Pakistan: A Study in Risk Analysi s Salman Ahmad * Introducti on The financial sector in Pakistan has evolved over the years in response to the growth of the economy and the governments plans for the growth and development of the country. The sector as on 31 March 2002 comprises the tate !ank of Pakistan" # state$owned %anks" 2 newly privatised %anks" # specialised %anks" 1# private scheduled %anks" a%out 30 leasing companies" #& Modara%as" 1# investment %anks" 3 stock e'changes" &( insurance companies" and )overnment aving *enters. *ommercial %anks were nationalised in 1+,# and are now in the process of %eing privatised. Two nationalised commercial %anks have %een privatised since 1++0- Muslim *ommercial !ank was sold %y auction. negotiation" while /llied !ank was sold to its employees .The market share of the nationalised commercial %anks has %een declining with the introduction of new private %anks. The three nationalised commercial %anks in Pakistan 01a%i% !ank 2imited" 3nited !ank 2imited" and 4ational !ank of Pakistan0have a large %ranch network that allows them to e'pand and compete with the private %anks in deposit mo%ilisation. The newly privatised %anks have ac5uired the %ranch network that will allow them to e'pand and compete with the state$owned %anks. 6hile these institutions play an important role in financing short$ term credit re5uirements" their success in raising deposits ensures that they have a significant surplus of funds that can %e lent or invested in government securities. * The author is Professor of Economics, Lahore School of Economics.
6 5 The Lahore Journal of Economics, Vol.6, No.2 There are 1+ foreign %anks which have %een esta%lished from time to time. The foreign %anks in Pakistan are %ranch operations7 they are not separate legal entities. Much of their success can %e attri%uted to their superior management skills and %etter access to international financial markets. / large fraction of foreign currency deposits are with foreign %anks" partly %ecause of their marketing efforts and partly %ecause of their credi%ility as international %anks. 8oreign %anks are only allowed four %ranches within the country. 9n recent years" as e'ternal credit flows to developing countries have declined" the need for an efficient and via%le financial sector has grown. /n efficient financial system collects domestic savings and allocates the collective resources to the %est possi%le investment opportunities" allowing for %etter domestic resource mo%ilisation and utilsation" and reducing the reliance on e'ternal financing. Throughout the 1+(0s" Pakistan was a%le to achieve strong growth with low inflation. The %anking sector operated inefficiently %ecause of nationalisation. 1 8inancial intermediation was geared towards the financing needs of the government and those of the targeted sectors of the economy to which credit was officially directed. /lthough Pakistans economy opened up internationally in 1++0" the government had a tight hold on entry to financial markets. The situation changed during the +0s with the entry of private %anks in the country and su%se5uent competition. 8inancial reforms were initiated with a strong commitment to market$oriented development of the financial sector. These reforms are still underway and will need to %e strengthened su%stantially if modern" competitive financial markets are to %e developed. 8oreign %anks have shown the highest growth rate in terms of deposits in the last three years. tate$owned %anks averaged an annual growth of 12., per cent over the 1 Khan Bashir Ahmad, Financial Markets and Economic e!elo"ment in Pakistan# 1$%&'1$$(, ".))% 66 Salman Ahmad three$year period 1++($2001" while the recently privatised nationalised commercial %anks averaged 1(.2 per cent and foreign %anks" 31.# per cent. 8oreign %anks are also more efficient in terms of controlling administrative costs per %ranch. 8or e'ample" growth of administrative costs averaged 1+., per cent for state$owned %anks" 1(.3 per cent for the recently privatised %anks" and 1,.: per cent for foreign %anks. /s a percentage of assets" administrative costs are the lowest for foreign %anks. The gross revenues of foreign %anks have grown at a much faster pace than the rate of growth of their costs. 