Capital Adequacy and Financial Performance of Banks in Nigeria: Empirical Evidence Based on the Fgls Estimator

Transcription

1 Captal Adequacy and Fnancal Performance of Banks n Ngera: Emprcal Evdence Based on the Fgls Estmator Davd Umoru, PhD Department of Economcs, Bankng & Fnance, Benson Idahosa Unversty, Ngera. Joy O. Osemwege, MSc Department of Accountng, Benson Idahosa Unversty, Ngera. do: /esj.016.v1n5p95 URL: Abstract The study examnes the degree of sgnfcance of the captal adequacy rato n nfluencng the fnancal deeds of Ngeran banks by applyng the feasble GLS estmator technque on the pooled panel model for the perod of 007 to 015. Emprcal evdence supports the overrdng mpact of captal adequacy n enhancng the fnancal deeds of Ngeran banks. Nevertheless, the mpact of the estmated captal adequacy s below 30%. The polcy stance of the emprcs holds thus that depostor s money n the bankng sector has not been absolutely assured. Hence, the depost money banks mght not be able to fulfl ther labltes and rsk. In lght of the fndngs, we suggested a constant reassessment of the least amount of captal requred of banks by the CBN. Keywords: Captal adequacy, bank performance, Ngeran depost money banks JEL Classfcaton: A46, G8, D6 Introducton Banks are fnancal nsttutons for moblzng fnancal resources through ther ntermedaton role for productve nvestment (Athansasoglou, Brssms, and Dels, 006). Nonetheless, the fnancal deed of most of the banks n Ngera over the some years has been unmpressve. The case n pont s the fact that the proft before tax (PBT) of the bankng system n Ngera oscllated between 000 and 005, and has snce 006 declned progressvely. The PBT whch was 80.8 percent n 000 reduced wth a loss value of 13.95% (Abreu and Mendes, 00). Even when the PBT peaked at 87.6 percent n 007, t further declned to percent n 008 (Obamuy, 01). By ntuton, the 95

2 envronment for Ngeran banks to make profts s snkng. The declne n profts could be attrbuted to the worldwde economc crses although the n 005, the Central Bank of Ngera (CBN) ncreased the lowest amount of captal that s requred by banks to stay n busness to N5 bllon (Somoye, 008). So, takng the footsteps of Tomola (013) that Ngeran banks are yet to n terms of realzng optmal captal structure, the ntent of ths research s to analyze emprcally how sgnfcant s adequacy of captal rato n nfluencng the deeds of Ngeran banks. Dstnctvely, the study ntends to evaluate the role of captal adequacy on the performance Ngeran banks. The research hypothess thus holds that there s no sgnfcant effect of captal adequacy on Ngeran banks performance. Followng ths ntroducton, s a revew of recent emprcal studes on the subject matter. The theoretcal framework, model specfcaton, research sample, methodology and data measurement are dscussed n secton three, secton four analyzes the feasble GLS results and secton fve concludes wth emphass on polcy stance. Emprcal studes A plethora of studes nvestgated the lnk between captal adequacy and bank performance for dfferent countres. To Mpuga (00), the nadequacy of mnmum captal s a major cause of bank falure n Uganda. The emprcal evdence of Scott and Aras (011) supported the fact that captal to asset rato s a sgnfcant proftablty determnant of the banks n the Unted States of Amerca. Vong and Anna (009) nsttuted that the bank proftablty s a foremost factor for explanng captal strength of a bank. Accordng to Kosmdou (008), Saona (011), Rahman and Farat (01), there s a consderable postve lnk between bank proftablty and captal adequacy. To Tanaka (00), monetary channel of nfluence s watered down f banks are poorly captalzed. Smlarly, Sufan (009) put forth emprcally, the sgnfcance of captal adequacy of state commercal banks n Bangladesh. Although Razard and Mehar (013) found that captal adequacy decreases wth bank sze and proft s unswervngly lnked wth captal adequacy, Mathuva (009) found that bank proftablty s postvely lnked to the core captal rato and ter one rsk based captal rato. Ho and Hsu (010) reported that the lmts on captal adequacy have nduced banks rsky nvestment n Japan. In Turkey, Isk and Hassan (003a) ascertaned that captal rato s absolutely lnked to bank proftablty whereas nflaton has an nverse lnk. The emprcal evdence of Perera, Skully, and Chaudrey (013) holds that adequate captal was a sgnfcant drver of banks proftablty. Ths was corroborated by the emprcal evdence from Jordan as Al-Sabbah (004) found captal adequacy as the most sgnfcant and postve determnant of banks proftablty n Jordan. 96

