Regulatory capital and its effect on credit growth non performing loans and bank

Description
This paper aims to investigate the influence of the central bank’s regulatory capital on
commercial banks specific performance outcomes such as credit supply, interest rate spread (as a
measure of efficiency) and non-performing loans (NPLs).

Journal of Financial Economic Policy
Regulatory capital and its effect on credit growth, non-performing loans and bank
efficiency: Evidence from Ghana
Eric Osei-Assibey J oseph Kwadwo Asenso
Article information:
To cite this document:
Eric Osei-Assibey J oseph Kwadwo Asenso , (2015),"Regulatory capital and its effect on credit growth,
non-performing loans and bank efficiency", J ournal of Financial Economic Policy, Vol. 7 Iss 4 pp. 401
- 420
Permanent link to this document:http://dx.doi.org/10.1108/J FEP-03-2015-0018
Downloaded on: 24 January 2016, At: 21:53 (PT)
References: this document contains references to 48 other documents.
To copy this document: [email protected]
The fulltext of this document has been downloaded 243 times since 2015*
Users who downloaded this article also downloaded:
Nico B. Rottke, J ulia Gentgen, (2008),"Workout management of non-performing loans: A formal
model based on transaction cost economics", J ournal of Property Investment & Finance, Vol. 26
Iss 1 pp. 59-79http://dx.doi.org/10.1108/14635780810845163
Eric Osei-Assibey, Baimba Augustine Bockarie, (2013),"Bank risks, capital and loan supply:
evidence from Sierra Leone", J ournal of Financial Economic Policy, Vol. 5 Iss 3 pp. 256-271 http://
dx.doi.org/10.1108/J FEP-09-2012-0041
D.C. Smith, M. Bruyns, S. Evans, (2011),"A project manager's optimism and stress management and
IT project success", International J ournal of Managing Projects in Business, Vol. 4 Iss 1 pp. 10-27http://dx.doi.org/10.1108/17538371111096863
Access to this document was granted through an Emerald subscription provided by emerald-
srm:115632 []
For Authors
If you would like to write for this, or any other Emerald publication, then please use our Emerald
for Authors service information about how to choose which publication to write for and submission
guidelines are available for all. Please visit www.emeraldinsight.com/authors for more information.
About Emerald www.emeraldinsight.com
Emerald is a global publisher linking research and practice to the benefit of society. The company
manages a portfolio of more than 290 journals and over 2,350 books and book series volumes, as
well as providing an extensive range of online products and additional customer resources and
services.
Emerald is both COUNTER 4 and TRANSFER compliant. The organization is a partner of the
Committee on Publication Ethics (COPE) and also works with Portico and the LOCKSS initiative for
digital archive preservation.
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
1
:
5
3

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
*Related content and download information correct at time of
download.
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
1
:
5
3

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
Regulatory capital and its effect
on credit growth, non-performing
loans and bank effciency
Evidence from Ghana
Eric Osei-Assibey
Department of Economics, University of Ghana, Accra, Ghana, and
Joseph Kwadwo Asenso
Ministry of Finance, Accra, Ghana
Abstract
Purpose – This paper aims to investigate the infuence of the central bank’s regulatory capital on
commercial banks specifc performance outcomes such as credit supply, interest rate spread (as a
measure of effciency) and non-performing loans (NPLs).
Design/methodology/approach – Using specifc commercial bank-level panel data from2002-2012,
a system of equations was modeled that allows us to apply the system generalized methods of moment
approach and estimate the equations, while controlling for specifc bank level, industry and
macroeconomic variables.
Findings – The study fnds a positive relationship between a net minimum capital ratio and the net
interest margin. Although this is in contrast with the study expectations, the result suggests that a high
net minimum capital requirement would widen the spread between the lending and saving rates. The
study further fnds evidence to support the fact that high minimum capital requirement and excess
capital above the minimumrequired drive credit growth in the banking sector of Ghana. However, high
excess capital increases risk-taking activities of the banks, as excess capital is found to be associated
with high NPL ratios.
Practical implications – Given the economic benefts and costs of sharply increasing bank
regulatory capital, our results speak to the ongoing debates on the right level of capital, the effectiveness
of the Bank of Ghana policy rate (PR) and the high lending rates that appear to respond only slowly to
macroeconomic indicators such as the PR and the infation rate. The fnding also has practical
implications for the adoption of the Basel III accord.
Originality/value – The empirical literature has not paid enough attention to the impact of regulatory
capital on the three specifc bank-level outcomes – NPLs, interest rate spread and the nature of
interrelationships among these variables, particularly in the African context.
Keywords Banks, Regulatory change, Capital
Paper type Research paper
1. Background
There has been a growing wealth of literature on the importance of regulatory capital for
bank stability and soundness in recent times. This has even been reinvigorated by the
introduction of the global charter that regulates banks’ capital requirements – the Basel
The authors are particularly grateful to the International Growth Centre, UK, for providing
fnancial support to undertake this study.
The current issue and full text archive of this journal is available on Emerald Insight at:
www.emeraldinsight.com/1757-6385.htm
Regulatory
capital
401
Received16 March2015
Revised10 July2015
Accepted27 July2015
Journal of Financial Economic
Policy
Vol. 7 No. 4, 2015
pp. 401-420
©Emerald Group Publishing Limited
1757-6385
DOI 10.1108/JFEP-03-2015-0018
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
1
:
5
3

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
III Accord – which aims to make each bank’s capital holdings proportional or sensitive
to its potential credit losses (The Basel Committee on Banking Supervision, 2006)[1].
This is because capital has long been recognized as one of the key factors to be
considered when the safety and soundness of a particular bank is being assessed.
Greuning and Bratanovic (2000), analysing banking risk, observed that an adequate
capital base serves as a safety net for a variety of risks to which an institution is exposed
in the course of its business. Furthermore, while capital absorbs possible losses and thus
provides a basis for maintaining depositor confdence in a bank, it also serves as the
ultimate determinant of a bank’s lending capacity and risk-taking activities. Moreover,
according to these studies, capital availability determines not only the maximumlevel of
assets a bank holds but also the amount and cost of capital impact on a bank’s effciency
and its competitive position.
At the same time, high risk-based capital adequacy requirements, serving as a buffer
against risk, can accentuate the procyclicality of bank lending, which is damaging to all
economies – but particularly so for fragile developing ones, which are more vulnerable
to strong cyclical fuctuations of fnancing (Griffth-Jones and Persaud, 2008). This is
because as raising equity capital is costly or very diffcult in developing countries, like
those in Africa, and the cost of holding capital comes over to loan prices (Jokivuolle et al.,
2007), banks may be forced to scale back their lending activities, particularly, to the
agricultural and high-risk informal sectors, to maintain a minimum capital buffer on
less risky assets. Thus, while minimum capital regulations are certainly necessary,
overly strict or high capital adequacy regulations can also become a disincentive to
credit expansion, particularly, to the perceived highly risked agricultural and informal
sectors which are the bedrock of most African economies.
In this regards, several theoretical studies assessing and comparing the fat-rate
capital requirement under the previous Basel I regime and the Basel II regime, which has
a risk-based capital requirement and their respective impacts on some banking sector
outcomes and on the macroeconomic indicators, have emerged (Jokivuolle and Vesala,
2007; Griffth-Jones and Persaud, 2008; Pasiouras et al., 2009; Pennacchi, 2005; and
Kopecky and VanHoose, 2006). The fndings of the regulatory capital adjustments have
however been somewhat contradictory on varying indicators, such as banking
effciency, credit risk and lending.
Ghana is an interesting case study in Africa because of the recent regulatory capital
adjustments that have occurred in its banking sector and the various risk concerns that
have been expressed by regulators and researchers alike. The minimum capital
requirement or the regulatory capital has been increased substantially on three separate
occasions since 2003. These developments made the banking industry fairly
well-capitalized with stated capital increasing from GH¢16.2 million (US $22.5 million)
in 2001 to about GH¢1,700 million (US $1,016 million) in February 2012. By the end of
2013, the stated capital of the banks had increased to GH¢2,345.4 million (US $1,172.7
million). In each of the years, the growth rate was exponential. These high increases
were also due to the fact that banks had continually kept their capital well in excess of
the minimum required on their balance sheets. Figure 1 compares the deviation of the
actual stated capital of the top six banks to the total minimum required by the BOG.
The six banks, which are also known as the frst quartile banks, account for 51 per cent
of the total industry assets. The deviation of the stated capital from the minimum
JFEP
7,4
402
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
1
:
5
3