8or Pakistani %anks" the difference in the growth rates is e'tremely narrow" if not negative. !ank deposits have grown significantly in recent years primarily %ecause of the rapid growth of money supply" which increased %y around 1( per cent and 20.: per cent during 1+++$2000. Money has also flowed into the %anking system from the informal economy with the change in incentives following the recent slide of the stock market and the significant slump in the real estate market. ;emand deposits constitute a%out #,$&0 per cent of total %ank deposits which is much higher than other countries. This also e'plains why the cost of funds have %een low for commercial %anks. Profitability and Producti vi ty of Commerci al Banks The profit earned %y the commercial %anks is the difference %etween the deposit rate and the lending rate" called the spread. 9n nominal terms" the spread has risen steadily from & per cent in 1+(:$(, to :.2 per cent in 1++2$+3" %ut then fell to 2.( per cent in 1++, and increased to 3.3 per cent in 2000. 9n real terms" the weighted average yield on deposits fell from 3.+ per cent in 1+(:$(, to <#.: per cent in 1++0$+1" and <1.1 per cent in 1++2$+3.The weighted average yield on loans also declined7 after declining from (.+ per cent in 1+(:$(, to
6 7 The Lahore Journal of Economics, Vol.6, No.2 1.1 per cent in 1++0$+1.Thereafter" it clim%ed rapidly to &.1# per cent %y 1++2$+3. 2 8oreign %anks seem to %e the main %eneficiaries of the relatively large spreads" especially as the nationalised %anks have to deal with pro%lems of large overheads" increasing inefficiencies and nonperforming loans. The three$year =1++($2001> averages of the profita%ility indicators for nationalised commercial %anks and private commercial %anks are illuminating. 9n the case of nationalised commercial %anks" total administrative costs were 2 per cent of total assets compared with 0.(& per cent in the case of foreign %anks" 0.: per cent in the case of private %anks. Pre$ta' profits as a percentage of deposits are 0.: per cent for nationalised commercial %anks" & per cent for foreign" and 2.& per cent for new private %anks. ?ver the years" the capital %ase of nationalised commercial %anks has %een severely affected %y the poor 5uality of %ank loans made on political and uneconomic grounds. /s a result" the single most formida%le pro%lem facing the %anks is the heavy %urden of nonperforming loans. Pakistan introduced Prudential @egulations in 1++3 to ensure that credit is not misused and the infected portfolio was minimum. 1owever" the infected portfolio has increased to significant proportions = see Ta%le 1 %elow>. ?n Aune 30" 2001" the 4on$Performing 2oans =4P2s> amounted to @s. 2,+ %illion" i.e." (.2 per cent of );P" 1(.: per cent of domestic assets" and 32.& per cent of total credit made availa%le to the private sector and pu%lic enterprises. 4on$Performing 2oans of the *ommercial !anking sector were @s. 221 %illion"i.e" :.& per cent of );P and 22.1 per cent of total deposits. ?ut of 4P2s" the defaulted loans of the financial institutions and the commercial %anking sector were @s. 1,2 %illion and @s. 1#1 %illion respectively. !ecause of such a large infected portfolio" the spread %etween lending and deposit rate has ) *l +a,ue, -adeem, and Shahid Kardar, 1$$( . e!elo"ment of the Financial Sector in Pakistan., ". %%/. 68 Salman Ahmad remained high. Though rescheduling of loans is common" the total advances of nationalised commercial %anks categorised as %ad and dou%tful de%ts are @s. &: %illion of which @s.#: %illion are classified as advances related to the private sector. Aust under 23 per cent of the private sectors classified de%t pertains to advances under mandatory targets and concessional credit schemes. 9n 1++(" the tate !ank of Pakistan estimated that around 1#.2 per cent of the loan portfolio of nationalised commercial %anks were made up of nonperforming loans. Tabl e 1: on! Performi n" and #efaul ted $oans %Rs& billion' on! Performi n" $oans #efaul ted $oans (une 1))* 20,.+ 1#:.1 (une 1))) 212.1 1#3.1 (une +,,, 23+.& 1#(.1 (une +,,1 2,+.1 1,2.# -ebruary +,,+ 2,(.: 1:(.1 Source: 3npu%lished data" cited in Pakistan 1uman *onditions @eport 2002 The pro%lem of de%t recovery is not simply a technical issue. 4ot only have political pressures affected the 5uality of the loan portfolio of %anks" they have also %een instrumental in preventing %anks to proceed against persistent defaulters and have resisted attempts to improve the enforcement mechanism. .easures of Performance . / total of 1: nationalised" privatised" private" and foreign %anks were selected for the study. The standard tests used to measure the performance of commercial %anks are applied. There are four ratios used in @isk analysis.