3 The emprcs of Dore (013) found that captal adequacy and lqudty of banks are negatvely assocated wth bank proftablty. Neceur (003) standardzed a strong postve mpact of captalzaton on return on assets of the bank and found that captal adequacy ncreases the proftablty n Tunsa. To Yudstra (003), there s a strappng absolute lnk between bank captal and bank deposts n Indonesan. In ther research, Somoye (008), Osuka and Rchard (013) ascertaned that asset qualty has no lnk wth captal base of the Ngeran banks. Theoretcal framework, model specfcaton and research methodology The commercal loan theory drves ths study. The ntty-grtty of the theory s that bank funds should be nvested n short term loans otherwse known as self-lqudaton loans n order to boost workng captal. The pooled effect regresson model s thus specfed: k q Y Z u t 1 j jt f f t t j f 1 p (3.1) The ndex dctates the unt of observaton depost money bank, t dctates the tme perod, and j and f are used to dstngush between observed and unobserved regressors, q s the varable of return on asset [Ln(r)] or bank proft after tax [Ln(b)],Yj s the vector of captal adequacy rato [Ln(c)], lqudty [Ln(l)], asset qualty [Ln(a)], management effcency [Ln(m)] and nflaton [Ln(n)], whle the Zf varables are answerable to unobserved heterogenety and as such consttute a nusance component of the model. Snce the Zf varables are unobserved, we defne the unobserved effect, representng the combned mpact of the Zf varables on q as w. w Gven that error such that: p Z f f f 1 k t q Y w u t 1 j jt t t j (3.) w are contemporaneously exogenous to the condtonal E[ u w ] 0, t 1,,..., T (3.3) t We mpose strct exogenety on equaton (3.3) n order to avert the estmaton problem as follows. In ths regard, 97

4 E[ u y, y,..., y, w ] 0, t 1,,..., T t 1 T E[ q y, w ] Y w t t 1 j jt j k t (3.4) Thus, the pooled panel model s obtaned as: ' qt yt ut (3.5) Under the usual assumptons of consstency and unbasedness, the estmate of s obtaned by centerng the data between transformaton and hence specfed the between estmator model as: t 1 j jt t j 1 k q y u k q y u j j j (3.6) In effect, the coeffcents estmate of uses the cross-sectonal depost money bank data by gnorng the tme seres devaton. Thus, the pooled estmator becomes: N T 1 N T 1 N T [ q 1 t 1 q 1 t 1][ et 1 e 1 1] t t t t POLS N T 1 N T [ y ] 1 t t y 1 t t (3.7) 1 T 1 T N 1 1 T N N t1 1 POLS T N t1 1 qe t t y t (3.8) POLS 1 G (3.9) nt H G N H Q as N T jontly d s [0, G], H, Usng the pooled OLS resduals, we estmated the varance-covarance matrx as follows: 98