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
required capital has been increasing over the years, as depicted by the height of the
vertical bars.
These massive injections of capital in the banking industry was expected to
stimulate competition in the banking sector, promote effciency and drive down lending
rates in particular, which have remained high over a long period of time. While credit
supply to the private sector has increased substantially, lending rates and the
non-performing loan (NPL) ratios remained high raising concerns about the quality of
bank assets and the macroeconomic implications. Even though the central bank
constantly adjusts its policy rate (PR) as a nominal anchor to infation, banks lending
rates have remained largely unresponsive to these adjustments.
To the best of our knowledge, no study has explicitly studied the impact of
minimum capital adjustments and other central bank policy instruments on banks’
specifc performance outcomes (interest rate spread, NPLs and credit allocation) and
macroeconomic variables, as well as the nature of interrelationships among these
variables, particularly in the African context. This study therefore extends the
frontiers of literature by flling this void and contributes to the on-going debate by
seeking answers to the following specifc questions: to what extent, do net minimum
capital requirements (the difference between stated capital and the minimum capital
requirement) with mainstream banks risk-taking incentives, effciency and supply
of credit? In other words, does high stated capital of banks infuence excessive
risk-taking activities, supply of credit and interest rate spread? How do banks of
different ownership structure respond to regulatory capital and policy instruments
(such as the PR) adjustments? What is the nature of relationship between the
supply of credit, NPL, regulatory capital and interest rate spread as a measure of
effciency?
The remainder of the paper is organised as follows: Section 2 presents the
macroeconomic context and an overview of the banking regulatory environment.
Section 3 comprehensively reviews the literature while Section 4 presents the
conceptual framework and empirical model specifcation. Section 5 reports the
model estimation procedure and discusses the estimation results. Section 6
concludes with the summary of the main fndings and the policy implications of the
fndings.
-
100,000,000
200,000,000
300,000,000
400,000,000
500,000,000
600,000,000
700,000,000
800,000,000
900,000,000
2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013
A
m
o
u
n
t

(
G
H
c
'
m
i
l
l
i
o
n
)
year
Stated Capital
Minimum Capital
Source: Bank of Ghana Annual Reports and Banks’ Financial Statements (2003-2013)
Figure 1.
Stated capital of
some selected banks
and the minimum
capital requirement
403
Regulatory
capital
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
1
:
5
3

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
2. Regulatory environment and recent developments in the banking
sector
As the banking industry across Africa began bracing itself to adopting the new
risk-based supervision and a more complex capital adequacy framework, known as the
Basel III[2], the BOG responded by repealing the existing Banking Act (Banking Law,
1989 [PNDCL 225]) and replaced it with the Banking Act, 2004 (Act 673). This Act
incorporates the requirements of the core principles of Basel II. Therefore, the main
purpose of the Act was to amend and consolidate the laws relating to banking, to
regulate institutions which carry on banking business and to provide for other related
matters. Further, under this Act, to ensure that the local banking industry meets the
more rigorous requirements of Basel III and to make the banking sector well capitalized,
the minimum capital requirement and the CAR were adjusted upward and continue to
be reviewed.
As mentioned previously, the Bank of Ghana since 2013 had reviewed upwards the
minimum capital requirement for the country commercial banks on a number of
occasions. For example, the bank increased the minimum capital to GH¢7 million (US
$7.8) by the end of 2006 and further raised in 2009 to GH¢60 million (US $50m) all
foreign-owned banks and GH¢25 million US $21m) for the Ghanaian-owned banks. The
latest was the announcement made in 2013 that new commercial banks are required to
have a minimumstated capital of GH¢120 million (US $ 55.5 million). Existing banks are
only required to maintain a stated capital of GH¢60 million it set previously. However,
the understanding was that existing banks would voluntarily grow their capital to the
GH¢120 million in line with their business. Since then four existing banks have moved
their capital voluntarily to over GH¢120 million, according to the 2014 Ghana Banking
Survey report.
However, the knock-on effects of these massive capital injection changes have been
mixed for the country. On the positive side, for example, the total operating assets of the
industry had increased from GH¢2.3 billion in 2003 to about GH¢12.42 billion in 2009,
representing an annual growth rate of well over 66 per cent per cent over the period. This
had further increased to GH¢42.5 billion in 2013[3]. A Ghana Banking Survey Report
(2008) observes that a key result of compliance with the recapitalization directive in 2007
was that bank lending increasing fromGH¢1,055 million (US $1,2126 million) in 2003 to
GH¢2,464 million (US $2,6213 million) in 2007, representing a 133 per cent increase.
Thereafter, the growth of credit had been kept above 40 per cent until a dramatic
downturn in 2010 when there was a sharp decline to 13.2 per cent Figure 2). Credit to the
private sector picked up thereafter but declined by 28.6 per cent to a nominal amount of
GH¢14,757.2 million in 2013, compared with 34.1 per cent in 2012. The sharp downturn
in credit growth (Cr) in 2010 was believed to be a result of a tightening response to a
sharp upturn in impairment allowances for NPL, as shown in Figure 2. The NPL ratio
increased to 20 per cent in 2010 from 8.1 per cent in the corresponding period in 2009.
There is a concern of banks being tempted to engage in bad lending, as they have
capital on their books in volumes unprecedented in Ghana’s banking history, for which
they need to fnd proftable business to generate returns for providers of capital (Ghana
Banking Survey Report, 2008).
However, as Figure 3 shows, the country’s lending rates have remained high
(averaging about 27 per cent) over a long period of time. A wide spread between the
lending and deposit rates has been a key feature of the banking industry over the past
JFEP
7,4
404
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
1
:
5
3