6 9 The Lahore Journal of Economics, Vol.6, No.2 These are- 1& Ca/ital Risk *apital %ase of financial institutions facilitates depositors in forming their risk perception a%out the institution. /lso" it is the key parameter for financial managers to maintain ade5uate levels of capitalisation. !esides a%sor%ing unanticipated shocks" it signals that the institution will continue to honour its o%ligations. 9n order to protect the interest of depositors and shareholders of commercial %anks" the tate !ank of Pakistan introduced the risk$%ased system for capital ade5uacy in 4ovem%er 1++( and asked %anks to maintain ( percent *apital to @isk 6eighted /ssets =*@6/> ratio. This is the %enchmark set %y the !/2B =!ank upervision @egulation *ommittee> of !ank for 9nternational ettlements. /dditionally" %anks are re5uired to achieve a minimum paid$up capital of @s. 1 %illion vide !; circular 4o. 31 dated ;ecem%er :" 2000. *apital @isk is measured %y the ratio of B5uity *apital to Total /ssets. This ratio for our sample of 1: %anks is shown in Ta%le 2. / higher percentage means that the %ank is safer %ecause it can withstand a sharper decline in the value of its assets. The ta%le shows that this ratio has improved for most %anks over the year 2000 to 2001. 1owever" it is %elow the target of ( per cent for all e'cept 8aysal %ank" Prime *ommercial %ank" and Bmirates !ank. The ratio has %een low for nationalised commercial %anks and foreign %anks. This suggests a high degree of capital risk or inade5uate capitalisation for the e'isting level of lending. ?nly four %anks" 8aysal !ank" Prime *ommercial %ank" the !ank of PunCa%" and Bmirates !ank had this ratio greater than ( per cent. This lower ratio can %e attri%uted to a fall in yield of government securities" and hence fall in returns on %anks investment. !eing Dero$ risk weighted" disinvestment of government securities inevita%ly led to a slight fall in the capital ade5uacy ratio. 9n addition" higher provisioning against 4on$Performing 2oans =4P2>" which affects the capital %ase through 7 Salman Ahmad profit.loss accounts" has further contri%uted to the decline of this ratio. This had three implications. 8irst" lending rates pro%a%ly did not ade5uately reflect the prevailing risk premiums in the market and affected the spreads %etween lending and %orrowing rates. econd" loan recovery was poor and the rate of default high with a corresponding write$off of losses and lower earnings. Third" volume growth did not ade5uately compensate for the reduced spreads over the long run.
7 ! The Lahore Journal of Economics, Vol.6, No.2 Tabl e +: Ca/ital Ade0uacy of Commerci al Banks in Pakistan Ca/ital Risk % /er cent' +,,, +,, 1 1. 1a%i% !ank 2imited 3.(1 3.(# 2. 3nited !ank 2imited 3.,+ #.33 3. Muslim commercial !ank 2td. 3.20 3.:( #. /skari *ommercial !ank 2td. &.0: &.0& &. !ank /l$8alah 2td. #.0 3.2, :. !olan !ank 2td. 1.& 1.: ,. 8aysal !ank 2td. +.:# +.3( (. Platinum *ommercial !ank 2td. &.:& 1.1# +. Prime *ommercial !ank 2td. + 10.3 10. oneri !ank 2td. :.21 ,.&, 11. The !ank of PunCa% 10 11 12. 3nion !ank 2td. 3.:& 3.,( 13. /!4 /M@? !ank 4E 2.1# 1.# 1#. *iti !ank 4/ 1.3, 0.2# 1&. Bmirates !ank 9nternational 2td. +.: 10.2 1:. tandard *hartered )rindlays !ank 2td. 2 1 Source: *alculated from financial statements as on 31.12.2001. +& Credit Risk /sset 5uality determines the ro%ustness of financial institutions against loss of value in the assets. The deteriorating value of assets" %eing the prime source of %anking pro%lems" directly pour into other areas" as losses are eventually written <off against capital" which ultimately Ceopardises the earning capacity of the institution. 