5 1 T N G E 1 t1 y e t t... e () 1 T T1 T... N 1 E y11e11... y1 Te1 T y1e1... ye 1 E y11e11... y1 Te1 T y1e1... ye E cross products (3.10) In effect, the error term s a vector whte nose generatng process wth st Ee ( t ) 0 and E( ete ' s) where the covarance matrx s 0 s t postve defnte mplyng that the errors are serally uncorrelated but contemporaneously correlated. Gven that E[ e 1e] 0 due to contemporaneous correlaton, the mathematcal expectaton of the cross products are themselves not equal to zero. Ths prompted us to estmate the Newey-West (1987) postve sem-defnte, heteroskedastcty and autocorrelaton consstent varance-covarance matrx as follows: N p N ' ' G Y ' Y 1 e [ ] 1 1 t 1 et ytyt yt yt N K p The related t-statstc becomes: t 1 1 N ' N ' ' N ' Y 1 Y Y 1 e ey Y 1 Y The analyss s based on a research sample of eght selected Ngeran banks to be precse; Frst Bank of Ngera, Zenth Bank, Guaranty Trust Bank and Frst Cty Monument bank,. The ratonale for choosng the abovementoned banks s the fact that the selected banks consttute the foremost promnent banks n recent tme especally when these banks have survved the global economc meltdown, fnancal shock and bank sector reforms of the CBN. The study covered the perod from 007 through to 015. The measurement of varables s as follows: captal adequacy rato was measured as the rato of the sum of Ter-One and Ter-Two to rsk weghted assets; 99

6 management effcency of the banks was measured as the rato of expendture to totalty of ncome; asset qualty of the banks was calculated as the rato of non-performng assets to advances, lqudty was measured as the rato of lqud assets to totalty of assets. Research data were obtaned from twelve-monthly reports and fnancal statements of the abovementoned banks and those of the CBN. The study utlzed the feasble GLS technque to estmate the pooled model havng corrected for heteroskedastcty and seral correlaton. Estmaton and results The correlaton matrx shows that none of the varables n the study exhbted a correlaton coeffcent above 75%. Ths ndeed demses the problem of multcollnearty amongst the regressors. Accordng to the correlaton results, captal adequacy rato s postvely related to proft after tax of the depost money banks as well as the return on assets whle management effcency and nflaton are negatvely related to return on assets and proft after tax of the banks. Table I: Correlaton Matrx Varables Correlaton Matrx for Return on Assets Log(r) Log(m) Log(l) Log(c) Log(a) Log(n) Log(r) 1 Log(m) Log(l) Log(c).698* Log(a) Log(n) -.935* * Correlaton Matrx for Bank Profts after Tax Log(b) Log(m) Log(l) Log(c) Log(a) Log(n) Log(b) 1 Log(m) Log(l).695*.60 1 Log(c).775* Log(a) * Log(n).783* * *ndcates sgnfcance of 5% The feasble GLS regresson estmates are as detaled n Table II. The estmates for return on assets and bank proft after tax. The coeffcent of captal adequacy rato (0.136) s statstcally sgnfcant for return on assets. Ths mples that a 16% percent ncrease n captal adequacy enhances the performance of the depost money banks up to the tune of.176 percent. The estmated captal adequacy rato s below 30%, t ndcates that the bankng system of the Ngeran economy s yet to grow stronger and as such the depost 300

7 money banks mght not be able to fulfl the labltes, operatonal rsk, credt rsk and market rsk. Table II: Feasble GLS Regresson Results Regressors FGLS Estmates for Return on Assets FGLS Estmates for Bank Proft after Tax Coeffcents Probablty Coeffcents Probablty C -.685** (.70) (4.30) Ln(m) (.633) (7.514) Ln(l).350** **.000 (.11) (.59) Ln(c).136** **.004 (.048) (.019) Ln(a) -.56** (.4) (.08) Ln(n) * (.144) ** (.50).000 Adjusted R =.59, DW =.014, F = 06.5 Included observatons: 711 after adjustment Adjusted R =.54, DW =.564, F = 376 Included observatons: 711 after adjustment 1. The values n parentheses are the related Newey-West standard errors. *(**) ndcates statstcal (5%) 3. The varance-covarance matrx for the FGLS model s adjusted for data clusterng The asset qualty coeffcent (-0.56) s sgnfcant n ts adverse effect on Ngeran bank performance. Ths goes to show that the Ngeran banks have been confronted wth the crss of non-performng assets. In other words, the loans offered by the banks have not sgnfcantly yelded postve returns and hence hampered the deeds of these banks. The coeffcent of lqudty (0.50) passes the sgnfcance test n both regressons. Thus, t can be nferred that a 16% ncrease n the banks lqud assets enhances the actvtes of the banks by 4 percent. In effect, the emprcal evdence holds that lqud assets have sgnfcant nfluence on the deeds of banks n Ngera. The coeffcent of effcency (-0.95) s sgnfcant n mpedng the deeds of the Ngeran banks. As t were, 16% rse n the bank effcency reduces the performance of the banks by 14.8 percent. In effect, the emprcal evdence upholds the fact that the bank management s hghly neffcent and by extenson, the depost money banks have low fnancal strength to envelop all operatng expenses from earned ncome. The coeffcents of nflaton rate (-1.058) and (-1.050) are statstcally sgnfcant n slowng down the deeds of Ngeran banks. Ths goes to show 301