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
decade with the net interest spread recording over 12 per cent in December 2013. Banks’
lending rates have largely remained unresponsive to the BOG PR, which is supposed to
be an indicative rate around which all the other rates revolve. This development raises
macroeconomic stability concerns as even when both infation rate and the PR were
consistently declining over 2009-2011, the lending rates by banks did not follow suit.
3. Theoretical literature
The controversy on the exact impact regulatory capital has on banking outcomes
remains despite the theoretical and empirical interests it has generated for several
decades now. The strand of empirical literature linking these variables is mixed and
even more in confict than in the theory.
For example, while many studies fnd that higher or stricter capital requirements
reduce the proftability of future lending and/or banking effciency (Repullo and Suarez,
2008), others fnd that stricter capital requirements improve cost effciency and have a
Source: BOG Annual Reports (2006
-
2013)
12.8
7.3
8.4 8.1
20
14.2
13.2
12.3
42
41.7
59.9
46.9
13.2
16.9
34.1
28.6
2006 2007 2008 2009 2010 2011 2012 2013
NPL Ra?os Growth of Bank Credit to the Private sector
Figure 2.
Non-performing loan
ratios and rate of
growth of credit to
the private sector
Source: BOG, 2013
1 2 3 4 5 6 7 8 9 10 11 12 13
Av. In?a?on rate (%) 32.9 14.8 26.7 12.6 15.5 10.9 10.7 16.52 19.29 10.75 8.72 9.2 11.7
BOG Policy rate (%) 27 24.5 21.5 18.5 15.5 12.5 13.5 17 18 13.5 15 15 16
T-Bill rate (%) 28.5 26.6 18.7 17.1 11.4 10.7 10.6 24.7 22.5 12.7 12.3 19.9 23.1
Lending rate (Base rate) 36.5 35.5 32.75 28.8 22.5 21.3 21 27 25 25.8 22.8 25.7 25.6
0
5
10
15
20
25
30
35
40
p
e
r
c
e
n
t
Figure 3.
Trends in infation
rate, T-bill rate, PR
and lending rate,
2001-2013
405
Regulatory
capital
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
1
:
5
3

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
signifcant effect on bank allocative effciency (Pasiouras et al., 2009; Fare et al., 2004 and
Barth et al., 2004). In a more detailed study but with somewhat mixed outcomes,
Pasiouras et al. (2009) used the Stochastic frontier analysis (SFA) model to provide
international evidence on the impact of regulatory and supervisory framework on bank
effciency. They investigated the impact of regulations related to the three pillars of
Basel II (namely, capital adequacy requirements, supervisory power and market
discipline) on the cost and proft effciency of banks. Their fndings suggest that stricter
capital requirements improve cost effciency but reduce proft effciency, while
restrictions on bank activities have the opposite effect, reducing cost effciency but
improving proft effciency. Their fndings support Fare et al. (2004)’s conclusion that
the effect of risk-based capital requirements on the proft effciency performance of US
banks indicates that allocative ineffciency is a larger source of proft loss than technical
ineffciency and that the risk-based capital standards have a signifcant effect on bank
allocative effciency.
Closely related to this present study is that of Kwan and Eisenbeis (1997), which
provides evidence on the links between bank risk, capitalization and operating
effciency. Using data from 176 banks from the USA and applying a simultaneous
equation framework, the study tests hypotheses about the interrelationships among
bank interest rate and credit risk taking, capitalization and operating effciency. A
positive effect of ineffciency on risk taking was found and supports the moral hazard
hypothesis that poor performers are more vulnerable to risk taking than high
performance banking organizations. The study attributed the positive effect of
ineffciency on the level of capital to regulatory pressure on underperforming
institutions. The study further fnds frms with more capital to operate more effciently
than less well-capitalized banking organizations. Moreover, a U-shaped relationship is
detected between ineffciency and loan growth, indicating that operating effciency
improves at a decreasing rate as loan growth rate increases.
In another respect, Berger and Bouwman (2011) investigated howbank capital affect
the survival, proftability and market shares of banks during crises and normal times
using a logit panel regression. The results of the study show that higher capital
increases the survival, market shares and proftability of banks during both normal and
crises times. These results were achieved in separate panel regressions. It is important to
note that this study recognized the existence of potential endogeneity between proft and
market shares and this was addressed using their lagged values. More recently, Berger
et al. (2014) examined how regulatory intervention and capital support infuence banks’
risk-taking predispositions in a panel using IV and the two-stage regression approach.
The study uses data on all banks operating in Germany from 1999 to 2009. The results
of the study show that banks signifcantly reduce their risk taking after regulatory
intervention and capital support. The study results further indicate that liquidity
creation reduces after taking regulatory interventions, but this is not affected by capital
support.
In another study, Boudriga et al. (2009) empirically analyse the determinants of NPLs
and the potential impact of regulatory capital on credit risk exposure in a cross-country
study. They use aggregate banking, fnancial, economic and legal environment data for
a panel of 59 countries over the period 2002-2006. Empirical results indicate that higher
capital adequacy ratio and prudent provisioning policy seem to reduce the level of
problem loans. They also report a desirable impact of private ownership, foreign
JFEP
7,4
406
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
1
:
5
3

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
participation and bank concentration. Another study (Osei-Assibey and Baimba, 2013)
investigates the factors that infuence banks’ credit supply in Sierra Leone. Using
annual bank-level data on an unbalanced panel of 13 banks in the market for the study
period, and using a time and bank-specifc fxed effects model, the study confrms the
principal hypothesis that the level of risk premium infuences the share of loans to the
private sector in interest earning asset of banks. Additionally, the study fnds that NPLs,
Tier 1 capital ratio and local currency deposit levels positively and signifcantly affect
banks’ loan supply to the private sector, while a ratio of a bank’s gross loans and
advances to local currency deposits, as a measure of liquidity and bank size have
signifcant negative effect on the dependent variable.
In the Ghanaian context, literature on regulatory capital and bank-level outcomes is
scanty. To the best of our knowledge, no study has explicitly controlled for minimum
capital requirements – although there have been several attempts to gauge the effciency
of the banking sector either by using the SFA model or interest rate spread. Although
several studies (Amidu, 2006; and 2007; Abor, 2008; Osei-Assibey et al. 2012) have
focused on the determinants of capital structure of banks and enterprises in Ghana, not
much attention has been given to the impact of regulatory capital on bank-specifc
performance. In a quite related study, however, Bawumia et al. (2005), examining the
determination of interest rate spreads in Ghana in a single equation model, conclude that
the existence of major structural impediments such as market concentration and the
degree of contestability among banking institutions, among others, prevent
the fnancial system from reaching its full level of effciency. However, even though the
study controlled for NPLs and the existence of liquidity reserves, as well as high
operating costs, regulatory capital changes were not a subject in their model. A closely
related study by Aboagye et al. (2008) investigates the determinants of interest rate
spreads in Ghana. Although they did not explicitly control for the stringency of capital
requirements, they alluded to the importance of capital adequacy in reducing interest
rate spread. Similar conclusion was also reached by Barnor and Odonkor (2012).
Specifcally, the studies recommend that to help reduce interest rate margins, the central
bank should consider lowering the capital adequacy ratio and banks should be required
to pass the full extent of reductions or increases in the central bank’s prime rate onto
borrowers.
As so far reviewed, the existing literature on the impact of regulatory capital on
banking performance has mostly centred on the advanced economies, while the fndings
have been somewhat contradictory on varying indicators such as banking effciency,
credit risk and lending. Furthermore, the empirical literature has not paid enough
attention to regulatory capital and the three specifc bank-level outcomes – NPLs,
interest rate spread and credit supply. Besides, their counterfactual effects or the
possible interrelationships among these variables have not been discussed in much
detail. Most of the studies have been done in isolation of the other performance
indicators. Yet, Barth et al. (2004) write that while recognizing the advantages of tightly
focused studies, there is a growing evidence stressing that the salient issues in bank
regulation, supervision and specifc bank level outcomes are inextricably interrelated.
Thus, there are advantages to examining an array of regulatory indicators and
regulatory policies simultaneously to identify those that enjoy a strong, independent
relationship with banks’ outcomes such as effciency, risk taking and credit supply.
Apart from these shortfalls, studies on regulatory capital have paid little attention to
407
Regulatory
capital
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
1
:
5
3

