Measuring bank efficiency in Ghana: a joint application of parametric and non-parametric techniques

Abstract

This study examines two aspects of bank efficiency in Ghana, namely, levels of efficiency and determinants of efficiency. The Ghanaian banking industry, like those of many sub Saharan African countries, has undergone significant changes since the introduction of financial sector reforms in the late 1980s, all aimed at improving the performance of banks and the efficiency of the financial sector in general. Yet the Ghanaian financial sector is still characterised by relatively low financial depth, a low level of intermediation and issues of weak corporate governance, amongst other things, that continue to affect the efficient operations of banks. This study, therefore, provides evidence on bank efficiency in Ghana by defining and measuring three concepts of efficiency (Technical, Cost and Scale) using the two most popular frontier approaches, namely Stochastic Frontier Analysis (SFA) and Data Envelopment Analysis (DEA), in ways that conform to economic theory so as to provide useful information to policymakers. The study employed an unbalanced panel of 220 observations from 27 Ghanaian banks over the period 2007 to 2016 to examine whether efficiency scores derived from the two frontier techniques yield consistent results according to five criteria, namely, distributions, rankings, identification of best and worst performance, stability over time, and correlation with traditional accounting-based performance measures. The results suggest that parametric and non-parametric techniques neither rank nor identify best and worst practice banks consistently. The results also indicate that neither the efficiency scores generated by SFA nor those generated by DEA are highly correlated with the non frontier accounting-based measures of bank performance in Ghana, suggesting that the latter may contain some exogenous variables that make them poor proxies for measuring efficiency. The SFA approach however yielded efficiency measures that were relatively more correlated with accounting-based measures in bank performance analysis, and more stable over time than efficiency scores generated by the DEA approach. The comparatively higher degree of consistency between SFA and standard accounting-based methods, as well as the fairly stable efficiency scores under the SFA approach compared with that of the DEA approach, makes the SFA approach more realistic and hence more attractive and useful for regulatory purposes in Ghana. The study also investigates the determinants of bank efficiency in Ghana to enhance policy focus. The results indicate that inefficiency in the Ghanaian banking industry is influenced by a number of factors, including bank regulations, competition, bank-specific characteristics, and macroeconomic factors. Although, in some instances, the two frontier techniques presented conflicting relationships between efficiency and the variables enumerated above, the study finds that regulatory intensity exacerbates inefficiency in the Ghanaian banking industry. The study also finds that, while the Lerner index shows a negative relationship between competition and efficiency, supporting the efficient structure hypothesis, the Boone indicator shows a positive relationship between competition and efficiency, supporting the quiet life hypothesis. It is argued that the Lerner index tends to misrepresent the evolution of competition in markets where the number of firms is few with high market concentration. Therefore, considering the fact that four big banks dominate the Ghanaian banking industry, and the fact that the sample size is not that large, the study concludes in favour of the Boone indicator as a more realistic and attractive measure of the relationship between competition and bank efficiency in Ghana. The study also investigates economies of scale and technological progress among Ghanaian banks and finds that the efficiency of Ghanaian banks is related to the exploitation of scale economics and the adoption of new production technologies. The study finds widespread scale economies across different bank categories, with large, foreign and private banks recording larger scale economies compared with their smaller, local and public counterparts respectively. The findings in this study have some implications from a policy standpoint in Ghana. The regulatory authorities need to strike a delicate balance between ensuring a stable banking system and promoting competition in the banking sector by guaranteeing market contestability. This can be done through the facilitation of the entry of well-capitalised banks to break the dominance of the big banks and hence reduce concentration and market power. This can lead to healthy competition, and hence bank efficiency, without jeopardising the stability of the banking system. The negative relationship between credit risk and bank efficiency also implies that the Bank of Ghana can promote bank efficiency and financial inclusion by strengthening its supervision and monitoring of bank credit to reduce credit risk.