6ith this %ackdrop" the asset 5uality is gauged in relation to the level and severity of non$performing 72 Salman Ahmad assets" ade5uacy of provisions for %ad loans" recoveries of loans" distri%ution of assets" etc. /lthough the %anking system is infected with a large volume of 4on$Performing 2oans =4P2s>" the severity of this pro%lem has sta%ilised to some e'tent. This is not to say that the pro%lem of 4P2s has taken a secondary position. 3nfortunately" it still remains the most dominant factor affecting the earning capacity of %anks. Popular indicators include non$ performing loans to advances" loan default to total advances" and recoveries to loan default ratios. 6e have used total loans to assets ratio to Cudge credit risk. 9t shows what percentage of assets are more risky %ecause loans are the most risky assets on which default occurs. 8or a less risky %ank" this ratio should %e low and must decline overtime. The results for the years 2000 and 2001 are given in Ta%le 3.The ta%le shows that all the %anks had a%ove average credit risk" ranging from 30 to :0 per cent. Technically" a high credit risk should %e associated with higher returns. 9n the case of nationalised commercial %anks" this ratio has declined for 1a%i% !ank. 9t fell from &3.1+ in 2000 to &0.10 in 2001. This shows that fresh loans are %eing e'tended much more prudently than was the case earlier. 9n other words" the percentage of loans out of total assets given to clients has fallen. 9n the case of private and privatised %anks" this ratio has increased. The ratio for 3!2 increased from 3+.+& to #,.,,. 8or M*!" it increased from #2.& to &2.,3. 8or !ank /l$8alah" it increased from #+.13 to &&.2,. imilarly" for Prime *ommercial !ank" this ratio increased from #,.0, to &,.3&. The same is the situation for Bmirates !ank and tandard *hartered !ank. This ratio was very high for Platinum !ank" /!4$/mro" and *iti %ank" although it has declined over the year 2000 to 2001.This improvement is much more pronounced given their share in total 4P2s. 9t shows a marked improvement in recovery efforts %y the private %anks. The ratio is lowest for !ank of PunCa% and !olan !ank indicating that the %anks are risk averse.
7 " The Lahore Journal of Economics, Vol.6, No.2 Table 1: Asset 2uality of Commerci al Banks in Pakistan Credi t Risk % /er cent' +,,, +,,1 1. 1a%i% !ank 2imited &3.1+ &0.10 2. 3nited !ank 2imited 3+.+& #,.10 3. Muslim commercial !ank 2td. #2.& &2.,3 #. /skari *ommercial !ank 2td. #:.&3 #&.:+ &. !ank /l$8alah 2td. #+.13 &&.2, :. !olan !ank 2td. 33.+1 3&.&& ,. 8aysal !ank 2td. &2.10 &:.,& (. Platinum *ommercial !ank 2td. :&.3& 3+.,1 +. Prime *ommercial !ank 2td. #,.0, &,.3& 10. oneri !ank 2td. &#.3# #+.:& 11. The !ank of PunCa% 33.,, 30.#3 12. 3nion !ank 2td. #+.2, #:.03 13. /!4 /M@? !ank 4E :1 &,.3, 1#. *iti !ank 4/ #+.&: ##.0# 1&. Bmirates !ank 9nternational 2td. #(.:0 &2.13 1:. tandard *hartered )rindlays !ank 2td. #+.0# &3.1# Source: *alculated from financial statements as on 31.12.2001 1& $i0uidity Risk /n ade5uate li5uidity position refers to a situation where the institution can o%tain sufficient funds" either %y increasing lia%ilities or %y converting its assets 5uickly at a reasona%le cost. 9t is" therefore" generally assessed in terms of overall assets and lia%ility management" as mismatching of maturities of assets and lia%ilities gives rise to li5uidity risk. Bfficient fund management refers to a situation where a spread %etween rate sensitive assets =@/> and rate sensitive lia%ilities = @2> is maintained. The most commonly 7# Salman Ahmad used tool to evaluate interest rate e'posure is the gap %etween @/ and @2" while li5uidity is gauged %y li5uid assets to total asset ratio. 6e have used investment in short$term securities to deposits as a measure of li5uidity risk. / higher ratio shows that the %ank has li5uid assets availa%le to meet deposit withdrawals. 