8 that that the performance of depost money banks n Ngera have not been favoured by macroeconomc stablty as made evdent n hgh nflaton rate. Evdently, asde bank specfc or nternal factors, external factors that characterzed the Ngeran economy have not been postve effect on the Ngeran banks actvtes. In accordance wth the analyss, varaton n bank performance as measured by return on assets and proftablty after tax payments s to a large extent as a result of the varaton n captal adequacy rato and lqudty. However, the varaton due to management effcency and nflaton do not account postvely for bank performance n Ngera. The presence of seral correlaton was dened by the regresson for return on assets gven a Durbn-Watson test statstc of.014 whch s wthn the acceptable lmt for zero autocorrelaton. Hence, the polcy stance of the study s drven manly by the feasble GLS estmates for return on assets. The standard error test results are gven n Table III. Table III: Sgnfcance Test Based on Newey-West V-C Estmates Regressors S( ) Implcaton Decson 1 S( ) Ln(m) < Insgnfcant Ln(l) > 0.11 Sgnfcant Ln(c) > Sgnfcant Ln(a) > 0.4 Sgnfcant Ln(n) > Sgnfcant Calculated DW (d*) Lower lmt DW (d L) Upper lmt DW (d U) d*<d L and d U Decson No Accept H 0 From the table above, only captal base and lqudty rato of the banks are sgnfcant determnants of the varaton n return on equty at 5% level of sgnfcance, usng a two-tal test. The mplcaton of the result s that as captal base of the banks ncrease, t sgnfcantly ncreases the return on assets. The summary of the t-statstcs of the related Newey-West varancecovarance estmates shows sgnfcant parameter estmates for captal adequacy rato and lqud assets of the depost money banks. Gven that the alternatve hypothess s expressed n the form of 0, a two-tal crtcal regon was utlzed whereby each tal corresponds to half the chosen level of sgnfcance wth the area equal to 5%. The decson rule was to accept the null hypothess f the absolute t- statstc(s) s (are) greater than the estmated t-statstc(s), otherwse, we reject the null hypothess. The test shows that the captal adequacy rato and lqud 30

9 assets are the polcy varables for explanng the varaton n return on assets at fve percent error margn. The result of F dstrbuton test at fve percent shows that the estmated pooled model could stand the test of tme n explanng the varaton n return on assets due to the overrdng effect of the captal adequacy rato and lqud assets of the depost money banks. Usng the d* statstc at fve percent crtcal value, the null hypothess s accepted ndcatng that the varance-covarance estmates are consstent and consequently the resdual varances are serally ndependent. Concluson and polcy stance The study examnes the degree of sgnfcance of the captal adequacy rato n nfluencng the fnancal deeds of Ngeran banks. Gong by the feasble GLS estmates, the coeffcent of asset qualty though sgnfcant, had a negatve mpact on the deeds of Ngeran banks. Ths emprcal evdence could be sgnally the fact that the bankng system n Ngera has an enormous percentage of non-performng loans and ths s a constrant on fnancal performance. Further, the emprcal evdence supports the overrdng effect of captal adequacy rato and lqudty n enhancng the deeds of Ngeran banks. However, the mpact of estmated captal adequacy rato s below 30%. It ndcates that the depost money banks mght not be able to fulfl labltes and rsk. The upshot of the analyss s that depostor s money and the confdence of the bankng sector have not been absolutely assured. In the lght of our fndngs, the polcy stance of the study the followngs: Frst, the Central Bank of Ngera and the regulatory authorty of the bankng system n Ngera should regularly reassess the least captal that s requred of banks n order to enhance ther fnancal deeds n the country. Second, the CBN should be more proactve n mplementng polces that can cushon the rsk management framework of the all the depost money banks n Ngera. Thrd, there s need to mplement macroeconomc polcy of low nflaton as ths could enhance a favourable macroeconomc envronment conducve to fnancal performance n the country. References: Abreu, M. and Mendes, V. (00), Commercal Bank Interest Margns and Proftablty: Evdence from E.U Countres. Workng paper Seres, Porto Athansasoglou, P., Brssms, S. and Dels, M. (006), Bank-specfc, ndustry-specfc and macroeconomc determnants of bank proftablty. Journal of Internatonal Money and Fnance, Vol. 19, No. 6, pp