Hierdie studie het twee aspekte van bankdoeltreffendheid in Ghana ondersoek, naamlik die vlakke van doeltreffendheid en die determinante van doeltreffendheid. Die Ghanese bankbedryf, soos heelwat ander subSahara-Afrika-lande, het betekenisvolle veranderinge sedert die bekendstelling van hervormings van die finansiële sektor in die laat 1980’s ondergaan, wat daarop gemik was om die prestasie van banke en die doeltreffendheid van die finansiële sektor oor die algemeen te verbeter. Nietemin word die Ghanese finansiële sektor steeds deur relatief lae finansiële diepte gekenmerk, asook onder meer ’n lae vlak van bemiddeling en kwessies van swak korporatiewe beheer wat voortgaan om die doeltreffende bedrywighede van banke te beïnvloed. Hierdie studie bied dus bewyse van bankdoeltreffendheid in Ghana deur drie konsepte van doeltreffendheid (tegnies, koste en skaal) te omskryf en te meet deur die twee gewildste grensbenaderings te gebruik, naamlik stochastiese grensanalise (Stochastic Frontier Analysis – SFA) en data-ontwikkelingsanalise (Data Envelopment Analysis – DEA), op wyses wat deur ekonomiese teorie ondersteun word ten einde beleidmakers van nuttige inligting te voorsien. Die studie het ’n ongebalanseerde paneel van 220 observasies van 27 Ghanese banke vir die tydperk 2007 tot 2016 gebruik om vas te stel of doeltreffendheidstellings afkomstig van die twee grenstegnieke konsekwente resultate oplewer ooreenkomstig vyf kriteria, naamlik verspreidings, rangordes, die identifikasie van beste en slegste prestasie, stabiliteit oor tyd en korrelasie met tradisionele rekeningkundig-gebaseerde prestasiemaatreëls. Die resultate het aangedui dat parametriese en nieparametriese tegnieke nie beste en slegste praktykbanke konsekwent kan klassifiseer nie; asook dat nòg die doeltreffendheidtellings wat deur stochastiese grensanalise genereer is nòg die wat deur data-ontwikkelingsanalise gegenereer is, met die niegrens- rekeningkundig-gebaseerde maatreëls van bankprestasie in Ghana in ’n groot mate gekorreleer kan word, wat daarop wys dat die laasgenoemde ’n mate van eksogene veranderlikes kan bevat wat hulle swak aanduiders vir metingsdoeleindes maak. Die stochastiesegrensanalisebenadering het egter doeltreffendheidsmaatreëls opgelewer wat relatief meer met rekeningkundig-gebaseerde maatreëls van bankdoeltreffendheidanalise korreleer en oor tyd meer stabiel was as doeltreffendheidstellings wat deur die data-ontwikkelingsanalisebenadering gegenereer is.

Die betreklike hoër graad van konsekwentheid tussen stochastiesegrensanalise- en standaard rekeningkundig-gebaseerde metodes, asook die redelike stabiele doeltreffendheidstellings van die stochastiesegrensanalisebenadering in vergelyking met die data ontwikkelingsanalisebenadering, dui daarop dat die stochastiesegrensanalisebenadering meer realisties en derhalwe aantrekliker en nuttiger is vir regulatoriese doeleindes in Ghana. Die studie het ook die determinante van bankdoeltreffendheid in Ghana omdersoek ten einde die beleidsfokus te verbeter. Die resultate het aangetoon dat ondoeltreffendheid in die Ghanese bankbedryf deur ’n aantal faktore beïnvloed word, insluitend bankregulasies, mededinging, bankspesifieke eienskappe en makroekonomiese faktore. Hoewel die twee grenstegnieke in sommige gevalle teenstrydige verhoudings tussen doeltreffendheid en die veranderlikes wat hierbo genoem is opgelewer het, het die studie bevind dat regulatiewe intensiteit ondoeltreffendheid in die Ghanese bankbedryf vererger. Die studie het ook bevind dat alhoewel die Lerner-indeks ’n negatiewe verhouding tussen mededinging en doeltreffendheid wys wat die doeltreffendheidstruktuurhipotese ondersteun, die Boone aanwyser ’n positiewe verhouding tussen mededinging en doeltreffendheid aantoon, wat die quiet life-hipotese ondersteun. Daar word aangevoer dat die Lerner-indeks neig om die evolusie van mededinging in markte, waar die aantal firmas min met hoë markkonsentrasie is, wan voor te stel. Met die inagneming van die feit dat vier groot banke die Ghanese bankbedryf oorheers, asook die feit dat die steekproefgrootte nie baie groot was nie, het die studie derhalwe tot die slotsom gekom dat die Boone-aanwyser ’n meer realistiese en aantreklike maatstaf is van die verhouding tussen mededinging en bankdoeltreffendheid in Ghana. Die studie het ook die skaalvoordele en tegnologiese vordering van Ghanese banke ondersoek en bevind dat die doeltreffendheid van Ghanese banke met die uitbuiting van skaalvoordele en die aanneem van nuwe produksietegnologieë verband hou. Die studie het wydverspreide skaalvoordele in verskillende bankkategorieë uitgewys, waar groot, buitelandse en privaat banke groter skaalvoordele noteer in vergelyking met hulle kleiner, plaaslike en openbare eweknieë. Vanuit ’n beleidstandpunt in Ghana hou die bevindings van hierdie studie sekere implikasies in. Die regulerende owerheid moet ’n fyn balans handhaaf tussen die versekering van ’n stabiele bankstelsel en die bevordering van mededinging in die banksektor deur markbetwisbaarheid te waarborg. Dit kan gedoen word deur die fasilitering van die toetrede van goed gekapitaliseerde banke om die dominansie van die groot banke te verswak en sodoende die konsentrasie en mag van die mark te verminder. Dit kan tot gesonde mededinging lei en uiteindelik tot bankdoetreffendheid sonder om die stabiliteit van die bankstelsel in gevaar te stel. Die negatiewe verhouding tussen kredietrisiko en bankdoeltreffendheid beteken ook dat die Bank van Ghana bankdoeltreffendheid en finansiële insluiting kan bevorder deur sy toesighouding en monitering van bankkrediet te versterk om kredietrisiko te verminder.

