Capial Srengh and Bank Profiabiliy Seok Weon Lee 1 Asian Social Science; Vol. 11, No. 10; 2015 ISSN 1911-2017 E-ISSN 1911-2025 Published by Canadian Cener of Science and Educaion 1 Division of Inernaional Sudies, Ewha Woman s Universiy, 11-1 Daehyun-dong, Seodaemun-gu, Seoul, Korea Correspondence: Seok Weon Lee, Division of Inernaional Sudies, Ewha Woman s Universiy, 11-1 Daehyun-dong, Seodaemun-gu, Seoul, Korea. Tel: 82-2-3277-4456. E-mail: seoklee@ewha.ac.kr Received: November 15, 2014 Acceped: December 29, 2014 Online Published: April 20, 2015 doi:10.5539/ass.v11n10p120 URL: hp://dx.doi.org/10.5539/ass.v11n10p120 Absrac The objecive of his sudy is o examine he relaion beween capial srengh and profiabiliy of Korean banks. Employing a panel regression analysis for Korean banks during 2000-2008, his sudy finds ha higher capial-raio banks end o manage he beer facors of deermining bank profiabiliy. Tha is, he associaion beween major explanaory variables and bank profiabiliy is more significan and consisen wih he general predicions of finance lieraure in he higher capial-raio banks: Higher capial-raio banks end o relae larger asse size, higher capial raio, and higher operaing leverage o higher profis more significanly han lower capial-raio banks. We find a similar resul wih respec o loan raio, oo. This indicaes ha higher capial-raio banks may ransform heir advanageous posiion in capial srengh more effecively and successfully ino generaing higher profis han lower capial-raio banks. This resul may be aribued o he various advanages higher capial-raio banks have such as lower cos of capial, lower bankrupcy and financial disress coss, and less regulaory inerference from bank regulaor, ec. The policy implicaion ha we can derive from he empirical resuls of his sudy may be he following. To improve bank profiabiliy for banking secor s soundness and safey, bank regulaor needs o implemen a discriminaive regulaory policy beween higher and lower capial raio banks. Tha is, ignoring he degree of bank capial srengh and imposing uniformed regulaory policy may no be effecive in improving bank profiabiliy. Keywords: capial srengh, banking secor, profiabiliy, bank regulaion, regulaory policy 1. Inroducion As a srong and sable banking sysem is criical o he healhiness and developmen of a naion s economy, undersanding decision making process and deerminans of bank profiabiliy is very imporan for boh bankers hemselves and for various sakeholders including invesors in he banking indusry, governmens and bank regulaors, ec. Korean banking indusry experienced wo financial crises in he las couple of decades and suffered severe damages o he banking secor s profiabiliy during hese crises. To recover he banking secor s profiabiliy and vialize he Korean economy, Korean governmen made comprehensive srucural reform program in he banking indusry. The main mehod of resrucuring program was focused on srenghening he capial healhiness of he Korean banks. The banks whose capial raios fall below he BIS (Bank of Inernaional Selemen) capial adequacy requiremen of eigh-percen were shu down and acquired by financially healhier banks. Knowledge of he relaionship beween capial adequacy and bank profiabiliy would be a very imporan and ineresing issue o derive policy implicaions no only for he bank regulaor making reform program in he banking indusry by conrolling for bank capial adequacy, bu also for mainaining he soundness and profiabiliy of he banking indusry. The objecive of his sudy is o examine he relaion beween capial srengh and profiabiliy of Korean banks. Employing a panel regression analysis for Korean banks during 2000-2008, his sudy finds ha higher capial-raio banks end o manage he beer facors of deermining bank profiabiliy. Tha is, he associaion beween major explanaory variables and bank profiabiliy is more significan and consisen wih he general predicions of finance lieraure in he higher capial-raio banks: Higher capial-raio banks end o relae larger asse size, higher capial raio, and higher operaing leverage o higher profis more significanly han lower capial-raio banks. We find a similar resul wih respec o loan raio, oo. This indicaes ha higher capial-raio banks may ransform heir advanageous posiion in capial srengh more effecively and successfully ino generaing higher profis han lower capial-raio banks. This resul may be aribued o he various advanages 120
