Banking Sector Intermediation Development and Growth of a Developing Economy: An Empirical Investigation of Nigeria

In developing economies, banks act as a conduit for the ecient mobilization of nancial resources from the surplus sectors for effective allocation to the decit sectors for productive investment that will in turn lead to economic growth. Thus, the study is aimed at evaluating whether development in the banking sector intermediation process in the form of increase in the number of branches, credit to private sectors, intermediation eciency and total assets stimulates economic growth in Nigeria during the period of 1987 to 2018. The study employed the Johansen cointegration test, dynamic ordinary least square (DOLS) regression and error correction model in determining the relationship between the variables. The results of the cointegration test conrmed the existence of long-run relationship between banking sector development indicators and economic growth in Nigeria. Whereas, in the short run, only number of bank branches and bank’s total asset have a positive and signicant impact on economic growth signifying that much of Nigeria’s superior growth performance is attributed to increase in the number of bank branches and growth in bank’s assets. Credit to private sector has negative and insignicant relationship with economic growth while bank’s intermediation eciency has positive and insignicant relationship with economic growth.

Also, most banks in Nigeria have concentrated in opening up branches in the urban areas at the expense of the rural areas. Hence, huge amount of cash are lying idle in the rural areas and thus, being left out of the banking stream (Azolibe, 2019). These idle funds are supposed to have been mobilized by banks and channeled to the de cit sector of the economy for productive investment that will in turn drive economic growth.
Empirically, the relationship between banking and nancial sector intermediation development has been widely investigated in most developing and developed countries of the world. Limited evidence however exist in Nigeria and also, the few empirical studies conducted in Nigeria and other developing countries such as Ayunku and Etale (2014), Nwaeze, Michael and Nwabekee (2014), Iwedi, Okey-Nwala, Kenn-Ndubuisi and Adamgbo (2016), Murari (2017), Petkouski and Kjosevski (2014), Awdeh (2012) have concentrated mainly on using bank deposit, credit to private sector, interest rate and money supply as a measure of banking sector intermediation development. However, this study will improve on the variables employed by previous studies by using growth in the number of bank branches, intermediation e ciency and bank's total assets as a measure of banking sector intermediation development. Thus, this study is aimed at determining how banking sector intermediation development measured by the growth in the number of bank branches, growth in credit to private sector, growth in intermediation e ciency and growth in bank's total asset contributes to economic growth in Nigeria.
The rest of the article is structured as follows. The second section covers the Banking sector intermediation development indicators. The third section covers the theoretical and empirical framework. The fourth section covers analytical model, estimation strategy and analysis and discussion of results and nally, the Conclusion and recommendations are drawn in the fth section.

Banking Sector Intermediation Development Indicators
The banking sector intermediation development is measured by growth in the number of bank branches, growth in credit to private sector, growth in intermediation e ciency and growth in total asset in this study. They are discussed as follows;

Growth in the Number of Bank Branches
The banking sector in Nigeria has witnessed growth in terms of spread in the number of branches. According to CBN (2018), the number of bank branches, increased to 5,714 in 2017 from 5,570 in 2016. Before the introduction of the rural banking scheme in 1977, most banks in Nigeria concentrated mainly in opening up branches in the urban areas. However, Government through the rural banking scheme had made it compulsory for banks to open up branches in the rural areas. Although, most Deposit Money Banks complied with these government directives, they did so reluctantly and are still doing so in recent time. Generally, rural branches were considered unpro table by Deposit Money Banks for reasons ranging from heavy capital outlay, due to lack of infrastructure in these rural areas, to inadequate manpower to meet the needs of these rural branches because of mass expansion.
In Nigeria, the banking system credit to private sectors has increased tremendously right from the period of the post structural adjustment programme (SAP) which led to an increase in the number of banks. Credit to private sector increased from N21.08 billion in 1987 to N22,521.93 billion in 2018. However, the capacity of the banking system to nance the economy declined as the ratio of credit to private sector to GDP fell from 8.53 per cent in 1988, to 6.71 per cent in 1990 and from 10.10 per cent in 1993, to 6.22 per cent in 1995 and then in recent time, the ratio of credit to private sector to GDP fell from 20.77 per cent in 2016, to 17.63 per cent in 2018 (CBN, 2018) as shown in table 3 below. Banking sector intermediation e ciency is measured by the ratio of currency outside banks to broad money supply. Currency outside banks refers to those currencies that are physically used to conduct transactions between consumers and businesses and which are not stored in the bank, nancial institution or central bank. Broad money supply includes bank money and any cash held in easily accessible accounts. The current measure of broad money supply in Nigeria is M3. However, banking sector intermediation e ciency is the ease at which funds are transferred from the surplus sectors of the economy to the de cit sectors of the economy for productive investment that will spur economic growth.  (2018) observes that there was an improvement in intermediation e ciency indicator, measured by the ratio of currency outside banks to broad money supply, which stood at 7.63 per cent in 2018, from 7.97 per cent at end-December 2017.