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
how different ownership structures of banks respond to regulatory capital and
instruments by the central bank.
4. Theoretical framework and econometric model specifcation
Within the framework of the role of capital in fnancial institutions, Berger et al. (1995)
noted that regulatory capital requirements can have unintended consequences. This is
because, in response to an increase in its required equity-to-asset ratio, a bank might
increase its portfolio risk and raise its probability of failure. For example, changes in
regulatory capital requirements elicit interest rate and Cr adjustments and result in
changes in risk-taking tendencies by banks to stay competitive. However, the theory
further hypothesize that risk-based capital requirements that penalize increases in
portfolio risk can reduce such unintended consequences of capital requirements, but
these standards are imprecise, leaving open the possibility that some banks may
increase portfolio risks when capital standards are raised (Berger et al., 1995). Similarly,
Koehn and Santomero (1980) argue that banks will respond to regulatory actions to
increase their capital and reduce their leverage by increasing asset risk. However, the
theory further indicates that banks with more capital are found to operate more
effciently than banks with less capital, indicating that the level of capitalization is a
good proxy for performance (Kwan and Eisenbeis, 1997). On the other hand, effciency
could, in turn, be also affected by the level of bank risk (Berger and De Young, 1995)
Figure 4.
With regards to the supply and allocation of loans, it is argued that imperfections in
setting the level of required capital and the relative risk weights may lead to allocative
ineffciencies if capital requirements distort relative prices both among banks and
between banks and non-bank competitors, and divert fnancial resources from their
most productive uses (Berger et al., 1995). Moreover, Kopecky and VanHoose (2006)
hypothesize that the imposition of binding capital requirements on a previously
unregulated banking systemunambiguously increases the market loan rate and reduces
aggregate lending. For instance, high credit risks which require high capital buffer
Figure 4.
Regulatory capital
and its relationship
with bank-level
outcomes
JFEP
7,4
408
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
1
:
5
3

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
could lead to credit apprehension – and reduce credit expansion, particularly to the
perceived high risk sectors (like agriculture and households) – which eventually puts an
upward pressure on net interest margins (NIMs) (due to high risk premium). However,
wide interest rate margins can lead to adverse selection and moral hazards which could
exacerbate default rates (Stiglitz and Weiss, 1981).
In this regards, any research on the impact of banks’ regulatory capital should take
into account the inter-relationships among the various outcomes such as credit risk, the
volume of loans, allocation effects, effciency and actual capitalization as encapsulated
in the foregoing theoretical literature. It is clear that regulatory capital requirements
infuence bank-level outcomes which, in turn, trigger certain counter responses from
banks, suggesting that all these variables are endogenously dependent and jointly
determined. Thus, in the same spirit with studies by Kwan and Eisenbeis (1997),
Altunbas et al. (2007) and Barth et al. (2004), we model three separate equations using
the system generalised method of moments (GMMs). This approach has been adopted
due to the perceived endogeneity between some of the variables in our specifed model
while controlling for specifc bank level, industry and country characteristics. This
leads to the specifcation of the model below.
The three models to be estimated are specifed as follows:
NIM
it
? ?
10
? ?
12
MCR
it
? ?
13
PR
it
? ?
14
CR
it
? ?
15
PBD
it
? ?
16
FBD
it
? ?
17
Fees
it
? ?
18
NPL
it
? ?
1t
(1)
Cr
it
? ?
20
? ?
21
MCR
it
? ?
22
PR
it
? ?
23
GDPG
it
? ?
24
PBD
it
? ?
25
FBD
it
? ?
26
NIM
it
? ?
2t
(2)
NPL
it
? ?
30
? ?
31
MCR
it
? ?
32
PR
it
? ?
33
Cr
it
? ?
34
PBD
it
? ?
35
FBD
it
? ?
36
Fees
it
? ?
37
Cr
it
? ?
3t
(3)
where NIMit is net interest margin, Crit is credit growth, PRt is PR, NPLit is
non-performing loans, GDPGit is GDP growth, Feesit is fees charged by banks for the
delivery of a service, PBDit is public bank dummy (i.e. 1, if public, 0 if private), FBDit is
foreign bank dummy (1, if foreign, 0 if domestic) and MCRit is net minimum capital
requirement. The residual terms are represented by the ?’s and are assumed to be
serially uncorrelated but could be contemporaneously correlated across equations,
while the ?s are the impact coeffcients of all the variables on the right hand side.
4.1 Measuring regulatory capital
MCRit, as used in this study, is defned as the ratio of the difference between the
minimumcapital required and a bank’s stated capital position to its assets. This ratio is
intended to measure whether minimum capital requirements set by regulators are
proportional or optimal not only with the actual risks banks take but also with the very
growth and development of the banking industry and its intermediation effciency. It
assesses the degree to which banks’ stated capital deviates/differs from the minimum
required – if it converges, or diverges widely, what would be the effect on a bank’s
outcome? Are banks that keep their stated capital way above the minimum more likely
to be associated with higher level performance outcomes or vice versa? The capital in
question, which the Basel II accord refers to as Tier 1 capital, comprises equity shares,
409
Regulatory
capital
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
1
:
5
3

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
retained earnings, non-redeemable and cumulative preference shares. According to
Greuning and Bratanovic (2000), the level of this capital has a crucial bearing on proft
margins, effciency and banks’ ability to bear risk and stay competitive. Thus, too little
or too much of such capital in relation to the minimumrequired can have implication for
a bank’s performance outcomes. Banks with low MCRit will be risk averse and will
either invest in safe assets or charge above market rates when making loans.
The following paragraphs explain how regulatory capital affects each of the
dependent variables specifed in the model. The frst equation represents banks’
effciency, the second explains credit outlay and the third – credit risk.
4.2.1 The effect of regulatory capital on effciency, credit and NPL. Based on the
specifed equations above, we estimate the impact of regulatory capital on effciency,
credit and credit risk as discussed below:
4.2.2 Regulatory capital and effciency (NIMit). NIMit is measured as the ratio of the
difference between interest income and interest expenditure to total earning assets (or
interest rate spread). Although there are other indicators of effciency such as
intermediation costs (as a percentage of total assets) and the SFA model, this effciency
indicator was chosen because apart from it being a much simpler procedure and the
relative ease of getting data, it is a current policy relevant indicator. This is because
interest rate spreads sometimes remain high despite effciency gains owing to the need
to build loan-loss provisions or charge a risk premiumin lending to high-risk borrowers.
We also expect a positive relationship between MCR and NIMit, as high minimum
capital requirement may lead to high cost of equity funds which will intend lead to high
lending rates.
4.2.3 Regulatory capital and credit (Crit) allocation. Crit, which represents the growth
of credit, is measured as the natural log-difference of credit outlay.
MCRit is expected to have a positive relationship with credit, as banks are likely to
make loans when they have more excess reserves. Regarding Crit and NPL, as stated
earlier, high NPL levels will create credit trepidation, thus leading to few loans being
made. We therefore expect a negative relationship between the two variables. On the
contrary, as banks proft from high interest rate spreads, they are likely to make more
loans at wide interest rate margins and vice versa. This will establish a positive
relationship between NIMit and Crit.
4.2.4 Regulatory capital and NPLit. NPLit is the ratio of NPLs to total gross loans by
banks. It indicates the credit quality of bank loans, thereby serving as a measure of risk
taking incentives of banks.
We expect a positive relationship between MCRit and NPLit, as low excess capital,
for example, will push banks to make less risky loans, while excessive capital holdings
could tempt them to make “bad loans”, as they will feel pressure to make proft to
generate returns for providers of capital. With regards to the other dependent variables,
we expect NIMit to be positively related to NPLit, as wide interest rate margins or higher
lending rates can exacerbate incentive problems such as adverse selection and moral
hazards, thereby increasing the default risk probabilities among borrowers. In the case
of credit, large volumes of loans may bear greater risks and unbridled advances in loans
without the accompanying effective risk mitigation strategies. This will most likely
result in high NPL levels. If the growth rate is excessively high, it could be that best
practices are compromised which could lead to more bad loans being made. Thus, we
expect a positive relationship between Crit and NPLit.
JFEP
7,4
410
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
1
:
5
3