higher capial-raio banks have such as lower cos of capial, lower bankrupcy and financial disress coss, and less regulaory inerference from bank regulaor, ec. The policy implicaion ha we can derive from he empirical resuls of his sudy may be he following. To improve bank profiabiliy for banking secor s soundness and safey, bank regulaor needs o implemen a discriminaive regulaory policy beween higher and lower capial raio banks. Tha is, ignoring he degree of bank capial srengh and imposing uniformized regulaory policy may no be effecive in improving bank profiabiliy. There are very few empirical lieraures ha direcly examine he relaionship beween bank capial adequacy and profiabiliy. Berger (1995) finds ha bank capial raio is posiively relaed o bank profiabiliy measured as he reurn on equiy for he U.S. bank in he 1980s. Ben Naceur and Goaied (2008) find ha Tunisian banks which hold a relaively high amoun of capial end o exhibi higher level of ne-ineres margin and profiabiliy. Sufian and Habibullah (2009) also find ha capializaion has a posiive impac on he sae-owned commercial banks profiabiliy in China. Samadi (2012) finds a posiive relaionship beween capial srucure and bank profiabiliy. Daa, esimaion model and hypoheses are described in nex secion. Secion 3 presens es resuls, and secion 4 discusses concluding commens. 2. Daa, Esimaion Model and Hypoheses Daa source of empirical analysis in his sudy is he Saisics of Bank Managemen from he Korean Financial Supervisory Service. The sample of empirical analysis is composed of Korean commercial banks during 2000-2008. For his sample period, financial informaion variables such as oal asse size, capial-o-asse raio, loan-o-asse raio, fixed asse raio as well as he hree measures of bank profiabiliy; reurn on asse (ROA), reurn on equiy (ROE), ne ineres margin (NIM) is colleced for each bank. Table 1 provides -es resul for he difference of he mean value of he variables used in he sudy beween higher capial-raio and lower capial-raio banks. Every year each bank is caegorized ino eiher higher capial-raio or lower capial-raio banks based on is capial-o-asse raio. I is shown in he able ha he wo main measures of he bank profiabiliy (ROA and ROE) are significanly higher for he higher capial-raio banks. Anoher profiabiliy measure, NIM, is also higher for he higher capial-raio banks. Regarding he main explanaory variables for he bank profiabiliy, Table 1 shows ha higher capial-raio banks have significanly greaer asse size and capial raio han he lower capial-raio banks. Loan raio is greaer for he higher capial raio banks; oo, however, he difference beween he wo groups is no significan. Thus, hese resuls may sugges ha he larger asse size and higher capial raio would be associaed wih higher bank profi. Fixed asse raio is significanly greaer for he lower capial raio banks. The implicaions of hese resuls are explained in deail in he nex secion, in regression analysis afer conrolling for he effec of he oher variables excep for each explanaory variable. Table 1. T-ess for he difference of he mean value Higher capial raio banks Lower capial raio banks ROA (Reurn on asse) 0.8298 0.2124 5.18*** ROE (Reurn on equiy) 11.5493 2.5957 6.41*** NIM (Ne ineres margin) 15,693 9,162 3.68*** Asse size 718,219 503,057 2.95*** Capial-o-asse raio 0.0528 0.0402 1.84** Loan-o-asse raio 0.5341 0.5247 0.72 Fixed asse raio 29.19 36.07-1.49 Number of observaions 63 63 T-saisic for he difference of he means To examine wheher bank financial and operaing characerisic variables differenly affecs bank profiabiliy beween higher capial-raio and lower capial-raio banks, we esimae he following mulivariae panel regression equaion for he sample of Korean commercial banks over he period 2000-2008. To examine he 121