Growth in Total Assets
A typical bank's asset consists of all forms of personal and commercial loans, mortgages and securities. Central Bank of Nigeria grouped Deposit Money Bank's assets into the following categories: reserves, claims on Central Bank, foreign assets, claims on Central Government, claims on state and local Government, claims on private sector, nancial derivatives and unclassi ed assets.
However, bank's total assets in Nigeria increased from N49.83 billion in 1987 to N17,522.86 billion in 2009. It dropped slightly to N17,331.56 billion in 2010 and then rose sharply to N37,206.99 billion in 2018.

Banking Sector Intermediation Development and Economic Growth: Theoretical and Empirical Framework
A well-structured banking sector mostly decreases costs of transactions and constraints of credits, conditions that may delay the growth of economy in a country. A banking sector that is not effective can by its ineffectiveness result in little activity and economic growth. The banking sector does propel economic growth through its investment function. Banks invest its excess deposit on real investment such as real estate, partnering with private individuals in real production of goods, procuring and leasing of equipment etc.
The theory nancial intermediation that was rst put forth by Schumpeter (1911) and later supported by the works of Shaw (1973), McKinnon (1973), Gupta (1984), Fry (1988), Greenwood and Jovanovic (1990), Bencivenga and Smith (1991) among others, postulates that nancial expansion causes the economy to grow. The theory posits that a strong developed sector of nance facilitates vital services that reduce transaction, information and monitoring costs and enhance the effectiveness of intermediation. As such, it identi es and funds good business projects, mobilizes savings, enables trading and risks diversi cation, promotes exchange of services and goods, monitors the performance of managers. All these services result in effective allotment of resources; lead to a quick increase of human and physical capital; and enables faster technological innovation. This eventually brings the outcome into faster and long-term economic growth (Schumpeter, 1911). Previous empirical studies in Nigeria and in other countries of the world have found a positive relationship between nancial and banking sector development and economic growth. Some found a negative relationship. Ayunku and Etale (2014) examined the relationship between banking sector development and economic growth in Nigeria. The result of the study revealed that trade openness, domestic credit and interest rate have positive relationship with real GDP while credit to private sector and deposit liabilities have negative relationship with real GDP. The result of their study lends very strong support to the existence of a short and long-run relationship between banking sector development and economic growth in Nigeria. Similarly, Tripathy and Pradhan (2014) extended the study to the Indian economy by investigating the short run as well as long-run relationships and also the causality relationships between banking sector development and the economic growth between the period of 1960 and 2011. They found strong evidence that banking sector development caused economic growth in the Indian economy. In another related study, Aigbovo and Uwubamwen (2014) included the stock market segment in their study to examine the short-run and long-run relationships between nancial system development and economic growth in Nigeria. The Granger causality test was used to determine the direction of causality among the variables. The ndings of the study were that nancial development (measured by banking system and stock market development) positively in uenced economic growth in Nigeria; that causality runs from nance to growth in the nance-growth nexus.
In a cross country study, Petkouski and Kjosevski (2014) examined the relation between banking sector development and economic growth in 16 transitional economies from Central and South Eastern Europe and they showed that credit to the private sector and interest margin were negatively related to the economic growth while ratio of quasi money was positively related to economic growth. In the same vein, Murari (2017) explored the relationship between nancial development and economic growth, using a panel data of South Asian middleincome countries for the years 1980-2013. The results indicate that the domestic credit provided by the banking sector has a signi cant association with economic growth in both directions but domestic credit to the private sector is associated with the economic growth in forward direction only, which con rms dearth in credit allocation in the region and suggests pathetic nancial regulation and supervision.