10 Al-Sabbagh, Noor (004), Determnants of captal adequacy rato n Jordanan and Evdence, Journal of Monetary Economcs, Vol. 3, pp Dore, M. (013), An Emprcal Analyss of Bank Proftablty n Ghana: Evdence from Bank Specfc and Macroeconomc Factors. Eastern Medterranean Unversty (EMU). Ho, Shrley J. and Hsu, Su-Chu (010), Leverage, Performance and Captal Adequacy rato n Tawan s Bankng Industry, Japan and the World Economy, Vol., pp Isk, I., and Hassan, M.K, (003a), Effcency, Ownershp and Market Structure, Corporate Control and Governance n the Turksh Bankng Industry. Journal of Busness Fnance and Accountng, 30, Kosmdou, K. (008). The Determnants of Banks' Profts n Greece durng the Perod of EU Fnancal Integraton. Manageral Fnance 34(3): Mpuga, P. (00), The Bankng crses n Uganda: What was the Role of the New Captal Requrements? Internatonal Journal of Fnancal regulaton and complance, 10(3), Naceur B. S. (003), The determnants of the Tunsan bankng ndustry proftablty: panel Evdence, Paper presented at the Economc Research Forum (ERF) 10th Annual Conference, Marrakesh-Morocco, and December Newey, W. K., and K. D. West (1987), A smple, Postve Sem-defnte, Heteroskedastcty and Autocorrelaton Consstent Covarance Matrx. Econometrca 55: Obamuy, T. M., (011): Incessant Bank Dstress and the Polces of Central Bank of Ngera. Internatonal Journal of Fnance and Accountng, 1(1), Osuka, B.O., and Rchard, O.C., (013): The Determnants of Fnancal Performance of Quoted Banks n Ngera: A Study of Selected Depost Money Banks (DMBs) ( ). Internatonal Journal of Educaton and Research, Vol. 1, No. 10, pp.1-18, October Perera, S., Skully, M., and Chaudrey, Z. (013): Determnants of Commercal Bank Proftablty: South Asan Evdence. Asan Journal of Fnance and Accountng, 5(1), Raz, S., and Mehar, A. (013). The Impact of Bank Specfc and Macroeconomc Indcators 304

11 on the Proftablty of Commercal banks. Romanan Economc Journal, 16(47), Saona, P. H., (011), Determnants of the Proftablty of the US Bankng Industry. Internatonal Journal of Busness and Socal Scence, (), Scott, J. W., and Aras, J. C., (011): Bankng Proftablty Determnants. Busness Intellgence Journal, 4(), Somoye, R.O.C., (008), The performance of commercal banks n postconsoldaton perod n Ngera: An Emprcal Revew. European Journal of Economcs, Fnance and Admnstratve Scences, 14, Sufan, F., (009), Determnants of Bank Proftablty n a Developng Economy: Emprcal Evdence from Bangladesh. Journal of Busness Economcs and Management, 10(3), Tomola, M.O., (013), Determnants of Banks Proftablty n a Developng Economy: Evdence from Ngera. Organzatons and Markets n Emergng Economes, Vol.4, No. (8), pp