Lolu cwaningo lubheka imikhakha emibili yokusebenza kahle kwamabhange eGhana, okungukuthi, amazinga okusebenza kahle kanye nezincazelo zokwenza kahle. Imboni yamabhange yaseGhana, njengaleyo yamazwe amaningi asendaweni yezwekazi lase-Afrika eliseningizimu yeSahara., yenze ushintsho olukhulu selokhu kwethulwa izinguquko emkhakheni wezezimali ngasekupheleni kweminyaka yama-1980, konke lokhu kuhlose ukwenza ngcono ukusebenza kwamabhange nokusebenza kahle komkhakha wezezimali jikelele. Kodwa-ke umkhakha wezezimali waseGhana usabonakala ngokujula okuphansi kwezezimali, izinga eliphansi lokulamula kanye nezinkinga zokuphathwa kwezinkampani ezibuthakathaka, phakathi kokunye, okuqhubeka nokuthinta ukusebenza kahle kwamabhange. Lolu cwaningo, ngakho-ke, lunikeza ubufakazi ekusebenzeni kahle kwebhange eGhana ngokuchaza nokulinganisa imiqondo emithathu yokusebenza kahle (Ezobuchwepheshe, Izindleko kanye neSikali) kusetshenziswa izindlela ezimbili ezaziwa kakhulu zemingcele, okuyindlela yesifanekiso kwezomnotho (SFA) nendlela engabandakanyi noma ikuphi ukucatshangelwa mayelana nokusatshalaliswa kokuvamisile ocwaningweni lokusebenza nakwezomnotho ngokulinganiselwa kwemingcele yokukhiqiza (DEA), ngezindlela ezihambisana nombono wezomnotho ukuze unikeze imininingwane ewusizo kubenzi bezinqubomgomo. Ucwaningo lusebenzise iqembu elincane elingalingani labantu abangama – 200 bokuhlola abakhethwe ukweluleka, bathathe isinqumo elivela emabhange angama-27 aseGhana esikhathini sonyaka ka-2007 kuya onyakeni ka-2016 ukubheka ukuthi ngabe imiphumela yokusebenza kahle etholakala kulezi zindlela ezimbili zemingcele iveza imiphumela engaguquguquki ngokuya ngemigomo emihlanu, okungukuthi, ukwabiwa, amazinga, ukukhonjwa kokusebenza okuhle nokubi kakhulu, ukuzinza ngokuhamba kwesikhathi, kanye nokuhambisana nezindlela zokusebenza zendabuko ezisuselwa ekulinganiseni. Imiphumela iphakamisa ukuthi amasu okucabanga enani ngokwezinombolo ngenhloso yokuhlaziya kanye nawangabandakanyi noma ikuphi ukucatshangelwa mayelana nokusatshalaliswa kokuvamisile awaphakamisi noma akhombe amabhange asebenza kahle kakhulu futhi amabi kakhulu ngokungaguquguquki. Imiphumela ibuye angakuvezi ukuthi amaphuzu okusebenza kahle akhiqizwe yi-SFA noma lawo akhiqizwe yi-DEA ahlobene kakhulu nezinyathelo ezingezona ezasemgceleni ezincike ekubhekeni kwezimali ezenziwa yibhange eGhana, okuphakamisa ukuthi lokhu okwedlule kungaqukatha ukuguquguquka okumbalwa okubenza babe ngabenegunya elibi lokumela omunye umuntu, ikakhulukazi ekuvoteni ukulinganisa ukusebenza kahle.The Indlela ye-SFA nokho iveze izinyathelo zokusebenza kahle ebezihambisana kakhulu nezindlela ezincike ekubhekeni kwezimali ekuhlaziyweni kokusebenza kwasebhange, futhi ezinze ngokuhamba kwesikhathi kunamanani okusebenza kahle akhiqizwa indlela ye-DEA. Izinga eliphezulu ngokuqhathaniswa lokuvumelana phakathi kwe-SFA nezindlela ezijwayelekile ezisuselwa ekubalweni kwemali, kanye nezilinganiso ezisebenza kahle ngaphansi kwendlela ye-SFA uma kuqhathaniswa naleyo yendlela ye-DEA, yenza indlela ye-SFA ibe