In a more recent study, Constantinos, Sofoklis and Joseph (2018) assessed the relationship between the nancial sector and economic growth in 34 European and commonwealth of independent states economies. Their results suggested that there has been a link between nancial sector and the real economy.
Dritsakis and Adamopoulos (2001) empirically examined the causal relationship between the degree of openness of an economy, nancial development and economic growth by using a multivariate autoregressive VAR model in Greece for the period of 1960 quarter one to 2000 quarter four. The granger causality tests based on error correction models show that there is a causal relationship between nancial development and economic growth and also between the degree of openness of the economy and economic growth. Keho (2010) interacted in ation rate in their model to analyze whether the strength of the relationship between nance and growth depends on in ation rate using time series data of seven African countries. The empirical ndings did not provide signi cant evidence of nonlinearity in the nance-growth relationship. Financial development has no signi cant effect on economic growth regardless of the level of in ation. Also, Kyophilavong, Uddin, and Shahbaz (2016) examined the nexus between nancial development and economic growth by testing the supplyleading hypothesis and demand-following hypothesis using time series data from Lao PDR. Using the autoregressive distributed lag (ARDL) bound testing approach to cointegration, the result con rmed the feedback effect between both of the variables. Financial development promotes economic growth and as a result, economic growth leads to nancial development. Panicos and Siong (2006) evaluated the relationship between nance, institutions and economic development using data from 72 countries for the period 1978-2000. They found that nancial development has larger effects on GDP per capita when the nancial system is embedded within a sound institutional framework. Moreover, they found that nancial development is most potent in middleincome countries, where its effects are particularly large when institutional quality is high. Importantly, they also found that in low-income countries, the in uence of nancial development is at its weakest; in these countries, more nance without sound institutions may not succeed in delivering long-run economic bene ts. Moreover, the study conducted by Malarvizhi, Zeynali, Mamun and Ahmad (2018) on the relationship between nancial sector development and economic growth of ASEAN-5 countries (Malaysia, Indonesia, Singapore, Thailand and Philippines) from 1980 to 2011 also revealed that nancial development has a signi cant positive effect on economic growth. Awdeh (2012) studied the causality direction between banking sector development and economic growth in Lebanon over the period 1992-2011 and found a one way causality running from economic growth to banking sector measures such as deposit growth and credit to local private sector. In another causality approach, Helmi, Rashid and Bedri (2014) investigated the causal relationship between nancial development and economic growth in Gulf Cooperation Council (GCC) countries that is, Bahrain, Oman, Kuwait, Qatar, United Arab Emirates and Saudi Arabia over the period of 1980 to 2012. Using error correction model and cointegration techniques to detect the long-run and short-run causalities between the variables, their overall empirical results revealed that nancial sector development contributes signi cantly to economic growth in GCC countries. King and Levine (1993a) studied a sample of 80 countries and concluded a strong positive relation between nancial development and economic growth. Also, King and Levine (1993b) studied a sample of 70 countries and examined the impact of nancial development on economic growth, capital accumulation and economic factor productivity and found a strong link between nancial development and growth. Levine, Loayza and Beck (2000) evaluated the role of nancial development in motivating economic growth and found that higher banking sector development implies higher economic growth and total factor productivity growth. Levine (2005) studied the in uence of bank system development and stock markets on economic growth by sampling eleven Arab countries, he concluded that countries with undeveloped nancial systems affects negatively the economic growth of those countries. He stressed that a sound nancial system is vital. Hence Levine studied a negative link between banking development and growth of economy due to an inadequately developed nancial banking system. However, from the above empirical literature, credit to private sector, interest rate and bank deposit has been commonly used by most scholars as a measure of banking sector development. Our study is unique and contributes to the empirical debates by introducing new variables such as number of bank branches, bank's total assets and bank's intermediation e ciency as a measure of banking sector development.

