Risk management andBasel-Accord-implementation

in PakistanOmar Masood

Royal Docks Business School, University of East London, London, UK, and

John FryDepartment of Physics and Mathematics, Nottingham Trent University,

Nottingham, UK

Abstract

Purpose – Recent events demonstrate that problems in the banking system pose a significant threatto the health of the global economy. Despite several shortcomings the Basel Accord thus emerges as anattempt to protect banking systems. The purpose of this study is to shed light on potential barriers toimplementation of the Basel Accord in emerging countries. Several issues of wider interest to riskmanagement and financial regulation also emerge.

Design/methodology/approach – The paper maps implementation of the Basel Accord against thewider regulatory context. Against this backdrop, the Basel Accord appears well-motivated but islimited by several practical considerations. These factors, amidst other practical implications, areidentified as the paper applies rigorous statistical methods to novel primary survey data from riskmanagers.

Findings – The Basel Accord is generally well-received due its dual aims of improved capitaladministration and scientific risk management. Operational risk is a significant barrier toimplementation, with a number of further issues only partially addressed (see below). Equallysupported by both public and private sector banks the reasons for delay appear due to lack of technicalexpertise and the level of preparation. Results highlight credit risk, practical implementation issues (ITand HR), minimal capital requirements, data security and operational risk as issues of criticalimportance.

Originality/value – The originality of the contribution lies in the scientific treatment of novelprimary data from risk managers tasked with implementation of the Basel Accord. Findings suggestseveral important practical implications discussed above.

KeywordsRisk management, Basel Accord, Banking, Financial regulation, Emerging markets, Pakistan

Paper type Research paper

1. IntroductionAs long ago as the 1970s financial crises have prompted the Basel Committee onBanking Supervision (BCBS) to work towards global financial stability, most notablyvia the Basel Accord. The Basel Accord has wide support and should improve financialstability via its risk-sensitive methodologies (BIS, 1999; BCBS, 2004; Cumming and Nel,2005; van Rixtel et al., 2004; Jacobsohn, 2004) and the scientific alignment of risk andcapital. Such standards benefit the economy by protecting banks against risk and aimto guarantee the availability of capital throughout business cycles (BIS, 2004; HassanAl-Tamimi, 2008), guarding against systemic risk (Amidu, 2007). By rationalizingbanks’ risk appetites the Basel Accord can also, in principle, manage developmentspertaining to financial instruments and technologies (Mboweni, 2004).

The current issue and full text archive of this journal is available at

www.emeraldinsight.com/1358-1988.htm

Riskmanagement

293

Journal of Financial Regulation andCompliance

Vol. 20 No. 3, 2012pp. 293-306

q Emerald Group Publishing Limited1358-1988

DOI 10.1108/13581981211237981

In 1988 the first Accord, Basel I, focused upon credit risk reasoning that appropriatecapital levels would reduce systemic risk, with individual financial institutions betterprotected from credit and other losses. Basel I was accepted by over 100 countries andEngelen (2005) notes its general acceptance by most banks globally. Primarilydesigned for G10 countries the Accord was structured so it could be applied acrossboth developed and developing countries (Mboweni, 2004). In 2004 Basel II revised andextended Basel I. However, Basel II was promulgated during the onset of the globalfinancial crisis and relatively short-lived. Responding to several key failings, are-modification, Basel III, was initiated in 2010.

Wholesale acceptance of Basel I has helped to improve the resilience of theinternational banking system through improved capital standards (Rime, 2001;Cumming and Nel, 2005). However, despite providing some stability to the financialsector implementation of the Basel Accord does raise some concerns. Ong (2004)criticizes the Accord for a “one size fits all” approach to risk management, beinginsensitive to the distinction between credit and other risks (Hai et al., 2007). The Accordwas also over-shadowed by the increased globalization and integration of financialmarkets. Regulations also led to banks “cherry picking” certain practices, providingleverage as banks created high-risk portfolios of assets within particular risk categories(Cumming and Nel, 2005). Further barriers to international implementation of theAccord are discussed in Section 3.

The above sets up the following research questions for this paper. In the light ofseveral short-comings it is interesting to assess how risk managers view implementationof the Basel Accord, the technical obstacles still to be overcome and the widerimplications for financial regulation and risk management that then come to light.Second, the results of the study are significant due to its timing (after Basel II andfollowing the recent financial crisis and the commencement of Basel III). Third, againstthe backdrop of a varied international experience it is interesting to studyimplementation of the Basel Accord in Pakistan. Previous studies have examinedBasel Accord implementation in several different countries ( Jacobsohn, 2004; Cummingand Nel, 2005). However, no such study has yet been conducted in Pakistan.

