1.1 Background to the Study
The precise link and direction of causation between financial development and economic growth has remained at the centre of empirical debates for decades. The debate arguably gathered momentum with the empirical works of King and Levine (1993) who, in a cross country study comprising data from 77 countries over the period 1960-1989, found that the level of financial development stimulates economic growth. Deidda and Fattouh (2002) with the same data but a threshold regression confirm the positive relationship between the level of financial depth and economic growth for countries with high income per capita but no significant relationship for lower-income countries, which is consistent with the non-monotonic relationship implied in the model.
Again, Rousseau and Sylla (2001) in their cross-country study covering 17 countries over the period 1850-1997 also find evidence of a leading role for finance. Their result was further supported by Rousseau and Wachtel (1998) who, examining the links between the financial and real sectors for five countries that underwent rapid industrialization over the 1870-1929 period, are able to confirm that financial intermediation Granger-cause real output, especially before the Great Depression, with little evidence of feedback from output to intermediation.
Allesandra (2010) has argued that the strongest critique to all these studies comes from Arestis and Demetriades (1997). The authors, using King and Levine’s (1993:3) data underline that the question of causality cannot be satisfactorily addressed in a cross-section framework. More specifically, they conclude that:
…we have warned against the over-simplified nature of results obtained from cross-country regressions in that they may not accurately reflect individual country circumstances such as the institutional structure of the financial system, the policy regime and the degree of effective governance. The econometric evidence we have reviewed using time-series estimations on individual countries suggests that the results exhibit substantial variation across countries, even when the same variables and estimation methods are used. Thus, the ‘average’ country for which cross-country regressions must, presumably, relate to may well not exist.(Allesandra,2010:2)
Some scholars have also approached the subject from the perspective of time series in a bid to find a common ground of consensus but here also, the results have been contentious. For instance, Harrison, Sussman and Zeira (1999) using a panel of data for 48 US states from 1982-1994, find a feedback effect between the real and the financial sector that helps to explain intra-national differences in output per capita. Luintel and Khan (1999) using the VAR technique on 10 developing countries with yearly data from the 1950s to the mid-1990s find two co-integrating vectors identified as long-run financial depth and output relationship linking financial development to economic development. They also find causality between the level of financial development (depth) and growth in per capita income in all sample countries. This confirms the findings of Demetriades and Hussein (1996) who, with data on 16 developing countries, with 30 to 40 yearly observations from the 1960s, find that in most countries evidence favours bi-directional causality and in quite a few countries economic growth systematically causes financial development.
Also Shan, Morris and Sun (2001), using quarterly data from the mid-70s to 90s for 9 OECD countries, find evidence of reverse causality, namely from growth to financial development, in some countries and bi-directional causality in others, but no evidence of one-way causality from financial development to growth.
Allessandra (2010) further argued the fact that many time-series studies yield unreliable results due to the short time spans of typical data sets cannot be ignored. It was for this reason that Christopoulos and Tsionas (2004) analyze 10 developing countries but resorted to a panel context that increases the sample size. With panel unit root tests and panel co-integration analysis the authors find a single a unique co-integrating vector, implying one-way causality from financial development to economic growth. From the foregoing, it seems that despite works on the contrary, there is a broad consensus that financial development spurs economic growth.
1.2 Statement of the Problem
Economic growth has long been considered an important goal of economic policy with a substantial body of research dedicated to explaining how this goal can be achieved. One of the earliest works on banking performance and economic growth was by Schumpeter (1959) who argued that financial (banking) services are paramount in promoting economic growth. In his view production requires credit to materialize and one can only become an entrepreneur by previously becoming a debtor. What the entrepreneur first wants is credit. The entrepreneur according to Schumpeter, is the typical debtor in a capitalist society.
Based on this strong background laid by Schumpeter, a lot of empirical works have been conducted especially in advanced economies to ascertain the relationship between banking sector performance and economic growth. Most of these empirical studies focused on explanatory variables selected on the basis of their relevance to policymakers or because of other theoretical predictions (see for instance, Barro, 1991; Levine and Renelt, 1992). Indeed, it could be said that empirical literature/works on the purported relationship between banking sector performance and economic growth is broad in advanced economies; transition economies of Central and Eastern Europe and the Baltics.
In Nigeria, empirical works that focused explicitly on banking sector performance and economic growth have yielded mixed results. Some of these works suggest that banking sector performance has impacted positively and significantly on economic growth (see; Adelakun,2010) while others reported an insignificant relationship between banking sector performance and economic growth (see. Ekpeyong & Acha,2011; Odeniran & Udeaja,2010 ). A major problem in these works are the authors’ selection of explanatory variables that do not explicitly underpin banking sector performance. An example is Balogun’s (2007) work on banking industry performance and the Nigerian economy where bank branches were used as one of the explanatory variables in his modelling. Given multiple channels of accessing banking services such as internet banking; telephone banking; mobile banking; and use of automated teller machines and point of sale machines; the relevance of the number of bank branches as a determinant of economic growth is clearly uncertain. Therefore, a case can be made for a more robust empirical modeling with variables that are more broad based and that underpin actual banking performance. Ayadi et., al. (2013) also suggest that financial development has been intensively studied in developed countries, with result indicating a strong and positive relationship between growth and financial sector development. They also affirm that studies in developing countries are sparse and where they exist, tend to support a negative and insignificant relationship between banking sector performance and economic growth. Given the foregoing, there still exist a research gap for an empirical evaluation of the impact of banking sector performance on economic growth using more robust and broad based explanatory variables.