Analytical Model
The regression model that was used in this study comprised of one dependent variable, four independent variables and three control variables. Data for the variables were compiled from World Bank, World Development Indicators and Central Bank of Nigeria (CBN) statistical bulletin 2018. Dependent variable was economic growth proxied by real Gross domestic product (RGDP). While variables that were independent include: number of bank branches, ratio of credit to private sector to GDP (CPS/GDP), banking sector intermediation e ciency proxied by ratio of currency outside bank to broad money supply (COB/M2) and ratio of bank's total assets to GDP (BTA/GDP).
The control variables include lending interest rate, export and Government expenditure. The lending rate refers to the rate at which Commercial banks grant loans to the productive sector of the economy. Higher bank lending rate discourages people from borrowing for investment purposes and will in turn reduce the level of investment and economic growth and vice versa. Commercial bank lending rate to the productive sectors is usually in uenced by the Central Bank rate. Export represents goods or services that are sold abroad. Export trade is an instrument for growth as it increases foreign exchange earnings, improves balance of payment position, creates employment and development of export oriented industries in the manufacturing sector and improves government revenue through taxes, levies and tariffs. These bene ts will in turn enhance the process of growth and development in such economy. Finally, Government expenditure on infrastructure such as road, transport, defense power supply, health etc. will increase investment and economic growth.
However, the regression model is been shown as follows: Y=α+β1 X1+β2 X2+β3 X3+ β4 X4 + β5 X5 + β6 X6 + β7 X7+ Transforming the above equation into natural logarithm, we have LogY=α+β1 LogX1+β2 LogX2+β3 LogX3+ β4 LogX4 + β5 LogX5 + β6 LogX6 + β7 LogX7+ Where: LogY= Log of Economic growth which was measured by the real GDP, = Error term normally distributed about the mean of Zero and β1, 2, 3, 4, 5, 6, 7 = Regression coe cients of the variations to determine the volatility of each variable to economic growth in the regression model. The variables were logged because most time series data have an unequal variance (heteroskedastic in nature), thus the natural logarithm helps to stabilize the variance within the sample, which helps to improve our analysis.

Population and Sample Size
Population according to Onwumere (2005), represents a universe or elements with similar characteristics, hence it is a census of all relevant elements and may be nite or in nite while a sample is a group of variables or items derived from a relevant population for the purpose of examination or analysis. Based on this, the population of this study will comprise of all banks in Nigeria such as Deposit Money Banks, Micro nance banks, Merchant banks and Non-interest banks. These banks are involved in nancial intermediation in one form or the other. Sequel to the fact that there may be obvious di culties in studying the entire population due to the pattern and size of distribution, su cient knowledge of the entire population will be gotten from studying a sample of the population.
The sample of this study shall be the Deposit Money Banks in Nigeria. The choice of these banks is based on the fact that they are the dominant institution of nancial intermediation in Nigeria and hence, holds the largest proportion of household savings and also based on the availability of data on the variables. The sample size will also cover a period of 1987 to 2018. The choice of this period is based the fact that it was the period of the post structural adjustment programme (SAP) during which the nancial sector was deregulated and thus paved way for the entrance of new banks. There were more banks intermediating between the surplus sectors and the de cit sectors.

Estimation Procedure
In the determination of the relationship between banking sector intermediation development and economic growth in Nigeria, the study employed dynamic ordinary least square (DOLS) regression analysis. The DOLS model is a robust single equation approach which corrects for endogeneity and correlation by including lags and leads of rst-difference variables. Descriptive statistics of the variables were presented and unit root test was conducted to ensure that the time series analysis is free from stationarity defects. Also, Johansen cointegration test and error correction model were estimated. In the event that variables in the model exhibit a long-run harmonious relationship (cointegrated), it may be necessary to determine the speed of adjustment of economic towards a steady state in response to systemic change. From the table above, the time series residual variable data of all the variables are normally distributed as the probability of the Jarque-Bera statistic is absolutely greater than the critical value of 0.05 hence the null hypothesis (H0) is rejected in favour of the alternative (H1) that the residual of the distribution of the model is normally distributed.