In Pakistan banks are regulated by the Banking Supervision Department of theState Bank of Pakistan (SBP). The growth of international banking and the entrance ofmultinational banks have made Pakistan’s Banking Sector more diverse and makesBasel Accord implementation increasingly significant. The SBP has set up a timetablefor Basel Accord implementation in-line with BCBS guidelines. However,implementation of the Accord in Pakistan is over two years behind schedule andlags behind that of other countries (see Section 3).

This paper is laid out as follows. Section 2 reviews the Basel Accord in its variousincarnations. Section 3 discusses international implementation of the Basel Accord.Section 4 discusses the survey methodology. Section 5 introduces the statisticalmethodology used and gives the empirical results. Section 6 concludes.

2. The Basel Accord: an overviewThe first Accord, Basel I, aligned regulatory capital against credit risk. Basel IIextended the scope of Basel I by distinguishing between credit, operational and marketrisks placed under the auspices of Pillar I: risk management. Two further pillars wereintroduced; supervisory review and market discipline.

JFRC20,3

294

Basel II lays out risk-management techniques, of varying complexity. Credit riskcan be measured using the Standardized Approach (SA) or the Internal Ratings Based(IRB) Approach. IRB approaches consist of the Foundation IRB (FIRB) and theAdvanced IRB (AIRB) approaches (BCBS, 2006). Methods proposed to measureoperational risk are the basic, standardized and internal measurement approaches.Market risk and evaluation of trading-book positions can be conducted using theStandardized Measurement Method or the Internal Models Approach.

The second pillar, supervisory review, aims to ensure that banks’ internal riskassessment procedures are sufficiently resourced (BCBS, 2001, 2006). The Accordstipulates that banks should develop their own assessment procedures and that thecalculation of capital targets is continuously updated (BCBS, 2001). Supervisors werealso given the authority to intervene by revising and upgrading capital requirementsand other risk-management procedures and processes as appropriate (BCBS, 2001). Thethird pillar, market discipline, seeks to ensure full disclosure of banks’ capital adequacyvia public reports (BCBS, 2006). By monitoring bank activities, and their ability tomanage risk, market participants can reward banks that do so effectively, penalizingthose that do not (Makwirmiti, 2008; BCBS, 2001).

The financial crisis highlighted several key failings in Basel II. These includeover-estimates of common equity, a failure to distinguish between simple and complexpositions in trading-book exposures, leverage, macro-prudential stability (banks’impact upon the wider financial system) and systemic risk. These failings have begunto be addressed in a third incarnation – Basel III (Folpmers, 2010). Modifications madeinclude more conservative micro-prudential and macro-prudential capital rules andadditional liquidity requirements to counter banks building up excessive leverage.

3. Implementation of the Basel Accord: international comparisonsThe perceived wisdom is that the Basel Accord should improve financial stability viathe risk-management techniques it contains. Primarily intended for G10 countries, theAccord was devised so it could be adopted by both developed and developing countries(Mboweni, 2004). However, some discrepancies in the Accord’s internationalimplementation are inevitable as banks in developing and emerging economies haveto work harder to meet international practices and procedures (Dupuis, 2006) and applythe Accord’s more advanced risk-management methodologies (see Section 2).

Several factors limit international implementation of the Accord. Crucially, Basel IIfailed to account for the effects of an international financial crisis – leading directly tothe need for Basel III. There are procyclicality issues (IMF, 2005), which may aggravateinstabilities in the banking system and may have more pronounced effects upondeveloping countries (Griffith-Jones and Spratt, 2001; Reisen, 2001; Ward, 2002).In economic downturns, increased risks lead to increased capital charges. Increases inthe cost of capital may then reduce the level of bank lending. Shocks may thus passfrom the financial sector to the real economy (Jacobsohn, 2004; Heid, 2003).

Successful implementation of the Accord depends critically on the resources offinancial regulators worldwide (Cornford, 2006). For instance, discrepancies betweeninternational and local financial standards may occur if the home country of a bankrequires higher IRB approaches than the host country prescribes due to limitations insupervisory capacity. Further, rating agencies can manipulate Credit Risk

Riskmanagement

295

Management in developing countries (IMF, 2005). There is thus a need for supervisorsin developing countries to monitor rating agencies.