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
5. Controlling for other bank-specifc characteristics, industry and
macroeconomic indicators
5.1 Bank characteristics
The bank characteristics include fees and commissions charged for rendering services
(feesit). In addition to this, bank ownership structure type, which is a dummy made up
of public and private banks (PBD) or foreign and domestic banks (FBD), is included to
assess how the ownership structure affects bank level outcomes.
5.2 Industry and macroeconomic variables
This category is made up of policy rate (PRit) and GDP growth rate (GDPGit). GDPGit
is expected to have a positive impact on credit through, as per capita income would have
increased alongside it, all other things being equal. We also expect a feedback effect that
higher Cr can lead to GDPG. However, a test results for endogeneity shows it was not
severe.
The BOG policy rate (PRit) is another instrument the monetary authority uses to
regulate money supply. This is the rate at which the central bank lends to commercial
banks. As the main operational target, the PR also infuences short- and medium-term
money market rates for open market operations, deposit money banks’ holdings of
excess reserves and, indeed, their own lending and deposit rates. However, the degree of
banks’ responsiveness to this instrument in recent times has been a matter of great
concern. We expect a positive relationship between PRit and NIMit, as banks use PRit as
a reference point when setting their base rates. Ahigher PRit will drive banks to increase
the cost of borrowing to consumers. Again, PRit is expected to have a negative
relationship with Crit because of the fact that the former is a borrowing cost to banks. A
high PRit restricts the availability of loanable funds and, subsequently, credit. We
expect PRit to also have negative a relationship with NPLit for the same reasons as
above.
5.3 Data sources
The study uses secondary banking sector data that spans 2003-2012, mainly due to data
availability and the need to include as many banks as possible for the study results to
refect the situation on the ground. The number of banks used for the study was based
on data availability fromall sources. Those banks that had data gaps were eliminated to
avoid “near single matrix” errors. Furthermore, some prominent existing banks did not
exist in 2003 or had just started operations. Bank-specifc data were sourced from the
Ghana Banking Survey by Price Water House Coopers and the annual fnancial
statements of the banks, while Treasury bill rates (T-bill rates), PRs, reserve
requirements and bank categories were drawn from the BOG annual reports.
6. Estimation procedure and results discussions
We identify the system GMMs estimation technique as the appropriate parameter
estimation technique for the estimations due to the characteristics of the model. The
likelihood of endogeneity issues, individual time-invariant fxed effects heterogeneity,
autocorrelation and the fact that the cross-sections are greater than the time periods for
the available data makes the GMM technique a more appropriate and robust technique
ahead of other available techniques such as the seemingly unrelated regression,
panel-corrected standard error estimates and instrumental variable and the two-stage
least square.
411
Regulatory
capital
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
1
:
5
3

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
Essentially, the system GMM procedure is preferred to the other estimation
techniques for this study because:
• it overcomes the problem of endogeneity through the use of lagged values of
explanatory variables as instruments;
• it eliminates the problem of information loss in cross-sectional regressions, as it
allows for multiple observations for each bank across time;
• it allows for the use of level and lagged values of the variables in the estimation
equation; and
• it is able to give consistent estimates even when T (time periods in years) is small
and N (number of banks) is large.
Additionally, using systemGMMis appropriate for at least two reasons. First of all, the
variables used to describe a banks’ business model are potentially endogenous. Second,
differencing the regression equation to eliminate the bank-specifc effects could lead to
a correlation between the lagged dependent variable and the error term. The system
GMM estimation procedure resolves these problems by instrumenting the
predetermined and endogenous variables with their own lags. As the estimates
produced are biased in the presence of too many instruments, we instrument the lagged
endogenous variable with its frst lags and the bank-specifc variables with their second
lag, as remote lags are unlikely to be informative instruments (Bond and Megshir, 1994).
Because lagged levels provide only weak instruments for frst differences when the time
series are persistent, the system GMM is used instead of the Arellano Bond (AR) GMM
estimator, also known as the differenced GMM (Blundell and Bond, 1998). The model is
estimated with two-step system GMM, as proposed by Arellano and Bover (1995) and
Blundell and Bond (1998) with Windmeijer’s (2005) fnite sample correction. This
estimation technique is particularly suitable for small T and large N samples, as it
applies to this study.
The validity of the instruments is tested using the Sargan test for over-identifying
restrictions. In all cases, the test statistic accepts the null hypothesis that the
instruments are indeed exogenous. We further use the AR test to control for serial
correlation in the residuals. The null hypothesis is not rejected in all cases, indicating
that there is no second-order autocorrelation in the frst difference regression. All test
results are reported at the bottom of the results in each regression in Table II.
6.1 Diagnostic tests
6.1.1 Endogeneity test. The study used the Durbin–Wu–Hausman (DWH) test to verify
the presence of endogeneity among some of the variables. This is an augmented
regression test which is applied to the residuals of each of the endogenous explanatory
variables as a function of all other exogenous variables (Yokoyama and Alemu, 2009). If
the null hypothesis holds, then basic regression techniques such as the pooled ordinary
least squares (OLS), and the panel fxed and randomeffects would be appropriate in the
sense that the variables are not correlated with the error term. Otherwise, the rejection of
the null hypothesis indicates that the variables are endogenous; hence, OLS estimators
would be inconsistent. The signifcant results of the DWH test provided in the table
below indicates that variables such as credit supply (Cr), net interest rate ratio (NIM),
NPLs, net minimum capital ratio (MCR) and the PR are endogenous in the model, while
JFEP
7,4
412
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
1
:
5
3

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
GDPGand fees are not. Again, as indicated earlier, the lagged values of the endogenous
variables are used as instruments, and the appropriateness of instruments are tested
using the AR test for autocorrelation and the Sargan test for over-identifying
restrictions.
6.2 Fixed versus random effects
The DWH test was performed to determine which of fxed or random effects models ft
the data for estimation. If the test favours fxed effects, it implies the existence of
heterogeneity across banks and therefore gives the indication that the unobserved
bank-specifc effects indeed vary across banks. The DWH test results in Table I reject
the null hypothesis of randomeffects for all the three regression models, indicating that
the system GMM estimates are therefore consistent, as the system GMM procedure
requires that the data fts the fxed effects model to yield consistent results. The DWH
test results reject the null hypothesis of random effects for all the three regression
models, indicating that the system GMM estimates are therefore consistent.
6.3 Estimation results and discussions
The study applies the system GMM estimation approach, an approach that ensures
unbiased and consistent estimates of regression parameters in the presence of
endogeneity and dynamic panel bias. The model controls for specifc bank level,
industry and country characteristics. Three regression equations are estimated for each
one of the dependent variables: interest rate margin (NIM), supply of credit (Cr) and
NPLs. The results for all the three models are shown in Table II and subsequently
discussed as follows:
6.4 NIM as dependent variable
The study results with the NIM (or interest spread), which proxy banking effciency, as
a dependent variable are generally consistent with behavioural expectations of the
independent variables. The results, however, establish a positive relationship between
MCRand the NIM, albeit at the 10 per cent signifcance level. Although this is in contrast
with the study expectations, the result suggests that a high net minimum capital
requirement could widen the spread between the lending rate and the saving rates. A
Table I.
Durbin–Wu–
Hausman (DWH) test
for endogeneity
Variables Test for Residuals ?
2
(1) statistic p-value (p ??
2
)
Credit growth (Cr) res_cr ?0 12.02 0.001
NIM res_nim ?0 4.31 0.038
NPL res_npl ?0 4.27 0.039
MCR res_MCR ?0 2.79 0.095
GDPG res_gdpg ?0 0.23 0.634
TBILL res_tbill ?0 2.07 0.150
Bank fees res_fees ?1 0.89 0.345
PR res_pr ?2 3.91 0.048
Notes: The residuals of the variables are obtained after regressing the variables on their instruments,
including the exogenous variables. The residuals are then tested for signifcance
Source: Authors’ estimation
413
Regulatory
capital
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
1
:
5
3

