16466
Accepted
Accepted
PC does not Granger Cause PAT
PAT does not Granger Cause PC
0.62326
0.59690
0.54500
0.55882
Accepted
Accepted
Far East Research Centre www.fareastjournals.com
procedure are necessary attempt needed to enhance earnings from credit as the single largest
activity of a commercial bank.
Furthermore undue emphasis of regulatory authorities on bail-out; recapitalization and merger of
banks often designed as exit-attempts to correct incidences of financial ill- health in Nigeria
might not after all be necessary where policy structure are designed to reduce high mortality of
business and worrisome infrastructural decay, currently the bane of Nigerian manufacturing
sector.
Perhaps more often than policies and infrastructural development is the need for personnel
training and retraining in Nigeria banks. These will enhance articulate research and development
in new product and market, as well as cost minimization efforts.
REFERENCES
African Report Group Jeune African (2009); The African Reports On Nigerian Bank.
Ahmed N.H (2003); Credit Risk Determinants: BY Institutional Type Proceedings of Malaysian
Finance Association Conferences (2003).
Ahmed N.H and Ariff M. (2001); Multi-country study of Bank Credit Determinants
“International Journal of Banking and Finance Vol.5 Issue I Article 6 (2007).
Athanasoglou P; Brissimis S. and Lelis M. (2005); “Bank – Specific, industry specific and
Macroeconomic Determinants of Bank Profitability”. Bank of Greece Working paper No. 25.
Al-Thamimi H. and Al-Mazrooei F. (2007); Bank Risk Management. Comparison Study Of
UAE National And Foreign Banks “The Journal Of Risk Finance Vol.8 Issue 4 pp.394-409.
Ariff M. and Marrisetti V.B (2001); A New Approach To Modeling Multi-country Risk
Premium Using Panel Data Test Method “Proceeding of MFS Conference In Cuprus (2003).
Basle Committee on Banking Supervision (2003) Risk Management Principles for Electronic
Banking, July. http//www.bis.org/Publ/bcbs98.pdf.
Berger, A. N. Daiq : Ongena S. and Smith D.C (2003); “To what Extent is the Banking Industry
Globalized?. A study of Banks, National Banking and Research in 20 European nations. Journal
of Banking and Finance Vol. 27 (3) March 2003 pp: 383-415.
Bidani S.N Mitra P.K and Praniod Kumar (2004); Taxmann?s Credit Risk Management.
Taxmann Allied Services Ltd New Delhi, India.
Casu B; Claudia G. and Philip M. (2006) Introduction to Banking. Pearson Education Ltd
Prentice Hall 100-273-69302- pp 258-269.
www.cenbank.org.ng (2010).
Cebenoyan A.S and Strahan P.E (2004); Risk Management, Capital Structure and Lending at
banks “Journals of Banking And Finance Vol. 28 pp: 19-43.
Far East Journal of Marketing and Management Vol. 2 No. 1 April 2012
51
Chen B, Cheng X and Wu L. (2005); Dynamic Interactions Between Interest Rate Credit and
Liquidity Risk “Theory and Evidence From Term Structure Of Credit Default and Swap spreads.
Working Paper Series Aug. 8 2005.
Central Bank of Nigeria (Various Years); Annual Reports and Statement of Accounts CBN.
Dickey D. and Fuller W. A (1998); Likelihood Ratio Statistics for Auto Regressive Time Series
with Unit Root. Econometrica: 0574972.
Diamond D.W (1984); Financial Intermediation and Delegated Monitoring. Review of
Economics Studies 51(3)393-414.
Demirguk – Kunt (1989); Deposit Institution Failures: “A Review Of the Empirical Literature”
Federal Reserves Of Cleveland Economic Review Qtr 4, 1998.
Fatemi A. and Fooladi I. (2006); Credit Risk Management. A Survey Of Practices “Managerial
Finance Vol. 32, Issue 3, pp:227-233.
Granger C.W.J. and Newbold P. (1974); Spurious Regression In Econometrics Pp:111-120.
Hassan M.K (2009); Risk Management Practices of Islamic Banks of Brunci Darusallam “The
Journal Of Risk Finance, Vol. 10, pp:23-37.
Koizol C. and Lawrence J. (2008); What Makes a Bank Risk? Insights from the Optimal Capital
Structure Of Banks “Journal of Banking and Finance. Vol. 33 pp:861-873.
Meyer L. (2000); Why Risk Management is Important for Global Financial Institutions. Risk
Management of financial Institutions; UN Conference Center Bangkok, Thailand.
Onyiriuba L. (2004) Analysing and Managing Risk of Banks Lending. Malthouse press Ltd,
Lagos Nigeria.
Okey N. (2010); Sanusi: Still fighting one year after. Nigeria Business World Intelligence, June
6-13 (2010).
Ogubunka U.M (2011); An Analytical Review of Cost Trend in The Banking Industry as a
reflection of the Cost Pressure in the Nigerian Economy. The Nigerian Banker, March 2011,
ISSN 0197-6679.
Onaolapo A (2007) “An Evaluation of the Effects of Recapitalization On the Financial Heath of
Nigerian Commercial Banks. A Ph.D thesis submitted to the school of Postgraduate Studies,
LAUTECH, Nigeria.
Onyiriuba L. (2004) Analysing and Managing Risks of Bank Lending. Malthouse Press Ltd,
Lagos, Nigeria.
Onwumere J.U.J (2005) Banking Industry Reforms and Developing the Nigerian Economy. The
Nigerian Bankers Journal of the CIBN, Oct. – Dec. (2005) ISSN 0197-6679.
Peter S. R. and Sylvia C. H. (2008) Bank Management and Financial Services, 7
t h
Edition,
McGraw Hill, 978-007-125938-4.
Far East Research Centre www.fareastjournals.com
Phillp P. C and Peron P. (1988); Testing for Unit Root in Time Series Regression biometrica 75
335-346.
Satchidananda S.S (2006); Multi-Step Method for Analysis of Credit Risk of Bank Loan, Center
for Banking IT Working Paper WP. (2006).
Saunders A. and Cornett M.; (2003); Financial Institutions Management. A Risk Management
Approach, 4
t h
Edition New York McGraw Hall Irwin.
Sinkey J. (1998); Commercial Bank Financial Management, 5
t h
Edition, London Prentice Hall
International.
Woheren E.E (2001); Information Technology in the Nigerian Banking Industry. Spectrum
Books Ltd; Ibadan Nigeria.
doc_317263103.pdf