As discussed below the international experience varies greatly. Prior toimplementing the Accord the South African Government worked for around 11 yearsto ensure that several fiscal and macroeconomic preconditions were met. Basel II wasfinally implemented across the whole banking sector from January 2008 (Neville, 2005).In contrast, India’s transition to the Basel Accord is very slow (Mehra, 2010). ByDecember 2010 Indian banks were still following initial approaches, with a ReserveBank of India notification stating that more advanced approaches are to be implementedfrom 2013.

In Switzerland, although all major banks started to apply approaches by 2006,typically only large international banks had the resources to apply methodologies likeIRB (Swiss Federal Banking Commission, 2005). Similar to South Africa, Switzerlandintegrated all three pillars into its regulatory system and adopted all the approachesavailable in the Basel II framework. No alterations were made to existingrisk-management methodologies (Swiss Federal Banking Commission, 2005),although the Swiss implementation provides two additional ways of calculatingcredit risk (separate methodologies for domestic and international banks). The initialapproaches were introduced by January 2007 with more advanced methodologiescoming into force by January 2008. This two-step process allowed enough time for asmooth transition (Makwirmiti, 2008). In Brazil, the first implementations of the BaselAccord were due to commence in 2004, later postponed to 2007. The process ofimplementation is expected to be completed in 2013, with the Central Bank of Brazilhaving put in place a detailed timetable for full implementation. Similarly, Basel Accordimplementation in Jordon is not at an advanced stage. Jordanian banks have met anumber of preconditions. However, a lack of supervisory authority compromises thewhole process as no mechanism for the external rating of credit exposures exists(Barakat, 2009).

In adopting the Basel Accord the USA followed a two-step process. The USAoutlined specific criteria assigning Basel I for non-core banks and Basel II for corebanks (Federal Reserve Board, 2007). Accordingly, all US banks adopting Basel II areobliged to promulgate advanced approaches for capital allocations (Federal ReserveBoard, 2007; Gordon-Hart, 2004). However, the recent financial crisis disruptedimplementation. Currently, Basel III does not seem to constitute a major issue for USBanks as they seem more cushioned in terms of liquidity, leverage and bufferrequirements (Folpmers, 2010). However, it remains unclear as to how to phase inimplementation of Basel III in-line with international standards.

In Pakistan the SBP has taken several steps towards Basel Accord implementationand in 2005 set the following schedule:

(1) SA for credit risk and basic indicator/SA for operational risk from 1 January2008.

(2) IRB approaches to credit risk from 1 January 2010.

According to annual reports submitted by Pakistan’s banks none have yet adopted IRBapproaches. Thus, Pakistan is already around two years behind schedule. SincePakistan has made less progress than other countries it is of interest to determine thereasons for this delay.

JFRC20,3

296

4. Survey methodologyThe questionnaire, available from the authors by request, was specially adaptedfrom a study into Basel Accord implementation by the Bank of Serbia. It asks aboutthe significance of the Basel Accord (Q1-2), overall attitudes to implementation (Q3-8)and several wider issues. These include questions regarding practical implementationof the Accord (Q10-11, Q17-20, Q31), knowledge and training of staff (Q12-14), riskmanagement (Q15-16), whether or not practices currently comply with the Accord(Q21-22, Q32), further changes necessitated by the Accord (Q23-24), dataprotection/security (Q25-29), minimal capital requirements (Q9, Q30), credit risk(Q34-37), market risk (Q38-40), operational risk (Q41-43), the supervisory review process(Q44-48) and market discipline (Q49-Q52). The survey thus highlights the followingareas of interest:

. Type of bank – are there differences between public and private sector banks?

. Inclination of the bank towards the Basel Accord.

. External training – has the bank provided external training relating to the BaselAccord?

. The level of expertise amongst employees in the Risk Management Department.

. The effectiveness of plans for Basel Accord implementation.

. Changes required in the system to ensure Basel Accord compliance?

. The number of years covered by default time series data – the amount ofavailable data indirectly measures the compliance and inclination of the banktowards the Accord and related standards.

. Compliance with market discipline – the extent of compliance with disclosurerequirements as per International Financial Reporting Standards and Pillar 3.

. Basel customer awareness – are banks raising customer awareness of BaselAccord compliance? Such steps are necessary in the light of Data Sharing, anissue highlighted in the Accord, and other international experiences of thefinancial crisis.

There are around 38 banks operating in Pakistan. We have collected data from five keypersonnel from 12 of these banks. These 12 banks own about 70 per cent of marketshare in both deposit and asset categories. Further, banks in our study have bothdifferent ownership types and capital adequacy ratios (CARs) showing differentpatterns – some increasing, some decreasing. Summary statistics of the sampled banksare shown in Table I. CARs calculated from published accounts are shown in Table II.