1owever" there is a tradeoff %etween li5uidity and profita%ility. / %ank that maintains higher li5uidity is not investing its funds in long$ term and risky proCects. This is shown in Ta%le #. 8oreign !anks have the lowest li5uidity risk. 8or tandard *hartered" the li5uidity ratio increased from #1.# to &0.+&" for *iti%ank" it increased from 12.+ to 1+.,&" for /!4 /M@?" it increased from 1& to 2&. The ratio decreased for private commercial %anks showing increased li5uidity risk. 8or !ank /lfalah the ratio fell from 32 to 2(.:. 8or !olan !ank the ratio fell fram #( to 32. 8or Prime *ommercial !ank the ratio also fell from 3# to 2:.&. 1owever it increased for !ank of PunCa% from 1+ to 33. 8or the %anking sector as a whole" the ratio is 5uite low due to the after effects of freeDing of 8oreign *urrency /ccounts as the %anks had less resources to invest in li5uid funds. /n alternative e'planation is decreasing yields on short$term securities which render investment in such securities unattractive. )enerally" declining ratio would imply that a smaller percentage of deposits are invested in li5uid assets" thus raising the li5uidity risk. Tabl e 3: $i0uidity of Commerci al Banks in Pakistan $i0uidity Risk % /er cent' +,,, +,, 1 1. 1a%i% !ank 2imited 1: 1: 2. 3nited !ank 2imited +.#, (.&# 3. Muslim commercial !ank 2td. 12 1# #. /skari *ommercial !ank 2td. 1(.: 21.1 &. !ank /l$8alah 2td. 32 2(.: :. !olan !ank 2td. #( 32 ,. 8aysal !ank 2td. 13.:: 13.# &
7 5 The Lahore Journal of Economics, Vol.6, No.2 (. Platinum *ommercial !ank 2td. 11 2, +. Prime *ommercial !ank 2td. 3# 2:.& 10. oneri !ank 2td. #+.2, &0.0 & 11. The !ank of PunCa% 1+ 33 12. 3nion !ank 2td. 12.: 1(.3 + 13. /!4 /M@? !ank 4E 1& 2( 1#. *iti!ank 4/ 12.+ 1+., & 1&. Bmirates !ank 9nternational 2td. 3(.+ 32.& 1:. tandard *hartered )rindlays !ank 2td. #1.# &0.+ & Source: *alculated from financial statements as on 31.12.2001 3& Interest Rate Risk The diversified nature of %ank operations makes them vulnera%le to various kinds of financial risks. ensitivity analysis reflects the institutions e'posure to interest rate risk" foreign e'change volatility and e5uity price risk. @isk sensitivity is mostly evaluated in terms of the managements a%ility to monitor and control market risk. 8or interest risk we have used a ratio of interest sensitive assets to interest sensitive lia%ilities. This is shown in Ta%le &. @/s has diverged from @2s" in a%solute terms. 1igher interest rate sensitivity of spread is reflected in less than 100 value of ratio %etween @/ and @2. 8or most of the %anks" this ratio declined %etween 2000 and 2001 indicating a rise in interest rate sensitivity. ?nly three %anks had a ratio of more than 100. The highest is &(( per cent for /skari *ommercial !ank during 2001. 8aysal !ank had 200 per cent during 2000" the !ank of PunCa% had 100 per cent during 2001 and oneri !ank had 111 per cent during 2000. Table 4: .arket Risk of Commerci al Banks in Pakistan Interest Rate Risk % /er cent' 76 Salman Ahmad +,,, +,,1 1. 1a%i% !ank 2imited +3 (+ 2. 3nited !ank 2imited +#., (&.# 3. Muslim commercial !ank 2td. +& (& #. /skari *ommercial !ank 2td. &3#.: &(( &. !ank /l$8alah 2td. (,.+( (+.3, :. !olan !ank 2td. (3 ,1 ,. 8aysal !ank 2td. 200 11# (. Platinum *ommercial !ank 2td. (1 ,2 +. Prime *ommercial !ank 2td. (2.+ (+ 10. oneri !ank 2td. 111 10, 11. The !ank of PunCa% +& 100 12. 3nion !ank 2td. 3#.#+ 2(.#+ 13. /!4 /M@? !ank 4E (1 (: 1#. *iti !ank 4/ (2.21 (( 1&. Bmirates !ank 9nternational 2td. :( :, 1:. tandard *hartered )rindlays !ank 2td. :3 (0 Source: *alculated from financial statements as on 31.12.2001 / discussion of all the four risk ratios for the 1: %anks shows that the !ank of PunCa% is performing well as far as the risk dimension is concerned.