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
plausible reason could be that as the cost of raising capital is high the country, banks are
compelled to charge high lending rates to make high enough returns for their common
equity holders, thereby widening the NIM. This is consistent with the fndings by Elliott
(2009) who fnds that higher minimum capital requirement will lead to high equity or
total funding costs which would then be passed on to borrowers in whole or part.
The result for the BOG PR is counter-intuitive but is in line with the experience in
Ghana. Under normal circumstances, the relationship between PR and NIM is positive,
as banks would be expected to respond to PR adjustments by adjusting their own
lending rates. However, the study results establish a negative relationship between PR
and NIM. This could be due to a number of factors. Firstly, NIM is defned as a ratio of
the difference between interest income (interest on lending) and interest expenditure
(interest on deposits) to total earning assets. What this means is that even if PR is
reduced by the monetary authority and the banks follow with a lower lending rate,
banks could still make proft by making more loans, which would in turn earn them
more interest income, particularly if the interbank interest rates do not change with the
PR, as has been the case in many instances in Ghana. Alternatively, banks could cut
interest expenditure, leading to much higher net interest incomes than used to be the
case. Furthermore, banks could reduce lending rates but also reduce savings rates even
more to create a higher interest rate spread and net interest incomes, all other things
being equal. Thus, the reduced PR would have been matched by a positive NIM.
Secondly, the negative relationship between PR and NIM could be justifed by the fact
that Ghanaian banks respond to PR changes with a lag at best, especially if it is a
downward review. There have been many instances where PR was reviewed
downwards and banks have been reluctant to follow by reducing their lending rates.
Table II.
System GMM
regression results
Variables NIM Cr NPL
MCR 0.336* (2.15) 6.136** (2.31) 2.319* (1.88)
PR ?0.004* (?1.86) ?0.170*** (?5.15) 0.063** (2.30)
GDPG 0.003 (1.69) 0.065** (2.88) ?0.030* (?1.89)
PBD 0.094* (2.20) ?0.628 (?1.48) ?0.787 (?1.58)
FBD 0.061 (1.07) ?0.627*** (5.69) ?1.367** (?3.04)
Bank fees ?0.043 (?1.02) 1.411*** (8.14)
Credit supply (Cr) ?0.044 (?0.30)
NPLs 0.030 (1.31)
NIM ?17.295*** (?3.98)
Constant 0.18 (1.01) 4.506*** (5.81) ?3.198* (?2.09)
Number of banks 11 11 11
Number of observations 78 78 78
F-test (p-value) 0.000 0.000 0.000
AR (2) test (p-value) 0.504 0.166 0.693
Sargan test (p-value) 0.868 0.186 0.264
Hausman test (p-value) 0.062 0.012 0.046
Notes: The dependent variables in the three regression models are NIM, Credit Supply (Cr) and NPL,
respectively. Figures in parentheses are t-statistics, and *, ** and *** indicates statistical
signifcance at the 10, 5 and 1% levels, respectively
Source: Authors’ estimation
JFEP
7,4
414
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
1
:
5
3

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
There have even been a few cases where banks have increased lending rates in the face
of downward PR reviews. This fnding corroborates the earlier observation from the
overview of developments in the banking system, which suggests that bank lending
rates are unresponsive or weakly responsive to the BOG PR.
The results also established a positive relationship between the public banks dummy
and NIM. This suggests that public banks are associated more with a higher NIM and
thus less effcient than private banks. Thus, public banks are more likely to charge
higher lending rates on loans while offering lower deposit rates to customers, resulting
in high NIMs.
6.5 Credit as dependent variable
The estimation with the growth of credit as the dependent variable has most of the
coeffcients being signifcant, as shown in Table II. In concord with theoretical
expectations, MCRhas a positive relationship with Cr. In general, as banks create excess
capital over the minimum capital, they are able to take on additional risks by mainly
advancing more credits to businesses and households. The fndings further suggest that
a higher PR appears to have a depressing effect on the supply of credit to the private
sector, as the previous result also suggests. Table II indicates that an increase in the
BOG PR reduces credit advancement to the private sector. This is because PR is a cost
to banks, and as it increases, banks will be averse to borrow from the central bank or
among themselves, leading to lowcredit supply. The result is consistent with theory, as
a high PR could mean a high lending rate and less credit due to the fact that the cost of
borrowing would have increased.
Growth in GDP positively impacts Cr, as the borrowing public would feel at ease to
contract loans when they are assured of high income fows both now and in the future.
The study fnds a negative relationship between the foreign banks dummy on Cr. This
means that the growth of credit by foreign banks is lower than that of their domestic
counterparts. This could be attributed to the cautious approach which they attach to
credit supply. The NIM has a negative effect on Cr. This could mean that banks are
unable to make more loans, even though they might want to, due to the high borrowing
cost to customers.
6.6 NPLs as dependent variable
The results with NPLas dependent variable, as shown in Table II, are mixed. In line with
expectations, the BOG PR shows a positive and statistically signifcant effect on NPLs.
This means that as the PRrises, lending rates generally increase which induces banks to
lend more at an increased cost to borrowers, which subsequently increases the chances
of loans going bad. The coeffcient for MCR is positive in relation to NPL, suggesting
that banks create more loans when they have excess capital over the requirement,
leading to high loan impairment. Although the recent increases in banks’ NPL ratios
have been partly attributed to government’s inability to pay contractors on time, the
banks might have been pushed to give “toxic loans” (loans which have high risk of
default) or sub-optimal credit decisions made by banks, as they have too much capital on
their books as a result of minimum capital upward adjustments. Elliott (2009) argues
that bank managers in such situations are likely to take bad risks in an attempt to keep
profts up in the face of the cost pressures as previously mentioned.
415
Regulatory
capital
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
1
:
5
3