difference in he profi generaing mechanism beween he wo group banks, he regression equaion is esimaed for each group separaely. PROFIT i, 0 1 LNASSET i, 2 CAPITAL i, 3 LOAN i 4 FIXED 5 NAREDUMMY 5 CPI i, i, Following he implicaions of he finance lieraure, his paper employs hree differen measures for bank profiabiliy which is he dependen variable of he regression equaion. They are he reurn on asse (ROA), reurn on equiy (ROE), and ne ineres margin (NIM) of he bank. ROA measures he reurn on he bank oal asse, while ROE measures he reurn on he shareholders equiy. NIM measures he bank average ineres incomes minus average ineres coss. Independen variables deermining bank profi include bank asse size (LNASSET), capial-o-asse raio (CAPITAL), loan-o-asse raio (LOAN) and fixed asse-o-asse raio (FIXED). Consumer price index (CPI) is included as a conrol for he influence of inflaion rae on bank profi. A dummy variable (NAREDUMMY) as o wheher a bank is regional one or naionwide one is included. NAREDUMMY akes he value of 1 if he bank is a regional bank and 0 if i is a naionwide bank. The expeced sign of he regression coefficien beween he explanaory variables and bank profi is summarized below. The expeced sign on he coefficien of bank asse size wih respec o profi is wo-way. A posiive relaion beween asse size and bank profi resuls from he effec of economies of scale. Economies of scale would be associaed wih lower producion and managemen coss, and his will resul in higher profi. On he oher hand, generally larger asse size is associaed wih higher degree of bank asse diversificaion, which would resul in lower overall risk, and herefore, lower profi. The ne effec of asse size on bank profi depends on which effec is greaer. The relaionship beween bank capial raio and profi is also wo-way. Higher capial raio could be effecive in increasing bank profi because invesors required reurn on he bank wih higher capial raio would be lower due o lower possibiliy of bankrupcy, which would resul in lower cos of capial. On he oher hand, if a lower capial raio or higher deb raio induces he bank managers o have moral hazard pursuing a risky sraegy, a negaive associaion beween capial raio and bank profi is possible if he higher risk aking is successful. Loan raio is included o measure he overall bank managemen abiliy on he bank asse porfolios. This is based on he fac ha loans are he larges caegory and one of he main sources of bank earnings. Fixed asse raio is used as he measure of bank abiliy o manage is operaional leverage. If he bank operaional leverage is well managed, a posiive relaion beween he fixed asse raio and profi is expeced. 3. Empirical Resuls 3.1 Esimaion Resuls for ROA, ROE and NIM Table 2 shows he esimaion resul regarding he dependen variable, ROA, for he wo differen groups, higher capial-raio and lower capial-raio banks. Higher and lower capial raios are pariioned a he median value of capial-o-asse raio each year for all he sample banks. The same esimaion is conduced wih respec o ROE in Table 3. I is presened from boh ables ha bank asse size is significanly posiively relaed o boh ROA and ROE for he group of higher capial-raio banks. I is no significan for he lower capial-raio banks. This resul may indicae here is a posiive effec of large asse size on achieving economies of scale and cos reducion only for higher capial-raio banks. Capial raio has also a significanly posiive relaion wih ROA only in he group of higher capial-raio banks. I is posiive wih respec o ROE a he significance level of 13% for higher capial-raio banks. Bu i is insignifican for boh ROA and ROE in he group of lower capial-raio banks. Thus, he posiive effec of higher capial raio on increasing bank profi hrough he effec of decreasing cos of capial and bankrupcy cos is observed only for higher capial-raio banks. Regarding he loan raio, he coefficien is posiive for boh ROA and ROE for he higher capial-raio banks, bu is no significan. For he lower capial-raio banks, i is negaive. This resul presens some weak evidence ha higher capial-raio banks end o manage loan porfolios beer han he lower capial-raio banks. Bu i is no significan o generae higher profi. i, i,, (1) 122