To obtain a representative sample banks were surveyed as follows:. two questionnaires were completed by senior executives to gauge the extent of

Basel Accord awareness at higher management levels;. one questionnaire was completed by an executive of the Risk Management

Division; and. two questionnaires were completed by a branch manager at each bank to gauge

the extent of Basel Accord awareness at lower management levels.

Riskmanagement

297

Nam

eof

the

ban

kO

wn

ersh

ipty

pe

Tot

alem

pa

Ris

kd

ivis

ion

emp

aC

red

itri

skem

pb

Tot

alri

skw

eig

hte

das

sets

aT

otal

risk

wei

gh

ted

cred

itas

sets

Tot

ald

epos

itsa

KA

SB

Ban

kL

ocal

1,32

116

1245

.745

42.4

9243

.807

SC

Ban

kF

orei

gn

5,04

214

891

180.

268

136.

300

206.

958

Ask

ari

Ban

kL

ocal

7,27

069

5816

5.59

614

4.79

320

5.97

0A

llie

dB

ank

Loc

al11

,690

351

315

238.

437

193.

031

328.

875

NIB

Ban

kF

orei

gn

6,38

547

3810

1.95

786

.115

93.9

20M

CB

Ban

kL

ocal

9,39

741

038

033

7.41

724

3.71

236

7.60

4N

atio

nal

Ban

kG

over

nm

ent

16,2

4830

828

662

4.88

145

9.75

572

6.46

5B

ank

Alf

alah

For

eig

n7,

462

130

106

213.

840

185.

162

324.

760

Fay

sal

Ban

kF

orei

gn

2,04

211

799

103.

420

83.0

1312

3.65

5U

nit

edB

ank

For

eig

n8,

738

185

163

485.

958

386.

466

503.

832

Bof

Pu

nja

bc

Gov

ern

men

t4,

811

100

8517

2.34

5c15

1.00

0c19

5.07

3H

abib

Ban

kL

ocal

13,1

2252

449

058

6.89

447

1.90

265

3.45

2

Sources:

aA

nn

ual

fin

anci

alst

atem

ents

are

avai

lab

leon

the

web

site

ofS

tate

Ban

kof

Pak

ista

n;

bav

erag

eof

resp

onse

ofal

lsu

rvey

resp

ond

ents

inea

chb

ank

;c Ban

kof

Pu

nja

bh

asn

otb

een

issu

ing

its

ann

ual

rep

orts

for

the

last

two

yea

rsd

ue

toca

pit

alad

equ

acy

pro

ble

ms

(wit

hth

ep

erm

issi

onof

Sta

teB

ank

ofP

akis

tan

);th

efi

gu

res

list

edar

eex

pec

ted

fig

ure

sas

obta

ined

from

surv

eyre

spon

den

ts

Table I.Summary statistics ofbanks in the sample

JFRC20,3

298

The survey data results in a complicated data set to analyse. This necessitates astrategic approach and the use of relatively advanced statistical methods (see Section 5).The over-rising research questions of this paper are the attitude of managers towardsimplementation, the technical obstacles that remain to be overcome and the widerimplications for financial regulation and risk management.

In terms of gauging attitudes towards implementation we test the followinghypothesis:

H1. Pakistani banks are positively inclined towards Basel Accordimplementation.

If H1 is accepted this leads naturally to a second hypothesis:

H2. Support for the Basel Accord is reflected by a number of keys steps Pakistanibanks have taken, e.g. upgrading of employee skills and technology.

H1 and H2 suggest that implementation may depend upon several key technicalissues. On a related theme we might expect to find important differences betweenpublic sector and private sector banks. For instance, private sector banks may be moreinternational in their outlook, have better technical infrastructure and overall viewimplementation more positively. Hence we test a third hypothesis:

H3. Progress on Basel Accord implementation is uniform across the bankingsector and there are no significant differences between publicly and privatelyowned banks.

5. Statistical analysis of the survey data5.1 H1We analyse the data by producing a score measuring respondents’ attitude (positive ornegative) towards implementation (Questions 3-8) as follows. If respondents stronglyagree or agree with the proposition they score 1, if they strongly disagree or disagreethey score 21 and if they neither agree nor disagree they score 0. If the proposition hasnegative implications regarding implementation of the Accord (Question 4) this score ismultiplied by21. Finally, an overall score is obtained by adding respondents’ scores foreach of the Questions 3-8. As constructed, high values indicate favourability towardsimplementation, negative values indicate opposition and zero indicates neutrality.