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
The same observation holds for bank fees, which suggests that the NPL level of
banks increases when the fees charged by banks on services in the intermediation
process are higher. It is expected that a rise in fees would raise the incentive for banks to
lend more to clients with the hope of increasing income, all other things being equal.
Such practice could lead to very little due diligence being done thereby, culminating in
loan impairment. GDPGhas a negative effect on NPLs. This could be the case, as higher
income levels in the country would enable borrowers to meet their loan commitments to
the banks. The FBDshows a negative and statistical signifcant effect on NPL, implying
that the effect of the other explanatory variables on NPL is lower for foreign-owned
banks than for domestic-owned banks.
7. Concluding remarks
In this paper, we investigate the infuence of regulatory capital and the central bank
policy instruments on bank-specifc outcomes such as credit supply, interest rate spread
(as a measure of effciency) and NPLs (as a measure of risk-taking behaviour of the
commercial banks). We model a system of equations that allows us to apply the system
GMM approach and estimate the equations, while controlling for specifc bank level,
industry and macroeconomic variables.
We fnd a positive relationship between MCR and the NIM. Although this is in
contrast with the study expectations, the result suggests that a high net minimum
capital requirement would widen the spread between the lending rate and the saving
rates. A plausible reason could be that as the cost of raising capital is high the country,
banks are compelled to charge high lending rates to make high enough returns for their
common equity holders. Furthermore, increased excess capital over the required levels
may not have the expected impact on NIM if the intention is towards meeting higher
capital requirements. Banks would, thus, discourage lending by increasing the cost of
borrowing to the client. The study also fnds a negative relationship between interest
rate spread (proxy by net interest income) and the BOG PR. This negative relationship
could largely be justifed by the fact that Ghanaian banks respond to the BOG policy
changes with a lag at best, especially if it is a downward review. We also fnd that the
effect of factors affecting the NIMin the model is more pronounced for public banks than
private banks.
Further, increasing BOG, PR appears to have a depressing effect on supply of credit
to the private sector. This is because as the PR rises the interest rate on government
securities such as the T-bill rates often rises faster than the lending rates and being less
risky, these government securities are preferred.
We fnd evidence to support the fact that high minimum capital requirement and
excess capital above the minimum required drive higher Cr in the banking sector of
Ghana. However, high excess capital increases risk-taking activities of the banks, as
excess capital is found to be associated with high NPL ratios. The positive coeffcient
suggests that banks create more loans when they have excess capital over the
requirement, leading to high loan impairment. As banks minimum capital is raised as a
buffer against risk, they may be able to fnd avenues to take on additional risks by
mainly advancing more credits to businesses and households. However, because of the
equity cost pressures, they are tempted to make more profts by giving out bad loans or
bad credit decisions.
JFEP
7,4
416
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
1
:
5
3

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
Once one accepts that there will be signifcant economic costs to sharply higher
capital requirements, then a useful debate can take place about the right level of capital,
given the trade-offs and howbest to achieve it. In fact, this is the debate that much of the
policymaking and academic community has been involved in for some years.
In sum, our results speak to the ongoing debates on the right level of capital,
effectiveness of the BOG PR and the high lending rates that appear to respond only
slowly to macroeconomic indicators such as the PR and the infation rate. More
specifcally, the fndings raise issues about excessive minimum capital regulations/
requirements, high cost of borrowing and high NPLs in the economy. While risk-based
minimum capital requirement improves stability and is associated with increased
lending, the evidence suggests that if it is too high or if banks keep capital high in excess
of the minimum required, it can increase the cost of borrowing as the cost of acquiring
this capital is high. Besides, high and strict capital requirement can increase risk-taking
incentives of banks and increase the NPL ratios, as the evidence above suggests. This
fnding has practical implications for the adoption of the Basel III accord, as some
studies (Slovik and Cournède, 2011), studying the macroeconomic impacts of the Basel
III, have found a negative impact on output growth. According to the study, economic
output would be mainly affected by an increase in bank lending spreads, as banks pass
a rise in bank funding costs, due to higher capital requirements, to their customers[4].
Another plausible explanation is also that although increase in NPL ratios in recent
times has been partly attributed to government inability to settle contractors in time and
the diffcult economic conditions, banks are being pushed to give “toxic loans” (loans
which has a high risk of default) or not-so-good credit decisions in recent times because
of excessive capital in their books. The industry watchers believed that there is a
temptation of bad lending, as banks have capital on their books in volumes
unprecedented in Ghana’s economic history, for which they need to fnd proftable
business to generate returns for providers of capital without enough risk mitigation
tools (Ghana Banking Survey Report, 2008) – thus leading to a further deterioration of
asset quality and increase in non-performing ratios.
Notes
1. The First Basel accord (Basel 1 accord) was published in 1988 and was revised to the Basel II
in 2004 to overcome some of the problems associated with the Basel I accord, which required
that all corporate debts having 100 percent risk be changed to a regulatory capital based on
credit rates. The Bank of Ghana, which is the primary regulator of banks in Ghana, has
currently adopted the Basel II accord in determining capital requirement. Basel III, which
implementation spans from 2013 to 2019, was supposed to strengthen bank capital
requirements by increasing bank liquidity and decreasing bank leverage.
2. Unlike Basel I and Basel II, which focus primarily on the level of bank loss reserves that banks
are required to hold, Basel III focuses primarily on the risk of a run on the bank by requiring
differing levels of reserves for different forms of bank deposits and other borrowings.
3. At the same time, the number of banks and branches had increased substantially largely due
to the infux of foreign banks. For example, as of 2000, there were only 17 banks with about
300 branches, but by the end of 2013, the number of banks had increased to 27 with 904
branches. Out of this, 15 were foreign owned (or majority shareholders) and 12 had local
ownership (with government having majority shares in only three of these banks).
417
Regulatory
capital
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
1
:
5
3

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
4. To meet the capital requirements under Basel III originally effective in 2015, banks were
estimated to increase their lending spreads on average by about 15 basis points. Capital
requirements effective as of 2019 (7 per cent for the common equity ratio, 8.5 per cent for the
Tier 1 capital ratio) could increase bank lending spreads by about 50 basis points.
References
A World Bank. (1995), World Development Report 1989: Financial Systems and Development,
Oxford University Press, New York.
Aboagye, A.Q.Q., Akoena, S.K., Antwi-Asare, T.O. and Gockel, A.F. (2008), “Explaining interest
rate spreads in ghana”, African Development Bank Journal, Vol. 20 No. 3, pp. 378-399.
Abor, J. (2008), “Determinants of the capital structure of Ghanaian frms”, African Economic
Research Consortium, Research Paper RP_176, Nairobi, Kenya.
Amidu, M. (2006), “Credit risk, capital structure and lending decisions of banks in Ghana”, Banks
and Bank Systems, Vol. 1 No. 1, pp. 93-101.
Amidu, M. (2007), “Determinants of capital structure of banks in ghana: an empirical approach”,
Baltic Journal of Management, Vol. 2 No. 1, pp. 67-79.
Altunbas, Y., Carbo, S., Gardener, P.M.E and Molyneux, P. (2007), “Examining the relationships
between capital, risk and effciency in European banking”, European Financial
Management, Vol. 13 No. 1, pp. 49-70.
Arellano, M. and Bover, S. (1995), “Another look at the instrumental variable estimation of
error-components models”, Journal of Econometrics, Vol. 68, pp. 29-51.
Bank of Ghana (BOG) (2013), “2013 BOG annual report”, available at: www.bog.gov.gh/index.php?
option?com_content&view?article&id?1734%3Aannual-report-2013&catid?102%.
Barnor, C. and Odonkor, T.A. (2012), “Capital adequacy and the performance of ghanaian banks”,
Journal of Business Research, Vol. 6 Nos 1/2.
Barth, J.R., Caprio, G. and Levine, R. (2004), “Bank regulation and supervision: what works best?”,
Journal of Financial Intermediation, Vol. 13, pp. 205-248.
Basel Committee on Banking Supervision. (2006), “International convergence of capital
measurement and capital standards: a revised framework”, Bank for International
Settlements, June.
Bawumia, M., Belnye, F. and Ofori, M. (2005), “The determination of bank interest spreads in
ghana: an empirical analysis of panel data”, Bank of Ghana Working Paper Series, WP/
BOG-2005/09.
Berger, A.N. and Bouwman, C.H.S. (2011), “How does capital affect bank performance during
fnancial crises?”, Journal of Financial Economics, available at SSRN:http://ssrn.com/
abstract?1739089 orhttp://dx.doi.org/10.2139/ssrn.1739089.
Berger, A.N., Bouwman, C.H.S., Kick, T.K. and Schaeck, K. (2014), “Bank Risk taking and
Liquidity creation following regulatory interventions and capital support”, available at:http://ssrn.com/abstract?1908102
Berger, A.N. and De Young, R. (1995), “Problem loans and cost effciency in commercial banks,”
Working Paper, Offce of the Comptroller of the Currency.
Berger, A.N., Herring, R.J. and Szego, P.G. (1995), “The role of capital in fnancial institutions”,
Journal of Banking and Finance, Vol. 19 Nos 3/4, pp. 393-430.
Blundell, R. and Bond, S. (1998), “Initial conditions and moment restrictions in dynamic panel data
models”, Journal of Econometrics, Vol. 87, pp. 115-143.
JFEP
7,4
418
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
1
:
5
3