ngokoqobo ngokwengeziwe futhi ngenxa yalokho ihehe futhi ibe wusizo ngezinjongo zokulawula eGhana. Ucwaningo luphinde luphenya nezimpawu zokusebenza kahle kwamabhange eGhana ukuthuthukisa ukugxila kwenqubomgomo. Imiphumela ikhombisa ukuthi ukungasebenzi kahle embonini yamabhange aseGhana kuthonywa yizici eziningi, kufaka phakathi imithethonqubo yasebhange, ukuncintisana, izici eziqondene namabhange, kanye nezici ze zomnotho jikelele. Yize, kwezinye izimo, izindlela zombili zemingcele ziveza ubudlelwano obuphikisanayo phakathi kokusebenza kahle kanye nokuguquguqukayo okubalulwe ngenhla, ucwaningo luthola ukuthi amandla okulawula akhulisa ukungasebenzi kahle embonini yamabhange aseGhana. Ucwaningo luthola nokuthi, ngenkathi inkomba yeLerner ikhombisa ubudlelwano obungebuhle phakathi kokuncintisana nokusebenza kahle, kusekela isakhiwo esisebenzayo sesiphakamiso esenziwe njengesisekelo sokucabanga, ngaphandle kokucabanga

okweqiniso laso, inkomba yeBoone ikhombisa ubudlelwane obuhle phakathi kokuncintisana nokusebenza kahle, kusekela impilo yokuthula yesiphakamiso esenziwe njengesisekelo sokucabanga, ngaphandle kokucabanga okweqiniso laso. Kuthiwa inkomba yeLerner ithambekele ekumeleleni kabi ukuvela komncintiswano ezimakethe lapho inani lamafemu lincane ngokugxila emakethe ephezulu. Ngakho-ke, ngokubheka iqiniso lokuthi amabhange amane amakhulu aphethe imboni yamabhange aseGhana, nokuthi ubukhulu besampula abubukhulu kangako, ucwaningo luphetha ngokuvumela inkomba yeBoone njengesilinganiso esingokoqobo nesikhangayo sobudlelwano phakathi komncintiswano nokusebenza kahle kwebhange eGhana. Ucwaningo luphenya nezinga lezomnotho kanye nenqubekela phambili kwezobuchwepheshe phakathi kwamabhange aseGhana futhi luthola ukuthi ukusebenza kahle kwamabhange aseGhana kuhlobene nokuxhashazwa kwezinga lezomnotho nokwamukelwa kobuchwepheshe obusha bokukhiqiza. Ucwaningo luthola umnotho wezinga elibanzi emikhakheni eyehlukene yamabhange, ngamabhange amakhulu, angaphandle nawangasese aqopha umnotho wezinga elikhulu uma kuqhathaniswa nozakwabo abancane, basekhaya kanye nabomphakathi ngokulandelana Okutholakele kulolu cwaningo kunomthelela othile kumbono wenqubomgomo eGhana. Iziphathimandla ezilawulayo zidinga ukufeza isimo esibucayi lapho izinto ezehlukene zenzeka ngamanani alinganayo noma afanele phakathi kokuqinisekisa uhlelo oluzinzile lwamabhange kanye nokukhuthaza ukuncintisana emkhakheni wamabhange ngokuqinisekisa ukuncintisana kwemakethe. Lokhu kungenziwa ngokwenza lula ukungena kwamabhange anemali eningi ukuze kunqandwe ukubusa kwamabhange amakhulu futhi ngalokho kunciphise amandla okuxakaniseka namandla emakethe. Lokhu kungaholela ekuncintisaneni okunempilo, ngakho-ke ukusebenza kahle kwebhange, ngaphandle kokubeka engozini ukuzinza kohlelo lokubhanga. Ubudlelwano obubi phakathi kobungozi besikweletu nokusebenza kahle kwebhange busho nokuthi iBhange laseGhana lingakhuthaza ukusebenza kahle kwebhange nokufakwa kwezezimali ngokuqinisa ukugadwa kwalo nokuqapha isikweletu sasebhange ukunciphisa ubungozi besikweletu.