Unit Root Test
In statistics, a unit root test tests whether a time series variable is non-stationary and possesses a unit root. The null hypothesis is de ned as the presence of a unit root and the alternative hypothesis is de ned as stationarity or variables have no unit root. The present study uses augmented Dickey Fuller (ADF) unit root test to examine the stationarity of the data series. It consists of running a regression of the rst difference of the series against the series lagged once, lagged difference terms and optionally, a constant and a time trend. This can be expressed as follows: The additional lagged terms are included to ensure that the errors are uncorrelated. In this ADF procedure, the test for a unit root is conducted on the coe cient of Y t-1 in the regression. If the coe cient is signi cantly different from zero, then the hypothesis that Y t contains a unit root is rejected. Rejection of the null hypothesis implies stationarity. Precisely, the null hypothesis is that the variable Y t is a non-stationary series (H 0 : α 2 = 0) and is rejected when α 2 is signi cantly negative (H a : α 2 < 0). If the calculated value of ADF statistic is higher than McKinnon's critical values, then the null hypothesis (H 0 ) is not rejected and the series is non-stationary or not integrated of order zero, I(0). Alternatively, rejection of the null hypothesis implies stationarity. Failure to reject the null hypothesis leads to conducting the test on the difference of the series, so further differencing is conducted until stationarity is reached and the null hypothesis is rejected.
If the time series (variables) are non-stationary in their levels, they can be integrated with I(1), when their rst differences are stationary. From the above unit root test result, all variables were found to be stationary at rst difference and as such we can go on and use the variables for further analysis. The optimal lag for ADF test is selected based on the Akaike info Criteria (AIC).

Cointegration Test
This study employs VAR based approach of Johasen and Juselius (1990) test which proposes the use of two likelihood ratio tests. The cointegration test was performed to ascertain the long run relationship between the variables that were captured in the model.
The Trace test: The trace statistic for the null hypothesis of co-integrating relations is computed as follows: Maximum eignvalue static tests the null hypothesis of r co-integrating relation against r + 1 co-integrating relations and is computed as follows: Г max (r, r + 1) = -log (1-λ r + 1 ) Page 13/18 The trace values and the maximum eigen values are then compared with the critical value at 5% level. If at most one trace value and the maximum eigen value exceeds the critical value at 5% level, then the null hypothesis of no co-integration is rejected, hence there is cointegration and vice versa. Cointegration of variables is an indication that they move together in harmony and are most likely to converge in the long run, which augurs well for policy making. Cointegration results are reported in Table 7 and indicate that maximum Eigen-statistic and trace-statistic reject the rst null hypothesis of no cointegrating relation among variables at the 5 per cent level of signi cance. The second hypothesis of at most one cointegrating relationship is also rejected. It follows that two or more cointegrating relationships exist among the variables; hence they are most likely to converge in the long run. The maximum Eigen-statistic and trace-statistic show similar results, con rming that the unit root coe cients of cointegrating vectors are signi cant at the 5 per cent level. This implies the variables are indeed cointegrated and results are not signi cantly sensitive to choice of test statistic. Since the variables are all found to be stationary in rst differences and also cointegrated, the model is estimated by employing DOLS and ECM estimation techniques. From the DOLS result, it can be observed that it is only number of Deposit Money Bank branches and bank's total asset that have positive and signi cant relationship with economic growth in Nigeria. An increase in number of Deposit money banks branches and ratio of bank total asset to GDP led to a 118% and 84% increase in economic growth of Nigeria respectively. This positive and signi cant relationship between number of bank branches and economic growth in Nigeria is due to the fact that branch network expansion remains one of the traditional and most effective methods used by Nigerian banks in mobilizing funds from the public for onward lending to the de cit economic units. As stated earlier, banks in developed countries such as USA, Canada, Australia, Germany, United Kingdom e.t.c, uses the electronic channels such as the Automated teller machine (ATM) to mobilize more deposit from the public as their ATM can perform both deposit and withdrawal operations. But in a developing country like Nigeria, the ATM is currently deemed as simply a cash dispensing machine as only withdrawals can be made. Hence, bank branches remain the only channel where cash deposit transactions can be made by the public. The branches are usually located in areas with high population growth and business activities for enhanced deposit mobilization.