Results show that managers generally support Basel Accord implementation andthe mean score is positive. A one-sample t-test (Snedecor and Cochran, 1989, Chapter 5)gives a t-statistic of 4.502 on 59 degrees of freedom giving a p-value of 0.000. A kerneldensity plot (Venables and Ripley, 2002, Chapter 5) of the scores is shown in Figure 1and corroborates the finding that managers generally support implementation.

Year KASB SCB ACB ABL NIB MCB NIB BALF FBL UBL BOP HBL

2005 9.34 6.45 11.3 12.17 11.61 12.54 11.51 8.66 12.01 9.26 12.78 9.932006 8.95 8.14 12.37 12.8 17.44 18.65 16.5 9.48 11.42 11.1 10.09 12.812007 12.18 9.45 13.2 10.26 3.33 16.73 18.05 9.85 10.27 10.32 9.69 11.62008 9.05 9.99 12.1 10.9 19.52 16.28 16.9 8.03 10.84 9.96 1.92 12.332009 3.53 11.57 13.5 13.47 19.53 19.07 17.23 12.46 12.36 13.18 7.68 13.07

Table II.CARs of banks in the

sample from 2005 to 2009

Riskmanagement

299

Figure 1 shows there is a high probability (clearly greater than 0.5) of obtaining apositive score and the modal or “most likely” score is positive. However, the probabilityof a negative score is significantly greater than zero and may in part reflect concernsraised about implementation discussed above. On a related theme there is no evidencefor different levels of support between publicly and privately owned banks. Anunpaired pooled two-sample t-test (Snedecor and Cochran, 1989, Chapter 6) gives at-value of 21.597 and a p-value of 0.116.

5.2 H2We test H2 by regressing the sentiment scores against explanatory variablespertaining to the survey data (Q9-52) described above.

Model selection is performed using stepwise regression, based on sequential F-tests(Bingham and Fry, 2010, Chapter 6), starting from the baseline model:

yi ¼ b0 þ 1i; ð1Þ

where yi is the score of the ith survey respondent, b0 is the underlying average scoreand 1i is a random error term. In the sequel we fit the model:

yi ¼ b0 þ b1x1;i þ · · · þ bpxp;i þ 1i; ð2Þ

where x1, . . . ,xp are explanatory variables pertaining to the survey data. H2 is acceptedif p . 0 in equation (2). Results obtained for the final model selected are shown inTable III. The corresponding survey questions are shown in Table IV. The R 2 valueshows that the model explains a substantial amount, 82 per cent, of the variability inthe data. Tables III and IV confirm that p . 0 in equation (2) and H2 is accepted.Moreover, the results of the regression model suggest several significant findings.

The regression highlights the following as critical factors affecting financialregulation and risk management: credit risk, practical implementation issues (IT andHR), minimal capital requirements, data security and operational risk. As the level

Figure 1.Kernel density plot ofsurvey scores

Score

Den

sity

0.15

0.10

0.05

0.00

0 5–5

JFRC20,3

300

of technical expertise increases, support for the Accord increases. Results show thathigh scores are associated with local development of credit risk models. Further, lowscores are associated with the standard measurement (simplest) approach to modellingcredit risk. Similarly, results show that as the level of preparedness increasesfavourability towards the Accord increases. Scores increase as the number of years ofavailable data on credit risk (historical default) increases. Higher scores are alsoobtained if a specialist Basel Accord Implementation Department exists and if thecurrent IT structure supports Basel Accord requirements for operational risk.

However, results also suggest that managers who support implementationnonetheless appreciate a number of wider issues that may arise. Higher scores are

Issue Variable EstimateEstimated

standard error t-value p-value

Intercept 26.023 1.294 24.658 0.000 * * *

Credit risk Number of years data onhistorical default (Q36a)

0.870 0.105 8.299 0.000 * * *

Local development of credit riskmodels (Q37)

3.753 1.391 2.698 0.010 * *

Practicalimplementationissues

Basel Accord implementationdepartment set-up (Q17)

1.617 0.506 3.196 0.002 * *

Q11 (agree) 1.807 0.839 2.155 0.036 *

Minimal capitalrequirements

Q9 (neither agree nor disagree) 1.564 0.791 1.978 0.054Q9 (agree) 1.769 0.359 4.931 0.000 * * *

Data security Q27 (agree) 1.340 0.451 2.974 0.005 * *

Operational risk Standardized measurementapproach (Q41)

21.553 0.628 22.473 0.017 *

Unknown measurementapproach (Q41)

3.822 1.372 2.785 0.008 * *

Q30b (neither agree nordisagree)