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
Bond, S. and Megshir, C. (1994), “Dynamic investment models and the frm’s fnancial policy”,
Review of Economic Studies, Vol. 61, pp. 197-222.
Boudriga, A., Boulila, N. and Jellouli, S. (2009), “Does bank supervision impact non-performing
loans: cross-country determinants using aggregate data?”, MPRA Paper No. 18068,
available at:http://mpra.ub.uni-muenchen.de/18068/
Dugan, J. and Jennifer, XI. (2011), US Implementation of Basel II: Final Rules Issued by no
Supervisory Approvals to Date, Covington & Burling LLP, Washington, DC.
Elliott, D. (2009), Quantifying the Effects of Lending Increased Capital Requirements, The
Brookings Institution, Washington, September, available at: www.brookings.edu/papers/
2009/0924_capital_elliott.aspx
Fare, R., Grosskopfz, S. and Weber, W. (2004), “The effect of risk-based capital requirements on
proft effciency in banking”, Applied Economics, Vol. 36, pp. 1731-1743.
Greuning, V.H. and Bratanovic, B.S. (2000), Analyzing Banking Risk: A Framework for Assessing
Corporate Governance and Financial Risk Management, The World Bank, Washington,
DC.
Griffth-Jones, S. and Persaud, A. (2008), “The pro-cyclical impact of basel ii on emerging markets
and its political economy”, Capital Market Liberalization and Development, Vol. 27,
pp. 262-288.
Jokivuolle, E., Kiema, I. and Vesala, T. (2007), “Portfolio effects and effciency of lending under
Basel II”, Bank of Finland Research Discussion Papers 13.
Koehn, M. and Santomero, A.M. (1980), “Regulation of bank capital and portfolio risk”, Journal of
Finance, Vol. 35, pp. 1235-1244.
Kopecky, K.J. and VanHoose, D. (2006), “Capital regulation, heterogeneous monitoring costs, and
aggregate loan activity”, Journal of Banking and Finance, Vol. 30, pp. 2235-2255.
Kwan, S. and Eisenbeis, R.A. (1997), “Bank risk, capitalization, and operating effciency”, Journal
of Financial Services Research, Vol. 12 Nos 2/3, pp. 117-131.
Osei-Assibey, E. and Baimba, A. (2013), “Bank risks, capital and loan supply: evidence fromsierra
leone”, Journal of Financial Economic Policy, Vol. 5 No. 3.
Osei-Assibey, E., Bokpin, A.G. and Twerefou, D.K. (2012), “Microenterprise fnancing preference:
testing POH within the context of Ghana’s rural fnancial market”, Journal of Economic
Studies, Vol. 39 No. 1, pp. 84-105.
Pasiouras, F., Tanna, T. and Zopounidis, C. (2009), “The impact of banking regulations on banks’
cost and proft effciency: cross-country evidence”, International Review of Financial
Analysis, Vol. 18, pp. 294-302.
Pennacchi, G.G. (2005), “Risk-based capital standards, deposit insurance, and procyclicality”,
Journal of Financial Intermediation, Vol. 14, pp. 423-465.
PWC, Ghana Banking Survey Report (2008), Raising the Bar: Increase in the Minimum Capital
Requirements and Implications on The Industry, Pricewaterhouse Coopers and Ghana
Association of Bankers, Ghana, available at: www.pwc.com/en_GH/gh/pdf/ghana-
banking-survey-2008.pdf
Repullo, R. and J. Suarez (2008), “The procyclical effects of Basel II”, CEPR Discussion Paper
No. 6862.
Slovik, P. and Cournède, B. (2011), “Macroeconomic impact of Basel III”, available at: www.oecd-
ilibrary.org/economics/oecd-economics-department-working-papers_18151973;
jsessionid?24qndegq4q31x.x-oecd-live-03 OECD Economics Department Working Papers
No.: 844, Pages 16, February.
419
Regulatory
capital
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
1
:
5
3

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
Stiglitz, E.J. and Weiss, A. (1981), “Credit rationing in markets with imperfect information”, The
American Economic Review, Vol. 71 No. 3, pp. 393-410.
Windmeijer, F. (2005), “A fnite sample correction for the variance of linear effcient two-step
GMM estimators”, Journal of Econometrics, Vol. 126, pp. 25-51.
Yokoyama, K. and Alemu, M.A. (2009), “The impacts of vertical and horizontal export diversifcation
on growth: an empirical study on factors explaining the gap between Sub- Sahara Africa and
East Asia’s Performances”, Ritsumeikan International Affair, Vol. 17 No. 41.
Further reading
Arellano, M. and Bond, S. (1991), “Some tests of specifcation for panel data: monte carloevidence
and an application to employment equations”, Review of Economic Studies, Vol. 58,
pp. 277-297.
Bank of Ghana (B.O.G) (2009), “Annual report”, available at: www.bog.gov.gh/privatecontent/
Publications/Annual_Reports/2009.pdf
Brownbridge, M. and Gockel, S.A.F. (1996), The Impact of Financial Sector Policies on Banking in
Ghana, Research Department of Bank of Ghana, Ghana.
Modigliani, F. and M. Miller (1958), “The cost of capital, corporation fnance, and the theory of
investment”, American Economic Review, Vol. 48 No. 3, pp. 261-297.
Osei-Assibey, E. and Baah-Boateng, W. (2012), “Interest rate deregulation and private investment:
revisiting the mckinnon – shaw hypothesis in Ghana”, The IUP Journal of Applied
Economics, Vol. 11 No. 2, pp. 12-30.
Santomero, A. and Watson, R. (1977), “Deter mining an optimal capital standard for the banking
industry”, Journal of Finance, Vol. 32, pp. 1267-1282.
Schliephake, E. and Kirstein, R. (2010), “Strategic effects of regulatory capital requirements in
imperfect competition”, FEMM Working Paper No. 12, available at: www.ww.uni-
magdeburg.de.
VanHoose, D. (2007), “Theories of bank behavior under capital regulation”, Journal of Banking
and Finance, Vol. 31, pp. 3680-3697.
Wooldridge, J.M. (2003), Introductory Econometrics: A Modern Approach, 2nd ed., Thomson
Learning, South-Western.
Zellner, A. (1962), “An effcient method of estimating seemingly unrelated regression and tests for
aggregation bias”, Journal of the American Statistical Association, Vol. 57 No. 298,
pp. 348-368.
Corresponding author
Eric Osei-Assibey can be contacte at: [email protected]
For instructions on how to order reprints of this article, please visit our website:
www.emeraldgrouppublishing.com/licensing/reprints.htm
Or contact us for further details: [email protected]
JFEP
7,4
420
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
1
:
5
3

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)

doc_631075057.pdf