Regression Analysis and Interpretations
Contrary to expectation, credit to private sector has negative and insigni cant relationship with economic growth in Nigeria. A one percent increase in credit to private sector led to 13% decrease in economic growth of Nigeria. These negative relationship between credit to private sector and economic growth could also be attributed to the fact that most individuals in Nigeria who are opportune to obtain credit from banks tend to divert the loan to unproductive ventures such as acquiring expensive cars, private apartment and other luxuries rather than investing it in productive projects that will likely spur economic growth. Also, it could be attributed to the ine ciency of the Nigerian banking sector and the poor quality of bank's risk selection policy. The negative relationship between ratio of credit to private sector to GDP and economic was in support of the ndings of Ayunku and Etale (2014) and Petkouski and Kjosevski (2014). While it disagreed with the ndings of Murari (2017) whose results indicate that the domestic credit provided by the banking sector has a signi cant association with economic growth.
Also, the ratio of currency outside bank to broad money supply which is a measure of bank's intermediation e ciency has positive and insigni cant relationship with economic growth in Nigeria. A one percent rise in bank's intermediation e ciency led to 27% increase in economic growth of Nigeria. Finally, the control variables export and Government expenditure have negative and signi cant relationship with economic growth at 10% level of signi cance while lending interest rate had negative and insigni cant relationship with economic growth in Nigeria.
The adjusted R2 indicates all variables together account for 92 per cent of systemic variation in economic growth of Nigeria. Note: Robust standard errors are in bracket. * and ** denotes signi cant at 1% and 5% level respectively.
The main focus in the estimation of error correction model is an adjustment of economic towards equilibrium when a systemic change occurs as indicated by the coe cients of ECM-1 in the table 9. The adjustment coe cient of −0.578 for indicates that economic growth adjusts by 58 per cent per year. This implies that full adjustment to a steady state takes about 2years and 2months.

Conclusion
The prime objective of this study is to determine the relationship between banking sector intermediation development and economic growth by adding lending interest rate, export and Government expenditure as a control variable from 1987 to 2018 with evidence from Nigeria. The stationary of data is checked by the Augmented Dickey Fuller (ADF) unit root test and this test con rms integrated order at rst difference. We have applied the Johansen cointegration to explore the cointegration between variables. The results con rm the existence of the long-run relationship between banking sector intermediation development and economic growth in Nigeria. However, in the short run, only number of bank branches and bank's total asset have a positive and signi cant impact on economic growth. Credit to private sector has negative and insigni cant relationship with economic growth while bank's intermediation e ciency has positive and insigni cant relationship with economic growth. The control variables export and Government expenditure have negative and signi cant relationship with economic growth while lending interest rate has negative and insigni cant relationship with economic growth in Nigeria.
The result of the analysis has proven that number of Deposit money Bank branches and bank's total asset is the major driver of economic growth in Nigeria. This shows that much of Nigeria's superior growth performance is attributed to increase in the number of bank branches and growth in bank's assets. However, growth in credit to private sector was found to be negative and insigni cant. This shows that the credit given to private sectors is not effectively utilized for productive investment that will spur economic growth. Most entrepreneurs who are opportune to obtain credit from banks in Nigeria tend to divert it to unproductive ventures and thus reducing the level of investment in the country. The calls for policy makers to develop good policy measures that will ensure that bank credit are utilized for productive investment that will spur economic growth. The study contributes to knowledge as it provides a good policy recommendation that will help address the negative relationship between credit to private sectors and economic growth in Nigeria and in other developing countries of the world. Extant literatures have not been able to address this challenge.