20.167 0.359 20.467 0.643

Q30b (agree) 1.912 0.647 2.956 0.005 * *

Note: Significant at: *5, * *1 and * * *0.1 per cent levels

Table III.Regression output for the

final model

Q9 Do you think that the Basel Accord will lead to lower capital requirements for some banks?Q11 Do you think information technology, and HR problems might hinder implementation of the

Basel Accord?Q16b How many employees are engaged in the risk management process in your bank? (operational

risk)Q17 Has your bank set up any special Basel Accord implementation department?Q27 Do you think that your bank should plan to resolve data protection issues under the Basel

Accord by taking the consent declaration of its customers?Q30b The current IT structure supports Basel Accord requirements for operational risk?Q36a How many years are covered by the time series that you have created for credit risk assessment

(of default)?Q37 If you are using models for measuring economic capital for credit risk, how were these models

developed?Q41 Which approach(es) are you planning to use for measuring operational risk in your bank?

Table IV.Survey questions

pertaining to the finalregression model

Riskmanagement

301

associated with those who recognise the importance of data protection (Q27). Similarly,results show that more favourable managers acknowledge that both IT and HR issuesmay hinder implementation and that the Accord may actually lower capitalrequirements for some banks.

5.3 H3Here, we search for features which distinguish between privately and publicly ownedbanks. This is achieved via a logistic regression model (McCullagh and Nelder, 1989,Chapter 13; Bingham and Fry 2010, Chapter 8). As before, we regress observedoutcomes against explanatory variables pertaining to the survey data (Q9-Q52)described above.

Let pi denote the probability that the ith bank is privately owned. Lethi ¼ log( pi/1 2 pi). We start with the basic model:

hi ¼ b0; ð3Þ

where b0 is related to the underlying average proportion of privately owned banks[1] Inthe sequel, using equation (3) as a benchmark, we fit the model:

hi ¼ b0 þ b1x1;i þ · · · þ bpxp;i; ð4Þ

where x1, . . . ,xp are explanatory variables pertaining to the survey data.H3 is rejected ifp . 0 in equation (4). Model selection is performed using stepwise model selection basedon x 2 likelihood-ratio tests for nested models (Bingham and Fry, 2010, Chapter 8). Theresults obtained for the final model selected are shown in Table V. The correspondingsurvey questions are shown in Table VI. Tables V and VI show that p . 0 in equation (4)and H3 is rejected.

Although none of the variables included in the final model are individuallystatistically significant, as measured by univariate t-tests, x 2-tests clearly demonstratethat the aggregate effects of the variables is statistically significant. Results indicate aclear delineation between publicly and privately owned banks, with private banksseemingly better prepared overall for implementation. Private banks are more likely to

Coefficient Estimate Estimated standard error t-value p-value

Intercept 19.01 5,770.84 0.003 0.997Question 18 (neutral) 35.77 6,522.90 0.005 0.996Question 18 (agree) 18.16 6,656.54 0.003 0.998Internal model for credit risk 235.69 6,302.67 20.006 0.995No. of years historical data on market risk 17.12 2,533.97 0.007 0.995

Table V.Final logistic regressionmodel: determinants ofprivate ownership

Q18 Do you think that the Basel Accord implementation department is highly functional in yourbank?

Q34 Have you developed methodology for identifying and measuring credit risk for your internalneeds?

Q39 How many years are covered by the time series that you have created for market riskassessment?

Table VI.Survey questionspertaining to the finallogistic regression model

JFRC20,3

302

have a highly functioning Basel Accord implementation department and are also morelikely to have used additional amounts of historical data to assess market risk.Results suggest that private banks are less likely to have developed internal models forcredit risk.

Simple summary statistics reinforce the results of the above regression model.Table VII shows that a functioning Basel Accord implementation department is morereadily associated with the private sector. Table VIII shows that public sector banksare more likely to have developed internal credit-risk methodologies. Figure 2 showsthe availability of historical data on market risk for publicly and privately ownedbanks, with private banks more likely to have longer records.

6. ConclusionsThis paper examines implementation of the Basel Accord in Pakistan. Results havewider implications for emerging markets and for risk management and financialregulation. Adopting the Basel Accord is intended to safeguard global financial

Privately owned (%) Publicly owned (%)

Agree 22 0Neutral 26 0Disagree 52 100

Table VII.Survey responses: is the

Basel Accordimplementation

department in your bankhighly functional?