Regarding he fixed asse raio, he coefficien is posiive for boh ROA and ROE for higher capial-raio banks. The coefficien on ROE is significan. However, i is negaive for boh ROA and ROE for he lower capial-raio banks. This resul implies ha managemen of operaional leverage using fixed asse o generae higher profi is beer for he higher capial-raio banks. Table 4 shows he esimaion resuls for he case where NIM is used as he dependen variable. I is shown ha effec of larger asse size on increasing NIM is posiive and significan only for higher capial-raio banks. The oher variables do no presen any significan differences beween he wo groups wih differen capial raios. Table 2. Regression resuls for he full sample (Bank profiabiliy measure: ROA) Higher capial raio banks Lower capial raio banks Coefficien - saisic Coefficien - saisic INTERCEPT -4.4350*** -3.4715 5.0099 1.5259 LNASSET 0.2957*** 3.8159-0.1571-0.7423 CAPITAL 24.9488*** 3.5382 22.6240 1.0390 LOAN 0.4230 0.4961-2.1708-1.0622 FIXED 0.0060 1.0467-0.0338*** -3.6361 NAREDUMMY 0.3997* 1.8832 0.3411 0.7216 CPI -12.4047-1.5428-49.6180*** -3.3803 R 2 0.46 0.5135 F-saisic 8.4080*** 9.5008*** Table 3. Regression resuls for he full sample (Bank profiabiliy measure: ROE) Higher capial raio banks Lower capial raio banks Coefficien - saisic Coefficien - saisic INTERCEPT -212.2028*** -3.8574 203.9824*** 2.1941 LNASSET 14.3054*** 4.1696-7.6070-1.2695 CAPITAL 478.1689 1.5315 285.2036 0.4626 LOAN 34.4923 0.9135-58.8600-1.0172 FIXED 0.4298* 1.6999-1.1705*** -4.4518 NAREDUMMY 26.0177*** 2.7685 0.2519 0.0188 CPI -618.4984* -1.7372-1433.26*** -3.4485 R 2 0.3198 0.5181 F-saisic 4.6232** 9.6770*** Table 4. Regression resuls for he full sample (Bank profiabiliy measure: NIM) Higher capial raio banks Lower capial raio banks Coefficien - saisic Coefficien - saisic INTERCEPT -0.0360*** -3.4267-0.0043-0.3816 LNASSET 0.0021*** 3.3487 0.0000 0.0161 CAPITAL 0.2836*** 4.8925 0.2113*** 2.8599 LOAN 0.0228*** 3.2463 0.0240*** 3.4655 FIXED 0.0001** 1.9235 0.0001* 1.9446 NAREDUMMY 0.0100*** 5.7497 0.0039** 2.4403 CPI -0.0914-1.3829-0.0389-0.7818 R 2 0.6506 0.6948 F-saisic 18.3063*** 20.4852*** 123
3.2 Sep-wise Regression Resuls Table 5 shows he esimaion resuls for he sep-wise regressions afer omiing he insignifican independen variables one by one from he above regression resuls, and examine wheher he resuls mainain consisency. For he sample of higher capial-raio banks, loan raio is omied firs, and boh loan raio and fixed asse raio are omied second. For he group of lower capial-raio banks, asse size is omied firs, and hen boh asse size and capial raio are omied, finally all he hree variables; asse size, capial raio, loan raio are omied. I is found from he able ha he regression resuls for he independen variables ha are significan in he resuls of Table 3 mainain consisency. We repor he resuls for only ROA in his able. The resuls for he oher profiabiliy measures are available from he auhors upon reques. One ineresing resul in Table 5 is ha in all he hree regressions for lower capial-raio banks wih asse size being omied, he coefficien on NAREDUMMY urns ou o be significanly posiive unlike he resuls in Table 2. We conjecure ha his resuls from a high correlaion beween asse size and NAREDUMMY for he group of lower capial-raio banks. I is -0.45. In Table 2, we concluded ha his correlaion is no high o such an exen as o cause a mulicollineariy problem. We checked on he esimaion resuls for all he regressions by omiing he NAREDUMMY from he equaion because i is no one of our main independen variables in his sudy, and we found ha he regression resuls are nearly he same as wha we repored above in Tables 2-4. These resuls are available from he auhors upon reques. Table 5. Sep-wise regression resuls for he full