7 7 The Lahore Journal of Economics, Vol.6, No.2 References !erger" /llen.4." ;. 1ancock" and ;avid !. 1umphrey" 1++3" F !ank Bfficiency ;erived from the Profit 8unctionG" Aournal of $an%in& and 'inance 1," pp. 31,$3#,. !erger" /llen and ;avid !. 1umphrey" 1++," F Bfficiency of 8inancial 9nstitutions- 9nternational survey and ;irections for 8uture @esearch"G Euro(ean Journal of )(era*ional +esearch +( =+2>" pp. 1,&$212. !erger" /llen 4." and A. Mester" 1++," F Bfficiency and Productivity *hange in the 3.. *ommercial !anking 9ndustry- / comparison of the 1+(0s and 1++0s"G 8ederal @eserve !ank of Philadelphia 6orking Paper +,$&" May" Philadelphia. !erger" /llen 4." and A. Mester" 2001" F B'plaining the ;ramatic *hanges in the Performance of 3.. !anks- Technological *hange" ;eregulation" and ;ynamic *hanges in *ompetition"G 8ederal @eserve !ank of Philadelphia 6orking Paper 01$:" /pril" Philadelphia. )il%ert @. /lton" and Paul 6. 6ilson" 1++(" F Bffects of ;eregulation on the Productivity of Horean !anks"G Journal of Economics and $usiness " Eol.&0"pp. 133$ 1&&. )ra%owski" @ichard" 4anda @angan" and @asoul @eDvanian" 1++#" FThe Bffects of ;eregulation on the Bfficiency of 3.. !anking 8irms"G Journal of $usiness and Economics" Eol.#:" pp. 3+$&#. 1ardy" ;aniel *. and Bmilia !onaccorsi di Patti" 2001" F!ank @eform and !ank Bfficiency in Pakistan"G 9M8 6orking Paper wp.01.13(" 6ashington- 9nternational Monetary 8und 1empel" )eorge 1. and imonson" ;onald ). =1+++> !ank Management" & th ed. Aohn 6iley I ons. 4.J. 78 Salman Ahmad 1umphrey" ;avid !." and 2awrence !. Pulley" 1++," F !anks" @esponses to ;eregulation- Profits" Technology and Bfficiency"G Journal of ,one-, .redi* and $an%in&, Eol.2+" pp. ,1$+3. Hhan" !ashir /. 1+++ 8inancial Markets and Bconomic ;evelopment in Pakistan- 1+#,$1++& in Hhan" hahrukh @afi" 8ifty Jears of Pakistans Bconomy" ?'ford 3niversity Press. Hlein" Michael 3. 1++2" F *ommercial !anking in PakistanG" in 8inancing Pakistans ;evelopment in the 1++0" /nCum 4asim" Harachi- 2ahore 3niversity of Management ciences and ?'ford 3niversity Press. Mahmood /mir and Hhan" !ashir /. 1++:" F 9ndustry tructure and the *onduct of the Pakistani !anking 9ndustryG" 6orking Paper" 2ahore 3niversity of Management ciences. tate !ank of Pakistan" /nnual @eport 2000$2001. 3l 1a5ue" 4adeem" and hahid Hardar" 1++&" F ;evelopment of the 8inancial ector in Pakistan"G in hahid 4. Kahid =ed.>"1++&" 8inancial ector ;evelopment in /sia- *ountry tudies" /sian ;evelopment !ank" Manila.