Yes (%) No (%)

Publicly owned 100 0Privately owned 70 30

Table VIII.Survey responses: have

you developed models forcredit risk internally

Figure 2.Number of years data

on market risk byownership type

0 1 2 3 4

Number of years of historical data on market risk

Per

cent

age

of r

espo

nden

ts

80

60

40

20

0

Publicly OwnedPrivately Owned

Riskmanagement

303

stability and poses a number of specific challenges for both developed and emergingeconomies. Pakistan faces a number of unique challenges in this regard due to recentexpansions in its banking sector and the fact that the rate at which the Accord is beingadopted lags behind that of other countries. This paper throws light on this and relatedissues due to the novelty of the primary survey data from risk managers (at variouslevels of seniority across different institutions) coupled with rigorous statisticalanalyses.

The Basel Accord is generally well regarded due to its underlying aims of improvedcapital standards and administration coupled with the scientific treatment of risk. Thisis reflected by survey results demonstrating that the Basel Accord is generallysupported by managers with no discernible differences in the level of support betweenprivately and publicly owned banks. In contrast, the level of support and the slow paceof implementation appear due to lack of technical expertise (advanced modelling ofcredit and operational risk) and the level of preparedness (availability of historicaldata, having a dedicated Basel Accord implementation department, IT infrastructure).

The analysis in Section 5 is potentially wide-ranging in scope and identifies anumber of critical issues in global banking from a regulatory and risk managementperspective. These include credit risk, practical implementation issues (IT and HR),minimal capital requirements, data security and operational risk. This paper hasdiscussed a number of concerns raised by the implementation of the Basel Accord.These concerns are reflected at least in part by the survey results and even thosesupportive of implementation acknowledge data protection, IT and HR considerationsand the determination of appropriate capital levels as important issues that remain tobe addressed.

Finally, alongside issues of preparedness and technical expertise we examineddifferences between publicly owned and privately owned banks. Results suggestthat in several ways privately owned banks appear better prepared forimplementation. Private banks are more likely to have both a highly functioningBasel Accord implementation department and to have longer historical records toassess market risk.

Note

1. If p is the probability that a bank is privately owned, b0 ¼ log( p/1 2 p).

References

Amidu, D. (2007), “Determinants of capital structure of banks in Ghana: an empirical approach”,Baltic Journal of Management, Vol. 2, pp. 67-79.

Barakat, A. (2009), “Banks Basel II norms requirement regarding internal control: field study onJordan banks”, Delhi Business Review, Vol. 10, pp. 35-48.

BCBS (2001), The New Basel Capital Accord: An Explanatory Note, BCBS, available at: www.bis.org/publ/bcbsca01.pdf (accessed 2 April 2011).

BCBS (2004), Internal Convergence of Capital Measurement and Capital Standards: A RevisedFramework, BIS, available at: www.bis.org/publ/bcbs107a.pdf (accessed 25 May 2011).

BCBS (2006), International Convergence of Capital Measurement and Capital Standards:A Revised Framework, BIS, Basel, Comprehensive Version, June.

Bingham, N.H. and Fry, J.M. (2010), Regression: Linear Models in Statistics, Springer, London.

JFRC20,3

304

BIS (1999), “Capital requirements and bank behaviour: the impact of the Basel Accord”,BCBS Working Paper No. 1.

BIS (2004), “G10 central bank governors and heads of supervision endorse the publication of therevised framework”, press release, available at: www.bri.org/press/p040626.htm (accessed23 March 2011).

Cornford, A. (2006), “The global implementation of Basel II: prospects and outstandingproblems”, SSRN, available at: http://ssrn.com/abstract¼1278049 (accessed 17 May 2011).

Cumming, S. and Nel, H. (2005), “Capital controls and the lending behaviour of South Africanbanks: preliminary findings on the expected impact of Basel II”, South African Journal ofEconomics, Vol. 73, pp. 641-56.

Dupuis, J. (2006), The Basel Capital Accords, Library of Parliament-Canada, Economic Division,Parliamentary Information and Research Service, available at: www.parl.gc.ca/information/library/PRBpubs/prb0596-e.pdf (accessed 15 January 2011).

Engelen, K.C. (2005), “BIS birthday blues: uneasy times as the bank for international settlementsturns 75”, International Economy, Vol. 19, pp. 46-51.

Federal Reserve Board (2007), Risk-based Capital Standards: Advanced Capital AdequacyFramework – Basel II, available at: http://edocket.access.gpo.gov/2007/pdf/07-5729.pdf(accessed 12 May 2011).

Folpmers, M. (2010), “Decoding Basel III: buffers, benefits and bonuses breaking down Basel III’scomplex new capital and liquidity rules for banks, and exploring their short term and longterm macroeconomic impact”, Risk Professional, December, pp. 30-4.