sample (Bank profiabiliy measure: ROA) INTERCEPT LNASSET CAPITAL LOAN FIXED NARE DUMMY CPI Higher capial raio banks Dependen Dependen variable: ROA variable: ROA Loan raio is omied -4.2697*** (-3.4838) 0.2977*** (3.8722) 25.6495*** (3.7345) 0.0054 (0.9763) 0.4353** (2.1945) -12.4067 (-1.5528) Loan raio and fixed asse raio are omied -3.5615*** (-3.6064) 0.2655*** (3.8241) 22.4244*** (3.7245) 0.4168** (2.1115) -11.1449 (-1.4140) Lower capial raio banks Dependen Dependen variable: ROA variable: ROA Asse size is omied 2.7478** (2.2577) 22.9755 (1.0597) -2.1380 (-1.0507) -0.0306*** (-3.7078) 0.6397*** (2.5889) -49.1882*** (-3.3673) Asse size and capial raio are omied 3.3316*** (3.0665) -0.9933 (-0.5754) -0.0360*** (-5.4707) 0.6970*** (2.8880) -52.20*** (-3.6392) Dependen variable: ROA Asse size, capial raio and loan raio are omied 2.7734*** (5.7038) -0.0344*** (-5.7688) 0.6371*** (2.9434) -51.8903*** (-3.6415) R 2 0.4586 0.4500 0.5085 0.4985 0.4955 F-saisic 10.1681*** 12.4814*** 11.3837*** 13.9183*** 18.6664*** The numbers inside parenhesis represen he T-saisic for he regression coefficien. 4. Concluding Commens The objecive of his sudy is o examine he relaion beween capial srengh and profiabiliy of Korean banks. Employing a panel regression analysis for Korean banks during 2000-2008, his sudy finds ha higher capial-raio banks end o manage he beer facors of deermining bank profiabiliy. Tha is, he associaion beween major explanaory variables and bank profiabiliy is more significan and consisen wih he general predicions of finance lieraure in he higher capial-raio banks: Higher capial-raio banks end o relae larger 124
asse size, higher capial raio, and higher operaing leverage o higher profis more significanly han lower capial-raio banks. We find a similar resul wih respec o loan raio, oo. This indicaes ha higher capial-raio banks may ransform heir advanageous posiion in capial srengh more effecively and successfully ino generaing higher profis han lower capial-raio banks. This resul may be aribued o he various advanages higher capial-raio banks have such as lower cos of capial, lower bankrupcy and financial disress coss, and less regulaory inerference from bank regulaor, ec. The policy implicaion ha we can derive from he empirical resuls of his sudy may be he following. To improve bank profiabiliy for banking secor s soundness and safey, bank regulaor needs o implemen a discriminaive regulaory policy beween higher and lower capial raio banks. Tha is, ignoring he degree of bank capial srengh and imposing uniformed regulaory policy may no be effecive in improving bank profiabiliy. References Berger, A. N. (1995). The Relaionship beween Capial and Earnings in Banking. Journal of Money, Credi and Banking, 27, 432-456. hp://dx.doi.org/10.2307/2077877 Hirle, B. J., & Siroh, K. J. (2007). The Reurn o Reail and he Performance of Commercial Banks. Journal of Banking and Finance, 31, 1101-1133. hp://dx.doi.org/10.1016/j.jbankfin.2006.10.004 Naceur, S. B., & Goaied, M. (2008). The Deerminans of Commercial Bank Ineres Margin and Profiabiliy: Evidence from Tunisia. Froniers in Finance and Economics, 5, 106-130. Samadi, M. T. (2012). An Empirical Sudy on he Impac of Operaing Risk on Srucure Capial and Profiabiliy in Iranian Banking Secor. Managemen Science Leers. hp://dx.doi.org/10.5267/j. msl.2012.04.020 Sufian, F., & Habibullah, M. S. (2009). Bank Specific and Macroeconomic Deerminans of Bank Profiabiliy: Empirical Evidence from China Banking Secor. Froniers of Economics in China, 4, 274-291. hp://dx.doi.org/10.1007/s11459-009-0016-1 Williams, B. (2003). Domesic and Inernaional Deerminans of Bank Profis: Foreign Banks in Ausralia. Journal of Banking and Finance, 27, 1185-1210. hp://dx.doi.org/10.1016/s0378-4266(02)00251-0 Copyrighs Copyrigh for his aricle is reained by he auhor(s), wih firs publicaion righs graned o he journal. This is an open-access aricle disribued under he erms and condiions of he Creaive Commons Aribuion license (hp://creaivecommons.org/licenses/by/3.0/). 125