Gordon-Hart, S. (2004), “Basel II: the risk to the global consensus”, Balance Sheet Journal, Vol. 12,pp. 22-6.

Griffith-Jones, S. and Spratt, S. (2001), “Will the proposed new Basel Capital Accord have a netnegative effect on developing countries?”, unpublished report, Institute for DevelopmentStudies.

Hai, S.S., Minnhaj, S.Q. and Ahmed, R. (2007), Implementation of Basel II: Issues, Challenges andImplications for Developing Countries, preprint, available at: www.hull.ac.uk/php/ecskrb/GDP2007/SESARO-58110.pdf (accessed 27 April 2011).

Hassan Al-Tamimi, H.A. (2008), “Implementing Basel II: an investigation of the UAE banks’Basel II preparations”, Journal of Financial Regulation and Compliance, Vol. 16, pp. 173-87.

Heid, F. (2003), “Is regulatory capital pro-cyclical? A macroeconomic assessment of Basel II”,Deutsche Bundesbank working paper.

IMF (2005), Implementation of Basel II – Implications for the World Bank and the IMF, IMF,Washington, DC.

Jacobsohn, O. (2004), Impact of Basel II on the South African Banking System, MagisterCommercii in Business Management, Rand Afrikaans University, Johannesburg.

McCullagh, P. and Nelder, J.A. (1989), Generalised Linear Models, 2nd ed., Chapman & Hall,London.

Makwirmiti, A.M. (2008), “The implementation of the new capital accord (Basel II): a comparativestudy of South Africa, Switzerland, Brazil, and the United States”, unpublished Master ofCommerce thesis, Department of Economics and Economic History, Rhodes University,Grahamstown, available at: http://eprints.ru.ac.za/view/creators/Makwiramiti¼3AA¼2E_M¼2E_¼3A¼3A.html (accessed 21 April 2011).

Mboweni, T.T. (2004), The South African Banking Sector: An Overview of the Past 10 Years, BIS,available at: www.bis.org/review/r041231f.pdf (accessed 5 April 2011).

Riskmanagement

305

Mehra, Y.S. (2010), Operational Risk Management in Indian Banks: Impact of Ownership and Sizeon Range of Practices for Implementation of Advanced Measurement, preprint, availableat: www.igidr.ac.in/money/OPERATIONAL%20RISK%20MANAGEMENT%20IN%20INDIAN%20BANKS.pdf (accessed 21 April 2011).

Neville, L. (2005), “Basel II: a step up the ladder”, Risk Magazine, November.

Ong, M.K. (2004), Introduction to the Basel II Handbook: A Guide for Financial Practitioners,Basel II Handbook, available at: www.baselalert.com (accessed 18 April 2011).

Reisen, H. (2001), “Will Basel II contribute to convergence in international capital flows?”,unpublished report, OECD Development Centre, Paris.

Rime, B. (2001), “Capital requirements and bank behaviour: empirical evidence for Switzerland”,Journal of Banking & Finance, Vol. 25, pp. 789-805.

Snedecor, G.W. and Cochran, W.G. (1989), Statistical Methods, 8th ed., Iowa University Press,Iowa, IA.

Swiss Federal Banking Commission (2005), Annual Report: Key Themes, available at: www.ebk.admin.ch (accessed 27 January 2011).

van Rixtel, A., Alexopoulou, I. and Harada, K. (2004), “The new Basel capital accord and itsimpact on Japanese banking: a qualitative analysis”, in Gup, B.E. (Ed.), The New BaselCapital Accord, South-Western, Cincinnati, OH.

Venables, W.N. and Ripley, B. (2002), Modern Applied Statistics with S, 4th ed., Springer,New York, NY.

Ward, J. (2002), “The new Basel Accord and development countries: problems and alternatives”,ESRC Centre for Business Research, Working Paper No. 04.

Further reading

Bailey, R. (2005), “Basel II and developing countries: understanding the implications”,Development Studies Institute Working Paper Series No. 05-71.

About the authorsOmar Masood is the Director of the UEL Centre of Islamic Finance and Banking, University ofEast London.

John Fry is a Senior Lecturer in Mathematics in the Department of Physics and Mathematics,Nottingham Trent University. John Fry is the corresponding author and can be contacted at:john.fry@ntu.ac.uk

JFRC20,3

306

To purchase reprints of this article please e-mail: reprints@emeraldinsight.comOr visit our web site for further details: www.emeraldinsight.com/reprints

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.